Professional Documents
Culture Documents
com
www.Ebook777.com
Free ebooks ==> www.Ebook777.com
GENOMIC MEDICINE
www.Ebook777.com
OXFORD MONOGRAPHS ON MEDICAL GENETICS
general editors:
Judith G. Hall
Peter S. Harper
Louanne Hudgkins
Evan Eichler
Charles J. Epstein (deceased 2011)
Arno G. Motulsky (resigned 2011)
SECOND EDITION
EDITED BY EDITED BY
1
Free ebooks ==> www.Ebook777.com
1
Oxford University Press is a department of the University of
Oxford. It furthers the University’s objective of excellence in research,
scholarship, and education by publishing worldwide.
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
This material is not intended to be, and should not be considered, a substitute for medical or other
professional advice. Treatment for the conditions described in this material is highly dependent on the
individual circumstances. And, while this material is designed to offer accurate information with respect
to the subject matter covered and to be current as of the time it was written, research and knowledge about
medical and health issues is constantly evolving and dose schedules for medications are being revised
continually, with new side effects recognized and accounted for regularly. Readers must therefore always
check the product information and clinical procedures with the most up-to-date published product
information and data sheets provided by the manufacturers and the most recent codes of conduct and
safety regulation. The publisher and the authors make no representations or warranties to readers, express
or implied, as to the accuracy or completeness of this material. Without limiting the foregoing, the
publisher and the authors make no representations or warranties as to the accuracy or efficacy of the drug
dosages mentioned in the material. The authors and the publisher do not accept, and expressly disclaim,
any responsibility for any liability, loss or risk that may be claimed or incurred as a consequence of the use
and/or application of any of the contents of this material.
1 3 5 7 9 8 6 4 2
Printed in the United States of America
on acid-free paper
www.Ebook777.com
“To our Parents and Families,
In the foreword of the previous first edition of this book It is extraordinary that, within only a few years after
(see page ix), Dr. Francis Collins had welcomed the reader the completion of the sequencing of the first genomes, we
to the genome era, and has eloquently introduced the have the capacity to determine genomic variability rapidly
genome into the medical practice. It is my great pleasure and within reasonable cost constraints; however, we are far
and honor to be able to comment on the significance of the from understanding the functional significance of the vast
genome analyses in medicine. majority of variants, and we are also far from determining,
It has become abundantly clear that there are two monitoring, and understanding the environmental impacts
main etiological components when considering health of the same.
and disease: the genomic variation, and the environmen- The physician who practices genetic medicine now
tal insults. In other words, the majority of disorders result has an “organ” or field of expertise similar (or superior?)
from the interaction between individual genomic compo- to those of other disciplines: the genome! As the cardi-
sition (and history of the subsequent somatic mutations), ologist is an expert in the cardiovascular system, the neu-
and the environmental variables. Thus it is of paramount rologist in the nervous system, the endocrinologist in the
importance to study the genomic variation of each indi- endocrine glands, and the ophthalmologist in the eye, the
vidual in order to begin to understand his or her disease, geneticist is becoming an expert in the genome’s anatomy,
improve the diagnostic possibilities, and introduce intel- variability, pathogenicity, function, inheritance, history,
ligent treatments. and evolution. The geneticist using all this information
The progress in human genomic sciences continues and knowledge participates as a primary actor in the diag-
at a remarkable pace: the HapMap Project and the 1000 nosis and treatment of individuals, and becomes a notable
Genomes Project have provided the opportunity to assess voice in the chorus for family and health planning, and a
the common and rare variations in genomes from differ- whole litany of ethical, legal, social, financial, and educa-
ent geo-ethnic groups; the ENCODE Project has pro- tional issues.
vided initial information on the functional elements of Remarkably, our knowledge of the individual genomic
the genome; the genome-wide association studies have variations, and the access to information through the
provided genomic signals for functional and diagnostic Internet and the social media, have made the physician only
studies related to almost all the complex human disorders one aspect of the health management and have elevated the
and traits with measurable heritability; the work in model patient/consumer to a position of partnership. Medicine is
organisms has provided the basis for the functional analysis becoming more participatory, and the physician less pater-
of genomic variation by taking advantage of evolutionary nalistic regarding diagnostic strategies and therapeutic
conservation; the discovery of the pathogenic mutations of options. In addition, several other disciplines well versed in
a large number of (near)-Mendelian disorders has extended computational biology, information technology, and statis-
the important list of rare disorders; and the exploration tics participate in the diagnostic and therapeutic schemes,
of de novo genomic variants points to genes involved in and their expertise is indispensable in the gathering the vast
complex phenotypes and underscores the importance of amounts of data and distilling the relevant information that
the unexplored genetics of sporadic cases. All of the above guides medical decisions.
advances were possible thanks to the truly extraordinary Genome sequencing has already become an important
and unexpected progress in sequencing technologies, the diagnostic tool. Yet, could the reading of each individual’s
availability of bioinformatic tools, and the collaboration of genome solve the majority of the diagnostic problems?
scientists worldwide. Undoubtedly not, even if we achieve a full functional
vii
understanding of each nucleotide. So-called personalized The revised and rich contents of this remarkable book,
medicine will need: edited by our outstanding colleagues, Professors Dhavendra
Kumar and Charis Eng, and written by an equally outstand-
(a) the sequence of the genomes of key somatic cells ing roster of authors, deal with all aspects of genomic medi-
that have become malignant, or have developed a cine. The book combines general principles and specific
recognizable pathological phenotype; aspects of clinical practice, all in the light of genome analysis
and the current laboratory methodologies. The genome era
(b) the sequence of the genomes of billions of bacterial
of medicine is well into development, and, as Shakespeare
and other microbes that we all host in our bodies
wrote in The Tempest: “What’s past is prologue.”
(microbiomes);
(c) the epigenetic modifications in different cell types; Stylianos E. Antonarakis
Professor and Chairman of Genetic Medicine,
(d) the repeated analysis of transcriptomes of various cell
University of Geneva
types; and
President of Human Genome Organization
(e) the repeated analysis of other “-omics” components. Geneva, Switzerland
June 23, 2013
The list is not exhaustive, and the exact battery of
genome-based tests will be determined on an individual basis.
viii • F o r ewo r d
Free ebooks ==> www.Ebook777.com
A scant twenty years have passed since the word “genom- thousand inherited conditions. With recent rapid advances
ics” was coined by Victor McKusick, Frank Ruddle, and in the understanding of human genetic variation, the spe-
Tom Roderick to describe a new discipline. The suffix of cific hereditary contributions to common diseases like dia-
the word derives from the Greek ome meaning all, and aptly betes, heart disease, cancer, and mental illness are emerging
conveyed an intention to transition the study of heredity at an unprecedented rate. The very real possibility of offer-
from a focus on single genes (genetics) to the more global ing individuals who are currently healthy a personalized
perspective of all of the hereditary material. A proliferation prediction of future risks of illness is no longer a distant
of other “omics” disciplines has subsequently erupted— dream. And given that many of the common disorders for
including proteomics, metabolomics, transcriptomics, gly- which predictions are becoming possible are associated
comics, microbiomics, and many more. with proven means of reducing risk through diet, exercise,
But genomics remains the foundation of the rest, reflect- lifestyle change, medical surveillance, or pharmacotherapy,
ing as it does a comprehensive analysis of the DNA instruc- the real likelihood of widespread individualized programs
tion book. The success of the Human Genome Project has of preventive medicine grows by the day. Similarly, the abil-
now laid that instruction book wide open. As a result, the ity to make predictions about the possibility of a beneficial
life sciences have been catapulted forward, and biology has or undesirable response to drug therapy, the field of phar-
now taken its rightful place alongside physics and chemistry macogenomics, is advancing rapidly, and will soon require
as a truly digital and quantitative science. health care providers to determine the genotype before
It is the application of genomics to medicine that car- writing the prescription, at least for certain drugs. Many of
ries its greatest promise of benefit to humankind. Thus, the us predict that the complete genome sequence of an indi-
publication of this first textbook of “Genomics and Clinical vidual will become part of that person’s medical record
Medicine” marks a milestone, a coming of age. Here in the within about ten years, at a cost of $1000 or less. And the
early years of the third millennium we can see the emerg- therapeutics that we use in the future will likely be heav-
ing outlines of a new synthesis of the noble tradition of the ily dependent upon an understanding of the genomic basis
healing arts with an increasingly precise way of understand- of illness, leading to interventions that are both more accu-
ing the causes of disease, based on an understanding of the rately targeted to the underlying problem and less likely to
human genome. cause side effects.
For some in the clinical medicine community, how- All of these advances should be welcomed by anyone
ever, this textbook may come as a surprise. After all, there interested in the alleviation of human suffering. Yet a num-
are still many practicing physicians who would say they see ber of major ethical, legal and social challenges lie along
no evidence of genetics or genomics as part of their daily the path if this vision is going to be realized. In the United
medical practice. Surely, however, that reveals a problem States, for example, we still lack effective federal legisla-
with the successful communication of rapid new develop- tion to prevent discriminatory uses of predictive genetic
ments in this field, not the facts of the matter. For in these information. Major challenges also lie ahead with regard
forty-two chapters, a vast array of genomic implications for to ensuring equitable access to new genomic technologies,
nearly every condition that affects humankind is laid out in especially as our medical care system seems to undervalue
elegant and comprehensive fashion. opportunities for preventive medicine, focusing instead on
The pace of progress in genomics has been astound- treating disease once it has already appeared. But perhaps
ing. Over just the last fifteen years, largely as a consequence the greatest barrier, and the one which this book admirably
of the tools made available through the Human Genome seeks to address, is an educational one. Most members of
Project, genes have been identified for more than two the public are interested in genomics, but relatively unsure
ix
www.Ebook777.com
of the details. Seeking advice, they generally turn to their that includes both principles and specific applications. The
health care providers, but many of those professionals are introduction of this textbook, with its distinguished and
poorly prepared to become practitioners of this new art. authoritative list of contributors, thus arrives in the nick of
After all, most physicians have had little or no training in time. Welcome to the genome era.
genetics or genomics, and will be hard pressed to quickly
acquire the scientific principles, the medical knowledge, and Francis S. Collins, M.D., Ph.D.
the psychosocial skills that will be necessary for the success- National Human Genome Research Institute
ful introduction of genomic medicine. Busy practitioners National Institutes of Health
will desperately need an authoritative source of information Bethesda, MD, USA
x • F o r ewo r d to t h e F i r s t E d i t i o n
PREFACE TO T HE FIR ST EDIT I ON
Although the science of genetics is only 150 years old, to modern medical science. Rapid developments in global
genetics as inheritance has been a concept discussed since gene analysis, gene product analysis, medical bioinformat-
ancient times. The evolution and natural-selection theories ics, and targeted molecular genetic testing are destined to
put forward by Charles Darwin had clear overtones that are change the practice of modern medicine. However, many
reflected in some of our present-day concepts of the genetic practicing clinicians perceive developments in genomics as
basis of biological life. Gregor Mendel’s laws of inheritance primarily confined to the research arena, with little clini-
and successive discoveries in various aspects of genetics laid cal applicability. But DNA- and RNA-based methods of
the foundations of a number of disciplines covering differ- disease-susceptibility screening, molecular-based disease
ent areas within the science of genetics. Human genetics diagnosis and prognosis, and genomics-based therapeutic
was no exception. However, this was heavily shrouded by choices and prediction of treatment outcomes are some of
the dark clouds of the so-called eugenics movement (ster- the key areas that are likely to influence the practice of mod-
ilization of the “unfit,” etc.) of the early twentieth century, ern clinical medicine.
when history recorded one of the worst practical applica- Undoubtedly the science of genomics holds tremen-
tions of modern science on fellow human beings under the dous potential for improving human health. The World
pretext of scientific research. Health Organization (WHO) has made several recommen-
It has taken almost sixty years to arrive at our present state dations on the scope and application of genomics on global
in the science of genetics. The future now appears bright, health. It is acknowledged that the information generated
opening up many opportunities on the horizon. Clinical by genomics will provide major benefits in the prevention,
genetics is now a recognized medical specialty among sev- diagnosis, and management of communicable and genetic
eral disciplines composing the current spectrum of modern diseases as well as other common medical diseases, includ-
medicine. The basis of clinical genetics is grounded in the ing cardiovascular diseases, cancer, diabetes, and mental
sound knowledge and understanding of medical genetics illnesses (Cardon and Bell, 2001). Together, these consti-
that emerged as a spinoff of “human genetics.” tute the major global health burden, as reflected in chronic
Fifty years after the discovery of the double-helix struc- ill-health and mortality. In addition, a number of infectious
ture of the deoxyribonucleic acid (DNA) molecule (Watson diseases are associated with genomic mutations, manifest-
and Crick, 1953), characterization of the complete sequence ing in the form of increased susceptibility, clinical sever-
and organization of the human genome was successfully ity, favorable or unfavorable responses to anti-microbial
accomplished (Lander et al., 2001; Venter et al., 2001). This therapy—or in conferring protection. It is possible that the
major scientific achievement laid the foundation of “human protective effect of a microbial vaccine might be influenced
genomics”; the section of the biological sciences that studies by genomic variation.
variations, mutations, and functions of genes and control- The sequence of the entire human genome is now
ling regions, and their implications for human variations, complete—but each person carries a distinct sequence.
health, and disease. This is strengthened by developments The variation among all humans is reflected in varia-
in the other areas of genomics relating to bacteria, vectors, tion within the human genome. The genomic variation
parasites, animals, and plants. between individuals, together with environmental factors,
The identification of all human genes and their regu- probably determines each person’s disease susceptibil-
latory regions provides the essential framework for our ity, and is important in drug efficacy and side effects for
understanding of the molecular basis of disease. This that person (Holden, 2000; Chakravati, 2000). The key to
advance has also provided a firm foundation for the future genomic variation lies in finding single-nucleotide poly-
development of genomic technologies that can be applied morphisms (SNPs) and their use in disease-association
xi
studies (Stephens et al., 2001). The positional cloning and many more. Some of these areas are included in
(identifying the gene by location, followed by functional this book. Whatever the basis of distinction might be,
analysis) of the disease susceptibility loci will depend on the driver of all these terms is GENOMICS—the study of
the successful application of haplotype associations. In genomes in its entirety.
addition, these will be important in clinical studies to find Genomics is not just about genome sequencing. Apart
individuals in whom a drug is likely to be efficacious. The from full-length cDNAs and their sequences, copies of
use of SNPs in pharmacogenetics is currently restricted mRNAs that actually exist and code for different proteins
to studying genes for drug-metabolizing enzymes, such as are probably more important. The study of proteins thus
P450s, and variations in genes that target drug receptors. derived falls within the broad field of proteomics, a likely out-
The newly emerging dynamic field of pharmacogenomics is come of functional genomics and probably a true compan-
an exciting application of genomic variation in drug dis- ion to genomics. It is likely that proteomics will eventually
covery and drug development. have more practical applications in clinical medicine. This is
The recent cloning of real disease-susceptibility genes rapidly moving ahead with the completion of the HapMap
for multifactorial diseases is encouraging: for example, the Project (Nature, 2005) and the future “functional-variant
identification of NOD2 as a susceptibility gene for Crohn’s database,” a natural outcome of the HapMap Project (Gibbs,
disease (Hugot et al., 2001; Ogura et al., 2001). This is a 2005).
major development in understanding the pathophysiology It is vital that existing gaps in our knowledge about
of inflammatory bowel disease. Similar studies are likely to various “omics” disciplines be filled to ensure efficient use
unravel the genetic mechanisms in other complex medical of the valuable information emerging from research. It is
diseases. A comprehensive SNP map will allow the cloning also important that the gap between “genetic profession-
of other susceptibility alleles. However, this will depend als” and the primary-care community, as well as the “pub-
upon population sample and size, the method employed, lic health community,” be narrowed (Khoury et al., 2003).
linkage disequilibrium, or association studies, rather than Integration of this knowledge into the medical education
on the technology used (Cardon and Bell, 2001). Some curriculum and the continued professional education pro-
of the best genetic studies of this kind include studies of grams is urgently required to ensure applications of genom-
susceptibility to infectious disease; for example, of an asso- ics in the provision of health care.
ciation between chemokine receptors (CCR5) and HIV During the last two decades, the practice of medical
susceptibility, and between the bacterial transporter pro- genetics or clinical genetics has found its niche within the
tein Nramp and resistance to macrophage-infecting bac- broad purview of clinical medicine. Genetic services now
teria such as Mycobacterium tuberculosis. Similarly, various constitute a small, albeit important, component of mod-
alleles at the G6PDH locus determine malaria susceptibil- ern medical practice and public health. Currently, genetic
ity (Tishkoff et al., 2001). services focus on providing information on chromosomal
These kinds of studies, and clinical applications of the and single-gene diseases, with limited contributions to
resulting outcomes, are not without ethical concerns. Some multifactorial/polygenic diseases. How would this then
of the questions and concerns are related to “ownership” of be different from genomics? Already there is tremendous
the genes and the freedom to use collected DNA for such enthusiasm for the recently introduced term of “genomic
studies. These are complex and emotional issues, especially medicine.” In a primer on genomic medicine, Guttmacher
when we are dealing with populations who may have been and Collins (2002) viewed “genetics as the study of single
exploited or are perceived to have been exploited. These genes and their effects” and genomics as “the study not
issues should always be dealt with carefully under the statu- just of single genes, but of the functions and interactions
tory requirements and rules. of all the genes in the genome.” In simple terms, there is a
There has been a tremendous surge in various subspe- quantitative difference between the two fields—the study
cialties and technologies with names ending in -omics. of multiple genes as opposed to one gene. Thus genetics can
We are rapidly moving into the “omics” era. In addition be seen as part of genomics. However, there is a qualitative
to genomics, several new specialist fields with an “-omics” difference between genetics and genomics in medical and
suffix have recently appeared; for example, pharmacoge- health applications, ranging from the concept of disease in
nomics, nutrigenomics, metabonomics, transcriptomics, genetics to the concept of information in genomics (Khoury
proteomics, microbiomics, glycomics, toxicogenomics, et al., 2003).
xii • P r e fac e to t h e F i r s t E d i t i o n
The practice of medical genetics has traditionally Personalized medicine will not only encompass com-
focused on conditions that result from specific alterations mon medical diseases, but could also include a wide
or mutations in single genes (e.g., inborn errors of metab- range of preventable diseases. Genetic testing for future
olism, Duchenne muscular dystrophy, and Huntington’s disease-susceptibility using multiple genomic variants
disease); in parts of, or whole, chromosomes (e.g., trisomy will be possible and affordable with the application of
21 in Down syndrome); or associated with congenital “high-throughput” microarray-based genetic testing.
malformations and developmental disabilities. The exist- A wealth of information on genomics is rapidly being
ing model of medical genetic services for these conditions acquired, with the potential for a major impact on human
includes laboratory diagnosis and genetic counseling and health. However, these data and this information are scat-
management. This is supported by public health mea- tered throughout several scientific journals, reviews, and
sures to ensure the delivery of genetic services and genetic state-sponsored reports and bulletins. A clinician or health
screening (e.g., newborn screening or screening the professional often has difficulty in accessing and assimilat-
high-risk population). On the other hand, the practice ing this information for application in her or his medical
of genomics in medicine and public health will focus on and public health practice. More importantly, an inability
information resulting from variations at one or multiple to assimilate and interpret this information can lead to
loci and strong interactions with environmental factors; frustration, and therefore avoidance of potentially useful
for example, diet, drugs, infectious agents, chemicals, information.
physical agents, and behavioral factors (Khoury et al., In view of the above developments and the rapidly
2003). increasing gulf between the practitioners and the avail-
What medical and public health applications could one able literature resources, the need for a dedicated book on
foresee following the completion of the human genome genomic medicine was appreciated. Writing such a book
sequence in 2003? How could these be applied and deliv- was obviously a nearly impossible task for a single author.
ered to the 95% of human diseases that do not fall under Several leading experts in different fields of the genome
the rubric of “genetic disorders”? These are some of the science and technology therefore offered to contribute.
likely questions related to genomic medicine. Medical The views and opinions reflected in their individual chap-
and public health professionals urgently need to make ters are largely influenced by each author’s experience,
the changes necessary to accommodate rapid identifica- perception, and interpretation of the available data and
tion and characterization of the numerous genomic vari- information.
ants at multiple loci that increase or decrease the risks for This book provides a wide-ranging coverage of the
various diseases, singly or in combination with other genes, subject, from the historical progress to general aspects of
and with various chemical, physical, infectious, pharma- genomics, and describes in some detail the medical and
cological, and social factors (Khoury, 1999). This genetic health applications. Generally, all chapters follow the
and genomic information is crucial in assessing the disease same format and are written by experts in their respective
susceptibility of healthy individuals, and in personalized fields of research and clinical expertise. Each chapter pro-
primary- and secondary-prevention planning. Collins and vides a detailed and comprehensive account of its subject.
McKusick (2001) stated that, However, it is likely that some gaps might exist, due to the
inevitable time-lag between the time of writing and appear-
By the year 2010, it is expected that predictive ing in print. This is due to rapid developments in each field.
genetic tests will be available for as many as a dozen However, all efforts have been made to provide the reader
common conditions, allowing individuals who wish core information on the basic principles, scientific facts,
to know this information to learn their risks for current and likely future applications, useful relevant ref-
which interventions are or will be available. Such erences, and information on Internet-based resources that
interventions could take the form of medical surveil- should be helpful in exploring the subject further.
lance, lifestyle modifications, diet, or drug therapy. It is hoped that this book will facilitate acquiring fac-
Identification of persons at highest risk for colon tual information on genomics, developing concepts about
cancer, for example, could lead to targeted efforts to the genomic basis of human disease, and provide a practical
provide colonoscopic screening to those individuals, base for any interested clinicians and health professionals
with [the] likelihood of preventing many premature to develop an understanding of applications of genom-
deaths. ics in clinical medicine and health. It is aimed at a wide
P r e fac e to t h e F i r s t E d i t i o n • xiii
range of scientists, clinicians, and health professionals who Genovations—the advent of truly personalized healthcare. Available at
http://www.genovariations.com.
are engaged in research, teaching, and training in medi- Gibbs R (2005). Deeper into the genome. Nature. 437:1233–1234.
cal and health applications of genome-based science and Guttmacher AE, Collins FS (2002). Genomic medicine: a primer. N
technology. Engl J Med. 347:1512–1520.
Holden AL (2000). The SNP consortium: a case study in large pharma-
Finally, the practice of medicine is an art based on sound ceutical company research and development collaboration. J Com
scientific principles. It would be appropriate to quote Sir Biotech. 6:320–324.
William Osler’s remark, “If there were no individual vari- Hugot JP et al. (2001). Association of NOD2 leucine-rich variants with
susceptibility to Crohn’s disease. Nature. 411:599–603.
ability, medicine would have been science, not an art.” Khoury MJ (1999). Human genome epidemiology: translating advances
Genomics in this context provides the basis of individual in human genetics into population-based data for medicine and
variability, and the modern post-genomic clinician will public health. Genet Med. 1:71–73.
Khoury MJ, McCabe LL, McCabe ER (2003). Population screening in
need to ensure that this is applied as an art. the age of genomic medicine. N Engl J Med. 348:50–58.
Lander ES et al. (2001). Initial sequencing and analysis of the human
Dhavendra Kumar genome. International Human Genome Sequencing Consortium.
Nature. 409:860–921.
Institute of Medical Genetics Nature (2005). A haplotype map of the human genome—report from
Cardiff University, Wales the International HapMap Consortium. Nature. 437:1299–1320.
United Kingdom Ogura Y et al. (2001). A frameshift in NOD2 associated with susceptibil-
ity to Crohn’s disease. Nature. 411:603–606.
Stephens C et al. (2001). Haplotype variation and linkage disequilibrium
in 313 human genes. Science. 293:489–493.
Tishkoff SA et al. (2001). Haplotype diversity and linkage disequilib-
RE F ERENCES rium at the human G6PDH: recent origin of alleles that confer
malarial resistance. Science. 293:455–461.
Cardon LR, Bell, JI (2001). Association study designs for complex dis- Venter JC et al. (2001). The sequence of the human genome. Science.
eases. Nat Rev Genet. 2:91–99. 291:1304–1351.
Chakravati A (2000). To a future of genetic medicine. Nature. Watson JD, Crick FHC (1953). Molecular structure of nucleic acids.
409:822–823. Nature. 171:737–738.
Collins FS, Guttmacher AE (2001). Genetics moves into medical main- World Health Organization (2002). Genomics and world health—report
stream. JAMA. 286:2322–2324. from the Advisory Committee on Health Research. Geneva: WHO.
Collins FS, McKusick VA (2001). Implications of the Human Genome
Project for medical science. JAMA. 285:540–544.
xiv • P r e fac e to t h e F i r s t E d i t i o n
PR EFACE
Since the publication of Kumar and Weatherall’s Genomics gene discoveries, unravelling novel molecular mechanisms,
and Clinical Medicine (2008), the science of genomics and identifying critical focal points in molecular pathways
has made tremendous progress (Kumar, 2008; Hamburg in designing and developing targeted molecular therapy
and Collins, 2010). The preface for the first edition (see models. Inevitably, this has led to intense debate on practi-
pages. . . .) set out a challenging and visionary goal for cal matters related to disclosure and applications of perti-
authors and editors. This was successfully accomplished, as nent and incidental findings (Green, Berg et al., 2013).
reflected in positive and constructive reviews, personal or Several major global initiatives are being pursued to
published (Feero, Guttmacher et al., 2010). Exciting new curate and annotate the enormous amount of genomic data
developments in biotechnology and bioinformatics have and information from new genomic technological advances.
opened horizons that were inconceivable only a few years The common theme is genotype–phenotype correla-
ago. The chatter about next-generation sequencing is not tion, which is vital for clinical care. Leaders in this type of
restricted to post-doctoral trainees and young investiga- multi-institutional approach include the Human Variome
tors. It is evident everywhere and is now firmly ingrained Project (www.humanvariomeproject.org), the GenPhen
in the minds and souls of genetic and genomic researchers project (www.Gen2Phen.Org), and the recently launched
and healthcare professionals (Berg, Khoury et al., 2011). Global Genome Alliance (http://www.ebi.ac.uk/about/
Indeed, even as we write this second edition of (probably) news/press-releases/Global-Alliance). Successful outcomes
the first book on genomic medicine, many tertiary genetics of these projects might offer clarification and evidence that
centers are utilizing whole-exome sequencing for routine could be applied to personalizing healthcare and wellness.
clinical care (Lupski, Reid et al., 2010; Green and Guyer, However, there is sufficient evidence supporting the argu-
2011). Unravelling the complexities of the RNA molecules ment for genomic applications for enhancing the diagnostic
has made a huge impact in molecular and experimental and probably the prognostic potential of genomic medicine
biology. Challenging and controversial stem cell genomic and health (Khoury, Gwinn et al., 2007). Promising new
research captured the headlines and was applauded by the therapeutic developments have followed, particularly the
awarding of the 2012 Nobel Prize. This is truly the begin- discovery and development of new drugs and the phar-
ning of a promising phase for applied and translational macogenetic/pharmacogenomic evidence necessary for
genomics. The sky is the limit. personalizing pharmacotherapy; or at least we have made a
Enormous genomic data and information generated good start (Chin, Andersen et al., 2011).
by genome-wide association studies (GWAS), deciphering So how do genomics and all the related genome tech-
the complex phenotypes by copy-number variations and nologies affect medicine and health? Do we have enough
single-nucleotide polymorphisms, and applying knowledge data, understanding, and robust evidence, especially of clin-
gained from genetic and genomic analysis in the so-called ical outcomes, to apply and translate into practicing effec-
rare Mendelian disorders, which affect more than 18 million tive and efficient clinical medicine? It is probably safe for
Americans alone, have offered fine molecular understanding us to gently move into the next phase of genomic medicine
of the underlying pathogenic mechanisms, as well as imple- and personalized healthcare.
mentation in clinical care of molecular diagnostics, risk The thoroughly revised and practically wholly new text
assessment, genetic counseling, management, and predic- in this second edition aims to address the above questions
tive testing of at-risk relatives (Chin, Andersen et al., 2011). and dilemmas. We are pleased to present this edition under a
There is a lot of enthusiasm for applying next-generation new title, Genomic Medicine: Principles and Practice. Several
sequencing methods (alongside Sanger sequencing, given of our colleagues and professionals in related networks
the necessity of validating of next-generation data) in new might agree on this ambitious title and probably share the
xv
view that genetics and genomics knowledge is ripe for judi- RE F ERENCES
cious clinical practice, but determining the practical clinical
implementation remains the challenge. Part of the challenge Berg JS, Khoury MJ, et al. (2011). Deploying whole genome sequencing
in clinical practice and public health: meeting the challenge one bin
is bringing all clinicians some minimum of practical knowl- at a time. Genet Med. 13(6):499–504.
edge of genomic advances so that genomics-enabled clini- Chin L, Andersen JN, et al. (2011). Cancer genomics: from discovery
cal practice can be understood, embraced, and leveraged science to personalized medicine. Nature Med. 17(3):297–303.
Feero WG, Guttmacher AE, et al. (2010). Genomic medicine—an
for the delivery of value-based healthcare. Our second edi- updated primer. N Engl J Med. 362(21):2001–2011.
tion seeks to reach this lofty goal. The views and opinions Green ED, Guyer MS (2011). Charting a course for genomic medicine
from base pairs to bedside. Nature. 470(7333):204–213.
expressed herein reflect each individual author’s content Green RC, Berg JS, et al. (2013). The American College of Medical
expertise and knowledge, interpretation, and determina- Genetics (ACMG) recommendations for reporting of incidental
tion for puutting forward their views and opinions on the findings in clinical exome and genome sequencing. Genet Med.
15(7):565–74; doi:10.1038/gim.2013.73.
future of medicine in the genome era. The editors have sim- Hamburg MA, Collins FS (2010). The path to personalized medicine.
ply facilitated the process to present the material in the best N Engl J Med. 363(4):301–304.
possible and deliverable manner. Naturally, all those who Khoury MJ, Gwinn M, et al. (2007). The continuum of translation
research in genomic medicine: How can we accelerate the appropri-
worked in developing and producing the second edition of ate integration of human genome discoveries into health care and
this unique genomic medicine textbook will be anxious to disease prevention? Genet Med. 9(10):665–674.
know how medical students, post-doctoral fellows, genome Kumar D (2008). Clinical medicine in the genome era: an introduction.
Genomics Clin Med. (53):145.
scientists, and genetic/genomic physicians would rank the Kumar D, Weatherall DJ (2008). Genomics and Clinical Medicine.
new edition. Oxford, UK: Oxford University Press. New York
Even with the genome in our palm, we remain humbly Lupski JR, Reid JG, et al. (2010). Whole-genome sequencing in a
patient with Charcot-Marie-Tooth neuropathy. N Engl J Med.
aware that we continue to strive to be better healers, as we 362(13):1181–1191.
have for 4,600 years:
Dhavendra Kumar
and
Charis Eng
May 2014
xvi • P r e fac e
ACK NOWLED G MEN TS
Nelson Mandela once said, “It always seems impossible (Cardiff, Wales), Reed Pyeritz (Philadelphia, Pennsylvania),
until it’s done.” This probably applied to this mammoth Michael Parker (Oxford, England), Jane Kaye (Oxford,
book project. On a number of occasions, it did not seem England), Dan Arking (Baltimore, Maryland), Kenneth
possible that the book would ever come to be. Nevertheless, Mills (Belfast, Northern Ireland), Bill Cookson (London),
we are proud and very pleased to present the second edi- Sarra Jamieson (Perth, Australia), Graeme Black
tion of Kumar and Weatherall’s Genomics and Clinical (Manchester, England), Karen Avraham (Tel Aviv, Israel),
Medicine largely rewritten, extensively revised, presented in Eugene Healy (Southampton, England), Julian Sampson
an entirely different style and format, and wrapped with a (Cardiff, Wales), Ben Lim (Singapore), Neil Robertson
new title of “Genomic Medicine: Principles and Practice.” (Cardiff, Wales), Julian Knight (Oxford, England), and
Developing and working on the second and revised edi- Stylianos Antonorakis (Geneva, Switzerland).
tion of the Oxford genomic medicine textbook has been Apart from having a team of world-class authors, the
a rewarding and learning experience. The project had the publishing team at the Oxford University Press (OUP)
blessings of Sir David Weatherall, who inspired and guided in New York worked hard and demonstrated unmatched
the original version of this book (Genomics and Clinical patience for managing delays, handling few unacceptably
Medicine, Oxford University Press, 2008) and continued to large and poorly presented manuscripts, and applying their
advise and mentor us on this entirely new version from con- superb editorial and technical skills in the production of
ception to completion. We are fortunate to have the support this new edition. The OUP team, led by Catherine Barnes
of a fresh team of dedicated experts and authors who have and ground managed by Chad Zimmerman and Meredith
produced the finest possible text, presented in a number of Keller, deserve praise and gratitude for bringing this excep-
chapters. Whilst we have retained a few selected authors tional book to reality and allowing it a place in the presti-
and contributors from the first edition, several of our cur- gious series, Oxford Monographs on Medical Genetics.
rent authors are entirely new to the emerging and chal- All clinicians work and live for serving patients and their
lenging field of genomic medicine. We will always remain families—this book is dedicated to all our patients. Finally,
indebted and grateful for their support and contribution. no small or large project could be completed without the
A few names deserve special thanks: notably Andrew blessings and support of the family, particularly those close
Read (Manchester, England), Stephen Pennington (Dublin, to our heart and soul. We are deeply indebted and grateful
Ireland), Richard Festenstein (London), Kevin White to our parents and families for their untiring and infinite
(Chicago, Illinois), Patrick Stover (Cornell, New York), support in the completion of this book.
Rino Rappuoli (Novartis, Italy), Teri Monolio (National
Institutes of Health, Bethesda, Maryland), Angus Clarke Dhavendra Kumar and Charis Eng
xvii
Free ebooks ==> www.Ebook777.com
CON T EN TS
xix
www.Ebook777.com
25. Genetics and genomics in clinical hematology, I: 38. Genomic perspectives of clinical immunology 591
Hemostasis and thrombosis 393 Cornelius L. Verweij
John H. McVey 39. Genomic applications in clinical pediatrics 603
26. Genetics and genomics in clinical hematology, II: Vinod Cherian Varghese, Sian Morgan,
Inherited disorders of hemoglobin 404 and Ian Frayling
Sir David J. Weatherall 40. Genetics and genomics in clinical ophthalmology, I:
27. Genetic and genomics in clinical hematology, III: The spectrum of genetic eye disease 623
Acute leukemias 412 Graeme Charles M. Black
Kenneth Mills 41. Genetics and genomics in clinical ophthalmology,
28. Genetics and genomics of chronic inflammatory II: Glaucoma 636
disorders, I: Inflammatory bowel disease 431 Roshanak Sharafieh, Anne H. Child, and
Saad Pathan and Derek P. Jewell Mansoor Sarfarazi
29. Genetics and genomics of chronic inflammatory 42. Genetics and genomics in clinical ophthalmology,
disorders, II: Rheumatoid arthritis and related III: Age-related macular degeneration 652
arthropathies 448 Mark E. Kleinman and Jayakrishna Ambati
Kate McAllister and Stephen Eyre 43. Genomic applications in audiological medicine 663
30. Genetics and genomics of chronic inflammatory Daphne Karfunkel-Doron, Zippora Brownstein, and
disorders, III: Bronchial asthma 462 Karen B. Avraham
William Cookson 44. Genetics and genomics of skin diseases I: Atopic
31. Genetics and genomics of neuro-psychiatric diseases, dermatitis and other skin complex diseases 683
I: Seizure disorders 473 Nilesh Morar
William Owen Pickrell 45. Genetics and genomics of skin diseases, II: Genomics
32. Genetics and genomics of neuro-psychiatric diseases, of pigmentation and skin cancer 696
II: Multiple sclerosis 487 Eugene Healy
Katharine Harding and Neil Robertson 46. The genetic and genomic practice of reproductive
33. Genetics and genomics of neuro-psychiatric diseases, medicine 713
III: The common dementias 508 Dhavendra Kumar
Amy Gerrish, Rebecca Sims, and Julie Williams 47. Stem cell genomics: Developmental competence 741
34. Genetics and genomics of neuro-psychiatric diseases, Kyle M. Loh, Bing Lim, and Lay Teng Ang
IV: Schizophrenia and bipolar disorder 521
48. Genomic applications in critical care medicine 766
Jinbo Fan
Matthew C. Frise, Charles Hinds, and
35. Genetics and genomics of neuro-psychiatric diseases, Julian C. Knight
V: Learning and behavioral disorders 530
49. Molecularly targeted therapy for Mendelian disorders 781
F. Lucy Raymond
Mark Davies and Julian Roy Sampson
36. Clinical cancer genomics 541
50. Glossary for genetic and genomic medicine 788
Joanne Ngeow and Charis Eng
Dhavendra Kumar
37. Genomics and infectious diseases: susceptibility,
resistance, response, and antimicrobial therapy 565
Michaela Fakiola, Wei Lu, Sarra E. Jamieson, Index 801
and Christopher S. Peacock
xx • C o n t e n t s
CON T R IBU TOR S
xxi
Eyre, Stephen Harding, Katharine, BM, BCh, MRCP
Arthritis Research UK Epidemiology Unit, Centre for Institute of Psychological Medicine and Clinical Lecturer
Musculoskeletal Research, Institute of Inflammation in Neurology
and Repair, and NIHR Manchester Musculoskeletal Cardiff University School of Medicine
Biomedical Research Unit, MAHSC, Stopford Building University Hospital of Wales
The University of Manchester Cardiff, Wales, UK
Manchester, England, UK
Healy, Eugene, PhD FRCP
Fakiola, Michaela, PhD Department of Dermatology
Cambridge Institute for Medical Research Southampton University Hospitals NHS Trust
Addenbrooke’s Hospital University of Southampton
Cambridge, England, UK Southampton
England, UK
Fan, Jinbo, PhD
Departments of Epidemiology and Biostatistics Hinds, Charles, MBBS, FRCP, FRCA
and Psychiatry Professor of Intensive Care Medicine
Case Western Reserve University School of Medicine West Smithfield
Cleveland, Ohio, OH London, England, UK
Festenstein, Richard, MD Jamieson, Saara E., PhD
Gene Control Mechanisms and Disease Group Telethon Institute for Child Health Research
Department of Medicine Centre for Child Health Research
Imperial College University of Western Australia
Hammersmith Hospital Subiaco, Australia
London, England
Jewell, Derek P., FRCP, Dphil
Frise, Matthew C., BM, BCh, MRCP Nuffield Department of Clinical Medicine
Specialty Registrar in General Internal Medicine John Radcliffe Hospital,
and Intensive Care Medicine The University of Oxford
Oxford University Hospitals NHS Trust Oxford, England, UK
John Radcliffe Hospital
Karfunkel-Doron, Daphne, PhD
Oxford, England, UK
Department of Human Molecular Genetics
Frayling, Ian, PhD FRCPath and Biochemistry
All Wales Genetics Laboratory Service Sackler Faculty of Medicine
Institute of Medical Genetics Sagol School of Neuroscience
University Hospital of Wales Tel Aviv University
Cardiff, Wales, UK Tel Aviv, Israel
Gerrish, Amy, PhD Kaye, Jane, DPhil
Research Fellow HeLEX—Centre for Health, Law and Emerging
Medical Research Council (MRC) Centre for Technologies at Oxford
Neuropsychiatric Genetics and Genomics Nuffield Department of Population Health
Cardiff University School of Medicine University of Oxford
Cardiff, Wales, UK Oxford, England, UK
Gu, Zhenglong, PhD Kleinman, Mark E., MD
Division of Nutritional Sciences Departments of Ophthalmology and Visual Sciences
Cornell University and Physiology
Ithaca, New York, NY University of Kentucky
Lexington, Kentucky, KY
xxii • C o n t r i b u to r s
Knight, Julian C., DPhil, FRCP Lu, Wei, PhD
Senior Clinical Research Fellow in Genomic Medicine School of Pathology and Laboratory Medicine
Wellcome Trust Centre for Human Genetics Faculty of Medicine and Dentistry
University of Oxford The University of Western Australia
Oxford, England, UK Nedlands, Australia
Krishnan, Neeraja M., PhD McAllister, Kate, PhD
Strand Life Sciences Arthritis Research UK Epidemiology Unit, Centre for
Bangalore, India Musculoskeletal Research, Institute of Inflammation
and Repair, and NIHR Manchester Musculoskeletal
Kumar, Dhavendra, MD FRCP FACMG
Biomedical Research Unit, MAHSC
Institute of Cancer & Genetics
The University of Manchester
Cardiff University School of Medicine,
Manchester, England, UK
All Wales Medical Genetics Service
University Hospital of Wales, Cardiff McHale, Duncan
& Global Exploratory Development
Genomic Policy Unit, UCB Pharma
The Faculty of Life Sciences & Education Stockport
The University of South Wales, England, UK
Pontypridd, UK
McVey, John H., PhD, FRCP
Law, Piu Pik MRC Centre for Transplantation
Gene Control Mechanisms and Disease Group Innate Immunity Section
Department of Medicine King’s College London, Guy’s Hospital
Imperial College London, England, UK
Hammersmith Hospital
Madore, Steven J., PhD
London, England, UK
Coriell Institute for Medical Research
Lewis, Santiago Uribe Camden, New Jersey, NJ
Gene Control Mechanisms and Disease Group
Manolio, Teri A., MD, PhD
Department of Medicine
National Human Genome Research Institute
Imperial College
National Institutes of Health
Hammersmith Hospital
Bethesda, Maryland, MD
London, England, UK
Margarit, Immaculada, PhD
Lim, Bing, PhD
Novartis Vaccines and Diagnostics
Genome Institute of Singapore, Stem Cell
Siena, Italy
and Developmental Biology Group
Singapore Mauri, Marta
Harvard Medical School, Department of Medicine, and Gene Control Mechanisms and Disease Group
the Beth Israel Deaconess Medical Center, Department Department of Medicine
of Hematology/Oncology Imperial College
Boston, Massachusetts, MA Hammersmith Hospital
London, England, UK
Loh, Kyle M., PhD
Stanford University School of Medicine Maxwell, Alexander P., PhD, FRCP
Department of Developmental Biology Centre for Public Health
Stanford Institute for Stem Cell Biology School of Medicine Queens University
and Regenerative Medicine Belfast, Northern Ireland, UK
Stanford, California, CA Middleton, Anna, PhD
Wellcome Trust Sanger Institute
Wellcome Trust Genome Campus
Cambridge, England, UK
C o n t r i b u to r s • xxiii
Mills, Kenneth, PhD Panda, Binay, PhD
Professor of Experimental Haematology Ganit Labs, Bio-IT Centre
Centre for Cancer Research and Cell Biology (CCRCB) Institute of Bioinformatics and Applied Biotechnology
Queen’s University Belfast Bangalore, India
Belfast, Northern Ireland, UK
Parker, Michael, DPhil
Mohan, Viswanathan, MD, DSc The Ethox Centre, Department of Public Health
Diabetes Specialties Centre University of Oxford
Madras Diabetes Research Foundation Oxford, England, UK
ICMR Centre for Advanced Diabetes Research
Pathan, Saad, MD
Dr. M.G.R. Medical University
Senior Consultant
University of Madras
Qatar Biomedical Research Institute
Chennai, Tamilnadu
Qatar Foundation
India
Qatar
Morar, Nilesh, FRCP, DPhil
Peacock, Christopher S., PhD
Department of Dermatology
School of Pathology and Laboratory Medicine
Chelsea and Westminster Hospital
Faculty of Medicine and Dentistry
Fulham Road
The University of Western Australia
London
Nedlands; Telethon Kids Institute
England, UK
The University of Western Australia
Morgan, Sian, BSc (Hons), MRCPath Subiaco, Australia
All Wales Genetics Laboratory Service
Pennington, Stephen R.
Institute of Medical Genetics
University College Dublin Conway Institute
University Hospital of Wales
School of Medicine and Medical Science
Cardiff, Wales, UK
Dublin, Republic of Ireland
Morrissey, Brian
Penny, Michelle, Phd
UCD Conway Institute, School of Medicine
Translational Medicine Unit
and Medical Science
Eli Lilly and Company
University College Dublin
Indianapolis
Dublin, Republic of Ireland
Indiana
Natisvili, Theona USA
Gene Control Mechanisms and Disease Group
Pickrell, William Owen, MRCP, MEng
Department of Medicine
Clinical Research Fellow
Imperial College
Neurology and Molecular Neuroscience
Hammersmith Hospital
College of Medicine, Institute of Life Science
London, England, UK
Swansea University
Ngeow, Joanne, MBBS, MRCP Swansea, Wales, UK
Genomic Medicine Institute
Pyeritz, Reed E., MD, PhD
Cleveland Clinic, Cleveland, Ohio, USA
Departments of Medicine and Genetics
Oncology Academic Program, Duke-NUS Graduate
Perelman School of Medicine
Medical School and National Cancer Centre,
The University of Pennsylvania
Singapore
Philadelphia, Pennsylvania, PA
Ong, Albert C. M., DM, FRCP
Academic Nephrology Unit and Bateson Centre
University of Sheffield Medical School
Sheffield, England, UK
xxiv • C o n t r i b u to r s
Venkatesan Radha Sarfarazi, Mansoor, PhD
Diabetes Specialties Centre Molecular Ophthalmic Genetics Laboratory
Madras Diabetes Research Foundation University of Connecticut Health Center
ICMR Centre for Advanced Diabetes Research Farmington, Connecticut, USA
Dr. M.G.R. Medical University
Segal , Jeremy, MD, PhD
University of Madras
Division of Genomic and Molecular Pathology
Chennai, Tamilnadu
Department of Pathology
India
The University of Chicago
Ramesh, Aravind S. Maryland Ave.
Gene Control Mechanisms and Disease Group Chicago, IL 60637
Department of Medicine USA
Imperial College
Sharafieh, Roshanak, PhD
Hammersmith Hospital
Molecular Ophthalmic Genetics Laboratory
London, England, UK
University of Connecticut Health Center
Rappuoli, Rino, PhD Farmington, Connecticut, CT
Novartis Vaccines and Diagnostics
Sims, Rebecca, PhD
Siena, Italy
Research Fellow
Raymond, F. Lucy, FRCP, PhD MRC Centre for Neuropsychiatric Genetics
Department of Medical Genetics and Genomics
Addenbrooke’s Hospital Cardiff University School of Medicine
Cambridge, England, UK Cardiff, Wales, UK
Read, Andrew P., PhD, FMedSci Staunton, Lisa
Department of Genetic Medicine University College Dublin Conway Institute
University of Manchester School of Medicine and Medical Science
St. Mary’s Hospital Dublin, Republic of Ireland
Manchester, England, UK
Stover, Patrick J., PhD
Robertson, Neil, DM, FRCP Cornell University
Institute of Psychological Medicine and Clinical Division of Nutritional Sciences
Neuroscience Ithaca, New York, NY
Cardiff University School of Medicine
Varghese, Vinod Cherian, MD, MRCPCH
University Hospital of Wales
All Wales Medical Genetics Service
Cardiff, Wales, UK
Institute of Medical Genetics
Rötig, Agnés University Hospital of Wales,
INSERM U781 Cardiff, Wales, UK
Hôpital Necker–Enfants Malades
Verweij, Cornelius L., PhD
Université Paris Descartes Paris, France
Department of Immunology
Sadee, Wolfgang, Dr.rer.nat. Academic Medical Centre
Center for Pharmacogenomics University of Amsterdam,
College of Medicine The Netherlands
Ohio State University
Weatherall, Sir David J., DM, FRCP
Columbus, Ohio, OH
Weatherall Institute of Molecular Medicine
Sampson, Julian Roy, DM FRCP University of Oxford
Institute of Cancer & Genetics John Radcliffe Hospital
Cardiff University School of Medicine Oxford, England, UK
University Hospital of Wales
Cardiff
Wales, UK
C o n t r i b u to r s • xxv
White, Kevin, PhD Yamada, Yoshiji, MD, PhD
Department of Human Genetics Department of Human Functional Genomics
Department of Ecology & Evolution Life Science Research Center
Institute for Genomics and Systems Biology Mie University
The Pritzker School of Medicine Tsu, Japan
The University of Chicago, IL
Yandim, Cihangir
USA
Gene Control Mechanisms and Disease Group
Williams, Julie, PhD Department of Medicine
MRC Centre for Neuropsychiatric Genetics Imperial College
and Genomics Hammersmith Hospital
Cardiff University School of Medicine London, England, UK
Cardiff, Wales, UK
Ye, Kaixiong
Wise, Anastasia L., PhD Cornell University
National Human Genome Research Institute Division of Nutritional Sciences
National Institutes of Health Ithaca, New York, NY
Bethesda, Maryland, MD
Wright, Caroline F., PhD
Wellcome Trust Sanger Institute
Wellcome Trust Genome Campus
Cambridge; The Ethox Centre,
Department of Public Health
University of Oxford
Oxford, England, UK
xxvi • C o n t r i b u to r s
GENOMIC MEDICINE
PA RT I
3
thousand years, various descriptions and explanations have and survival. This is obviously more relevant to millions
been put forward to define the physical shape and func- of people in the developing and less-developed countries,
tional nature of hereditary factors. From ancient times, and where limited resources and lack of infrastructure limit the
in almost every civilization, intense debate and arguments optimal use of the science of genetics and genomics in appli-
have failed to arrive at a consensus. Most arguments focused cations to eradicate poverty and ensure optimal health. The
on whether the hereditary factor was a creation by God, a reader will find cross-references to separate chapters in the
new product fresh from the soil and water, or something in book containing detailed information and further discus-
the blood and in the semen. The symbolic representation of sion of each subject.
the phallus in ancient sculptures and paintings of the Indian
subcontinent is an example of the concept that the phallus,
and thus semen, is a key factor in the creation and transmis- G E N E S , G E N ET I C S , A N D G E N O M I C S
sion of individuals’ (including families’) physical traits and
behavior characteristics. A detailed description of the basic principles of genetics and
In the historical context, the concept of the gene was human genetic diseases is beyond the scope of this chapter.
introduced only recently as the most acceptable answer to These facts are explained in subsequent chapters and vari-
explain one of the hereditary factors. It is unclear when and ous other information resources on basic genetics and medi-
by whom this term was first introduced. It does not matter, cal genetics (see “Further Reading”). However, some basic
as the term gene (from the Greek genos, race) is now uni- principles and relevant information are outlined in this sec-
versally accepted and used in the context of understanding tion to assist the reader with limited understanding of basic
heredity, and is probably the single most important biologi- genetics.
cal factor regulating biological life, ranging from single-cell Living organisms are divided into two large classes—
organisms to multicellular mammals. Rapid and extraordi- the eukaryotes and prokaryotes. The cells of the eukaryotes
nary scientific progress made during nineteenth and twen- have a complex compartmentalized internal structure, the
tieth centuries has led to the development of genetics, the nucleus; these include algae, fungi, plants, and animals.
science of heredity. This has now been transformed into Prokaryotes, on the other hand, are single-celled micro-
the broader field of genomics that includes all genes with all organisms without any specific part harboring the genetic
possible biologically active, heritable, regulatory and evolu- material or genome; examples include bacteria and other
tionary genetic elements, whether recent or extending back related microorganisms. The other types of living organisms
through several thousand years of life on our planet. are viruses, which are intracellular obligate parasites living
In biological terms, genes, genetics, and genomics are in both eukaryotes and prokaryotes, and composed of short
keys to procreation, development, growth, function, and dispersed nucleic acid (DNA or RNA) sequences.
survival. The health of any living organism is judged by its Genetic information is transferred from one generation
physical and functional well-being. Thus, genes, genetics, to the next by small sections of the nucleic acid, deoxyri-
and genomics are central to all forms of biological health, bonucleic acid (DNA), which is tightly packaged into sub-
including that of humans. Human health depends not only cellular structures called chromosomes. Prokaryotes usually
on its own genetic or genomic constitution, but on that of have a single circular chromosome, while most eukaryotes
other organisms whose well-being is also essential to human have more than two, and in some cases up to several hun-
health—for example, food (plants, fish, and animals), shel- dred. In humans, there are 46 chromosomes arranged in
ter (homes made of wood from trees), the environment 23 pairs, with one of each pair inherited from each parent
(water, trees, and plants), protection (clothes from cotton (Figure 1.1, a & b). Twenty-two pairs are called autosomes,
and animal skin), and transportation (animals and vehicles and one pair is called sex chromosomes, designated as X and
made of wood from trees). From a medical perspective, the Y; females have two X chromosomes (46, XX) and males
science of genetics or genomics offers deep insight into and have an X and a Y (46, XY).
evidence for a number of human diseases, including infec- A chromosome consists of a tightly coiled length of
tious diseases resulting from either lack of protection and/ DNA and the proteins (e.g., chromatins) that help define
or failure in controlling the spread of microbial infections its structure and level of activity. DNA consists of two long
or parasitic infestations. This chapter introduces the reader strands of nucleotide bases wrapped round each other along
to some of the basic facts about genes, genetics, and genom- a central spine made up of phosphate and sugar (Figure 1.2).
ics, and discusses how these impact human health and that There are four bases: adenine (A), guanine (G) cytosine
of the plants, crops, and animals necessary for human health (C), and thymine (T). Pairing of these bases follows strict
Centromere 1 2 3 4 5
1 2 3 4 5 6 7 8 9
6 7 8 9 10 11 12
a
13 14 15 16 17 18
10 11 12 13 14 15 16 17 18
19 20 21 22 X 23Y or
Chromatid Y
b 19 20 21 22 X X X
Figure 1.1 Human chromosomes: a) Diploid set in a male (46, XY); b) Complete set of human chromosomes map.
C C
A C T
C A A A TA
C A T T AA
C T A G AA
FLANKING IVS IVS 2 Gene
NC GT AG GT AG NC
5’ 3’ mRNA Precursor
AAAA A Translation
Ribosome UAA
Transfer RNA
Amino
Acid Growing
Chain
Finished
Chain
Figure 1.3 The synthesis of the peptide chain from the coding sequences in the exon (Turnpenny and Ellard, 2011).
www.Ebook777.com
senescence. Some of these may be evolutionarily conserved possible pairs of individuals are then compared to see how
across species, but relevant to human health. Mutations and often each nucleotide differs. When this is done for a sam-
alterations in several of these genomic elements are linked ple of humans, the result is that individuals differ, on aver-
to a broad range of medical conditions. age, at only about one in 1,300 DNA base pairs. In other
words, any two humans are about 99.9% identical in terms
of their DNA sequences (see Chapter 2).
THE HUM AN GENOME PROJECT During the past several years, a new type of genetic vari-
ation has been studied extensively in humans: copy-number
The advent of recombinant DNA technology in the 1970s variants (CNVs) —DNA sequences of 1,000 base pairs or
revolutionized our ability to characterize and capitalize larger are fairly distributed across the genome.8 In some
on the molecular basis of human genetic disease. This laid instances, CNVs could be deleted, duplicated, or inverted
the foundation of eventually mapping and deciphering the in some individuals with mild phenotypical effects. Several
DNA sequence of all the structural and functional genes of thousand CNVs have been discovered in humans, indicat-
the human genome. The Human Genome Project (HGP) ing that at least 4 million nucleotides of the human genome
was, therefore, a natural progression from all previous (and perhaps several times more) vary in copy-number
developments in the field of human genetics. Such a mam- among individuals. CNVs thus are another important class
moth task could not have been accomplished without the of genetic variation and contribute to at least an additional
international collective efforts supported by generous fund- 0.1% difference, on average, between individuals. Despite
ing from governmental and nongovernmental sources.6 significant progress, the medical and health implications of
The project (HGP) has helped map and provide nucle- CNVs are not entirely clear.9
otide sequences of around 23,000 nuclear genes, which, Comparisons of DNA sequences can be done for pairs
along with a number of other sequence variations, compose of individuals from the same population or for pairs of
the whole human genome (see Chapter 2). Although a individuals from different populations. Populations can be
large number of the nuclear genes have been assigned with a defined in various ways; one common way is to group indi-
structural or functional link, the precise roles of other parts viduals into populations according to a continent of origin.
of the genome are not yet fully understood. However, HGP Using this definition, individuals from different popula-
provides the basis for “functional genomics” to explore fur- tions have roughly 10% to 15% more sequence differences
ther the genome’s functional role, and understand the com- than do individuals from the same population (this estimate
plex mechanisms through which genes and their products is approximately the same for both SNPs—see below—and
interact to affect biological function and influence disease CNVs). In other words, people from different populations
processes. The development of new therapeutic agents is are slightly more different at the DNA level than are people
now possible on the basis of genomic arrangement and its from the same population. The slightness of this difference
designated functional role. This approach also helps char- supports the conclusion that all humans are genetically
acterize the genomes of various pathogens and other organ- quite similar to one another, irrespective of their geographic
isms, an invaluable tool in realizing the full potential of this ancestry.10
field to improve human health.7 Because it is still fairly expensive to assess DNA sequences
on a large scale, investigators often study genetic variations
at specific sites that are known to vary among individuals.
H U M A N G E N O M E VA R I AT I O N Suppose that a specific site in the DNA sequence harbors
AND HUM AN DISE ASE an A in some individuals’ DNA sequences, and a G in oth-
ers. This is a single nucleotide polymorphism (SNP), where
polymorphism refers to a genetic site that exists in multiple
M E A S U R I N G G E N ET I C A N D G E N O M I C
forms. The proportion of individuals who have an A and
VA R I AT I O N
the proportion with a G give the frequency of each form,
The most direct way to measure genetic differences, or or allele, and this frequency can be estimated for a sample
genetic variation, is to estimate how often two individu- of individuals from a population. If the frequencies of A in
als differ at a specific site in their DNA sequences—that three different populations are .10,.20, and .50, the genetic
is, whether they have a different nucleotide base pair at a distance between the first two populations is smaller than
specific location in their DNA. First, DNA sequences are that between the third population and the first two. On
obtained from a sample of individuals. The sequences of all the basis of this assessment, the first two populations are
Figure 1.6 Informatics as the central dogma for systems biology and genome sciences.
• Applications of genomic biotechnologies in the study world will be judged on applications in a number of areas,
and monitoring of environmental health including bio-fuels, accelerated breeding of crops and live-
stock, personalized health products, pharmaceutical effi-
cacy, and genomic monitoring of environmental health.
C O N C LU S I O N S
F U RT H E R R E A D I N G
Developments in genetics and the subsequent sequencing
of the human and other genomes have provided us with an Readers who wish to enhance their knowledge or seek more
opportunity to review the role of genes and genomes in all information are advised to consult the following books.
aspects of health and disease. Human health, including cau- Turnpenny and Ellard, Emery’s Elements of Medical Genetics,
sation of disease, is not exclusively dependent on the human Churchill Livingstone, 2011.25
genes and genome. Evolutionary links with other genomes Genomics and World Health: Report of the Advisory Committee
and ecologically relevant and beneficial parts of other on Health Research 2002, World Health Organization.26
genomes play crucial roles in the maintenance of human Harper, A Short History of Medical Genetics, Oxford University
health and, to some extent, in morbidity and mortality. Press, New York, 2008.27
Understanding genomes of microbes, parasites, animals,
plants, and crops is an acknowledged priority of current
biomedical and biotechnology research. REFERENCES
Conventionally, the causation of human disease
includes malformations, trauma, infection, immune dys- 1. Temple LK, et al. Defining disease in the genomics era. Science.
function, metabolic abnormality, malignancy, and degen- 2001;293(5531):807–808.
2. Lay JO Jr., et al. Problems with the “omics.” TrAC.
erative conditions associated with aging. Genetic factors 2006;25(11):1046–1056.
have long been recognized in all of these disease groups. 3. Feero WG, et al. Genomic medicine—an updated primer. N Engl J
The traditional genetic categories of diseases include chro- Med. 2010;362(21):2001–2011.
4. Kumar D, Weatherall DJ. Genomics Clin Med. 2008: Oxford
mosomal disorders, single-gene or Mendelian diseases, University Press, Oxford.
and several forms of multifactorial/polygenic conditions. 5. Fisher RA. The Genetical Theory of Natural Selection: A Complete
In addition, somatic genetic changes and mutations of the Variorum Edition. 1999: Oxford University Press, Oxford.
6. Collins FS, Morgan M, Patrinos A. The Human Genome Project: les-
mitochondrial genome probably account for a small, albeit sons from large-scale biology. Science. 2003;300(5617):286–290.
important, number of diseases. These groups of disorders 7. Collins FS, McKusick VA. Implications of the Human Genome
are well recognized and have an established place in the Project for medical science. JAMA. 2001;285(5):540–544.
8. Conrad DF, et al. Origins and functional impact of copy number
classification of human disease. Recent developments in variation in the human genome. Nature. 2009;464(7289):704–712.
genome research have provided vast data indicating differ- 9. Pinto D, et al. Functional impact of global rare copy number variation
ent genomic mechanisms to explain complex pathogenesis in autism spectrum disorders. Nature. 2010;466(7304):368–372.
10. Freeman JL, et al. Copy number variation: new insights in genome
in some disorders. The spectrum of these disorders is wide diversity. Genome Res. 2006;16(8):949–961.
and includes both acute and chronic medical and surgical 11. Akey JM, et al. Interrogating a high-density SNP map for signatures
of natural selection. Genome Res. 2002;12(12):1805–1814.
diseases. Perhaps it is reasonable to identify these disorders 12. Puffenberger E. Genetic heritage of the Old Order Mennonites of
on the basis of underlying molecular pathology, including southeastern Pennsylvania. In Am J Med Genet C: Sem Med Genet.
genomic imprinting, genomic rearrangements, and gene– 2003: Wiley Online Library.
13. Harris MI, et al. Is the risk of diabetic retinopathy greater in
environment interactions involving multiple genes and non-Hispanic blacks and Mexican Americans than in non-Hispanic
genomic polymorphisms. whites with type 2 diabetes? A US population study. Diabetes Care.
This chapter has reviewed the genetic and genomic 1998;21(8):1230–1235.
14. Harris MI, et al. Racial and ethnic differences in glycemic control of
approaches to human health and disease. The genomic adults with type 2 diabetes. Diabetes Care. 1999;22(3):403–408.
approaches to understanding and managing human disease 15. Yaspo, M.-L. Taking a functional genomics approach in molecular
are rapidly being incorporated in the practice of clinical medicine. Trends MolMed. 2001;7(11):494–501.
16. Steinmetz LM, Davis RW. Maximizing the potential of functional
medicine. In addition, applications of genome science and genomics. Nature Rev Genet. 2004;5(3):190–201.
technology are also reforming biotechnologies in a number 17. Rehm B. Bioinformatic tools for DNA/protein sequence analy-
of industries, including pharmaceutical, agricultural, and sis, functional assignment of genes and protein classification. Appl
Microbiol Biotech. 2001;57(5–6):579–592.
ecological bioengineering. The enormous impact of genome 18. Gehlenborg N, et al. Visualization of omics data for systems biology.
sciences and technologies on the economy of the developing Nature Methods. 2010;7:S56–S68.
www.Ebook777.com
19. Burke W, et al. Translational genomics: seeking a shared vision of 23. Singer PA, Daar AS. Harnessing genomics and biotechnology to
benefit. Am J Bioethics. 2008;8(3):54–56. improve global health equity. Science. 2001;294(5540):87–89.
20. Karp PD, et al. Pathway Tools version 13.0: integrated software 24. Kumar D. Genomics and Health in the Developing World.
for pathway/genome informatics and systems biology. Briefings in 2012: Oxford University Press, New York.
Bioinformatics. 2010;11(1):40–79. 25. Turnpenny PD, Ellard S. Emery’s Elements of Medical Genetics.
21. Khoury MJ, et al. Population sciences, translational research, and 2011: Churchill Livingstone, Edinburgh.
the opportunities and challenges for genomics to reduce the burden 26. WHO Advisory Committee on Health Research. Genomics and
of cancer in the 21st century. Cancer Epidemiol Biomarkers Prev. World Health: Report of the Advisory Committee on Health Research
2011;20(10):2105–2114. 2002. 2002: World Health Organization, Geneva, Switzerland.
22. Kumar D. Clinical medicine in the genome era: an introduction. 27. Harper PS, A Short History of Medical Genetics. 2008: Oxford
Genomics Clin Med. 2008(53):145. University Press, New York.
13
3
2 2
p
2 2
1 1 1 1
1 1
1 1
1 1
1 1
2
2 2
q 2 2
3 2
4 3 3 3
1 2 3 4 5 6
2
2 2
p 1
1 1 1
1 1
1 1 1 1
1 1
2 2
q
2 2 2
3 3 2
7 8 9 10 11 12
1 1 1
p 1 1 1
1 1 1 1
1 1
2 2 2
q 2 2
2
3 3
13 14 15 16 17 18
p 1 1 1 1
1 1
1 1
q 1 1 2 1 1
19 20 21 22 Y X
Chromosomes have two general functions. On a large All that matters is that each chromosome should be a stable
scale, they are necessary to ensure accurate partitioning of DNA package with a single centromere and protected ends.
the replicated DNA between daughter cells at mitosis, and On the micro scale, the way the naked DNA is packaged by
to allow the more complicated events of meiosis. For this a variety of proteins and small RNA molecules is a crucial
purpose, the gene content and DNA sequence are irrelevant. determinant of its function, as will be described below.
The stunning success of the Human Genome Project has A functioning chromosome must have one, and only one,
had one small negative effect. By focusing attention on centromere. During cell division, spindle fibers pull chroma-
DNA sequences written as extended lines of A, G, C, and tids apart, and centromeres nucleate the formation of kinet-
T letters, it encourages the unwary seeker to forget that our ochores, the structures to which the spindle fibers attach.
genome functions, not as extended naked DNA, but as Thus chromosomes without centromeres cannot move to the
highly folded coils and loops of chromatin, a DNA–protein daughter cell nuclei, while chromosomes with two or more
complex. Looping brings together DNA sequences that centromeres risk being pulled apart by conflicting fibers.
appear widely separated in the linear DNA sequence, allow- Human centromeres are marked by megabase-sized tracts of
ing them and their bound proteins to interact. repetitive DNA consisting of tandem arrays of highly similar
Watson and Crick showed that successive base pairs 171 bp α-satellite sequences.2 However, the defining feature
along the DNA double helix are 0.34 nm apart. This allows of centromeres is not the DNA sequence but a modification
us to calculate that, at metaphase, 1 meter (3 × 109 bp) of of the nucleosomes. Histone H3 is replaced by a variant,
DNA is packed into 23 chromosomes with a total length Centromere protein A (CENPA), and this is both necessary
of approximately 100 μm. During interphase, the DNA and sufficient to form a functional centromere. Sequencing
is more extended, but still highly organized, with loops long, highly repetitive tracts of DNA is extraordinarily dif-
of packaged DNA occupying defined territories whose ficult, because it can be nearly impossible to work out the
location (central vs. peripheral) and proximity to one correct assembly of the individual nearly identical clones.
another are believed to be important factors in control- Because of this difficulty, and because they are believed not
ling gene expression. Proteins are the main packaging to contain any genes, centromeres are not included in the
agents of chromatin, with some involvement of small current human genome sequence.
RNA molecules.
The basic level of packaging is into a string of nucleo- TELOMERES
somes. One hundred and forty-seven base pairs of naked The ends of chromosomes require a special structure for
DNA wrap around an octamer of histone proteins (two two reasons:
molecules each of H2A, H2B, H3, and H4) to form a
nucleosome. Successive nucleosomes are separated by • Because of the detailed enzymology of DNA replication,
10–80 base pairs of spacer DNA. Nucleosomes are rela- it is not possible to replicate the extreme 3′-end of a DNA
tively stable structures that nevertheless must permit poly- strand. Each round of replication shortens a chromosome
merases and other progressive enzymes to move along a by 50–100 nucleotides (nt). Within the life of a multicel-
DNA strand. Adenosine triphosphate (ATP)-powered lular organism, that would be tolerable, provided no vital
chromatin remodeling complexes of proteins assist in this gene is located near the end of any chromosome. However,
process, while the DNA of active gene-regulatory sequences over evolutionary time, it would be disastrous. Some spe-
is often relatively devoid of nucleosomes. cial mechanism is needed to restore chromosome ends, at
Adjacent nucleosomes are linked by H1 histones. They least once every generation of the whole organism.
can associate tightly or loosely to form “open” or “closed” • As part of their mechanism for repairing DNA damage,
chromatin. Large tracts of closed, genetically inactive chro- cells check for loose DNA ends and join together any
matin are called heterochromatin, whereas open, poten- that they find. Chromosome ends need protection from
tially active chromatin is euchromatin. This level of packing this mechanism.
is determined by chemical modification of the histones and
DNA. Mammalian DNA is modified by methylation of Telomeres of human chromosomes carry long arrays of
the 5-position of cytosine bases. Histones are subject to a tandemly repeated (TTAGGG)n sequence. These contract
great variety of modifications, including methylation, acet- during successive rounds of somatic cell replication. Germ
ylation, phosphorylation, and ubiquitination of specific cells have a special RNA–protein complex, telomerase, that
residues. These modifications of the DNA and histones is able to restore telomeres to full length by non-templated
constitute the epigenetic marks discussed below. Different addition of the telomeric repeat.3 Specific proteins bound
combinations of epigenetic marks produce a repertoire of to telomeres protect the DNA end, which is formed into
“chromatin flavors” that define different functional regions a non-standard looped structure that does not trigger the
of DNA. DNA damage response.
Table 2.1 DNA AND GENE CONTENT OF THE REFERENCE HUMAN GENOME (GRCH37, FEB. 2009)
CHROMOSOME LENGTH (BP) PROTEIN-CODING GENES PSEUDOGENES RNA GENES GENES PER
(KNOWN + NOVEL) MB
1 249,250,621 2037 1131 672 8.17
2 243,199,373 1259 947 526 5.18
3 198,022,430 1066 719 430 5.38
4 191,154,276 758 698 363 3.97
5 180,915,260 874 675 343 4.83
6 171,115,067 1042 726 358 6.09
7 159,138,663 907 800 350 5.70
8 146,364,022 731 568 288 4.99
9 141,213,431 803 714 260 5.69
10 135,534,747 762 498 295 5.62
11 135,006,516 1320 774 290 9.78
12 133,851,895 1051 582 336 7.85
13 115,169,878 326 323 173 2.83
14 107,349,540 652 472 310 6.07
15 102,531,392 605 471 329 5.90
16 90,354,753 867 384 229 9.60
17 81,195,210 1197 255 273 14.74
18 78,077,248 277 56 157 3.55
19 59,128,983 1418 180 198 23.98
20 63,025,520 546 213 189 8.66
21 48,129,895 233 150 69 4.84
22 51,304,566 455 308 105 8.87
X 155,270,560 836 780 351 5.38
Y 59,373,566 53 327 44 0.89
Total 3,286,906,385 21,099 15,520 11,960 6.42
The overall totals are derived from a slightly different analysis from the individual chromosome totals, so the figures do not exactly add up.
SOURCE: Data from Ensemble Release 66, March 20, 2012.
www.Ebook777.com
Primary transcript
ATG TAA
5’UT 3’UT
5’ Intron 1 Intron 2 Intron 3 3’
Figure 2.3
Structure of a typical gene. This gene has four exons. The positions of the translation start (ATG in DNA, AUG in the mRNA) and stop
(TAA in DNA, UAA in the mRNA) signals are shown. The part of exon 1 upstream of the translation start is the 5′ untranslated region (5′UT),
and the part of exon 4 downstream of the translation stop is the 3′ untranslated region (3′UT). The primary transcript includes all the exons and
introns; in the mature mRNA, the introns are removed and the exons spliced together. In a real gene, the introns would probably be considerably
larger than the exons.
Table 2.2 HUMAN GENE STATISTICS downstream of the 5′-end (the genome-wide aver-
age is 300 nt). The 5′ untranslated region binds the
Average exon number 9
ribosomes, which slide along until they encounter
Average exon size 145 bpa
an AUG codon embedded in a suitable context (a
Average intron size 3365 bp Kozak sequence, consensus GCCRCCAUGG, where
Average size of transcription unit 27,000 bp R means a purine [A or G] and the initiator codon
Average size of mRNA 2600 bp is underlined). At the 3′ end, the stop codon (UAA,
a
This is the average size of internal exons. The 3′ exon of a gene is often consider- UAG, or UGA) is usually several hundred bases or even
ably larger than the internal exons. kilobases upstream of the physical end of the mRNA.
SOURCE: Data from International Human Genome Sequencing Consortium 3′ untranslated sequences contain important elements
(2001).1
that regulate the activity and turnover of mRNAs. Note
that unlike introns and other regulatory elements (see
that is exported to the cytoplasm. A large multimolecular below), the 5′ and 3′ untranslated regions form part of
machine, the spliceosome, cuts out the introns and splices the mature mRNA.
together the exons. A special nucleotide structure, the cap,
• Introns—introns are usually considerably larger than the
is added to the 5′ end of the RNA, and a string of a few hun-
dred A nucleotides (the polyA tail) is attached to the 3′ end. exons they separate.
• Gene regulatory sequences—one reason for the increase
THE EXTENDED GENE in genome size as we move up the evolutionary scale is
the greater complexity and sophistication of gene regula-
Human genes are much larger than their coding sequences. tion in higher organisms. Table 2.3 lists some of the play-
If we define a gene as a functional unit of DNA, extra ele- ers in human gene regulation.
ments include:
The ENCODE consortium of laboratories5 is pur-
• 5′ and 3′ untranslated regions—the AUG ini- suing a large-scale effort to define the nature and action
tiation codon of an mRNA is located some distance of these regulatory elements across the entire human
ELEMENT COMMENTS
Promoter The 500 bp of DNA immediately upstream (5′) of a gene usually includes a number of different motifs that
attract and bind the transcription factor proteins that recruit and assemble an RNA polymerase complex.
Enhancer A sequence that increases transcription of the gene by binding proteins that help attach or activate the
RNA polymerase. Enhancers are defined by histone modifications closely resembling those of promoters.
They may be upstream or downstream of the promoter, and may be a considerable distance away from
the gene they regulate. Although present in every cell, they usually act only in specific tissues, presumably
because the protein(s) they bind are present only in that tissue.
Silencer Similar to an enhancer, but with a negative action.
Insulator (boundary element) Sequences that limit the extent of influence of a regulatory element. Insulators have to be located between
the regulatory element and the DNA that they are protecting. They act by preventing the spread of chro-
matin modifications along the DNA.
G E N E S E N C O D I N G F U N C T I O NA L R NA S
P S EU D O G E N E S
Not all RNA molecules in a cell are mRNA. Ribosomal
RNA and transfer RNA have long been known, but there Researchers analyzing genome sequences to identify genes
are many other non-coding RNAs (ncRNAs).6 These can often come across sequences that at first sight appear to be
be divided into “classical” ncRNAs and long intergenic functional genes, but that on closer examination are seen
ncRNAs (lincRNAs). The classical ncRNAs are small mole- to harbor changes that make them nonfunctional. These
cules, typically 16–30 nt, derived by processing much longer are pseudogenes. They are often the result of evolutionary
precursors. There has been an explosion in our knowledge duplication of a gene: when there are two copies, one is free
of the numbers and classes of these molecules, and this is to acquire mutations without affecting the function of the
still a very active research area. The main well-established other. Large gene families such as the olfactory receptor or
classes are: piRNA families are especially rich in pseudogenes.
1 3 5 6
9 11 16 19
Figure 2.4
Examples of segmental duplications in the human genome. The central line represents chromosome 7, with the small rectangle representing
the centromere. Duplicated regions at least 10 kb long and 98% identical are connected by lines, either within chromosome 7 or linked to other
chromosomes (nos. 1, 3, 5, 6, 9, 11, 16, and 19). Adapted with permission from http://humanparalogy.gs.washington.edu/build37/ figures/blowup/starburst.S10000.P0.98/chr7_10kb_98perc.pdf.
OH PH
D-loop
NA Phe 16S
7S D Val
Thr 23S
CYB
Pro Leu
H
STRAND PL ND1
Glu lie
ND6 Gln f-Met
ND2
ND5 Ala
OL Asn
Cys Trp
Tyr
Leu
Ser
His CO1
Ser
ND4
Asp
L ND4L
STRAND CO2
ND3
CO3 Lys
Arg ATPase 8
Gly
ATPase 6
Figure 2.5
The mitochondrial genome. The heavy (H) and light (L) strands of the circular 16,659 bp double helix are shown. Protein-coding genes are
shaded; transfer RNAs genes are shown as short lines with the name of the amino acid. There are no introns. OR, OL, and the heavy arrows indicate
the origins and directions of replication of the two strands. PR, PL, and the light arrows show the promoters and the direction of transcription
of the two multicistronic transcripts that are subsequently cleaved into individual mRNAs. Adapted with permission from Figure 9.1 of Strachan T, Read A (2004), Human
Molecular Genetics (3rd ed.), Garland.
normally. Presumably, with only 13 protein-coding genes, (matrilineal inheritance), but most genetic diseases where
the mitochondrial system could tolerate mutations that there is mitochondrial dysfunction are caused by mutations
modified the coding scheme in a way the main genome in nuclear-encoded genes, and so follow normal Mendelian
could not. patterns. As cells contain many copies of the mitochondrial
Mutations in mtDNA are important causes of disease, genome, they can be heteroplasmic, containing a mix of dif-
and perhaps also of aging.15 Phenotypes caused by variation in ferent sequences. Unlike mosaicism for nuclear variants, het-
mtDNA are transmitted exclusively down the maternal line eroplasmy can be transmitted by a mother to her children.
www.Ebook777.com
THE FUNCTIONAL GENOME: THE are mutational hot spots. All cytosines, methylated or not,
E P I G E N O M E , T R A N S C R I P TO M E , have a tendency to deaminate spontaneously. Deamination
A N D P R OT E O M E of unmethylated cytosine produces uracil. Cells recognize
this as an unnatural base in DNA and repair the damage
The genome is fixed and the same in every cell, apart from by replacing uracils with cytosine. However, deamination
the effects of mitotic errors and somatic mutations (espe- of 5-methyl cytosine produces the natural DNA base thy-
cially important in cancer) and the special rearrangements mine. Cells are therefore unable to recognize such events.
of immunoglobulin and T-cell receptor genes in lympho- Over evolutionary time, the majority of methylatable cyto-
cytes. But all these other “-omes” are variable, specific to a sines have mutated to thymine, hence the scarcity of CpG
tissue, a cell type, and a point in time. They are responsible sequences in the human genome.
for development, differentiation, and the response to exter- Although most of the human genome is depleted in CpG
nal changes. sequences, approximately 1% consists of “CpG islands” where
the cytosines are usually unmethylated, and so have not been
lost. CpG islands are associated with the 5′ ends of approxi-
ONE GENOME, MANY EPIGENOMES
mately 60% of human protein-coding genes, particularly
Patterns of cell- and tissue-specific gene expression are estab- the ubiquitously expressed housekeeping genes. Typically,
lished and maintained by the patterns of epigenetic marks islands are a few hundred base pairs long, and lie immediately
on the genome. As mentioned above, these consist of DNA upstream of the gene, or overlap the first exon. Methylation
methylation and a variety of specific covalent modifications of the island is associated with pathological gene silencing.
of histones. The marks are established by a large series of
“writers”: DNA methyltransferases, histone methyltrans-
T H E T R A N S C R I P TO M E
ferases and demethylases, histone acetyltransferases and
deacetylases, histone kinases and phosphatases, and so on. The transcriptome is the totality of RNA transcripts pres-
In some cases, small RNA molecules help ensure sequence ent in a cell at a given time. Next-generation sequencing of
specificity. The effects on gene expression are mediated by bulk cDNAs has allowed detailed studies of transcriptomes.
“readers,” which include methylated DNA-binding pro- All transcripts can be catalogued, and the number of times a
teins, chromodomain and bromodomain proteins that bind given transcript is sequenced is a measure of its abundance.
methylated and acetylated histones respectively, and a large Transcription of a gene depends on assembling an initia-
number of other proteins. tion complex upstream of the gene. This includes the RNA
As a result of these modifications, chromatin exists in polymerase, but also a whole suite of transcription factors
a variety of epigenetic “flavors.” The basic distinction is and co-activators that provide the specificity and control
between heterochromatin (inactive, repressed) and euchro- of transcription. Sequences to be transcribed are identified
matin (potentially active), but subtypes define transcrip- in the DNA both by the chromatin flavor and by specific
tional activity and regulatory elements such as promoters, small-sequence motifs that bind transcription-factor pro-
enhancers, and insulators. The flavor depends on a combi- teins. Whether such sequences are actively transcribed or
nation of types and relative quantities of marks rather than not depends on the availability of the necessary proteins,
a simple histone code. For example, in an analysis of nine and the absence of inhibitory proteins, in the particular cell
different epigenetic marks in nine human cell types, Ernst at that particular time. Transcripts often vary considerably
et al.16 characterized 15 different flavors. in abundance between people, and much of this variation
is heritable.17 Presumably, this variation explains much of
human individuality.
CPG ISLANDS
Both the human transcriptome and the proteome are
A striking feature of human DNA is the scarcity of G considerably larger than our total of 21,000 genes would
nucleotides directly downstream (3′) of a C nucleotide. suggest. Contrary to earlier views, it turns out that the great
The overall GC content is 41%, so we might expect 4.2% majority of all our DNA is transcribed, at least in some cells
(0.205 x 0.205) of all dinucleotides to be CG. The observed and at some times.5 Often transcripts are found from both
frequency is one-fifth of this. The explanation lies in DNA strands of the double helix. Thus cells are awash with RNA
methylation. DNA methyltransferases specifically attach a molecules of unknown function. How much of this perva-
methyl group to the 5-position of cytosines in CG (tradi- sive transcription is functional, and how much is just “tran-
tionally written CpG) sequences. Such methylated cytosines scriptional noise,” is unknown.
27
techniques, Matrix Assisted Laser Desorption Ionization separation—isoelectric focusing (IEF) and SDS-PAGE—
(MALDI) and Electrospray Ionization (ESI), enabled ioniza- this technique provided the capacity of separating up to
tion of bio-macromolecules for subsequent analysis by MS, 1000 proteins.21 In the first dimension, proteins are sepa-
which until this time had been confined to small molecule/ rated according to their isoelectric point (pI) by IEF along a
chemical analysis.17,18 While this accelerated the development pH gradient, which is subsequently followed by SDS-PAGE,
of proteomics, its true potential was not fully realized until which separates proteins according to their molecular mass.
the genome-wide sequences (human, rat, mouse, and various After separation, proteins are visualized by staining, com-
human pathogens) became available. This genome sequence monly Coomassie Brilliant Blue or the more sensitive silver
data in the form of extensive and readily accessible sequence stain. Proteins resolved by 2-DE are then analyzed, digested,
databases provided the template that protein analysis by MS and identified using MS. This classical proteomic meth-
could use to match against experimentally generated peptide odology saw widespread application but was hampered
mass spectra to identify their constitutive proteins.19,20 This by experimental drawbacks such as poor reproducibility
development revolutionized proteomics, paving the way for and sensitivity. The introduction of commercially available
a new era in the study of human disease. Although numerous immobilized pH gradient gels for IEF aimed to overcome
techniques, including 2-DE and protein arrays, are used and the lack of reproducibility seen with 2-D maps. In 1997, in
used very effectively, MS-based proteomics has so far led the an attempt to improve the reproducibility and sensitivity of
way in the emergence of proteomics. MS instrumentation the original 2-DE protocol, a new technique, 2-dimensional
has been developed to incorporate new mass analyzers and difference in gel electrophoresis (2-D DIGE), was estab-
hybrid instruments designed to tackle the challenges associ- lished.22–24 This technique, which uses fluorescent cya-
ated with protein analysis. These developments have taken nine dyes for protein labelling prior to protein separation,
advantage of improvements in mass accuracy achieved with revolutionized quantitative proteomic analysis, with over
Fourier transform ion cyclotron resonance MS instruments, 2,700 citations, including primary research papers and
in the resolving power of time-of-flight instruments, and in reviews (GE Healthcare, 2011). The advantage of using cya-
the sensitivity and dynamic range of triple quadrupoles.13 nine dyes is that they are size- and charge-matched to the
This diversity of instrumentation and post-source manipula- amino acid lysine and so ensure negligible shift in pI during
tion of ions has ensured that MS remains the current pro- first-dimension separation, thus allowing multiple experi-
teomic technique of choice and is quite likely to continue to mental samples to be separated on a single gel, thus reducing
do so for the foreseeable future. However, there is much that the number of gels to be run in the experiment and allowing
MS as an analytical tool cannot as yet achieve, and approaches the user to include an internal standard; that is, equal quan-
that complement or perhaps even surpass its capabilities tities of all experimental samples into one pooled internal
remain highly desirable. As an illustration: a key limitation standard sample. The internal standard allows the user to
of current proteomics is that the widely used “bottom-up” apply a normalization factor to correct for experimental gen-
approach whereby proteins are cleaved to peptides prior to erated differences. The generation of images following the
MS analysis rarely gives 100% coverage of individual pro- CyeDye TM DIGE flours requires a laser scanner capable of
teins (compare this with the redundancy encountered in fluorescence excitation of the dyes, which can pose a finan-
genome sequencing), leaving potentially important “parts” cial disadvantage of the 2-D DIGE approach. The resulting
of the sequence of individual proteins refractory to analysis. improvement in experimental reproducibility is particularly
Top-down approaches, in which intact proteins are analyzed advantageous for clinical studies, where sample numbers can
and fragmented in the MS itself, are time consuming, require be much higher than in animal or cell model studies.
significant skill, lack sensitivity, and are not readily applied Following the generation of the gel images, specialist
to complex mixtures of proteins. Technologies that could software/bioinformatic packages are required to identify
address these and other important limitations are still sought. protein spots deemed to be differentially expressed. Such
software is readily available and can come as a package with
laser scanners, such as Quant XL, which is available with
T WO -D I ME N S I O N A L G EL the Typhoon laser scanner series; however, specialist bioin-
ELE C T R O P H O R ES I S formatic packages dedicated to gel imaging and statistical
analysis such as Progenesis SameSpots from Non Linear
First described in 1975, 2-DE was the answer to a widely Dynamics may also assist in differentially expressed pro-
recognized need for greater resolution in protein separa- teins. Protein spots of interest are then excised and digested
tion.11,21 By coupling two analytical methods of protein for protein identification via MS.
2 8 • P rincip l e s o f G e no m ic M e dicin e
However, a number of well-publicized limitations have G L O BA L P ROT EO M I C A NA LY S I S
seen the 2-DE technique fall out of favor with scientists
The global analysis of complex protein samples, frequently
in recent years.25 The technique in itself is laborious, with
referred to as “shotgun” proteomics, can conceptually be
reproduction still an arduous task, although automation
undertaken by three core experimental steps as shown in
has been attempted. Its separation capability is still limited
Figure 3.1. They consist of 1) protein extraction (which
when analyzing complex protein species, with both dynamic
can be optionally followed by protein fractionation prior
range and total numbers of proteins still a major challenge.
to enzymatic digestion to reduce sample complexity);
Moreover, the loss of hydrophobic protein species during
2) enzymatic digestion of protein into peptides (routinely
the IEF step and the inability of SDS-PAGE to resolve pro-
using trypsin), followed by upfront separation of peptides
teins with a pI at the extremities of the pH scale have led
by liquid chromatography (LC) (offline peptide separation
to much criticism of the technique.26 While 2-DE has now
can also be considered to further reduce sample complex-
largely been replaced by MS techniques, its utility in pro-
ity); and 3) peptide/protein quantification and identifica-
teomics should not be ignored. The ability to isolate protein
tion determined from the generation of MS1 or MS/MS
isoforms and post-translational modifications means it still
data respectively. MS1 data are generated when peptides
provides a significant advantage over alternative proteomic
eluting from the LC system “fly” through the mass spec-
strategies.27 This advantage was recently highlighted by
trometer intact, with the resulting peptide ion signal pro-
Angelica Gorg, one of the pioneers of 2-DE, who wrote,
portional to the peptide abundance in the sample. MS/
“Due to the wide diversity of protein abundance and prop-
MS data are generated when peptide ions are fragmented
erties in complex proteomes it is anticipated that no single
in a collision cell within the mass spectrometer, generating
proteome analysis will be able to effectively address all the
a series of fragmentation ions that are used to build MS/
proteome analysis required.”
MS spectra. Peptides are then identified by matching these
spectra against a theoretical spectral database generated
P ROT E I N S P OT I D E N T I FI C AT I O N from a genome sequence database. Identified peptides are
BY M A S S S P EC T RO M ET RY subsequently rolled up to generate a list of proteins identi-
MS identification of the protein gel spots consists of excis- fications with this process stemming the term “bottom–up
ing the spots of interest followed by in-gel digestion.28 The proteomics.” By matching peptides ion (MS1) with their
reduced complexity of protein gel spots enables protein fragmentation ions (MS/MS), proteins are quantified and
identification by MALDI and LC-ESI-based instruments, identified. This process is conceptually undertaken in most
with LC-ESI shown to provide a greater protein sequence LC-MS/MS workflows with the exception of iTRAQ label-
coverage.29 ling and spectral counting, where protein quantification is
determined from the MS/MS data.
While the choice of mass spectrometer is determined
M A SS S P E C T R O MET RY by a variety of parameters, the popularity of liquid chroma-
tography coupled to a mass spectrometer via ESI has largely
Although in the early days, MS-based proteomics largely overtaken MALDI-based approaches due to its ability to
provided a supporting role for 2-DE experiments, the tech- seamlessly connect the LC and MS, providing peptide sepa-
nology quickly moved to provide an independent means of ration in an automated fashion whereby 2500 proteins can
analysis for complex samples. Initially providing qualitative be identified in a mammalian cell over a 90-minute analy-
data, new methods for quantification of complex mixtures sis window.35 In 2011, Nagaraj and colleagues detailed the
and targeted approaches for individual proteins have con- in-depth proteomic and transcriptomic profiling (RNA-
tinued to develop and improve in recent years.30 This con- seq) of the Hela human cervical cell line to determine
tinued evolution in MS has resulted in a plethora of mass what depth of proteome coverage could be achieved.36
spectrometers run as single units and coupled, known as This experiment consisted of 288 hours of analysis time in
tandem mass spectrometry (MS/MS), which greatly vary in a LTQ-Orbitrap mass spectrometer, resulting in the iden-
analytical performance and experimental capabilities.31 The tification of 10,255 proteins encoded by 9,207 genes from
rapid development of MS has in turn forced the develop- the total 11,936 estimated genes in total. This undoubtedly
ment of supportive bioinformatics tools to analyses the demonstrates the deepest level of coverage of any human
increasingly complex data, a process that now requires as cell to date. However, this level of analysis may still be out-
much consideration as the MS analysis itself.32–34 side the capabilities of most proteomic research centers, and
H u m an P rot e o m ic s • 2 9
MS spectra
Peak
Quantification
Proteome
Abundance
Protein Peptide LC-MS+
Extract Extract LC-MS/MS Peptide/Protein
Identification
MS/MS
Protein Peptide spectra
Fractionation Fractionation
Database
search
m/z
it is extremely time consuming. Moreover, this experiment including O18 labelling, isotope-coded affinity tags (ICAT),
was limited to generating a list of protein identifications isobaric tags for relative, absolute quantification (iTRAQ)
for a single cell type. The ability to quantify the proteins and stable isotope labelling by amino acids in culture
either for an individual cell type or a more complex samples (SILAC).39–42
such as tissue lysates or body fluids would be a much greater O18 labelling is one of the simplest labelling approaches
challenge. available for shotgun proteomics. In this enzyme-catalyzed
Figure 3.1 outlines the core experimental steps for global process, O16 is replaced with O18 at the C-terminal of the
analysis of complex protein samples (shotgun proteomics), carboxyl group of proteolytical peptides, resulting in a 4Da
including: protein extraction, enzymatic digestion followed mass shift with subsequent relative quantification via the
by separation of peptides by liquid chromatography (LC), parent ion peak height/area.42 As labelling is performed
and peptide/protein quantification and identification during protein digestion, this technique is particularly use-
determined from the generation of MS1 or MS/MS data ful where metabolic labelling such as SILAC is not possible,
respectively. such as human tissue or serum providing a suitable labelling
approach for clinical studies.43,44 However, a number of lim-
itations have been reported, including a limited dynamic
L A B E L O R L A B E L-FR E E?
range and variability of uptake among different peptides.45
The process of generating qualitative and/or quantitative iTRAQ, first described by Ross et al., consists of a set
peptide data is by no means trivial, and one of the first con- of isobaric labels that are isotopically incorporated at the N
siderations before analysis can take place is to label or not termini and lysine side chain peptides in a digest mixture.41
to label. Both provide relative quantification of peptides These isobaric tags are indistinguishable at the MS level;
across samples with the choice of employing a labelled or however, following MS/MS fragmentation, the reporter
label-free approach, determined by cost, sample numbers, groups that contain variable mass (114-117 Da or 113-121
sample type, and the degree of required accuracy.37 Da) are released, with the peak area of the reporter ions
Labelling approaches utilize the incorporation of a used to determine the relative abundance of the peptides at
metabolic, enzymatic, or chemical isotopic tags that change the MS/MS level.38 Initially released as a 4-plex, and later
the mass of the peptides without changing its biochemical 8-plex allowing analysis of up to eight samples, the iTRAQ
properties, in a process referred to as differential stable iso- procedure has been applied to numerous biological stud-
tope labelling. By combining labelled and unlabeled samples ies of different samples, such as human saliva, fibroblasts,
in a single run on a mass spectrometer, both labelled and and mammary epithelial cells.46 However, given the limita-
unlabeled peptides can be identified and differentiated. tion of eight labels for analysis, this technique is incompat-
The measured signal coming from both peptides can then ible with large-scale clinical applications where numerous
be used to determine the relative quantitative differences patients or sample numbers are present.
between the two samples.38 Numerous labelling procedures Isotope-coded affinity tags (ICAT), first described by
have been developed that support shotgun proteomics, Gygi and Aberscold in 1999, is a chemical label consisting
3 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
of three functional elements: a specific chemical reactivity, problem, along with its dependency on the generation of
an isotopically coded linker, and an affinity tag (biotin).40 high-quality MS/MS data.38 The quantification of peptides
Labelling of reduced protein is achieved as the thiol-reactive via the ion peak intensity/area signal is based on the obser-
group is selective for the sulfhydryl group in the side chain vation that, as a peptide is eluted from the LC system, the
of reduced cysteine, which are then subsequently isolated resulting increase of observed ion signal in the mass spec-
via the biotin-avidin affinity enrichment.47 The label is com- trometer is proportional to the concentration of the peptide
pleted with an isotopically labelled linker available in two in question. By mapping this peptide elution profile, a peak
forms, heavy and light. The heavy label, which originally area is generated, with the resulting area under the curve
consisted of deuteriums, has subsequently been replaced used to determine the peptide peak abundance. Subsequent
with C13, resulting in a 9 Da mass shift between the heavy retention time alignment of peptide peaks across multiple
and the light labelled peptides.48 ICAT has been applied samples thereby enables the relative quantification of pep-
extensively in biological research with applications in whole tides, with no restriction on the number of samples that can
cell lysates,49 conditioned culture medium,50 and subcellu- be analyzed in any one experiment.37 However, this data
lar fractions such as mitochondria.51 However, the limited processing is not a trivial task, with time alignment, peak
availability of labels does restrict its use; moreover, as label- quantification, peak matching, peak identification, and
ling is achieved via cysteine binding and subsequent enrich- subsequent statistical analysis requiring sophisticated bio-
ment of labelled peptides, there is a resulting loss of peptide informatic tools.56 Fortunately, this bioinformatic support
devoid of the presence of cysteine. is readily available, with numerous free online applications
Stable isotope labelling by amino acids in culture available and commercial platforms with dedicated support
(SILAC), first established a decade ago, consists of the for persons of limited bioinformatic knowledge, further
incorporation of specific amino acids into (mammalian) opening the door of proteomics to the field of biology.57,58
proteins by culturing cells in media depleted of an essen-
tial amino acid and replacing them with an isotopically
labelled form of the amino acid.39 In this process, two TA R G ET E D P R OT E O M I C S
cell populations are grown, one cultured in medium with
heavy-labelled amino acids containing 2H instead of H, As global proteomics continued to produce increasingly
13
C instead of 12C, or 15N instead of 14N, and the other in complex data and identified even greater numbers of pro-
medium with the light amino acid (unlabeled). The result- teins deemed to be of biological interest, the need to vali-
ing known mass shift in the heavy-labelled peptides enables date was becoming increasingly evident. This bottleneck in
differentiation and quantification via the parent ion peak development still poses a significant problem in all areas
height/area.52 Although SILAC was initially established of proteomic research, but significantly so for biomarker
for the analysis of cell culture, this technique has now been research, where validation of multiple targets across large
extended to animal models, thereby increasing its applica- numbers is required. Traditionally, validation of proteins
bility to the study of human disease.53–55 was conducted by Western Blotting and/or enzyme-linked
immunosorbent assay (ELISA) techniques that require
a lot of protein; however, although these techniques still
Label-Free
demonstrate superior sensitivity over MS, their ability to
While labelling strategies are generally accepted as being multiplex is limited. Moreover, the development of anti-
more accurate, the associated cost of label reagents and bodies for every target identified can be a time-consuming
dedicated software, combined with limited availability of and expensive process and ultimately a death sentence for
labels, has driven the development of label-free strategies. the validation of biomarker candidates. An alternative
Label-free quantification can be undertaken by two distinct technique that could be considered the ELISA of the MS
processes: spectral counting and peptide ion peak intensity/ world is multiple reaction monitoring (MRM), which has
area signal. Spectral counting is based on the observation been used for testing chemical analytes for decades. The
that the number of MS/MS spectra generated from a sin- importance of targeted MS was highlighted in 2012 as
gle peptide is proportional to the abundance of that pep- the Method of the Year by Nature Methods.59 Unlike the
tide in the sample: that is, the more abundant, the greater shotgun approach, MRM has the ability to specifically tar-
chance it will be selected for MS/MS fragmentation. While get individual “proteotypic” peptides (peptides unique to
this method has demonstrated linearity over two orders one protein) with the measured signal representative of its
of magnitude, bias towards highly abundant proteins is a constitutive protein. This process, traditionally undertaken
H u m an P rot e o m ic s • 3 1
www.Ebook777.com
in a triple quadrupole mass spectrometer, consists of three performance in analyzing the sample, in what was a worrying
steps: 1) proteotypic peptides (parent ions) are selected outcome, many of the centers failed to identify all protein
in quadrupole (Q)1; 2) the selected parent ion is then species in the standard, and in many cases they identified
fragmented in the second quadrupole (Q2) by allowing proteins that were not present.68 Moreover as the popularity
it to collide with the molecules of an inert gas smashing; of proteomics continues to grow, the number of scientists
3) the resulting multiple fragments or “transitions” are active in the field with limited bioinformatic capabilities
then selected to allow them to pass through Q3 and reach is on the increase, thus requiring user-friendly supportive
a detector. The intensity of the transitions (fragments of software. Fortunately such software has been developed in
the parent ion) is a measure of the amount of the parent abundance in recent years to supply users with a range of
ion (peptide). This serial selection process provides a fast, abilities. The analysis of labelled or label-free shotgun data
sensitive, and cost-effective solution to validating multiple has been provided with numerous open-source packages
protein targets in large numbers of patient samples.60–62 In such as SuperHirn and Maxquant, and commercial pack-
a move to increase sensitivity, Fortin et al. demonstrated ages such as Progenesis LC-MS.37,69,70 The choice of software
the use of a new development, MRM(3).63 This technique can be difficult and is dependent on one’s bioinformatic
is similar to multiple reaction monitoring (MRM), with the skill, the cost, and the software’s performance. Moreover,
exception of the linear ion trap used in Q3. At this stage, the performance of data analysis can vary greatly from plat-
the transitions generated in Q2 are subjected to a second form to platform, as was recently highlighted by Mancuso
round of fragmentation in Q3, with the resulting fragmen- et al. when comparing nine different platforms using a
tation used for quantification. This process has been shown large Orbitrap data set.71 The development and analysis of
to increase sensitivity of the lower limits of detection three- MRM assays has also been supported with high-quality
to five-fold in plasma-based studies when compared to and in many cases intuitive bioinformatic software. While
MRM.64 However, the sensitivity of MRM when used for all mass spectrometer vendors provide machine-specific
measuring peptides in complex biological fluids continues software (e.g., Agilent-Mass Hunter, ABSciex-MultiQuant)
to be a challenge.65 In particular, the plasma proteome in for the post-analysis of MRM data,32 the development of
which proteins span at least 10 orders of magnitude in con- such assays can be a more complex process. Fortunately,
centration range continues to challenge mass spectrome- numerous free online platforms (Skyline72, MRMer73) for
trists with its analytical complexity.66,67 As the sensitivity the development and analysis of MRM have provided users
of MRM and MRM(3) continues to improve, and given with a single analysis tool for MRM assays.74 The developers
their current ability to multiplex protein measurements, it of such programs have significantly accelerated the analy-
is expected that these techniques will play a leading role sis of such data, preventing what would have been another
in studies to validate of biomarkers. Moreover, with the major bottleneck in the validation of many new candi-
inclusion of isotopically labelled peptide standards, this date biomarkers identified in global proteomic studies. In
technique can provide absolute quantification of peptides/ addition to the analysis of shotgun proteomic and target
proteins, providing a real alternative to ELISA and other proteomic data, other software tools that provide file con-
antibody-based quantification platforms. version capabilities, de novo sequencing, post-translational
modification and structure analysis, and more can be
sourced at www.ms-utils.org.
B I O I N F O R M AT I C S O F M A SS
S P E C T R O MET RY
P R OT E I N M I C R OA R R AYS
As the data output from mass spectrometers became
increasingly complex, the need for supportive bioinformat- Following the establishment of DNA microarray tech-
ics to decipher this information was critical. While tradi- nologies in the 1990s, protein microarrays soon followed
tionally bioinformatics, and to a lesser extent statistics, suit, with a plethora of array-based technologies now
were seen as a tax on proteomic scientists, the generation of available with applications in drug discovery, biomarker
increasingly complex data ensured that bioinformatic anal- research, and pathway analysis of disease networks.75,76
ysis now required as much careful consideration as choos- These microarrays primarily fall into two distinct catego-
ing the instrument of analysis. In a report from HUPO, ries: 1) functional microarrays that identify protein inter-
where a standard containing 17 proteins was sent to pro- actions, and 2) abundance microarrays used in protein
teomic research centers across the globe to determine their quantification. Functional protein microarrays comprise
3 2 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
purified proteins, protein domains, or peptides immobi- enrichment, affinity-based enrichment, and MS-based
lized on glass support slides.77 Such microarrays have been PTM analysis.93–96 Affinity-based techniques such as
used to identify protein interactions with other proteins, immobilized metal affinity chromatography (IMAC),97
nucleic acids, lipids, small molecules, and biomolecules. and titanium dioxide (TiO2)98 have allowed global char-
These microarrays have provided in-depth network signal- acterization of phosphorylated proteins in complex
ing information for proteins implicated in human disor- samples. Other PTMs such as ubiquitylation94 and glyco-
ders,78 protein–pathogen interactions,79 and the human sylation99 are now receiving attention due to the techni-
interactome.80 Abundance-based microarrays, used in cal advances in mass spectrometry. Standard PTM studies
quantitative protein-expression analysis, have been exten- focus either on PTMs of a single protein of interest, or
sively used in the areas of clinical trials and personalized on PTMs of a protein population. Singular protein PTM
medicine.81,82 These arrays typically consist of either cap- studies focus on antibody-specific precipitations and
ture molecules such as antibodies (antibody arrays), or the affinity chromatography in order to analyze the protein
sample of interest (tissue microarrays, lysate microarrays) of interest and identify its PTMs. In contrast, PTM map-
bound to the glass support, with subsequent detection ping of a protein population is a formidable task, as there
via direct sample-labelling or labelled detection antibod- is a need to systematically assess modifications on a large
ies.77 Antibody arrays and tissue microarrays have been number of proteins. Both “bottom up” and “top down”
widely applied in biomarker research for bladder cancer,83 proteomics followed by MS have been used to identify
breast cancer,84,85 and pancreatic cancer.86,87 Lysate/reverse different PTM combinations and interactions in order to
microarrays have proven to be a particularly useful tool in understand the interplay between different cellular stim-
the analysis of cell line models of disease88,89 or tissue lysate ulation and physiological responses.100
such as breast tissue aspirates.90 However, one of the inher-
ent limitations of these abundance-based microarrays is
the reliance on antibodies. The ability to produce spe- M A L D I-I M AG I N G
cific and reliable antibodies that work consistently under
experimental conditions can significantly affect reliabil- Matrix-assisted laser desorption ionization (MALDI)
ity.91 In addition, protein microarrays, as with MS-based imaging mass spectrometry (MALDI-MSI) is a pow-
proteomics, rely on genomic sequencing to code for pro- erful and diverse technology for analyzing the spatial
tein expression, which can prove problematic where a par- distribution of endogenous and exogenous compounds
ticular species is not sequenced. directly from a tissue section.101 MALDI imaging records
the spectra from thin tissue sections in order to produce
molecular-weight encoded images of the distribution
P O S T-T R A N SL AT I O N A L of constituent biomolecules, and it was first applied to
M O D I F I C AT I O N S the visualization of proteins in 1997 by Capriloli and
co-workers.102 The technique requires the sectioning of
Even with advancements in proteomic technologies such thin slices of tissue, which are then coated with an appro-
as MS and microarray analysis, the complexity of the pro- priate matrix solution. The laser cuts across the tissue
teome cannot be ignored due to gene splicing forming section, causing the solvent to evaporate. Upon solvent
different protein isoforms and various post-translational evaporation, the extracted molecules are co-crystallized
modifications (more than 200 available). Protein with the matrix. The matrix functions to absorb the laser
post-translational modifications (PTMs) play a role in the energy and facilitate desorption/ionization of the ana-
regulation of both the structure and the function of cel- lyte molecules. Mass spectra are then acquired across the
lular proteins,92 and include, but are not limited to, phos- tissue at defined coordinates, resulting in a dataset that
phorylation, acetylation, ubiquitinylation, methylation, contains hundreds to thousands of individual spectra,
and glycosylation. With advances in MS technologies consisting of all ions detected at each location of acqui-
allowing more sensitive profiling of complete proteomes, sition. Like 2-DE, MALDI imaging suffers from lack of
similar advances have allowed the characterization of reproducibility whereby sample preparation and han-
multiple PTMs. Phosphorylation, due to its role in a dling are crucial to obtaining good images despite incon-
wide variety of cellular processes, is the most widely stud- sistencies in matrix crystal formation resulting in errors in
ied PTM, with a number of experimental approaches analyte abundances and distribution making the need for
used: 2-DE gel-based PTM analysis, antibody-based normalization imperative for accurate results.
H u m an P rot e o m ic s • 3 3
www.Ebook777.com
C H EM I C A L P R OT E O M I C S proteomics, and metabolomics, which together provide
a better understanding of biological systems, creating a
Where LC-MS/MS, protein microarrays, and 2-DE have framework for investigating the complexity of biological
increased our understanding of protein expression within systems by defining the components of the system (see also
a cell, the focus has now turned to functional proteomics Chapter 6). With the development of new technologies
whereby chemical proteomics gives researchers the abil- such as high-throughput sequencing and MS, we can gain a
ity to assess the function of a protein within a cell or tissue. global profile of health and disease, allowing the combined
Activity-based probe-profiling (ABPP) looks at the enzymatic genomic, transcriptomic, and proteomic data to develop
activity of a particular protein family.103 ABPP experiments new approaches in personalized medicine.108 Systems biol-
are designed to covalently label the active site of an enzyme ogy has led to the development of systems medicine or “P4”
or enzymes of interest. Each probe contains a reactive group medicine: predictive, preventive, personalized, and partici-
and reporter tag.104 Activity-based probes can be used in both patory.109 This type of medicine will provide deep insights
a direct and indirect fashion. Direct probes can be developed into disease complexity, allowing the development of treat-
to target distinct enzyme families; for example, biotinylated/ ment strategies for disease due to networks perturbed either
fluorophore-tagged fluorophosphonates (FPs) target the ser- by genetic or by environmental cues. Although this type of
ine hydrolase superfamily.104 Indirect strategies can be used medicine is attractive, it does have downsides. The cost of
in the profiling of proteins lacking affinity labels. Libraries of an individual patient molecular fingerprint would be great;
candidate probes synthesized and screened against complex also, this type of medicine would also depend heavily on
proteomes can identify “specific” protein-labelling events.104 technology.108
Chemical proteomics can be used for targeted discovery
experiments and inhibitory experiments.
C O N C LU S I O N S
SYS T EMS B I O L O GY
R E C O MME N D E D R E A D I N G
Systems biology is an inter-discipline that makes use of the 1. Görg A, Drews O, Lück C, Weiland F, Weiss W. 2-DE with IPGs.
“-omics” datasets, including genomics, transcriptomics, Electrophoresis. 2009 Jun; 30 Suppl 1:S122–S132.
3 4 • P rincip l e s o f G e no m ic M e dicin e
2. Nagaraj N, Wisniewski JR, Geiger T, Cox J, et al. Deep proteome 4. Smejkal GB. Genomics and proteomics: of hares, tortoises and the
and transcriptome mapping of a human cancer cell line. Mol Syst complexity of tortoises. Expert Rev Proteomics (Internet). 2012
BiolMol Syst Biol. 2011;7:548. Oct;9(5):469–472. Available from: http://www.ncbi.nlm.nih.gov/
3. Bantscheff M, Lemeer S, Savitski MM, Kuster B. Quantitative mass pubmed/23194260
spectrometry in proteomics: critical review update from 2007 to the 5. Vangala RK, Ravindran V, Ghatge M, et al. Integrative bioinfor-
present. Anal Bioanal Chem. 2012;404(4):939–965. matics analysis of genomic and proteomic approaches to under-
4. Elliott MH, Smith DS, Parker CE, Borchers C. Current trends in stand the transcriptional regulatory program in coronary artery
quantitative proteomics. J Mass Spectrom. 2009;44:1637–1660. disease pathways. PloS One (Internet). 2013 Jan [cited Mar
5. America AH, Cordewener JH. Comparative LC-MS: a landscape of 8, 2013];8(2):e57193. Available from: http://www.ncbi.nlm.nih.
peaks and valleys. Proteomics. 2008;8:731–749. gov/pubmed/23468932
6. Lemoine J, Fortin T, Salvador A, et al. The current status of clinical 6. Altelaar AFM, Munoz J, Heck AJR. Next-generation pro-
proteomics and the use of MRM and MRM(3) for biomarker vali- teomics: towards an integrative view of proteome dynamics. Nature
dation. Expert Rev Mol Diagn. 2012;12:333–342. Rev Genet (Internet). 2013 Jan [cited Feb 28, 2013];14(1):35–48.
7. Gillette MA, Carr SA. Quantitative analysis of peptides and Available from: http://www.ncbi.nlm.nih.gov/pubmed/23207911
proteins in biomedicine by targeted mass spectrometry. Nature 7. Venter JC, Adams MD, Myers EW, et al. The sequence of the human
Methods. 2013;10:28–34. genome. Science. 2001 Feb;291(5507):1304–1351.
8. Picotti P, Abersold R. Selected reaction monitoring-based pro- 8. Lander ES, Linton LM, Birren B, et al. Initial sequencing and analy-
teomics: workflows, potential, pitfalls and future directions. Nature sis of the human genome. Nature. 2001 Feb;409(6822):860–921.
Methods. 2012;(9)6:555–566. 9. Collins F, Green E, Guttmacher A. A vision for the future of
9. Mueller LN, Brusniak MY, Mani DR, Aebersold R. An assessment genomics research. Nature (Internet). 2003 Apr [cited Jun
of software solutions for the analysis of mass spectrometry based 21, 2012];422(6934):835–847. Available from: http://www.
quantitative proteomics data. J Proteome Res. 2008;7, 51–61. nature.com/ng/journal/v36/n11s/full/doifinder/10.1038%2Fnat
10. Cham Mead JA, Bianco L, Bessant C. Free computational resources ure01626
for designing selected reaction monitoring transitions. Proteomics. 10. Klose J. From 2-D electrophoresis to proteomics. Electrophoresis
2010;10:1106–1126. (Internet). 2009 Jun [cited Mar 17, 2013];30 Suppl 1:S142–149.
11. Lee JR, Magee DM, Gaster RS, LaBaer J, Wang SX. Emerging Available from: http://www.ncbi.nlm.nih.gov/pubmed/19517494
protein array technologies for proteomics. Expert Rev Proteomics. 11. Klose J. Protein mapping by combined isoelectric focusing and elec-
2013;10(1):66–75. trophoresis of mouse tissues. Hum Genet. 1975;26(3):231–243.
12. Kamath KS, Vasavada MS, Srivastava S. Proteomic databases and 12. Moreira JM, Cabezón T, Gromova I, et al. Tissue proteomics of
tools to decipher post-translational modifications. J Proteomics the human mammary gland: towards an abridged definition of the
2011;75(1):127–144. molecular phenotypes underlying epithelial normalcy. Mol Oncol
13. Vickerman J. Molecular imaging and depth profiling by mass
(Internet). 2010 Dec [cited Apr 26, 2013];4(6):539–561. Available
spectrometry—SIMS, MADI or DESI? Analyst. 2011;136(11): from: http://www.ncbi.nlm.nih.gov/pubmed/21036680
2199–2217. 13. Domon B, Aebersold R. Mass spectrometry and protein analysis.
14. Miao Q, Zhang CC, Kast J. Chemical proteomics and its
Science. 2006;312(5771):212–217.
impact on the drug discovery process. Expert Rev Proteomics. 14. Cox J, Mann M. Is proteomics the new genomics? Cell.
2012;(9)3:281–291. 2007;130(3):395–398.
15. Rix U, Superti-Furga G. Target profiling of small molecules by 15. Pennington SR, Wilkins MR, Hochstrasser DF, Dunn MJ.
chemical proteomics. Nature Chem Biol. 2009;5(9):616–624. Proteome analysis: from protein characterization to biological func-
16. Ngounou Wetie AG, Sokolowska I, Woods AG, Roy U, Loo JA, Darie tion. Trends Cell Biol (Internet). 1997 Apr;7(4):168–173. Available
CC. Investigation of stable and transient protein-protein interac- from: http://www.ncbi.nlm.nih.gov/pubmed/17708936
tions: past, present, and future. Proteomics 2013;13(3–4):538–557. 16. Herbert AP, Riesen M, Bloxam L, et al. NMR structure of Hsp12,
17. Cox J, Mann M. Quantitative, high-resolution proteomics for
a protein induced by and required for dietary restriction-induced
data-driven systems biology. Annu Rev Biochem. 2011;80:273–299. lifespan extension in yeast. PloS One (Internet). 2012 Jan [cited Apr
9, 2013];7(7):e41975. Available from: http://www.pubmedcentral.
nih.gov/articlerender.fcgi?artid=3407059&tool=pmcentrez&rend
ertype=abstract
R EFE R E N C ES 17. Fenn J, Mann M, Meng C, Wong S. Electrospray ionization for mass
spectrometry of large biomolecules. Science (Internet). 1989 [cited
1. Legrain P, Aebersold R, Archakov A, et al. The human proteome Jun 21, 2012];246(4926):64–71. Available from: http://www.sci-
project: current state and future direction. Mol Cell Proteomics: encemag.org/content/246/4926/64.short
MCP. 2011 Jul;10(7):M111.009993. 18. Karas M, Bachmann D, Bahr U, Hillenkamp F. Matrix-assisted
1a. Wilhelm M, Schlegl J, Hahne H, et al. Nature. 2014 ultraviolet laser desorption of non-volatile compounds. Int J Mass
May;29;509(7502):582–587. doi: 10.1038/nature13319. Spectrom. 1987;78:53–68.
1b. Kim MS, Pinto SM, Getnet D, et al. Nature. 2014 19. Evans VC, Barker G, Heesom KJ, Fan J, Bessant C, Matthews DA.
May;29;509(7502):575–581. doi: 10.1038/nature13302. De novo derivation of proteomes from transcriptomes for transcript
2. Nilsson T, Mann M, Aebersold R, Yates JR, Bairoch A, Bergeron and protein identification. Nature Methods. 2012;9(12):1207–1211.
JJM. Mass spectrometry in high-throughput proteomics: ready 20. Hoopmann MR, Moritz RL. Current algorithmic solutions
for the big time. Nature Methods (Internet). Nature Publishing for peptide-based proteomics data generation and identifica-
Group; 2010 Sep [cited Mar 1, 2012];7(9):681–685. Available tion. Curr Opin Biotechnol (Internet). 2013 Feb [cited Mar
from: http://www.ncbi.nlm.nih.gov/pubmed/20805795 3, 2013];24(1):31–38. Available from: http://www.ncbi.nlm.nih.
3. Kropinski AM, Waddell T, Meng J, et al. The host-range, gov/pubmed/23142544
genomics and proteomics of Escherichia coli O157:H7 bacte- 21. O’Farrell PH. High resolution two-dimensional electrophoresis of
riophage rV5. Virology J (Internet). Jan 2013 [cited Apr 19, proteins. J Biol Chem. 1975;250(10):4007–4021.
2013];10(1):76. Available from: http://www.pubmedcentral. 22. Görg A, Wilhelm P, Siegfried G. The current state of two-
nih.gov/articlerender.fcgi?artid=3606486&tool=pmcentrez&r dimensional electrophoresis with immobilized pH gradients.
endertype=abstract Electrophoresis. 1988;9:531–564.
H u m an P rot e o m ic s • 3 5
23. Görg A, Wildgruber R. Review: The current state of two-dimensional 42. Yao X, Freas A, Ramirez J, Demirev PA, Fenselau C. Proteolytic
electrophoresis with immobilized pH gradients: Proteomics and 18O labeling for comparative proteomics: model studies with two
2-DE. Electrophoresis. 2000;21:1037–1053. serotypes of adenovirus. Analytical Chemistry. 2001 Jul;73(13):
24. Unlü M, Morgan ME, Minden JS. Difference gel electrophore- 2836–2842.
sis: a single gel method for detecting changes in protein extracts. 43. Zhang Z, Li M, Zhang G, et al. Identification of human gastric car-
Electrophoresis. 1997;18(11):2071–2077. cinoma biomarkers by differential protein expression analysis using
25. Meyfour A, Tavirani MR, Sadeghi MR. Common proteomic tech- 18O labeling and nanoLC-MS/MS coupled with laser capture
nologies, applications, and their limitations. Journal of Paramedical microdissection. Medical Oncology (Northwood, London, England).
Sciences (JPS). 2013;4:115–125. 2010 Jun;27(2):296–303.
26. Görg A, Drews O, Lück C, Weiland F, Weiss W. 2-DE with IPGs. 44. Qian W-J, Monroe ME, Liu T, et al. Quantitative proteome analysis
Electrophoresis. 2009 Jun;30 Suppl 1:S122–S132. of human plasma following in vivo lipopolysaccharide administra-
27. Rogowska-Wrzesinska A, Le Bihan M-C, Thaysen-Andersen M, tion using 16O/18O labeling and the accurate mass and time tag
Roepstorff P. 2D gels still have a niche in proteomics. J Proteomics approach. Mol Cell Proteomics: MCP. 2005 May;4(5):700–709.
(Internet). 2013 Jan 24 [cited Mar 5, 2013];1–10. Available 45. Ye X, Luke B, Andresson T, Blonder J. 18O stable isotope label-
from: http://www.ncbi.nlm.nih.gov/pubmed/23353020 ing in MS-based proteomics. Briefings in Functional Genomics &
28. Shevchenko A, Tomas H, Havlis J, Olsen JV, Mann M. In-gel Proteomics. 2009 Mar;8(2):136–144.
digestion for mass spectrometric characterization of proteins 46. Aggarwal K, Choe LH, Lee KH. Shotgun proteomics using the
and proteomes. Nature Protocols (Internet). 2006 Jan [cited Mar iTRAQ isobaric tags. Briefings in Functional Genomics & Proteomics.
5, 2012];1(6):2856–2860. Available from: http://www.ncbi.nlm. 2006 Jun;5(2):112–120.
nih.gov/pubmed/17406544 47. Smolka MB, Zhou H, Purkayastha S, Aebersold R. Optimization
29. Lim H, Eng J, Yates JR, et al. Identification of 2D-gel proteins: a com- of the isotope-coded affinity tag-labeling procedure for quantitative
parison of MALDI/TOF peptide mass mapping to μ LC-ESI tandem proteome analysis. Analytical Biochemistry. 2001 Oct;297(1):25–31.
mass spectrometry. J Am Soc Mass Spectrom. 2003 Sep;14(9):957–970. 48. Shiio Y, Aebersold R. Quantitative proteome analysis using
30. Bantscheff M, Lemeer S, Savitski MM, Kuster B. Quantitative mass isotope-coded affinity tags and mass spectrometry. Nature Protocols.
spectrometry in proteomics: critical review update from 2007 to the 2006 Jan;1(1):139–145.
present. Anal Bioanal Chem. 2012 Sep;404(4):939–965. 49. Shiio Y, Donohoe S, Yi EC, Goodlett DR, Aebersold R, Eisenman
31. Yates JR, Ruse CI, Nakorchevsky A. Proteomics by mass spectrom- RN. Quantitative proteomic analysis of Myc oncoprotein function.
etry: approaches, advances, and applications. Annu Rev Biomed Eng The EMBO Journal. 2002 Oct;21(19):5088–5096.
(Internet). 2009 Jan [cited Mar 9, 2012];11(c):49–79. Available 50. Tam EM, Morrison CJ, Wu YI, Stack MS, Overall CM. Membrane
from: http://www.ncbi.nlm.nih.gov/pubmed/19400705 protease proteomics: isotope-coded affinity tag MS identification of
32. Brusniak M-YK, Chu CS, Kusebauch U, Sartain MJ, Watts JD, undescribed MT1-matrix metalloproteinase substrates. Proceedings
Moritz RL. An assessment of current bioinformatic solutions for of the National Academy of Sciences of the United States of America.
analyzing LC-MS data acquired by selected reaction monitoring 2004 May;101(18):6917–6922.
technology. Proteomics. 2012 Apr;12(8):1176–1184. 51. Ramus C, Gonzalez de Peredo A, Dahout C, Gallagher M, Garin J.
33. Reker D, Malmström L. Bioinformatic challenges in targeted pro- An optimized strategy for ICAT quantification of membrane pro-
teomics. J Proteome Res. 2012 Sep;11(9):4393–4402. teins. Mol Cell Proteomics: MCP. 2006 Jan;5(1):68–78.
34. Gonzalez-Galarza FF, Lawless C, Hubbard SJ, et al. A critical 52. Mann M. Functional and quantitative proteomics using
appraisal of techniques, software packages, and standards for quan- SILAC. Nature Reviews. Molecular Cell Biology (Internet). 2006
titative proteomic analysis. OMICS. 2012 Sep;16(9):431–442. Dec;7(12):952–958. Available from: http://www.ncbi.nlm.nih.
35. Michalski A, Damoc E, Hauschild J-P, et al. Mass spectrometry-based gov/pubmed/17139335
proteomics using Q Exactive, a high-performance benchtop quad- 53. Everley P, Krijgsveld J, Zetter B. Quantitative cancer proteomics: sta-
rupole Orbitrap mass spectrometer. Mol Cell Proteomics: MCP ble isotope labeling with amino acids in cell culture (SILAC) as
(Internet). 2011 Sep;10(9):M111.011015. Available from: http:// a tool for prostate cancer research. Cellular Proteomics (Internet).
www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3284220&t 2004 Jul [cited Jun 21, 2012];3(7):729–35. Available from: http://
ool=pmcentrez&rendertype=abstract www.mcponline.org/content/3/7/729.short
36. Nagaraj N, Wisniewski JR, Geiger T, et al. Deep proteome and 54. Moser M, Ussar S, Thievessen I. SILAC mouse for quantitative pro-
transcriptome mapping of a human cancer cell line. Mol Syst Biol. teomics uncovers kindlin-3 as an essential factor for red blood cell
(Internet). 2011 Jan [cited Feb. 29, 2012];7(548):548. Available function. Cell (Internet). 2008 Jul [cited Jun 21, 2012];134(2):353–
from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid 364. Available from: http://www.sciencedirect.com/science/
=3261714&tool=pmcentrez&rendertype=abstract article/pii/S0092867408006958
37. Neilson KA, Ali NA, Muralidharan S, et al. Less label, more 55. Westman-Brinkmalm A, Abramsson A, Pannee J, et al. SILAC
free: approaches in label-free quantitative mass spectrometry. zebrafish for quantitative analysis of protein turnover and tissue
Proteomics. 2011 Jan;11(4):535–533. regeneration. J Proteomics. 2011 Dec;75(2):425–434.
38. Elliott MH, Smith DS, Parker CE, Borchers C. Current trends in 56. America AH, Cordewener JH. Comparative LC-MS: a landscape of
quantitative proteomics. J Mass Spectrom. 2009;44(12):1637–1660. peaks and valleys. Proteomics. 2008;8(4):731–749.
39. Ong S-E. Stable isotope labeling by amino acids in cell culture, 57. Mueller LN, Brusniak MY, Mani DR, Aebersold R. An assessment
SILAC, as a simple and accurate approach to expression proteomics. of software solutions for the analysis of mass spectrometry based
Mol Cell Proteomics. 2002 May;1(5):376–386. quantitative proteomics. Data Reviews. 2008;51–61.
40. Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH, Aebersold 58. Hoekman B, Breitling R, Suits F, Bischoff R, Horvatovich P.
R. Quantitative analysis of complex protein mixtures using msCompare: a framework for quantitative analysis of label-free
isotope-coded affinity tags. Nature Biotechnology. 1999 Oct;17(10): LC-MS data for comparative biomarker studies. Mol Cell Proteomics.
994–999. 2012;M111.015974.
41. Ross PL, Huang YN, Marchese JN, et al. Multiplexed protein quan- 59. Gillette MA, Carr SA. Quantitative analysis of peptides and pro-
titation in Saccharomyces cerevisiae using amine-reactive isobaric teins in biomedicine by targeted mass spectrometry. Nature Methods
tagging reagents. Mol Cell Proteomics: MCP. 2004 Dec;3(12): (Internet). 2013 Jan [cited Mar 1, 2013];10(1):28–34. Available
1154–1169. from: http://www.ncbi.nlm.nih.gov/pubmed/23269374
3 6 • P rincip l e s o f G e no m ic M e dicin e
60. Domanski D, Percy AJ, Yang J, et al. MRM-based multiplexed regulator intersectin 2 through a selective SH3 domain interaction †.
quantitation of 67 putative cardiovascular disease biomarkers 2007;3:9874–9885.
in human plasma. Proteomics (Internet). 2012 Apr [cited May 80. Rual J-F, Venkatesan K, Hao T, et al. Towards a proteome-scale map
14, 2012];12(8):1222–1243. Available from: http://www.ncbi. of the human protein–protein interaction network. Nature. 2005
nlm.nih.gov/pubmed/22577024 Oct;437(7062):1173–1178.
61. Da Costa LA, García-Bailo B, Borchers CH, Badawi A, El-Sohemy 81. Mueller C, Liotta LA, Espina V. Reverse phase protein microarrays
A. Association between the plasma proteome and serum ascor- advance to use in clinical trials. Mol Oncol. 2010 Dec;4(6):461–481.
bic acid concentrations in humans. The Journal of Nutritional 82. Yu X, Schneiderhan-Marra N, Joos TO. Protein microarrays for per-
Biochemistry. 2012 24(5):842–847. sonalized medicine. Clin Chem. 2010 Mar;56(3):376–387.
62. Garcı B, Brenner DR, Nielsen D, et al. Dietary patterns and eth- 83. Sanchez-Carbayo M, Socci ND, Lozano JJ, Haab BB, Cordon-Cardo
nicity are associated with distinct plasma proteomic groups 1–3. C. Profiling bladder cancer using targeted antibody arrays. Am J
2012;(3):352–361. Pathol. 2006 Jan;168(1):93–103.
63. Fortin T, Salvador A, Charrier JP, et al. Multiple reaction moni- 84. Carlsson A, Wingren C, Ingvarsson J, et al. Serum proteome profiling
toring cubed for protein quantification at the low nanogram/mil- of metastatic breast cancer using recombinant antibody microarrays.
liliter level in nondepleted human serum. Analytical Chemistry. Eur J Cancer (Oxford, England: 1990). 2008 Feb;44(3):472–480.
2009;81(22):9343–9352. 85. Abd El-Rehim DM, Ball G, Pinder SE, et al. High-throughput
64. Lemoine J, Fortin T, Salvador A, Jaffuel A, Charrier J-P,
protein expression analysis using tissue microarray technology
Choquet-Kastylevsky G. The current status of clinical proteomics of a large well-characterised series identifies biologically dis-
and the use of MRM and MRM(3) for biomarker validation. Expert tinct classes of breast cancer confirming recent cDNA expression
Rev Mol Diagn. 2012 May;12(4):333–342. analyses. Int J Cancer (Journal International du Cancer). 2005
65. Shi T, Su D, Liu T, et al. Advancing the sensitivity of selected reac- Oct;116(3):340–350.
tion monitoring-based targeted quantitative proteomics. Proteomics. 86. Orchekowski R, Hamelinck D, Li L, et al. Antibody microarray
2012 Apr;12(8):1074–1092. profiling reveals individual and combined serum proteins associated
66. Anderson NL, Anderson NG. The human plasma proteome: his- with pancreatic cancer. Cancer Res. 2005 Dec;65(23):11193–11202.
tory, character, and diagnostic prospects. Mol Cell Proteomics. 87. Chung GG, Yoon HH, Zerkowski MP, et al. Vascular endothelial
2002;1(11):845–867. growth factor, FLT-1, and FLK-1 analysis in a pancreatic cancer tis-
67. Anderson NL, Anderson NG, Pearson TW, et al. A human proteome sue microarray. Cancer. 2006 Apr;106(8):1677–1684.
detection and quantitation project. Mol Cell Proteomics: MCP. 88. Pirnia F, Pawlak M, Thallinger GG, et al. Novel functional profiling
2009 May;8(5):883–886. approach combining reverse phase protein microarrays and human
68. Bell AW, Deutsch EW, Au CE, et al. A HUPO test sample study 3-D ex vivo tissue cultures: expression of apoptosis-related proteins
reveals common problems in mass spectrometry-based proteomics. in human colon cancer. Proteomics. 2009 Jul;9(13):3535–3548.
Nature Methods. 2009;6(6):423–430. 89. Spurrier B, Ramalingam S, Nishizuka S. Reverse-phase protein
69. Cox J, Mann M. MaxQuant enables high peptide identification rates, lysate microarrays for cell signaling analysis. Nature Protocols. 2008
individualized p.p.b.-range mass accuracies and proteome-wide pro- Jan;3(11):1796–1808.
tein quantification. Nature Biotech (Internet). 2008 Dec [cited Feb 90. Rapkiewicz A, Espina V, Zujewski JA, et al. The needle in the
28, 2013];26(12):1367–1372. Available from: http://www.ncbi. haystack: application of breast fine-needle aspirate samples
nlm.nih.gov/pubmed/19029910 to quantitative protein microarray technology. Cancer. 2007
70. Mueller LN, Brusniak MY, Mani DR, Aebersold R. An assessment Jun;111(3):173–184.
of software solutions for the analysis of mass spectrometry based 91. Saviranta P, Okon R, Brinker A, Warashina M, Eppinger J,
quantitative proteomics data. J Proteome Res. 2008;7(01):51–61. Geierstanger BH. Evaluating sandwich immunoassays in microar-
71. Mancuso F, Bunkenborg J, Wierer M, Molina H. Data extraction from ray format in terms of the ambient analyte regime. Clin Chem. 2004
proteomics raw data: An evaluation of nine tandem MS tools using Oct;50(10):1907–1920.
a large Orbitrap data set. J Proteomics. 2012;75(17):5293–5303. 92. Choudhary C, Mann M. Decoding signalling networks by
72. Cham Mead JA, Bianco L, Bessant C. Free computational resources mass spectrometry-based proteomics. Nature Rev. Mol Cell Biol
for designing selected reaction monitoring transitions. Proteomics. (Internet). 2010 Jun [cited Mar 1, 2012];11(6):427–439. Available
2010 Mar;10(6):1106–1126. from: http://www.ncbi.nlm.nih.gov/pubmed/20461098
73. Martin DB, Holzman T, May D, et al. MRMer, an interactive open 93. Engholm-Keller K, Larsen MR. Technologies and challenges in
source and cross-platform system for data extraction and visual- large-scale phosphoproteomics. Proteomics (Internet). 2013 Feb 13
ization of multiple reaction monitoring experiments. Mol Cell [cited Mar 2, 2013];910–931. Available from: http://www.ncbi.
Proteomics: MCP. 2008 Nov;7(11):2270–2278. nlm.nih.gov/pubmed/23404676
74. Maclean B, Tomazela DM, Shulman N, et al. Skyline: an open 94. Sylvestersen KB, Young C, Nielsen ML. Advances in characterizing
source document editor for creating and analyzing targeted pro- ubiquitylation sites by mass spectrometry. Curr Opin Chem Biol
teomics experiments. Bioinformatics. 2010;26(7):966–968. (Internet). 2013 Feb [cited Mar 18, 2013];17(1):49–58. Available
75. Berrade L, Garcia AE, Camarero JA. Protein microarrays: novel from: http://www.ncbi.nlm.nih.gov/pubmed/23298953
developments and applications. Pharmaceut Res. 2011 95. Guo M, Huang BX. Integration of phosphoproteomic, chemical,
Jul;28(7):1480–1499. and biological strategies for the functional analysis of targeted pro-
76. Nand A, Gautam A, Pérez JB, Merino A, Zhu J. Emerging technol- tein phosphorylation. Proteomics (Internet). 2013 Feb [cited Mar
ogy of in situ cell free expression protein microarrays. Protein & Cell. 7, 2013];13(3–4):424–437. Available from: http://www.ncbi.nlm.
2012 Feb;3(2):84–88. nih.gov/pubmed/23125184
77. Wolf-Yadlin A, Sevecka M, MacBeath G. Dissecting protein func- 96. Mann M, Jensen ON. Proteomic analysis of post-translational mod-
tion and signaling using protein microarrays. Curr Opin Chem Biol. ifications. Nature Biotech (Internet). 2003 Mar;21(3):255–261.
2009 Oct;13(4):398–405. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12610572
78. Hesselberth JR, Miller JP, Golob A, Stajich JE, Michaud GA, Fields 97. Fedjaev M, Parmar A, Xu Y, et al. Global analysis of protein phos-
S. Comparative analysis of Saccharomyces cerevisiae WW domains phorylation networks in insulin signaling by sequential enrichment
and their interacting proteins. Genome Biol. 2006 Jan;7(4):R30. of phosphoproteins and phosphopeptides. Mol Biosyst (Internet).
79. Lim CS, Seet BT, Ingham RJ, Gish G, Matskova L. The K15 protein 2012 Apr [cited Apr 26, 2013];8(5):1461–1471. Available
of Kaposi’s sarcoma-associated herpes virus recruits the endocytic from: http://www.ncbi.nlm.nih.gov/pubmed/22362066
H u m an P rot e o m ic s • 3 7
98. Zhou H, Ye M, Dong J, et al. Robust phosphoproteome enrichment 104. Speers AE, Cravatt BF. Chemical strategies for activity-based
using monodisperse microsphere-based immobilized titanium (IV) proteomics. Chembiochem (Internet). 2004 Jan 3 [cited Mar 29,
ion affinity chromatography. Nature Protocols (Internet). 2013 Feb 2012];5(1):41–47. Available from: http://www.ncbi.nlm.nih.
7 [cited Mar 1, 2013];8(3):461–480. Available from: http://www. gov/pubmed/14695510
ncbi.nlm.nih.gov/pubmed/23391890 105. Zanivan S, Cascone I, Peyron C, et al. A new computational
99. Hitchen PG, Twigger K, Valiente E, Langdon RH, Wren BW, approach to analyze human protein complexes and predict novel
Dell A. Glycoproteomics: a powerful tool for characterizing protein interactions. Genome Biol (Internet). 2007 Jan [cited Aug
the diverse glycoforms of bacterial pilins and flagellins. Biochem 10, 2012];8(12):R256. Available from: http://www.pubmedcen-
Soc Transact (Internet). 2010 Oct [cited Apr 4, 2013];38(5): tral.nih.gov/articlerender.fcgi?artid=2246258&tool=pmcentrez
1307–1313. Available from: http://www.ncbi.nlm.nih.gov/ &rendertype=abstract
pubmed/20863304 106. Rajagopala SV, Sikorski P, Harry Caufield J, Tovchigrechko A,
100.
Lanucara F, Eyers CE. Top-down mass spectrometry for the Uetz P, Caufield JH. Studying protein complexes by the yeast
analysis of combinatorial post-translational modifications. Mass two-hybrid system. Methods (San Diego, Calif.) (Internet). 2012
Spectrom Rev. 2013 Jan–Feb;32(1):27–42. Available from http:// Jul 24 [cited Aug 2, 2012]; Available from: http://www.ncbi.nlm.
www.ncbi.nlm.nih.gov/pubmed/22718314 nih.gov/pubmed/22841565
101. Castellino S, Groseclose MR, Wagner D. MALDI imaging
107. Voshol H, Ehrat M, Traenkle J, Bertrand E, Van Oostrum J.
mass spectrometry: bridging biology and chemistry in drug Antibody-based proteomics: analysis of signaling networks using
development. Bioanalysis (Internet). 2011 Nov;3(21):2427– reverse protein arrays. FEBS J (Internet). 2009 Dec [cited Aug
2441. Available from: http://www.ncbi.nlm.nih.gov/pubmed/ 31, 2012];276(23):6871–6879. Available from: http://www.ncbi.
22074284 nlm.nih.gov/pubmed/19860827
102. Caprioli RM, Farmer TB, Gile J. Molecular imaging of bio- 108. Chen R, Snyder M. Systems biology: personalized medicine for
logical samples: localization of peptides and proteins using the future? Curr Opin Pharmacol (Internet). 2012 Jul 31 [cited
MALDI-TOF MS. Anal Chem (Internet). 1997 Dec 1;69(23): Oct 5, 2012]; Available from: http://www.ncbi.nlm.nih.gov/
4751–4760. Available from: http://www.ncbi.nlm.nih.gov/ pubmed/22858243
pubmed/9406525 109. Hood L, Flores M. A personal view on systems medicine and the
103. Cravatt BF, Wright AT, Kozarich JW. Activity-based protein pro- emergence of proactive P4 medicine: predictive, preventive, per-
filing: from enzyme chemistry to proteomic chemistry. Annu Rev sonalized and participatory. New Biotech (Internet). 2012 Oct
Biochem (Internet). 2008 Jan [cited Mar 9, 2012];77:383–414. 15 [cited Aug 31, 2012];29(6):613–624. Available from: http://
Available from: http://www.ncbi.nlm.nih.gov/pubmed/18366325 www.ncbi.nlm.nih.gov/pubmed/22450380
3 8 • P rincip l e s o f G e no m ic M e dicin e
4.
EPIGENETICS, EPIGENOMICS, AND HUMAN DISEASE
Aravind Ramesh, Cihangir Yandim, Theona Natisvili, Marta Mauri, Piu Pik Law,
Jackson P. K. Chan, Santiago Uribe Lewis, and Richard Festenstein*
39
DNA accessibility.27 It is very likely that these biochemical a human genome functions.54–59 In addition, genome-wide
processes are regulated by an “epigenetic code” written as analyses on identical twins have furthered our understand-
modifications in DNA and histones and “read” by factors ing of epigenetic differences.60,61
that specifically recognize single or combined modifica- Where genes reside in the nucleus also provides a mech-
tions.28–30 For example, histone acetylation, recognized by anism for epigenetic gene regulation. Genes may dynami-
proteins with bromodomains,31 as well as histone H3 lysine cally relocate to associate with nuclear structures rich in
4 methylation, recognized by WDR5 containing H3 lysine transcription factors required for expression,62 or be devel-
4 methyltransferase complexes,32 are generally associated opmentally inactivated by their relocation to heterochro-
with “open” chromatin and gene expression.14,33,34 On the matic pericentromeric clusters.63 In the phenomenon of
other hand, histone H3 lysine 9 or 27 methylation, recog- position effect variegation (PEV), genes are repressed in a
nized by the chromodomains of HP1 and polycomb pro- proportion of the cells when abnormally juxtaposed to het-
teins, respectively,35,36 and DNA methylation, recognized by erochromatin.64–67 That gene expression is affected by place-
methyl binding domain (MBD) proteins,37 associate with ment of genes near heterochromatin has focused attention
“condensed” chromatin and gene repression.38–40 Recently, on the associations between genes and their native cis-acting
it has also been uncovered that DNA can be hydroxy- regulatory elements. DNA sequences, such as locus control
methylated preferentially at cytosines of CpG residues regions68 or boundary elements/insulators,69 are thought to
via ten eleven translocation (TET) enzymes.41 The exact regulate gene expression by establishing “permissive” chro-
function of this mark has not been resolved yet; however, matin domains70 or by facilitating nuclear relocation and/
it is thought to be an intermediary step between unmeth- or regulating interactions between gene promoters and
ylated CpGs and methylated ones.42 In line with this, enhancer or silencer elements.69 Imprint control regions,
5-hydroxymethylcytosine residues were associated mainly typically found as DNA differentially methylated regions
with transcriptionally active euchromatin and polycomb (DMRs) in imprinted gene clusters, are conceived as sites
silencing of formerly euchromatic promoters, where it pos- from which chromatin structures are propagated bidirec-
sibly allows a rapid release of the repressed state in poised tionally to control the expression of genes within imprinted
genes.42,43 domains.71 Notably, recent advances now allow scientists to
Methylated DNA, through MBDs, may target histone study the three-dimensional organization of the genome
deacetylases (HDACs)44,45 and H3 lysine 9 methyltransfer- via chromosome-conformation capture (3C) approaches,
ases (e.g., Suv39h) to specific loci.46 Conversely, histone H3 which are based on the restriction enzyme digestion of
lysine 9 methylation may bring about DNA methylation crosslinked DNA and its subsequent ligation followed by
(by association with DNA de novo methyltransferases)47 as PCR.72 Within the human ENCODE project, a global
well as HP1.48,49 Therefore, both DNA and histone modi- long-range interaction map of gene promoters was revealed
fication machineries may cross-talk to generate condensed thanks to the fusion of chromosome confirmation-capture
chromatin (see Figure 4.1a and 4.1b). DNA methylation technique with high-throughput sequencing.73
and H3 lysine 9 methylation may, however, function inde- Disease states with an epigenetic basis are classified into
pendently49,50 and thus confer different degrees of epigen- those where changes in chromatin structure at the deregu-
etic plasticity. DNA methylation, thought to be more stable lated gene(s) result from mutations in DNA sequences in
throughout meiosis and mitosis, may provide long-term the same chromosome; i.e., in cis, and those where genetic
(repressive) epigenetic memory, whereas histone modifica- mutations affect the genes that encode for factors that estab-
tions are more labile epigenetic marks. lish or maintain chromatin structures; i.e., in trans. Table 4.1
Many aspects of chromatin explained so far have been illustrates disease states with a confirmed or possible epi-
evaluated using the chromatin immunoprecipitation genetic basis, some of which we describe in greater detail
(ChIP) assay, which involves the immunoprecipitation of in the text. Importantly, epigenetic mechanisms appear to
protein-bound DNA using specific antibodies, followed play a key role in the development of numerous different
by PCR.51,52 Until recently, results were only obtained in types of cancer, and given their potential reversible nature,
limited contexts. Importantly, though, development of they offer exciting therapeutic targets. It has been also sug-
modern genomics techniques allowed high-throughput gested that epigenetic mechanisms might differ between
sequencing of the genomic ChIP-DNA and thereby global genders based on sex chromosome complements and the
examination of chromatin.53 In line with this, the human male-determining gene “SRY”.74,75 These may explain dif-
ENCODE project aimed to reveal a detailed picture (e.g., ferent predisposition rates towards disease among different
histone modifications, transcription factor binding) of how genders.
4 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
(A) (B)
Suv39h Suv39h Suv39h
RNAPII
Triplex
MSH2/MSH3
(GAA)10–66
RNAPII
DNA methylation
Histone deacetylation
RNAPII
H3K4 demethylation
H3K9me2/me3
H3K27me3
CDKN1C
PHLDA2
NAP1L4
PHEMX
KCNQ1
ASCL2
TSSC4
CD81
IGF2
H19
L23
INS
TH
CTCF Enhancers
Cen. Tel.
ICR2 ICR1
(DMR-LIT1/KvDMR1) (H19DMR)
Figure 4.1
Chromatin and disease. (A) Schematic representation of chromatin condensation. DNA (thick black line) wraps around histone octamers
(gray oval) to form nucleosomes. Methylated (CH3) DNA is recognized by DNA MBD proteins that complex with histone deacetylases (HDAC)
and histone H3 methyltransferases (e.g., Suv39h) to deacetylate histones, and specifically methylate lysine 9 of histone H3. A methylated lysine 9 of
H3 is specifically recognized by heterochromatin protein 1 (HP1), which interacts with Suv39h and normally exists as a dimer. (B) HP1 dimers may
link nucleosomes methylated at lysine 9 of H3 by Suv39h and facilitate chromatin compaction. (C) Potential mechanism underlying Friedreich’s
ataxia. In the presence of a normal GAA repeat in intron 1 of the frataxin gene, the region around the repeat and the nearby promoter may lie
in loosely packaged chromatin. Therefore, the gene is actively expressed. (D) In the presence of an expanded GAA repeat within intron1ofthe
frataxin gene,densely packagedheterochromatin mayform. This may induce similar changes at the nearby promoter (arrow). Variegated expression
of frataxin may then occur. The proportion of frataxin-positive cells may depend on GAA repeat length and the degree of heterochromatinization.
In transgenic mice, an untranslated GAA repeat expansion linked to a hCD2 reporter gene was associated with variegated expression and
heterochromatin formation close to the repeat and at the more distant promoter (see text). (E) Schematic representation of the imprinted domains
1 and 2 in chromosome 11p15.5 associated with the Beckwith–Wiedemann syndrome. Domain 1 contains insulin (INS), insulin-like growth factor
2 (IGF2), and H19. A differentially methylated region (DMR) upstream of H19 (H19 DMR) is the imprint control region 1 (ICR1) that, when
bound by CTCF on the unmethylated maternal chromosome (open lollipop), acts as a chromatin boundary which prevents an IGF2-downstream
enhancers (diamonds) interaction. Domain 2 contains the KCNQ1OT1 gene expressed from the paternal allele, and the PHLDA2, SLC22A1L,
CDKN1C, KCNQ1DN, and the KCNQ1 genes expressed from the maternal allele. A DNA DMR at the 50 end of KCNQ1OT1 (DMR-LIT1 or
KvDMR1; methylated on the maternal allele 002D filled lollipop) is the imprint control region for domain 2 required for repression of genes in the
paternal chromosome and expression of those in the maternal chromosome. Arrows indicate transcriptional orientation. Genes with two arrows
represent non-imprinted genes (modified from Robertson, 2005).
www.Ebook777.com
Table 4.1 GENETIC MUTATIONS GENERATING “EPIGENETIC” DISEASE IN CIS OR TRANS
4 2 • P rincip l e s o f G e no m ic M e dicin e
and somatic mosaicism may play a role in such phenotypi- silencing was also modified by altering the dosage of HP1β,
cal variation. an important component of heterochromatin.99
In recent years, it has been shown that a similar het-
erochromatinization also takes place at the pathologically
FRATA X IN G E N E R E P R E S S I O N
silenced frataxin gene (Figure 4.1c and d). DNA methylation
I N FR DA
levels in lymphoblastoid cells, peripheral blood mononuclear
Abnormally expanded GAA repeats within the frataxin cells (PBMC),100–102 as well as nerve tissues in FRDA mouse
gene impair frataxin expression. This impairment could be models103 were found to be higher on the frataxin gene in
a direct result of physical blockage effects caused by unusual the presence of expanded GAA repeats. This result was also
conformations of DNA adopted by GAA triplets on the accompanied by an increase in heterochromatic histone marks
elongation of transcription (Figure 4.1c and d). The most (i.e., H3K9 di- and tri-methylation and H3K27 trimethyl-
common such structure is hairpin DNA resulting from ation) and an overall decrease in histone acetylation in the
unusual hydrogen bonding between G•G, G•A and A•A, flanking regions of expanded GAA repeats in patient-derived
and RRY (purine:purine:pyrimidine) triplexes, which is lymphoblastoid cells/PBMCs,96–98,102,104 fibroblasts,105 and
also known as “sticky” DNA.88 Consistent with the idea of FRDA mouse models.103 Indeed the pattern of heterochro-
transcriptional blockage effects, in vitro transcription assays matic marks on pathologically silenced alleles supports the
based on RNAse protection and northern blots revealed a hypothesis that heterochromatin is spreading bidirectionally
transcriptional elongation defect in the presence of expanded from the GAA repeats causing the silencing of the gene.88
GAA repeats.89–95 Moreover, ChIP assays performed on How exactly heterochromatin is triggered in FRDA
lymphoblastoid cell lines using antibodies against initiating is an important question that remains to be answered.
and elongating forms of RNAPII, as well as histone marks Interestingly, unusual DNA conformations adopted by
associated with transcriptional elongation (i.e., H3K36 expanded GAA repeats were shown to be recognized by the
and H3K79 methylation) also underlined a deficit in tran- cell’s mismatch repair mechanism. MSH2/MSH3 dimers
scriptional elongation on the frataxin gene.96–98 Whether were shown to be attracted by expanded GAA repeats
expanded GAA triplets affect the initiation of RNAPII at (Figure 4.1d) in various FRDA models, including iPS cells
the frataxin gene promoter, however, remains elusive. derived from patients.106–110 Some studies suggest hetero-
Importantly, expanded GAA repeats are also associ- chromatinization as a protective response against faulty
ated with the heterochromatinization of the frataxin gene. transcription, which may be caused by DNA damage.111–116
The first in vivo experiments to suggest that GAA-repeat Moreover, De Biase et al.105 reported increased antisense
expansions could trigger heterochromatin formation were transcription in pathological frataxin alleles and suggest
performed in transgenic mice. Here, a (GAA)200 repeat this as a potential trigger for the heterochromatinization.
expansion was linked to a human CD2 (hCD2) reporter. However, it is still unclear whether antisense transcription
The direct inhibitory transcriptional effect of GAA repeats may lead to heterochromatic silencing in mammals.117 De
on DNA transcription was excluded as the GAA repeat was Biase et al.105 also described a CTCF binding site near the
linked to the 3′ untranscribed region of the hCD2 trans- promoter of frataxin. CTCF is a chromatin insulator pro-
gene. In this transgenic mouse model, the hCD2 reporter tein associated with enhancer blocking or chromatin insu-
gene alone is sensitive to juxtaposition to constitutively lation activities.118,119 The effect of CTCF on frataxin has
tightly packaged DNA (heterochromatin), e.g., centro- not been clearly characterized yet; however, the study of De
meres, and results in variegated hCD2 expression on T cells, Biase et al. reports a depletion of CTCF binding on silenced
or PEV.66 In PEV, rather than gene expression being silenced alleles. Interestingly, knockdown of CTCF in healthy and
in all cells, a proportion of cells become silenced with the patient fibroblasts resulted in increased antisense tran-
remaining continuing to express. Linking a GAA repeat scription in the promoter of the gene. This may imply that
expansion to the hCD2 transgene also resulted in PEV99 but CTCF has an inhibitory function against the spreading of
importantly this occurred even when the transgene was sit- heterochromatin nucleated by GAA repeats.
uated in regions of the chromosome that are usually loosely One of the challenges in the FRDA field is limited access
packaged in euchromatin, suggesting that the presence of to primary nerve tissue, which is predominantly affected in
GAA repeats induced chromatin condensation and hetero- this disease. Most of the studies so far have presented results
chromatin formation. In T cells where hCD2 was silenced, obtained from Epstein Barr virus–transformed lymphoblas-
DNase I hypersensitive site analysis showed condensation toid cells derived from patients. Other sources of research were
of chromatin packaging at the promoter of the gene. This fibroblasts and primary peripheral blood mononuclear cells,
as well as mouse models, which only exhibit a mild disease repeat containing RNA transcripts cause sequestration
phenotype.120 Notably, the arrival of iPS technology in the of muscle blind protein (MBNL) and increase CUGBP/
recent years has allowed scientists to differentiate neuron-like Elav-like family member 1 (CELF1) protein activi-
cells using fibroblasts obtained from FRDA patients.110,121 ties,134–137 creating a spliceopathy, the CTG expansion has
A lot has been resolved about the pathological silenc- also been found to affect chromatin packaging of DNA.
ing of the frataxin gene in the last decade. Importantly, For example, in fibroblasts from myotonic dystrophy
uncovering different aspects of heterochromatin brought patients, the presence of a CTG expansion in the DMPK
up the possibility of treating FRDA with HDAC inhibi- gene is associated with condensation of chromatin, as indi-
tors, which could reduce histone deacetylation and thereby cated by nuclease resistance at the six5 enhancer present in
a subsequent methylation. Indeed, synthetically derived the 3′ region of the DMPK gene, rendering it inaccessible
HDAC inhibitor BML-210 and its derivatives were shown to transcription factors and causing downregulation of
to upregulate frataxin expression significantly in FRDA six5 expression.138,139 CTG repeats also efficiently recruit
cells.104,122–126 Moreover, a recent study from our group nucleosomes, the basic structural element of chroma-
showed that an HDAC class III (Sirtuin) inhibitor nicotin- tin.140,141 CTG expansions also behaved in a similar fashion
amide also upregulates frataxin in vitro, ex vivo and in vivo to pericentromeric heterochromatin, causing gene silenc-
(mouse model) (Chan et al., 2013).126a Clinical trials for ing and chromatin condensation in the hCD2 transgenic
BML-210 derivatives are currently ongoing. Importantly, a mouse model.99 Some features of myotonic dystrophy
recent phase IIa clinical trial with oral-dosing of high-dose (particularly cataract formation as seen in six5 knock-out
nicotinamide revealed significant up-regulation of the FXN mice142) may be secondary to deregulation of the six5 gene
gene bringing its expression level towards that found in located near the CTG repeat142–144; this deregulation may
asymptomatic carriers (Libri and Yandim et al. in press).126b be secondary to CTG repeat-induced chromatin conden-
Undoubtedly, identification of specific epigenetic modifiers sation. Evidence supporting this notion has emerged as
responsible for the silencing of frataxin will help scientists loss of CTCF binding to the regions flanking the CTG
to further develop such radical therapeutic strategies, which repeat expansion has been shown at the DM1 locus, which
specifically address the primary cause of this currently incur- might be associated with DNA hypermethylation of these
able disease rather than its symptoms. regions.145,146 Spreading of heterochromatin at the CTG
expanded allele was indicated by the enrichment of his-
tone H3 lysine 9 (H3-K9) methylation and heterochro-
MYOTO N I C DYS T RO P H Y
matin protein 1γ (HP1γ) recruitment where antisense
Myotonic dystrophy is an autosomal dominant disease that transcription of DMPK was activated. This might lead to
is the most common adult form of muscular dystrophy. a wider dysregulation of the mRNA and protein amount
It is a multisystem neuromuscular disorder characterized in DM1-affected cells.145 The exact mechanisms of CTG
by clinical manifestations, including myotonia (skeletal repeat-mediated chromatin remodeling are still uncertain;
muscle hyperexcitability), progressive muscular dystrophy, however, the repeat itself strongly stimulates nucleosome
cataracts, cardiac conduction defects, cognitive deficits, and formation.141 In addition, it has been suggested that DNA
endocrine anomalies. binding proteins that recognize DNA triplet repeats147–149
Two forms of the disease with similar features are caused contribute to the epigenetic changes seen at the DMPK
by different microsatellite expansions in two different gene locus.
loci. Myotonic dystrophy type 1 (DM1) is caused by a CTG Other diseases where DNA repeats are found in
repeat expansion located in the 3′ untranslated region untranslated regions, like spinocerebellar ataxia type 10150
(UTR) of the DMPK gene on chromosome 19q127–129 and myotonic dystrophy type 2,151 may share similar “epi-
while myotonic dystrophy type 2 (DM2) is caused by genetic” molecular pathogenic mechanisms.
expansion of a CCTG repeat in the intron 1 of the Zinc
finger 9 (ZNF9) gene on chromosome 3q21.130 Expanded
FR AG I L E X SY N D RO M E
CTG or CCTG repeats are highly unstable in both germ-
line and somatic tissues,130,131 and the length of the repeats Fragile X syndrome (FXS) is the most common inher-
is correlated with the severity of symptoms and the earlier ited form of mental retardation152 and is one of the
disease onset.132,133 best-characterized forms of autism spectrum disorder
Although DM1 is now thought to be largely mediated (ASD).153 It is an X-linked dominant disorder characterized
by an “RNA gain of function mechanism” in which CUG by variable penetrance. The name “fragile X” derives from
4 4 • P rincip l e s o f G e no m ic M e dicin e
www.Ebook777.com
the observation that the cytogenetic band Xq27.3, where A (TSA, an HDAC inhibitor) also resulted in moder-
the causative fragile X mental retardation 1 (FMR1) gene ate transcriptional activation of FMR1, suggesting that
resides, is a fragile site in individuals carrying the full muta- in fragile X, histones are deacetylated at the inactive
tion (FM).154 promoter. Furthermore, treatment of fragile X cells with
As FMR1 is located on the X chromosome, the degree HDAC inhibitors and 5-azadC synergistically activated
of cognitive disability is more severe in males, who possess transcription.170
only one X chromosome. Females (with two X chromo- To date, the exact mechanisms behind CGG expansion
somes) manifest a less severe phenotype than males, which and the consequent alteration in FMR1 transcription and
is correlated to the extent of X inactivation on the abnormal translation are still not fully understood. The generation of
chromosome.155 FXS has an estimated frequency of 1/5,000 stable cell lines harboring the FMR1 5′-UTR with varying
in males and 1/10,000 in females.156 CGG repeat lengths targeted to the correct gene locus have
The main disease manifestations are moderate to severe proven to be a useful model for studying FXS. The promoter
intellectual disability, autistic features, seizures, hypersen- with variable (CGG)n length has been fused to the coding
sitivity to sensory stimuli, attention deficit, hyperactivity, sequence of a reporter gene. This construct has shown that
motor incoordination, growth abnormalities, sleep distur- a full-mutation CGG repeat length inhibits reporter gene
bances, connective tissue dysplasia, craniofacial abnormali- expression, whilst a premutation CGG repeat does not
ties, and macroorchidism.157 affect reporter gene expression. Therefore this model could
The syndrome is a trinucleotide repeat disorder caused be a better tool to elucidate the molecular mechanisms of
by the expansion of the triplet CGG in the 5′ untranslated FMR1 deregulation in FXS.171
region (5′UTR) of the FMR1 gene.158 A CGG expansion Transcriptional silencing of the FMR1 gene due to
greater than 200 units results in hypermethylation at CpG hypermethylation of CpG islands and the consequent loss
sites at the FMR1 promoter region; this is responsible of FMRP expression is still considered to be the major
for the silencing of FMR1 and the subsequent loss of the cause of the disease.154 However, a mouse model of FXS
protein it codes for, the fragile X mental retardation pro- with mice carrying long CGG repeats of nearly 230 units
tein (FMRP).154 FMRP is an RNA binding protein that is showed high levels of the FMR1 mRNA, although low
able to bind to several mRNAs, including its own,159 and is levels of the FMRP protein, arguing against a purely tran-
believed to be involved in the transportation of these target scriptional deficit. Promoters in these mice do not show
mRNAs throughout neuronal dendrites and in the inhibi- the abnormal methylation described above, which suggests
tion of their translation upon stimulation of the metabo- that modeling FXS in mice requires more genetic manipu-
tropic glutamate receptor 5 (mGluR5) at the synapse.160 It is lation, and that perhaps the threshold number of repeats
ubiquitously expressed until day 14 of embryonic develop- in mice might be higher than the level observed in human
ment, after which its expression becomes restricted to the subjects.172
brain (specifically in neurons) and the gonads.161 Loss of In addition, within humans there are cases of males
FMRP is thought to be critical due to the important role it expressing FMR1 to some extent despite carrying the full
plays in neuronal function. Abnormal dendritic spines were mutation. Mosaicism of the gene promoter methylation
observed in both FXS patients and in FMR1 knockout pattern is thought to underlie this, allowing some transcrip-
mice, supporting FMRP involvement in synaptic matura- tion to occur. The presence of three types of mosaicism was
tion.162 More generally, the absence of FMRP appears to lead tested in cells derived from male expressing patients, and
to a global increase in brain protein synthesis,163 and several the data suggests that inter-cell mosaicism in DNA methyl-
studies in human patients are currently testing whether this ation patterns might explain the presence of FMR1 mRNA
observation could be a biomarker of the disease. in some FXS=affected individuals.173
If cells from FXS patients are treated with a DNA The CGG repeat region is unstable, and repeat length
methylation inhibitor (5-aza-2-deoxycitidine), the level can vary in unaffected individuals from 6–55 repeats. The
of CpG methylation decreases and FMR1 expression is instability of the region can result in an expansion of this
reactivated,164–168 suggesting that DNA methylation is the region upon maternal transmission to the next generation.
major factor causing FMR1 silencing, rather than the trip- A (CGG)n range between 55–200 is referred to as the
let expansion itself. DNA methylation has been to shown premutation (PM).174 It was previously thought that carri-
to also cause local histone deacetylation, creating another ers of the PM simply had a higher risk of developing FXS
mechanism of transcriptional silencing.169 Treatment of upon transmission of the PM allele to the next generation.
the same lymphoblastoid fragile X cells with trichostatin However, it has become evident that certain carriers show
4 6 • P rincip l e s o f G e no m ic M e dicin e
initially that, in healthy patients, D4Z4 repeat tracts are Early work suggested that FRG1, FRG2, and ANT1 seemed
heterochromatinized, and that variable spreading of het- to be upregulated in muscle cells from FSHD patients in a
erochromatin silences nearby genes. Loss of these repeats manner specific to FSHD (not found in other hereditary
may produce a more open chromatin configuration, result- myopathies) and also to muscle (not replicated in patient
ing in inappropriate de-repression of these nearby genes.204 lymphocytes).210 Subsequent studies failed to fully corrobo-
However, early ChIP analysis looking at H4 acetylation rate these findings, being unable to replicate any increase
of the FSHD locus in human lymphoid cells as well as in FRG1 or ANT1 expression205,211–213 and showing that
human-rodent somatic cell hybrids suggested that regions FRG2 was only upregulated in FSHD myoblasts and not
close to the D4Z4 repeats and of two nearby genes (FRG1 in mature myocytes.214 In fact, studies in certain FSHD
and ANT1) showed histones acetylated more in the pat- families have largely ruled out these genes as causative
tern of euchromatin than heterochromatin.205 Subsequent for the disease. In one family, the disease-associated allele
ChIP analysis has looked at other histone modifica- showed a large deletion in the D4Z4 repeat array, includ-
tions: both repressive, such as trimethylation at lysine 9 ing both the DUX4c and FRG2 genes, arguing against a role
and 27 on histone 3 (H3K9me3 and H3K27me3), as well for them in causing disease.215 Similarly, in another family,
as markers of transcriptional activation such as dimeth- the pathological allele was actually found on chromosome
ylation at lysine 4 on histone 3 (H3K4me2). This has in 10q26, where the short repeat array had been extended by a
fact shown that, in healthy controls, D4Z4 arrays display D4Z4 fragment translocated from 4q35. This translocated
both heterochromatic and euchromatic regions; while fragment included part of the distal D4Z4 fragment, allow-
in FSHD patients, there are significantly reduced levels ing stable DUX4 mRNA expression, but did not include
of H3K9me3, with unaltered levels of H3K27me3 and FRG1, ANT1 or DUX4c, again arguing against a role for
H3K4me2,206 indicating a relative reduction in repressive these genes,202 and suggesting that DUX4 is the key gene
modifications, a more open chromatin configuration as a involved in FSHD.
consequence, and perhaps gene upregulation. Early work looking at DUX4 was greatly hampered by
Along with histone modifications, DNA meth- the difficulty of detecting specific mRNA transcripts. This,
ylation patterns also reflect chromatin configuration. coupled with a lack of an identifiable polyadenylation site,
Heterochromatic regions are usually hypermethylated, and led to the theory that DUX4 was actually a non-functional
normal D4Z4 arrays reflect this. Contracted D4Z4 arrays, pseudogene.216 The major breakthrough arose with the
however, display hypomethylation,207 and, rather akin to identification of a mature mRNA transcript containing the
phenotype, the level of this hypomethylation correlates DUX4 ORF, with RT-PCR.217 This transcript was shown
with repeat length: the shorter the array, the lower the level to originate from the distal D4Z4 unit and extend to an
of methylation.208 However, hypomethylation is also seen in adjacent region conferring the polyadenylated tail, termed
some unusual asymptomatic individuals who carry both the pLAM1.217 There is evidence that DUX4 expression at high
contracted D4Z4 array as well as a permissive haplotype, levels in muscle has numerous effects, including inhibi-
perhaps suggesting that hypomethylation is not sufficient tion of differentiation,218 and induction of genes involved
for disease onset.207 In addition, more profound levels of in muscle atrophy, apoptosis and cell death, implying that
hypomethylation are seen at D4Z4 repeats in the immu- DUX4 overexpression could be responsible for FSHD.219,220
nodeficiency, centromere instability, and facial anomalies However, DUX4 mRNA abundance is very low; it was
(ICF) syndrome,207 which is phenotypically entirely differ- not always detectable in FSHD muscle biopsies, and only
ent from FSHD. In the same syndrome, H3K9me3 enrich- around one per 1,000 FSHD myoblasts expressed DUX4
ment is normal in at D4Z4,206 suggesting that it is the loss of in culture.221 Snider et al. (2010) have shown that the low
H3K9me3 that is crucial in FSHD, rather than any change abundance in muscle actually reflects a small number of
in DNA methylation. nuclei expressing abundant amounts of DUX4, and that
The above indicates that transcriptional upregulation DUX4 is highly expressed in human testis and germline
is key to FSHD pathogenesis, and given the clearly dem- cells.221 This has led to a developmental model being pro-
onstrated association with the chromosome 4q35 region, posed for FSHD, whereby in normal individuals, DUX4 is
a number of candidate genes found in that area have been expressed in early development and is heterochromatically
investigated as potentially causative. These include FRG1, silenced in mature tissues. In FSHD, there is a defect in this
FRG2, ANT1, and, more recently, DUX4c and DUX4. silencing mechanism, which leads to occasional escape from
FRG1 in particular seemed a promising candidate, as func- repression in muscle cells, with consequent DUX4 expres-
tional studies suggested a role in muscle development.209 sion and cell death.221,222
4 8 • P rincip l e s o f G e no m ic M e dicin e
been reported, but exogenously tagged murine Dnmt3b heterochromatin and to relocate away from these domains
co-localizes with pericentric heterochromatin in some cell upon gene activation.250 Centromere distribution within
types.225 The domain necessary for targeting DNMT3B to the nucleus of lymphoid cells has been shown to vary with
heterochromatin has not been determined, but it is likely to different stages of differentiation, suggesting that the distri-
be at the aminoterminus where there are two domains com- bution of heterochromatin can influence gene expression
monly found in other chromatin-associated proteins. One in trans.251,252 The 3D organization of intranuclear pericen-
of these is a PHD finger similar to that found in ATRX, tric heterochromatin has been shown to be abnormal in
a chromatin remodeling protein that is concentrated at ICF patient lymphoblastoid B cells, at least for chromo-
sites of heterochromatin and repetitive DNA sequences.245 some 1, and treatment with a demethylating agent partially
There is also hypomethylation of specific repetitive DNA induces this heterochromatic remodeling.253,254 Further
sequences in ATR-X patients (but hypermethylation of work is needed to see if particular genes also show altered
other sequences). However, there is no chromosome insta- spatial distributions in ICF cells in parallel to heterochro-
bility reported at the sites of hypomethylation in ATR-X matic changes.
cells, and little phenotypical overlap between ATR-X and Alternatively, the loss of DNA methylation at large
ICF syndromes is observed (see ATR-X below). The other arrays of satellite repeats may release or recruit protein
domain that may be involved in targeting DNMT3B is a complexes and affect the balance of regulatory complexes
PWWP (conserved proline and tryptophan) domain, throughout the genome. Interestingly, ICF lymphoblastoid
which binds DNA246 and, based on similarity to tudor and cell lines showed altered binding patterns of HP1, with the
chromodomains,247 may also recognize methylated proteins. formation of large foci containing HP1 and components of
A homozygous missense mutation in the PWWP domain promyelocytic (PML) nuclear bodies that co-localize with
of DNMT3B has been identified in ICF sibs, resulting in chromosome 1qh and 16qh DNA sequences.255 The altered
a serine to proline change that may have a profound con- pattern was, however, only observed during the G2 phase
sequence for the mutant protein’s structure.228 In addition, of the cell cycle and not in fibroblasts. These results suggest
the PWWP domain of murine Dnmt3b has been shown to that binding of HP1 is not dependent on DNA methyla-
bind DNA non-specifically and to be required for target- tion (since the abnormal HP1 distribution only occurs at
ing DNA methyltransferase activity to murine pericentric one stage in the cell cycle) and also indicate that cell type–
sequences.226 specific defects in the timing of heterochromatin packaging
Presumably, DNA hypomethylation in ICF syndrome may be major determinants of chromosomal abnormalities
leads to deregulation of genes that perturb craniofacial, and gene deregulation.255 The aggregation of such chroma-
cerebral, and immunological development. Microarray tin proteins in ICF may simply be a result of underconden-
analysis has been used to identify genes with significantly sation of heterochromatin at those specific loci, but it also
altered mRNA levels in ICF lymphoblastoid cell lines may have an effect in trans on gene expression elsewhere in
compared with controls.248 Many of the genes identified the genome.243
have known roles in immune function in both B and T Finally, in addition to hypomethylation and any
cell lineages that could account for the immunodeficiency changes in transcription, a recent study has shown that
consistently manifested in ICF syndrome. No alteration of DNA replication itself is altered in ICF, either as a result
DNA methylation was detected at the promoters of any of of altered transcription of genes involved in replication, or
the deregulated genes tested,248 consistent with the find- perhaps due to alteration of chromatin structure affecting
ings of the whole-genome scan,244 arguing against a direct the access of the replication machinery.256 Given that one
cis effect of a methylation abnormality. Furthermore, none of the key features of ICF is the chromosomal instabil-
of these genes is located on chromosomes 1, 9, or 16. This ity at certain locations, a parallel is seen here with other
raises the question of how hypomethylation of specific conditions in which DNA replication defects result in
repetitive DNA sequences in ICF patients can lead to chromosomal instability, such as certain cancers or cancer
altered expression of genes located at distant genomic sites. syndromes.256
One possibility is that the hypomethylation of satellite Since the mutation responsible for ICF was first
DNA alters their heterochromatin properties, and that it described, it has been clear that hypomethylation plays a
is the physical association of deregulated genes with these key role in this disorder. However, many questions remain
domains in the nucleus that is aberrant in ICF cells.249 unanswered. It is not clear why only certain regions of the
Silenced genes in human B and T lymphocytes have been genome are hypomethylated in ICF, and while it is plausi-
shown to co-localize with domains of pericentromeric ble that reduced methylation could lead to de-repression of
5 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
6 imprinted domain where Dlx5 and its non-imprinted ATRX may be involved in gene activation, suggested
neighbor Dlx6 are present. Dlx5 and Dlx6 encode proteins by the reduced expression of the α-globin locus in ATR-X
that regulate neurotransmitter production,285 and were syndrome. This, however, does not explain the additional
shown be upregulated two-fold in MeCP2-null mice brains phenotypical traits observed, and presumably deregulation
with accompanying changes in histone modifications and of many other loci gives rise to the complex phenotype.
lack of chromatin loop formation.286 The observation that diverse DNA methylation defects
Around 5–10% of individuals clinically diagnosed with (hypermethylation at DYZ2 Y-chromosome repeats and
Rett syndrome do not appear to have mutations in the hypomethylation of ribosomal DNA) are present in dis-
MeCP2 gene. In addition, MeCP2 mutations have been ease297 indicates ATRX is able to regulate chromatin struc-
found in other disorders, including neonatal-onset encepha- ture at several distinct loci. Direct (stimulatory) effects of
lopathy, autism, patients exhibiting Angelman phenotype, ATRX upon loci are also indicated by its association with
nonsyndromic X-linked mental retardation, and psychosis, the transcriptional regulatory death-associated protein 6
pyramidal signs, and macroorchidism (PPM-X) syndrome. (DAXX). The DAXX-ATRX-containing complex, levels
Recently, mutations in cyclin-dependent kinase 5 (CDKL5), of which are reduced in ATRX patient cell lines, is able to
which are shown to directly interact with MeCP2, and remodel nucleosomes in vitro,298 and patient mutations in
Netrin G1, have been found in patients with a very similar ATRX cause a reduction in this function.293 Furthermore,
phenotype to that of Rett syndrome.287–289 ATR-X associates with PML bodies, which are thought to
Recent work has confirmed the importance of epigene- function as regulatory (activator) factor reservoirs in the
tic mechanisms in Rett syndrome pathophysiology, but spe- nucleus.298,299
cific downstream targets are still to be determined. Perhaps On the other hand, chromatin remodeling by ATR-X
most significantly, the accepted hypothesis that Rett syn- may facilitate chromatin condensation and gene silenc-
drome is purely a disorder of neurodevelopment has been ing. The ADD domain of ATR-X has been shown to bind
recently called into question, with reversal of phenotype in directly to histone H3 trimethylated at lysine 9 (H3K9me3),
adult mouse models of disease. a hallmark of pericentric chromatin, and disease-causing
mutations impair this association.300 ATR-X was also found
to associate with the histone methyltransferase enhancer of
AT R-X SY N D RO M E : A C O N N EC T I O N
zeste homologue 2 (EZH2).301 EZH2 is part of the poly-
TO C H RO M AT I N R E MO D E L I N G
comb group repressor complexes (PRC) that methylate
The X-linked α-thalassemia mental retardation syndrome is histone H3 lysine 27 (PRC2) and histone H1 lysine 26
another example of genetic mutation in a factor involved (PRC3).302 H3 lysine 27 methylation is a mark recognized
in chromatin organization affecting disease loci in trans. by the chromodomain of the Polycomb protein (contained
The ATRX gene at chromosome Xq13.3 encodes for a within the PRC1 complex) implicated in chromatin con-
SNF2-like chromatin remodeling helicase. Functional densation and developmentally regulated gene silenc-
domains of ATRX include a PHD zinc finger–like motif ing.303–305 Methylation at H1 lysine 26 may also be involved
at its aminoterminus (homologous to the PHD motifs in in chromatin condensation, as HP1 has been shown to
DNMT3A and DNMT3B), an adjacent coil-coil motif specifically interact with this mark.306 These interactions
termed ATRX-DNMT3-DNMT3L (ADD), and a heli- reveal potential new pathways of gene regulation where the
case domain at its carboxyterminus.290 Mutations in ATRX HP1 and the polycomb group–silencing pathways may be
are clustered in these domains and are thought to impair its synergistic. For example, a novel polycomb group complex,
nuclear localization, protein–protein interactions, or chro- PRC4, was recently shown to be upregulated upon onco-
matin remodeling functions.245,291–293 genic transformation.307 PRC4 preferentially methylates
Affected individuals have low levels of α-globin subunits histone H1b lysine 26, particularly in the presence of native
that favor the formation of unstable β-globin tetramers that complex subunit isoforms (i.e. Eed2), and histone deacety-
precipitate within erythrocytes, causing varying degrees of lases (i.e. Sirt1). Excess of PRC4 might lead to increased
hemolysis and splenomegaly. Affected males have relatively lysine 26 methylation and subsequent recognition by HP1.
severe mental retardation together with facial and skeletal Dependent on the genomic context and associated com-
abnormalities, urogenital abnormalities, and microcephaly, plexes, HP1 may either silence or de-repress the affected
whereas heterozygous females are usually asymptomatic due loci.308 Furthermore, ATR-X and HP1 have been shown to
to a skewed pattern of X-chromosome inactivation prefer- co-localize by immunofluorescence245 and to directly inter-
entially silencing the mutated allele.294–296 act.309–311 These interactions of ATR-X with other molecular
www.Ebook777.com
chaperones may generate regulatory complexes that directly The cancer methylome is characterized by a global
affect chromatin condensation at pericentric sites. reduction in DNA methylation that enhances genomic
Mutations in ATRX are frequent at the PHD-zinc instability, and by focal gains of methylation, in particu-
finger motif, which is likely to be responsible for the tar- lar at promoters enriched in CpG dinucleotides (CpG
geting of ATR-X to pericentromeric heterochromatin.245 islands). DNA hypermethylation of promoter CpG islands
Inadequate targeting of catalytically active ATR-X may frequently is associated with repression of the associated
therefore result in ectopic binding at loci deregulated in dis- genes—more often than not of tumors’ suppressor loci.317
ease. Additionally, due to the predominant localization of Nevertheless, the function of DNA methylation is now
ATR-X to pericentromeric heterochromatin, ATR-X may understood to be dependent on where exactly it is located;
be indirectly regulating additional loci by modulating the it has been shown to also associate with active genes.318
nature of gene association with pericentromeric heterochro- It is becoming increasingly clear that an interplay exists
matin. Interestingly, different patients with identical muta- between DNA methylation and histone modification.319 In
tions in the ATRX gene exhibit significant phenotypical colonic tumors, increased DNA methylation is frequent at
variation.290,312 Recent ChIP sequencing data have suggested promoters of tumour suppressor genes, which in common
that ATR-X is found associated with variable number tan- with embryonic stem cells, are marked by both histone H3
dem repeats (VNTRs), and given that the size of these lysine 4 and lysine 27 trimethylation.320 This bivalency for
VNTRs varies between individuals, this may explain some histone H3 modification is therefore part of an instructive
of this variation.313 process that guides the DNA methylation machinery to
Acquired forms of alpha-thalassemia myelodysplastic the affected loci. Indeed, polycomb group complexes that
syndrome (ATMDS) with mutations in ATRX exhibit a methylate H3 lysine 27 are known to interact with the de
much more severe form of alpha-thalassemia compared to novo DNA methyltransferases.321
inherited forms even when the mutation itself is identical, Long non-coding RNA (>200bp) and small non-coding
again implicating epigenetic mechanisms in causing sever- RNA (~22bp) such as microRNAs are novel functional ele-
ity.314 The above observations plainly demonstrate that ments capable of regulating gene expression, and they seem
ATRX is clearly a chromatin modifier acting in trans, par- to be involved in tumorigenesis.322,323 For example, a long
ticularly at pericentric heterochromatin, with the potential non-coding RNA that is antisense to the CDKN2B (p15)
for either gene activation or silencing. However, the down- tumor suppressor locus can regulate the chromatin and
stream targets and the exact nature of modifying complexes DNA methylation of the p15 locus,324 and a similar effect in
formed are yet to be elucidated. cis was seen at the CDKN1A (p21) locus.325 Effects in trans,
where the non-coding RNA has an effect on a chromosome
other than the one where it originates, have been described
E P I G E N ET I C ME C H A N I SMS where large intergenic non-coding RNAs appear to target
IN CANCER chromatin-modifying machineries to affect the expression
of distant loci.326 MicroRNAs are very powerful regulatory
Cancer is a complex multifactorial disease.315 Genetic muta- elements327 grossly de-regulated in cancers323 and thus pro-
tions may promote tumorigenesis by hampering tumor vide potentially attractive targets for cancer therapy.328
suppressor activity or by hyper-activation of oncogenic Effects at a distance are also brought about by chroma-
pathways. The balance between these activities promotes tin looping, where higher-order chromatin architectures
gene expression programs that fuel neoplastic metabolic enable the interaction of physically distant regulatory ele-
states. How the genome responds to these signals in terms ments, enhancer–promoter interaction for example, or
of transcriptional output is thus central to the generation place loci within nuclear-chromosomal territories that
and/or maintenance of the cancer phenotype. facilitate or inhibit gene activity. Higher-order architec-
Access to genetic information and subsequent transcrip- tures such as those mediated by CTCF binding sites and
tional readout are modulated by epigenetic factors—DNA cohesin are thought to be part of networks of long-range
methylation, chromatin histone modifications, non-coding interactions that control developmentally regulated tran-
RNA, and higher-order chromatin structures. The epig- scriptional programs.329 Since CTCF binding is sensitive
enome is thought to constitute a “buffering” system that to DNA methylation, further refinement of the role that
regulates gene expression “noise”.This buffering system is long-distance interactions play in tumorigenesis is needed.
grossly disrupted in cancer, leading to increased noise and Unlike genetic mutation, epigenetic mutation/modi-
heterogeneity of gene expression.316 fication is potentially reversible and thus a very attractive
5 2 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
target for therapeutic intervention. Inhibitors of DNA underacetylation of centric heterochromatin in human metaphase
chromosomes. Chromosoma. Mar 1992;101:322.
methylation and histone deacetylation are already in use for 15. Peters AH, et al. Loss of the Suv39h histone methyltransferases
some types of tumours,319,328 and extensive research is under- impairs mammalian heterochromatin and genome stability. Cell.
way to identify further epigenetic targets and to understand Nov 2, 2001;107:323.
16. Maison C, et al. Higher-order structure in pericentric heterochro-
how the epigenetic modifications interrelate such that com- matin involves a distinct pattern of histone modification and an
binatorial therapies can be designed. Together with the RNA component. Nat Genet. Mar 2002;30:329.
technical advances that make genome-wide profiling of epi- 17. Maison C, et al. SUMOylation promotes de novo targeting
of HP1alpha to pericentric heterochromatin. Nat Genet. Mar
genetic modifications in individuals possible, the future of 2011;43:220.
epigenetic therapy in cancer is promising. 18. Schramke V, et al. RNA-interference-directed chromatin modifica-
tion coupled to RNA polymerase II transcription. Nature. Jun 30,
2005;435:1275.
19. Fukagawa T, et al. Dicer is essential for formation of the heterochro-
C O N C LU S I O N matin structure in vertebrate cells. Nat Cell Biol. Aug 2004;6:784.
20. Pal-Bhadra M, et al. Heterochromatic silencing and HP1 localiza-
tion in Drosophila are dependent on the RNAi machinery. Science.
It is clear that previously mysterious aspects of gene regu- Jan 30, 2004;303:669.
lation that can be grouped under the terms “epigenetic” 21. Eissenberg JC, Elgin SC. The HP1 protein family: getting a grip on
or “epigenomic” are finally yielding to molecular biology chromatin. Curr Opin Genet Dev. Apr 2000;10:204.
22. Festenstein R, et al. Modulation of heterochromatin protein 1
approaches and have revealed a new level of genome orga- dynamics in primary mammalian cells. Science. Jan 31, 2003;299:719.
nization and regulation. Already, the rapid increase in our 23. Cheutin T, et al. Maintenance of stable heterochromatin domains
by dynamic HP1 binding. Science. Jan 31, 2003;299:721.
understanding of the control of gene expression patterns has 24. Schmiedeberg L, Weisshart K, Diekmann S, Meyer Zu Hoerste G,
revealed potentially powerful new therapeutic avenues for an Hemmerich P. High- and low-mobility populations of HP1 in het-
ever-increasing number of human diseases, and a great vari- erochromatin of mammalian cells. Mol Biol Cell. Jun 2004;15:2819.
25. Dillon N, Festenstein R. Unravelling heterochromatin: competition
ety of human cancers, many of which are currently incurable. between positive and negative factors regulates accessibility. Trends
Genet. May 2002;18:252.
26. Bannister AJ, Kouzarides T. Regulation of chromatin by histone
modifications. Cell Res. Mar 2011;21:381.
R EFE R E N C ES 27. Hargreaves DC, Crabtree GR. ATP-dependent chromatin remodel-
ing: genetics, genomics and mechanisms. Cell Res. Mar 2011;21:396.
1. Holliday R. DNA methylation and epigenotypes. Biochemistry 28. Turner BM. Histone acetylation and an epigenetic code. Bioessays.
(Mosc). May 2005;70:500. Sep 2000;22:836.
2. Reik W, Walter J. Genomic imprinting: parental influence on the 29. Strahl BD, Allis CD. The language of covalent histone modifica-
genome. Nat Rev Genet. Jan 2001;2:21. tions. Nature. Jan 6, 2000;403:41.
3. Ferguson-Smith AC. Genomic imprinting: the emergence of an epi- 30. Narlikar GJ, Fan HY, Kingston RE. Cooperation between com-
genetic paradigm. Nat Rev Genet. Aug 2011;12:565. plexes that regulate chromatin structure and transcription. Cell. Feb
4. Hochedlinger K, Plath K. Epigenetic reprogramming and induced 22, 2002;108:475.
pluripotency. Development. Feb 2009;136:509. 31. de la Cruz X, Lois S, Sanchez-Molina S, Martinez-Balbas MA. Do
5. Takahashi K, et al. Induction of pluripotent stem cells from adult protein motifs read the histone code? Bioessays. Feb 2005;27:164.
human fibroblasts by defined factors. Cell. Nov 30, 2007;131:861. 32. Wysocka J, et al. WDR5 associates with histone H3 methylated at
6. Takahashi K, Yamanaka S. Induction of pluripotent stem cells from K4 and is essential for H3 K4 methylation and vertebrate develop-
mouse embryonic and adult fibroblast cultures by defined factors. ment. Cell. Jun 17, 2005;121:859.
Cell. Aug 25, 2006;126:663. 33. Bernstein BE, et al. Methylation of histone H3 Lys 4 in coding regions
7. Kornberg RD, Lorch Y. Twenty-five years of the nucleosome, of active genes. Proc Natl Acad Sci U S A. Jun 25, 2002;99:8695.
fundamental particle of the eukaryote chromosome. Cell. Aug 6, 34. Schubeler D, et al. The histone modification pattern of active genes
1999;98:285. revealed through genome-wide chromatin analysis of a higher
8. Thoma F, Koller T, Klug A. Involvement of histone H1 in the orga- eukaryote. Genes Dev. Jun 1, 2004;18:1263.
nization of the nucleosome and of the salt-dependent superstruc- 35. Lachner M, O’Carroll D, Rea S, Mechtler K, Jenuwein T.
tures of chromatin. J Cell Biol. Nov 1979;83:403. Methylation of histone H3 lysine 9 creates a binding site for HP1
9. Rattner JB, Hamkalo BA. Nucleosome packing in interphase chro- proteins. Nature. Mar 1, 2001;410:116.
matin. J Cell Biol. May 1979;81:453. 36. Fischle W, et al. Molecular basis for the discrimination of repressive
10. Tremethick DJ. Higher-order structures of chromatin: the elusive methyl-lysine marks in histone H3 by polycomb and HP1 chromo-
30 nm fiber. Cell. Feb 23, 2007;128:651. domains. Genes Dev. Aug 1, 2003;17:1870.
11. Hsu TC, Cooper JE, Mace ML Jr, Brinkley BR. Arrangement of 37. Hermann A, Gowher H, Jeltsch A. Biochemistry and biology
centromeres in mouse cells. Chromosoma. 1971;34:73. of mammalian DNA methyltransferases. Cell Mol Life Sci. Oct
12. Joseph A, Mitchell AR, Miller OJ. The organization of the mouse 2004;61:2571.
satellite DNA at centromeres. Exp Cell Res. Aug 1989;183:494. 38. Bird AP, Wolffe AP. Methylation-induced repression—belts, braces,
13. Guenatri M, Bailly D, Maison C, Almouzni G. Mouse centric and and chromatin. Cell. Nov 24, 1999;99:451.
pericentric satellite repeats form distinct functional heterochroma- 39. Lachner M, Jenuwein T. The many faces of histone lysine methyla-
tin. J Cell Biol. Aug 16, 2004;166:493. tion. Curr Opin Cell Biol. Jun 2002;14:286.
14. Jeppesen P, Mitchell A, Turner B, Perry P. Antibodies to defined 40. Dellino GI, et al. Polycomb silencing blocks transcription initiation.
histone epitopes reveal variations in chromatin conformation and Mol Cell. Mar 26, 2004;13:887.
www.Ebook777.com
41. Tahiliani M, et al. Conversion of 5-methylcytosine to
67. Henikoff S. Dosage-dependent modification of position-effect var-
5-hydroxymethylcytosine in mammalian DNA by MLL partner iegation in Drosophila. Bioessays. May 1996;18:401.
TET1. Science. May 15, 2009;324:930. 68. Grosveld F. Activation by locus control regions? Curr Opin Genet
42. Branco MR, Ficz G, Reik W. Uncovering the role of
Dev. Apr 1999;9:152.
5-hydroxymethylcytosine in the epigenome. Nat Rev Genet. Jan 69. West AG, Fraser P. Remote control of gene transcription. Hum Mol
2012;13:7. Genet. Apr 15, 2005;14:Spec No 1, R101.
43. Williams K, et al. TET1 and hydroxymethylcytosine in transcrip- 70. Festenstein R, Kioussis D. Locus control regions and epigenetic
tion and DNA methylation fidelity. Nature. May 19, 2011;473:343. chromatin modifiers. Curr Opin Genet Dev. Apr 2000;10:199.
44. Nan X, et al. Transcriptional repression by the methyl-CpG-binding 71. Soejima H, Wagstaff J. Imprinting centers, chromatin structure, and
protein MeCP2 involves a histone deacetylase complex. Nature. disease. J Cell Biochem. May 15, 2005;95:226.
May 28, 1998;393:386. 72. Gavrilov A, et al. Chromosome conformation capture (from
45. Ng HH, et al. MBD2 is a transcriptional repressor belonging to the 3C to 5C) and its ChIP-based modification. Methods Mol B.
MeCP1 histone deacetylase complex. Nat Genet. Sep 1999;23:58. 2009;567:171.
46. Fuks F, et al. The methyl-CpG-binding protein MeCP2 links
73. Sanyal A, Lajoie BR, Jain G, Dekker J. The long-range interaction
DNA methylation to histone methylation. J Biol Chem. Feb 7, landscape of gene promoters. Nature. Sep 6, 2012;489:109.
2003;278:4035. 74. Wijchers PJ, et al. Sexual dimorphism in mammalian autosomal
47. Fuks F, Hurd PJ, Deplus R, Kouzarides T. The DNA methyltransfer- gene regulation is determined not only by Sry but by sex chromo-
ases associate with HP1 and the SUV39H1 histone methyltransfer- some complement as well. Dev Cell. Sep 2010;19:477.
ase. Nucleic Acids Res. May 1, 2003;31:2305. 75. Wijchers PJ, Festenstein RJ. Epigenetic regulation of autosomal gene
48. Aagaard L, et al. Functional mammalian homologues of the
expression by sex chromosomes. Trends Genet. Apr 2011;27:132.
Drosophila PEV-modifier Su(var)3–9 encode centromere-associated 76. Harding AE. Friedreich’s ataxia: a clinical and genetic study of 90
proteins which complex with the heterochromatin component families with an analysis of early diagnostic criteria and intrafamilial
M31. Embo J. Apr 1, 1999;18:1923. clustering of clinical features. Brain. 1981;104:589.
49. Lehnertz B, et al. Suv39h-mediated histone H3 lysine 9 methylation 77. Campuzano V, et al. Friedreich’s ataxia: autosomal recessive dis-
directs DNA methylation to major satellite repeats at pericentric ease caused by an intronic GAA triplet repeat expansion. Science.
heterochromatin. Curr Biol. Jul 15, 2003;13:1192. 1996;271:1423.
50. Lewis A, et al. Imprinting on distal chromosome 7 in the placenta 78. Mirkin SM. Expandable DNA repeats and human disease. Nature.
involves repressive histone methylation independent of DNA meth- Jun 21, 2007;447:932.
ylation. Nat Genet. Dec 2004;36:1291. 79. Pandolfo M, Pastore A. The pathogenesis of Friedreich ataxia and
51. Ebralidse KK, Hebbes TR, Clayton AL, Thorne AW,
the structure and function of frataxin. J Neurol. Mar 2009;256
Crane-Robinson C. Nucleosomal structure at hyperacetylated Suppl. 1:9.
loci probed in nuclei by DNA-histone crosslinking. Nucleic Acids 80. Montermini L, et al. Phenotypic variability in Friedreich ataxia: role
Research. Oct 11, 1993;21:4734. of the associated GAA triplet repeat expansion. Ann Neurol.
52. Orlando V, Strutt H, Paro R. Analysis of chromatin structure by 1997;41:675.
in vivo formaldehyde cross-linking. Methods. Feb 1997;11:205. 81. Monros E, et al. Phenotype correlation and intergenerational
53. Shendure J, Ji H. Next-generation DNA sequencing. Nature Biotech. dynamics of the Friedreich ataxia GAA trinucleotide repeat. Am J
Oct 2008;26:1135. Hum Genet. Jul 1997;61:101.
54. Arvey A, Agius P, Noble WS, Leslie C. Sequence and chromatin 82. Filla A, et al. The relationship between trinucleotide (GAA) repeat
determinants of cell-type-specific transcription factor binding. length and clinical features in Friedreich ataxia. Am J Hum Genet.
Genome Res. Sep 2012;22:1723. Sep 1996;59:554.
55. EP Consortium, et al. An integrated encyclopedia of DNA elements 83. Durr A, et al. Clinical and genetic abnormalities in patients with
in the human genome. Nature. Sep 6, 2012;489:57. Friedreich’s ataxia. N Engl J Med. 1996;335:1169.
56. Djebali S, et al. Landscape of transcription in human cells. Nature. 84. Lamont PJ, Davis MB, Wood NW. Identification and sizing of the
Sep 6, 2012;489:101. GAA trinucleotide repeat expansion of Friedreich’s ataxia in 56
57. Harrow J, et al. GENCODE: the reference human genome annota- patients. Clinical and genetic correlates. Brain. 1997;120:673.
tion for The ENCODE Project. Genome Res. Sep 2012;22:1760. 85. Campuzano V, et al. Frataxin is reduced in Friedreich ataxia patients
58. Thurman RE, et al. The accessible chromatin landscape of the and is associated with mitochondrial membranes. Hum Mol Genet.
human genome. Nature. Sep 6, 2012;489:75. 1997;6:1771.
59. Wang J, et al. Sequence features and chromatin structure around 86. Geschwind DH, et al. Friedreich’s ataxia GAA repeat expan-
the genomic regions bound by 119 human transcription factors. sion in patients with recessive or sporadic ataxia. Neurology.
Genome Res. Sep 2012;22:1798. 1997;49:1004.
60. Wong CC, et al. A longitudinal study of epigenetic variation in 87. Gellera C, et al. Very late onset Friedreich’s ataxia without cardiomy-
twins. Epigenetics. Aug 16, 2010;5:516. opathy is associated with limited GAA expansion in the X25 gene.
61. Bell JT, Spector TD. A twin approach to unraveling epigenetics. Neurology. 1997;49:1153.
Trends Genet. Mar 2011;27:116. 88. Kumari D, Usdin K. Is Friedreich ataxia an epigenetic disorder?
62. Osborne CS, et al. Active genes dynamically colocalize to shared Clin Epigenet. 2012;4:2.
sites of ongoing transcription. Nat Genet. Oct 2004;36:1065. 89. Bidichandani SI, Ashizawa T, Patel PI. The GAA triplet-repeat
63. Brown K, et al. Association of transcriptionally silent genes with expansion in Friedreich ataxia interferes with transcription and may
Ikaros complexes at centromeric heterochromatin. Cell. 1997;91:845. be associated with an unusual DNA structure. Am J Hum Genet.
64. Locke J, Kotarski MA, Tartof KD. Dosage-dependent modifiers of Jan 1998;62:111.
position effect variegation in Drosophila and a mass action model 90. Grabczyk E, Usdin K. Alleviating transcript insufficiency caused
that explains their effect. Genetics. Sep 1988;120:181. by Friedreich’s ataxia triplet repeats. Nucleic Acids Res. Dec 15,
65. Cattanach BM. Position effect variegation in the mouse. Genet Res. 2000;28:4930.
Jun 1974;23:291. 91. Grabczyk E, Usdin K. The GAA*TTC triplet repeat expanded in
66. Festenstein R, et al. Locus control region function and
Friedreich’s ataxia impedes transcription elongation by T7 RNA
heterochromatin-induced position effect variegation. Science. Feb polymerase in a length and supercoil dependent manner. Nucleic
23, 1996;271:1123. Acids Res. Jul 15, 2000;28:2815.
5 4 • P rincip l e s o f G e no m ic M e dicin e
92. Krasilnikova MM, et al. Effects of Friedreich’s ataxia (GAA) 113. Luijsterburg MS, et al. Heterochromatin protein 1 is recruited to
n*(TTC)n repeats on RNA synthesis and stability. Nucleic Acids various types of DNA damage. J Cell Biol. May 18, 2009;185:577.
Res. 2007;35:1075. 114. Peng JC, Karpen GH. Heterochromatic genome stability requires
93. Sakamoto N, et al. Sticky DNA: self-association properties of long regulators of histone H3 K9 methylation. PLoS Genet, Mar,
GAA.TTC repeats in R.R.Y triplex structures from Friedreich’s 2009;5:e1000435.
ataxia. Mol Cell. Apr 1999;3:465. 115. Sun Y, et al. Histone H3 methylation links DNA damage detection
94. Ohshima K, Montermini L, Wells RD, Pandolfo M. Inhibitory to activation of the tumour suppressor Tip60. Nat Cell Biol. Nov
effects of expanded GAA.TTC triplet repeats from intron I of 2009;11:1376.
the Friedreich ataxia gene on transcription and replication in vivo. 116. Zarebski M, Wiernasz E, Dobrucki JW. Recruitment of heterochro-
J Biol Chem. Jun 5, 1998;273:14588. matin protein 1 to DNA repair sites. Cytometry A. Jul 2009;75:619.
95. Sakamoto N, Ohshima K, Montermini L, Pandolfo M, Wells RD. 117. Djupedal I, Ekwall K. Epigenetics: heterochromatin meets RNAi.
Sticky DNA, a self-associated complex formed at long GAA*TTC Cell Res. Mar 2009;19:282.
repeats in intron 1 of the frataxin gene, inhibits transcription. J Biol 118. Barkess G, West AG. Chromatin insulator elements: establish-
Chem. Jul 20, 2001;276:27171. ing barriers to set heterochromatin boundaries. Epigenomics. Feb
96. Kim E, Napierala M, Dent SY. Hyperexpansion of GAA repeats 2012;4:67.
affects post-initiation steps of FXN transcription in Friedreich’s 119. Bushey AM, Dorman ER, Corces VG. Chromatin insulators: reg-
ataxia. Nucleic Acids Research. Oct 1, 2011;39:8366. ulatory mechanisms and epigenetic inheritance. Mol Cell. Oct 10,
97. Kumari D, Biacsi RE, Usdin K. Repeat expansion affects both tran- 2008;32:1.
scription initiation and elongation in friedreich ataxia cells. J Biol 120. Puccio H. Multicellular models of Friedreich ataxia. J Neurol. Mar
Chem. Feb 11, 2011;286:4209. 2009;256 Suppl 1:18.
98. Punga T, Buhler M. Long intronic GAA repeats causing Friedreich 121. Liu J, et al. Generation of induced pluripotent stem cell lines from
ataxia impede transcription elongation. EMBO Mol Med. Apr Friedreich ataxia patients. Stem Cell Rev. Sep 2011;7:703.
2010;2:120. 122. Chou CJ, Herman D, Gottesfeld JM. Pimelic diphenylamide 106
99. Saveliev A, Everett C, Sharpe T, Webster Z, Festenstein R. DNA is a slow, tight-binding inhibitor of class I histone deacetylases.
triplet repeats mediate heterochromatin-protein-1-sensitive varie- J Biol Chem. Dec 19, 2008;283:35402.
gated gene silencing. Nature. Apr 24, 2003;422:909. 123. Rai M, et al. Two new pimelic diphenylamide HDAC inhibitors
100. Castaldo I, et al. DNA methylation in intron 1 of the frataxin gene induce sustained frataxin upregulation in cells from Friedreich’s
is related to GAA repeat length and age of onset in Friedreich ataxia patients and in a mouse model. PLoS One. 2010:5: e8825.
ataxia patients. J Med Genet. Dec 2008;45:808. 124. Rai M, et al. HDAC inhibitors correct frataxin deficiency in a
101. Evans-Galea MV, et al. FXN methylation predicts expression Friedreich ataxia mouse model. PLoS One. 2008;3:e1958.
and clinical outcome in Friedreich ataxia. Ann Neurol. Apr 125. Sandi C, et al. Prolonged treatment with pimelic o-aminobenzamide
2012;71:487. HDAC inhibitors ameliorates the disease phenotype of a
102. Greene E, Mahishi L, Entezam A, Kumari D, Usdin K. Repeat-induced Friedreich ataxia mouse model. Neurobiol Dis. Jun 2011;42:496.
epigenetic changes in intron 1 of the frataxin gene and its conse- 126. Xu C, et al. Chemical probes identify a role for histone deacety-
quences in Friedreich ataxia. Nucleic Acids Res. 2007;35:3383. lase 3 in Friedreich’s ataxia gene silencing. Chem Biol. Sep 25,
103. Al-Mahdawi S, et al. The Friedreich ataxia GAA repeat expansion 2009;16:980.
mutation induces comparable epigenetic changes in human and 126a. Chan PK, Torres R, Yandim C, Law PP, Khadayate S, Mauri M,
transgenic mouse brain and heart tissues. Hum Mol Genet. Mar 1, Grosan C, Chapman-Rothe N, Giunti P, Pook M, Festenstein R.
2008;17:735. Heterochromatinization induced by GAA-repeat hyperexpansion
104. Herman D, et al. Histone deacetylase inhibitors reverse gene in Friedreich's ataxia can be reduced upon HDAC inhibition by
silencing in Friedreich’s ataxia. Nat Chem Biol. Oct 2006;2:551. vitamin B3. Human Molecular Genetics. 2013;22:2662–2675.
105. De Biase I, Chutake YK, Rindler PM, Bidichandani SI. Epigenetic 126b. Libri V*, Yandim C*, Athanasopoulos S, Loyse N, Natisvili T, Law
silencing in Friedreich ataxia is associated with depletion of CTCF PP, Chan PK, Mohammad T, Mauri M, Tam KT, Leiper J, Piper
(CCCTC-binding factor) and antisense transcription. PloS One. S, Ramesh A, Parkinson MH, Huson L, Giunti P, Festenstein R.
2009;4(11):e7914. Epigenetic and neurological effects and safety of high-dose nicotin-
106. Bourn RL, Rindler PM, Pollard LM, Bidichandani SI. E. coli mis- amide in patients with Friedreich’s ataxia: an exploratory, open-label,
match repair acts downstream of replication fork stalling to stabi- dose-escalation study. The Lancet (in press-1st May 2014).
lize the expanded (GAA.TTC)(n) sequence. Mutat Res. Feb 10, 127. Fu YH, et al. An unstable triplet repeat in a gene related to myo-
2009;661:71. tonic muscular dystrophy. Science. Mar 6, 1992;255:1256.
107. Du J, et al. Role of mismatch repair enzymes in GAA{middle dot} 128. Mahadevan M, et al. Myotonic dystrophy mutation: an unstable
TTC triplet-repeat expansion in Friedreich ataxia induced plurip- CTG repeat in the 3′ untranslated region of the gene. Science.
otent stem cells. J Biol Chem. Aug 24, 2012;287:29861. Mar 6, 1992;255:1253.
108. Ezzatizadeh V, et al. The mismatch repair system protects against 129. Brook JD, et al. Molecular basis of myotonic dystrophy: expansion
intergenerational GAA repeat instability in a Friedreich ataxia of a trinucleotide (CTG) repeat at the 3′ end of a transcript encod-
mouse model. Neurobiol Dis. Apr 2012;46:165. ing a protein kinase family member. Cell. Apr 17, 1992;69:385.
109. Kim HM, et al. Chromosome fragility at GAA tracts in yeast 130. Liquori CL, et al. Myotonic dystrophy type 2 caused by a CCTG
depends on repeat orientation and requires mismatch repair. expansion in intron 1 of ZNF9. Science. Aug 3, 2001;293:864.
EMBO J. Nov 5, 2008;27:2896. 131. Gomes-Pereira M, Monckton DG. Chemical modifiers of unsta-
110. Ku S, et al. Friedreich’s ataxia induced pluripotent stem cells model ble expanded simple sequence repeats: what goes up, could come
intergenerational GAATTC triplet repeat instability. Cell Stem down. Mutat Res. Jun 25, 2006;598:15.
Cell. Nov 5, 2010;7:631. 132. Ashizawa T, Sarkar PS. Myotonic dystrophy types 1 and 2. Handb
111. Ayoub N, Jeyasekharan AD, Bernal JA, Venkitaraman AR. Clin Neurol. 2011;101:193.
HP1-beta mobilization promotes chromatin changes that initiate 133. Lee JE, Cooper TA. Pathogenic mechanisms of myotonic dystro-
the DNA damage response. Nature. May 29, 2008;453:682. phy. Biochem Soc Trans. Dec 2009;37:1281.
112. Goodarzi AA, et al. ATM signaling facilitates repair of DNA 134. Mankodi A, Lin X, Blaxall BC, Swanson MS, Thornton CA.
double-strand breaks associated with heterochromatin. Mol Cell. Nuclear RNA foci in the heart in myotonic dystrophy. Circ Res.
Jul 25, 2008;31:167. Nov 25, 2005;97:1152.
5 6 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
180. Tassone F, et al. Elevated levels of FMR1 mRNA in carrier males: a 201. Thomas NS, et al. A large patient study confirming that facioscapu-
new mechanism of involvement in the fragile-X syndrome. Am J lohumeral muscular dystrophy (FSHD) disease expression is
Hum Genet. Jan 2000;66:6. almost exclusively associated with an FSHD locus located on a
181. Tassone F, et al. Elevated FMR1 mRNA in premutation carriers is 4qA-defined 4qter subtelomere. J Med Genet. Mar 2007;44:215.
due to increased transcription. RNA. Apr 2007;13:555. 202. Lemmers RJ, et al. A unifying genetic model for facioscapulo-
182. Kenneson A, Zhang F, Hagedorn CH, Warren ST. Reduced humeral muscular dystrophy. Science. Sep 2010;329:1650.
FMRP and increased FMR1 transcription is proportionally associ- 203. Spurlock G, Jim HP, Upadhyaya M. Confirmation that the specific
ated with CGG repeat number in intermediate-length and premu- SSLP microsatellite allele 4qA161 segregates with fascioscapulo-
tation carriers. Hum Mol Genet. Jul 2001;10:1449. humeral muscular dystrophy (FSHD) in a cohort of multiplex and
183. Hagerman RJ, Hagerman PJ. Fragile X syndrome: a model of simplex FSHD families. Muscle Nerve. Nov 2010;42:820.
gene-brain-behavior relationships. Mol Genet Metab. 74:89. 204. Fisher J, Upadhyaya M. Molecular genetics of facioscapulohumeral
184. Hoem G, et al. CGG-repeat length threshold for FMR1 RNA muscular dystrophy (FSHD). Neuromuscul Disord. Jan 1997;7:55.
pathogenesis in a cellular model for FXTAS. Hum Mol Genet. Jun 205. Jiang G, et al. Testing the position-effect variegation hypothesis
2011;20:2161. for facioscapulohumeral muscular dystrophy by analysis of histone
185. Sheridan SD, et al. Epigenetic characterization of the FMR1 modification and gene expression in subtelomeric 4q. Hum Mol
gene and aberrant neurodevelopment in human induced pluripo- Genet. Nov 15, 2003;12:2909.
tent stem cell models of fragile X syndrome. PLoS One. 2011;6, 206. Zeng W, et al. Specific loss of histone H3 lysine 9 trimethylation
e26203. and HP1gamma/cohesin binding at D4Z4 repeats is associated
186. Wang T, et al. Genome-wide DNA hydroxymethylation changes with facioscapulohumeral dystrophy (FSHD). PLoS Genet. Jul
are associated with neurodevelopmental genes in the developing 2009;5, e1000559.
human cerebellum. Hum Mol Genet. Dec 2012;21:5500. 207. van Overveld PG, et al. Hypomethylation of D4Z4 in 4q-linked
187. Padberg G. In: Upadhyaya M, Cooper DN, eds. Facioscapulohumeral and non-4q-linked facioscapulohumeral muscular dystrophy. Nat
Muscular Dystrophy: Clinical Medicine and Molecular Cell Genet. Dec 2003;35:315.
Biology. New York: Garland Science/BI OS Scientific, Abingdon; 208. van Overveld PG, et al. Variable hypomethylation of D4Z4
2004: 41–54. in facioscapulohumeral muscular dystrophy. Ann Neurol. Oct
188. Tawil R, Figlewicz DA, Griggs RC, Weiffenbach B. Facioscapulo 2005;58:569.
humeral dystrophy: a distinct regional myopathy with a novel 209. Gabellini D, et al. Facioscapulohumeral muscular dystrophy in
molecular pathogenesis. FSH Consortium. Ann Neurol. Mar mice overexpressing FRG1. Nature. Feb 2006;439:973.
1998;43:279. 210. Gabellini D, Green MR, Tupler R. Inappropriate gene activation
189. Tawil R. Facioscapulohumeral muscular dystrophy. Curr Neurol in FSHD: a repressor complex binds a chromosomal repeat deleted
Neurosci Rep. Jan 2004;4:51. in dystrophic muscle. Cell. Aug 2002;110:339.
190. van Deutekom JC, et al. FSHD associated DNA rearrangements 211. Klooster R, et al. Comprehensive expression analysis of FSHD
are due to deletions of integral copies of a 3.2 kb tandemly repeated candidate genes at the mRNA and protein level. Eur J Hum Genet.
unit. Hum Mol Genet. Dec 1993;2:2037. Dec 2009;17:1615.
191. Wijmenga C, et al. Chromosome 4q DNA rearrangements associ- 212. Arashiro P, et al. Transcriptional regulation differs in affected
ated with facioscapulohumeral muscular dystrophy. Nat Genet. Sep facioscapulohumeral muscular dystrophy patients compared to asymp-
1992;2:26. tomatic related carriers. Proc Natl Acad Sci U S A. Apr 2009;106:6220.
192. Lunt PW. 44th ENMC International Workshop: Facioscapulo 213. Masny PS, et al. Analysis of allele-specific RNA transcription in
humeral Muscular Dystrophy: Molecular Studies. July 19–21, FSHD by RNA-DNA FISH in single myonuclei. Eur J Hum
1996, Naarden, The Netherlands. Neuromuscul Disord. Apr Genet. Apr 2010;18:448.
1998;8:126. 214. Rijkers T, et al. FRG2, an FSHD candidate gene, is transcription-
193. Richards M, Coppée F, Thomas N, Belayew A, Upadhyaya M. ally upregulated in differentiating primary myoblast cultures of
Facioscapulohumeral muscular dystrophy (FSHD): an enigma FSHD patients. J Med Genet. Nov 2004;41:826.
unravelled? Hum Genet. Mar 2012;131:325. 215. Deak KL, et al. Genotype-phenotype study in an FSHD fam-
194. Tawil R, et al. Evidence for anticipation and association of deletion ily with a proximal deletion encompassing p13E-11 and D4Z4.
size with severity in facioscapulohumeral muscular dystrophy. The Neurology. Feb 2007;68:578.
FSH-DY Group. Ann Neurol. Jun 1996;39:744. 216. Yip DJ, Picketts DJ. Increasing D4Z4 repeat copy number com-
195. Gabriels J, et al. Nucleotide sequence of the partially deleted D4Z4 promises C2C12 myoblast differentiation. FEBS Lett. Feb
locus in a patient with FSHD identifies a putative gene within each 2003;537:133.
3.3 kb element. Gene. Aug 5, 1999;236:25. 217. Dixit M, et al. DUX4, a candidate gene of facioscapulohumeral
196. Lyle R, Wright TJ, Clark LN, Hewitt JE. The FSHD-associated muscular dystrophy, encodes a transcriptional activator of PITX1.
repeat, D4Z4, is a member of a dispersed family of Proc Natl Acad Sci U S A. Nov 2007;104:18157.
homeobox-containing repeats, subsets of which are clustered on 218. Geng LN, et al. DUX4 activates germline genes, retroelements, and
the short arms of the acrocentric chromosomes. Genomics. Aug immune mediators: implications for facioscapulohumeral dystro-
1995;28:389. phy. Dev Cell. Jan 2012;22:38.
197. Wijmenga C, et al. Chromosome 4q DNA rearrangements asso- 219. Kowaljow V, et al. The DUX4 gene at the FSHD1A locus encodes
ciated with facioscapulohumeral muscular dystrophy. Nat Genet. a pro-apoptotic protein. Neuromuscul Disord. Aug 2007;17:611.
Sep 1992;2:26. 220. Vanderplanck C, et al. The FSHD atrophic myotube phenotype is
198. Tupler R, et al. Monosomy of distal 4q does not cause facioscapu- caused by DUX4 expression. PLoS One. 2011;6, e26820.
lohumeral muscular dystrophy. J Med Genet. May 1996;33:366. 221. Snider L, et al. Facioscapulohumeral dystrophy: incomplete sup-
199. Lemmers RJ, et al. Facioscapulohumeral muscular dystrophy is pression of a retrotransposed gene. PLoS Genet 6, e1001181 (Oct,
uniquely associated with one of the two variants of the 4q subtelo- 2010).
mere. Nat Genet. Oct 2002;32:235. 222. SM van der Maarel, Tawil R, Tapscott SJ. Facioscapulohumeral
200. Lemmers RJ, et al. Worldwide population analysis of the 4q and muscular dystrophy and DUX4: breaking the silence. Trends Mol
10q subtelomeres identifies only four discrete interchromosomal Med. May 2011;17:252.
sequence transfers in human evolution. Am J Hum Genet. Mar 223. Smeets DF, et al. ICF syndrome: a new case and review of the lit-
2010;86:364. erature. Hum Genet. Sep 1994;94:240.
www.Ebook777.com
224. Hansen RS, et al. The DNMT3B DNA methyltransferase gene is 246. Qiu C, Sawada K, Zhang X, Cheng X. The PWWP domain of
mutated in the ICF immunodeficiency syndrome. Proc Natl Acad mammalian DNA methyltransferase Dnmt3b defines a new family
Sci U S A. Dec 1999;96:14412. of DNA-binding folds. Nat Struct Biol. Mar 2002;9:217.
225. Bachman KE, Rountree MR, Baylin SB. Dnmt3a and Dnmt3b are 247. Maurer-Stroh S, et al. The Tudor domain “Royal Family”: Tudor,
transcriptional repressors that exhibit unique localization proper- plant Agenet, Chromo, PWWP and MBT domains. Trends
ties to heterochromatin. J Biol Chem. Aug 24, 2001;276:32282. Biochem Sci. Feb 2003;28:69.
226. Chen T, Tsujimoto N, Li E. The PWWP domain of Dnmt3a and 248. Ehrlich M, et al. DNA methyltransferase 3B mutations linked to
Dnmt3b is required for directing DNA methylation to the major the ICF syndrome cause dysregulation of lymphogenesis genes.
satellite repeats at pericentric heterochromatin. Mol Cell Biol. Oct Hum Mol Genet. Dec 1, 2001;10:2917.
2004;24:9048. 249. Bickmore WA, van der Maarel SM. Perturbations of chromatin
227. Wijmenga C, et al. Genetic variation in ICF syndrome: evidence structure in human genetic disease: recent advances. Hum Mol
for genetic heterogeneity. Hum Mutat. Dec 2000;16:509. Genet. 2003;12(2): R207.
228. Shirohzu H, et al. Three novel DNMT3B mutations in Japanese 250. Brown KE, Baxter J, Graf D, Merkenschlager M, Fisher AG.
patients with ICF syndrome. Am J Med Genet. Sep 15, 2002;112:31. Dynamic repositioning of genes in the nucleus of lymphocytes pre-
229. Okano M, Bell DW, Haber DA, Li E. DNA methyltransferases paring for cell division. Mol Cell. Feb 1999;3:207.
Dnmt3a and Dnmt3b are essential for de novo methylation and 251. Alcobia I, Quina AS, Neves H, Clode N, Parreira L. The spatial
mammalian development. Cell. Oct 29, 1999;99:247. organization of centromeric heterochromatin during normal
230. Matarazzo MR, De Bonis ML, Vacca M, Della Ragione F, human lymphopoiesis: evidence for ontogenically determined spa-
D’Esposito M. Lessons from two human chromatin diseases, tial patterns. Exp Cell Res. Nov 2003;290:358.
ICF syndrome and Rett syndrome. Int J Biochem Cell Biol. Jan 252. Gasser SM. Positions of potential: nuclear organization and gene
2009;41:117. expression. Cell. Mar 2001;104:639.
231. Ueda Y, et al. Roles for Dnmt3b in mammalian develop- 253. Jefferson A, et al. Altered intra-nuclear organisation of het-
ment: a mouse model for the ICF syndrome. Development. Mar erochromatin and genes in ICF syndrome. PLoS One. 2010;5,
2006;133:1183. e11364.
232. Jiang YL, et al. DNMT3B mutations and DNA methylation 254. Dupont C, et al. 3D position of pericentromeric heterochromatin
defect define two types of ICF syndrome. Hum Mutat. Jan within the nucleus of a patient with ICF syndrome. Clin Genet.
2005;25:56. Aug 2012;82:187.
233. Hagleitner MM, et al. Clinical spectrum of immunodeficiency, 255. Luciani JJ, et al. Subcellular distribution of HP1 proteins is altered
centromeric instability and facial dysmorphism (ICF syndrome). in ICF syndrome. Eur J Hum Genet. Jan 2005;13:41.
J Med Genet. Feb 2008;45:93. 256. Lana E, et al. DNA replication is altered in Immunodeficiency
234. Saito Y, et al. Overexpression of a splice variant of DNA meth- Centromeric instability Facial anomalies (ICF) cells carrying
yltransferase 3b, DNMT3b4, associated with DNA hypo- DNMT3B mutations. Eur J Hum Genet. Oct 2012;20:1044.
methylation on pericentromeric satellite regions during 257. Neul JL, et al. Rett syndrome: revised diagnostic criteria and
human hepatocarcinogenesis. Proc Natl Acad Sci U S A. Jul 23, nomenclature. Ann Neurol. Dec 2010;68:944.
2002;99:10060. 258. Girard M, et al. Parental origin of de novo MECP2 mutations in
235. de Greef JC, et al. Mutations in ZBTB24 are associated with Rett syndrome. Eur J Hum Genet. Mar 2001;9:231.
immunodeficiency, centromeric instability, and facial anomalies 259. Amir RE, et al. Rett syndrome is caused by mutations in X-linked
syndrome type 2. Am J Hum Genet. Jun 2011;88:796. MECP2, encoding methyl-CpG-binding protein 2. Nat Genet.
236. Chouery E, et al. A novel deletion in ZBTB24 in a Lebanese family Oct 1999;23:185.
with immunodeficiency, centromeric instability, and facial anoma- 260. Christodoulou J, Grimm A, Maher T, Bennetts B. RettBASE: The
lies syndrome type 2. Clin Genet. Nov 2012;82:489. IRSA MECP2 variation database-a new mutation database in evo-
237. Heyn H, et al. Whole-genome bisulfite DNA sequencing of a lution. Human Mutation. May 2003;21(5):466.
DNMT3B mutant patient. Epigenetics. Jun 2012;7:542. 261. Villard L, et al. Two affected boys in a Rett syndrome family: clini-
238. Gisselsson D, et al. Interphase chromosomal abnormalities and cal and molecular findings. Neurology. Oct 2000;55:1188.
mitotic missegregation of hypomethylated sequences in ICF syn- 262. Klose RJ, Bird AP. Genomic DNA methylation: the mark and its
drome cells. Chromosoma. Jul 2005;114:118. mediators. Trends Biochem Sci. Feb 2006;31:89.
239. Jeanpierre M, et al. An embryonic-like methylation pattern of clas- 263. Meehan RR, Lewis JD, Bird AP. Characterization of MeCP2,
sical satellite DNA is observed in ICF syndrome. Hum Mol Genet. a vertebrate DNA binding protein with affinity for methylated
Jun 1993;2:731. DNA. Nucleic Acids Res. Oct 11, 1992;20:5085.
240. Ji W, et al. DNA demethylation and pericentromeric rearrange- 264. Klose RJ, et al. DNA binding selectivity of MeCP2 due to a
ments of chromosome 1. Mutat Res. Sep 1997;379:33. requirement for A/T sequences adjacent to methyl-CpG. Mol Cell.
241. Prada D, et al. Satellite 2 demethylation induced by 5-azacytidine Sep 2, 2005;19:667.
is associated with missegregation of chromosomes 1 and 16 in 265. Buschdorf JP, Stratling WH. A WW domain binding region in
human somatic cells. Mutat Res. Jan 2012;729:100. methyl-CpG-binding protein MeCP2: impact on Rett syndrome.
242. Miniou P, et al. Abnormal methylation pattern in constitutive J Mol Med (Berl). Feb 2004;82:135.
and facultative (X inactive chromosome) heterochromatin of ICF 266. Muotri AR, et al. L1 retrotransposition in neurons is modulated by
patients. Hum Mol Genet. Dec 1994;3:2093. MeCP2. Nature. Nov 18, 2010;468:443.
243. Ehrlich M, et al. ICF, an immunodeficiency syndrome: DNA 267. Chahrour M, et al. MeCP2, a key contributor to neurological
methyltransferase 3B involvement, chromosome anomalies, and disease, activates and represses transcription. Science. May 30,
gene dysregulation. Autoimmunity. May 2008;41:253. 2008;320:1224.
244. Kondo T, et al. Whole-genome methylation scan in ICF syn- 268. Nikitina T, et al. Multiple modes of interaction between the meth-
drome: hypomethylation of non-satellite DNA repeats D4Z4 and ylated DNA binding protein MeCP2 and chromatin. Mol Cell
NBL2. Hum Mol Genet. Mar 1, 2000;9:597. Biol. Feb 2007;27:864.
245. McDowell TL, et al. Localization of a putative transcriptional reg-
269. Georgel PT, et al. Chromatin compaction by human
ulator (ATRX) at pericentromeric heterochromatin and the short MeCP2. Assembly of novel secondary chromatin structures
arms of acrocentric chromosomes. Proc Natl Acad Sci U S A. Nov in the absence of DNA methylation. J Biol Chem. Aug 22,
23, 1999;96:13983. 2003;278:32181.
5 8 • P rincip l e s o f G e no m ic M e dicin e
270. Long SW, Ooi JY, Yau PM, Jones PL. A brain-derived MeCP2 292. Cardoso C, et al. ATR-X mutations cause impaired nuclear loca-
complex supports a role for MeCP2 in RNA processing. Biosci Rep. tion and altered DNA binding properties of the XNP/ATR-X
Oct 1, 2011;31:333. protein. J Med Genet. Oct 2000;37:746.
271. Shahbazian MD, Antalffy B, Armstrong DL, Zoghbi HY. Insight 293. Tang J, et al. A novel transcription regulatory complex containing
into Rett syndrome: MeCP2 levels display tissue- and cell-specific death domain-associated protein and the ATR-X syndrome pro-
differences and correlate with neuronal maturation. Hum Mol tein. J Biol Chem. May 7, 2004;279:20369.
Genet. Jan 15, 2002;11:115. 294. Hendrich B, Bickmore W. Human diseases with underlying defects
272. Ballas N, Lioy DT, Grunseich C, Mandel G. Non-cell autonomous in chromatin structure and modification. Hum Mol Genet. Oct 1,
influence of MeCP2-deficient glia on neuronal dendritic morphol- 2001;10:2233.
ogy. Nat Neurosci. Mar 2009;12:311. 295. Ausio J, Levin DB, De Amorim GV, Bakker S, Macleod PM.
273. Kriaucionis S, Bird A. The major form of MeCP2 has a novel Syndromes of disordered chromatin remodeling. Clin Genet. Aug
N-terminus generated by alternative splicing. Nucleic Acids Res. 2003;64:83.
2004;32:1818. 296. Gibbons RJ, Higgs DR. Molecular-clinical spectrum of the ATR-X
274. Mnatzakanian GN, et al. A previously unidentified MeCP2 open syndrome. Am J Med Genet. 2000;97:204.
reading frame defines a new protein isoform relevant to Rett syn- 297. Gibbons RJ, et al. Mutations in ATRX, encoding a SWI/SNF-like
drome. Nat Genet. Apr 2004;36:339. protein, cause diverse changes in the pattern of DNA methylation.
275. Chen RZ, Akbarian S, Tudor M, Jaenisch R. Deficiency of Nat Genet. Apr 2000;24:368.
methyl-CpG binding protein-2 in CNS neurons results in a 298. Xue Y, et al. The ATRX syndrome protein forms a
Rett-like phenotype in mice. Nat Genet. Mar 2001;27:327. chromatin-remodeling complex with Daxx and localizes in promy-
276. Guy J, Hendrich B, Holmes M, Martin JE, Bird A. A mouse elocytic leukemia nuclear bodies. Proc Natl Acad Sci U S A. Sep 16,
MeCP2-null mutation causes neurological symptoms that mimic 2003;100:10635.
Rett syndrome. Nat Genet. Mar 2001;27:322. 299. Wang J, et al. Promyelocytic leukemia nuclear bodies associate
277. Guy J, Gan J, Selfridge J, Cobb S, Bird A. Reversal of neuro- with transcriptionally active genomic regions. J Cell Biol. Feb 16,
logical defects in a mouse model of Rett syndrome. Science. Feb 2004;164:515.
2007;315:1143. 300. Dhayalan A, et al. The ATRX-ADD domain binds to H3 tail
278. Giacometti E, Luikenhuis S, Beard C, Jaenisch R. Partial rescue of peptides and reads the combined methylation state of K4 and K9.
MeCP2 deficiency by postnatal activation of MeCP2. Proc Natl Hum Mol Genet. Jun 2011;20:2195.
Acad Sci U S A. Feb 2007;104:1931. 301. Cardoso C, et al. Specific interaction between the XNP/ATR-X
279. Robinson L, et al. Morphological and functional reversal of gene product and the SET domain of the human EZH2 protein.
phenotypes in a mouse model of Rett syndrome. Brain. Sep Hum Mol Genet. Apr 1998;7:679.
2012;135:2699. 302. Kuzmichev A, Jenuwein T, Tempst P, Reinberg D. Different
280. Cheval H, et al. Postnatal inactivation reveals enhanced require- EZH2-containing complexes target methylation of histone H1 or
ment for MeCP2 at distinct age windows. Hum Mol Genet. Sep nucleosomal histone H3. Mol Cell. Apr 23, 2004;14:183.
2012;21:3806. 303. Kuzmichev A, Nishioka K, Erdjument-Bromage H, Tempst P,
281. Tao J, et al. Phosphorylation of MeCP2 at Serine 80 regulates its Reinberg D. Histone methyltransferase activity associated with a
chromatin association and neurological function. Proc Natl Acad human multiprotein complex containing the Enhancer of Zeste
Sci U S A. Mar 2009;106:4882. protein. Genes Dev. Nov 15, 2002;16:2893.
282. Zhou Z, et al. Brain-specific phosphorylation of MeCP2 regulates 304. Cao R, et al. Role of histone H3 lysine 27 methylation in
activity-dependent BDNF transcription, dendritic growth, and Polycomb-group silencing. Science. Nov 1, 2002;298:1039.
spine maturation. Neuron. Oct 2006;52:255. 305. Czermin B, et al. Drosophila enhancer of Zeste/ESC complexes
283. Yasui DH, et al. Integrated epigenomic analyses of neuronal have a histone H3 methyltransferase activity that marks chromo-
MeCP2 reveal a role for long-range interaction with active genes. somal Polycomb sites. Cell. Oct 18, 2002;111:185.
Proc Natl Acad Sci U S A. Dec 2007;104:19416. 306. Daujat S, Zeissler U, Waldmann T, Happel N, Schneider R. HP1
284. Chang Q, Khare G, Dani V, Nelson S, Jaenisch R. The disease pro- binds specifically to Lys26-methylated histone H1.4, whereas
gression of MeCP2 mutant mice is affected by the level of BDNF simultaneous Ser27 phosphorylation blocks HP1 binding. J Biol
expression. Neuron. Feb 2006;49:341. Chem. Nov 11, 2005;280:38090.
285. Stühmer T, Puelles L, Ekker M, Rubenstein JL. Expression from 307. Kuzmichev A, et al. Composition and histone substrates of poly-
a Dlx gene enhancer marks adult mouse cortical GABAergic neu- comb repressive group complexes change during cellular differen-
rons. Cereb Cortex. Jan 2002;12:75. tiation. Proc Natl Acad Sci U S A. Feb 8, 2005;102:1859.
286. Horike S, Cai S, Miyano M, Cheng JF, Kohwi-Shigematsu T. Loss 308. Hiragami K, Festenstein R. Heterochromatin protein 1: a pervasive
of silent-chromatin looping and impaired imprinting of DLX5 in controlling influence. Cell Mol Life Sci. Nov 2, 2005;62(23):2711.
Rett syndrome. Nat Genet. Jan 2005;37:31. 309. Le Douarin B, et al. A possible involvement of TIF1 alpha and
287. Borg I, et al. Disruption of Netrin G1 by a balanced chromosome TIF1 beta in the epigenetic control of transcription by nuclear
translocation in a girl with Rett syndrome. Eur J Hum Genet. Aug receptors. Embo J. Dec 2, 1996;15:6701.
2005;13:921. 310. Lechner MS, Schultz DC, Negorev D, Maul GG, Rauscher FJ
288. Evans JC, et al. Early onset seizures and Rett-like features asso- III. The mammalian heterochromatin protein 1 binds diverse
ciated with mutations in CDKL5. Eur J Hum Genet. Oct nuclear proteins through a common motif that targets the
2005;13:1113. chromoshadow domain. Biochem Biophys Res Commun. Jun 17,
289. Tao J, et al. Mutations in the X-linked cyclin-dependent kinase-like 2005;331:929.
5 (CDKL5/STK9) gene are associated with severe neurodevelop- 311. Kourmouli N, Sun YM, van der Sar S, Singh PB, Brown JP.
mental retardation. Am J Hum Genet. Dec 2004;75:1149. Epigenetic regulation of mammalian pericentric heterochromatin in
290. De la Fuente R, Baumann C, Viveiros MM. Role of ATRX in chro- vivo by HP1. Biochem Biophys Res Commun. 2005;337(3):901–907.
matin structure and function: implications for chromosome insta- 312. Gibbons RJ, et al. Mutations in the chromatin-associated protein
bility and human disease. Reproduction. Aug 2011;142:221. ATRX. Hum Mutat. Jun 2008;29:796.
291. Gibbons RJ, et al. Mutations in transcriptional regulator ATRX 313. Law MJ, et al. ATR-X syndrome protein targets tandem repeats
establish the functional significance of a PHD-like domain. Nat and influences allele-specific expression in a size-dependent man-
Genet. Oct 1997;17:146. ner. Cell. Oct 2010;143:367.
6 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
5.
GENES, GENOME, AND DEVELOPMENTAL
MALFORMATIONS
Dhavendra Kumar
D
isorders that affect tissue differentiation, organo- The first group, which includes advances in gene map-
genesis, and morphogenesis constitute a signifi- ping, cloning, and identification of genes, has received
cant proportion of human genetic disease. These a tremendous boost from the Human Genome Project
disorders may result from any of the known genetic mecha- (HGP) (see Chapter 1). Advances in genome sequenc-
nisms and may occur singly, in combination with other ing, particularly exome genome sequencing, following
malformations, or as part of the multisystem complex phe- the successes of HGP and other projects, have helped
notype. Clinical and laboratory analyses of these disorders in identifying mutations and polymorphisms in spe-
have led to the emergence of clinical dysmorphology as a cific human genes associated with particular syndromes.
distinct discipline, now an integral part of medical genetics The HGP has also permitted comparisons between the
and clinical genomics. All practicing pediatricians, clinical human genome and the genomes of other organisms such
geneticists, and genetic counselors are required to have the as mouse and Drosophila, which has helped in the iden-
basic skills in order to deal with patients who present with tification of new genes, deciphering evolutionary con-
a developmental malformation that may occur either singly served genes and genomic regions, and the subsequent
or as part of a recognizable dysmorphic syndrome ( Jones study of their role in abnormal development. The second
and Smith, 2006). group includes a number of highly interactive molecular
During the past two decades, clinical dysmorphologists pathways that govern the developmental process. These
(specialist clinicians dealing with clinical management of molecular pathways include several specific genes and
malformations) have delineated a large number of mal- polymorphisms belonging to a particular molecular fam-
formation syndromes that comprise multiple malforma- ily; for example, RAS/MAPK, NOTCH signaling, and
tions affecting unrelated organs and tissues. Researchers many more. These advances have made it possible to
and clinicians continue to generate more information, develop a better understanding of development and mor-
and the number of related syndromes continues to phogenesis and to work out a plausible genotype–pheno-
increase. Dedicated dysmorphology databases (London type correlation.
Medical Databases, www.lmdatabases.com [London This chapter reviews some of the basic concepts in
Dysmorphology, LDDB and London Ophthalmic developmental biology, which are intricately related to
Genetics, GeneEye], Pictures of Standardized Syndromes genes and genomes. This chapter is not intended to pro-
and Undiagnosed Malformations [www.POSSUM.net. vide the reader with a comprehensive account of normal
au]) are now available to help clinical geneticists and and abnormal human development, as this field is enor-
other clinicians for carrying out comprehensive database mous and beyond the scope of this review. The interested
searches. Although considerable progress has been made in reader may refer to excellent texts available on this subject
understanding the developmental pathology and possible (see “Further Reading”). The main aim of this chapter is to
causes of some of these malformation syndromes, there introduce the subject from the perspective of genomics. It
remains a huge gap in our understanding of the abnormal is anticipated that the reader may find the contents of this
human development and clinical challenges. chapter helpful in understanding aspects of advances in
Recent advances in genetics and genomics have made genomic-related science and technology that are relevant
it possible to understand the molecular basis of develop- to understanding developmental malformations, which
mental malformations. These fall into two broad groups. may in turn influence their clinical practice.
61
www.Ebook777.com
G ENOME P RO J EC T S on the phenotype. Although their role is yet unclear, these
polymorphisms could affect the level or activity of certain
Since the 1990s, the efforts to sequence a number of eukary- classes of proteins through direct or epigenetic influence on
otic genomes have been successful, resulting in the avail- a range of genes or alleles. In humans, some of these dif-
ability of complete genome sequences. In addition, several ferences may be disease-causing or result in developmental
prokaryotic organisms have been sequenced (http://www. malformations. Genomic polymorphisms could be bio-
tigr.org/). Apart from completion of the human genome logically important by virtue of being physically within the
sequence (Lander, Linton, et al., 2001; Istrail, Sutton et al., coding region (exons) of genes, outside of or close to the
2004), genomes that have been sequenced in their entirety promoter region of genes, altering or modifying the func-
include mouse, Drosophila melanogaster (Goldstein and tion of non-coding RNAs, or influencing the structure
Gunawardena, 2000), worm (Caenorhabditis elegans) (Li, or function of the imprinting control regions with major
Stoeckert, et al., 2003), and budding yeast (Saccharomyces epigenomic outcomes. In addition, these sequence poly-
cerevisiae) (Krogan, Cagney, et al., 2006). Despite limited morphisms could be extremely important in the identifica-
information on the precise gene content and gene order tion of both single loci and sets of loci that produce disease.
in the human genome, most of the approximately 23,000 The most common type of polymorphic genomic variation
human genes (see Chapter 2) are comparable to the corre- includes single nucleotide polymorphisms (SNPs) and
sponding mouse genome. It is widely agreed that the avail- copy number variations (CNVs). Several research insti-
ability of genome sequences is of paramount importance tutions and organizations are engaged in collecting data
to our understanding of fundamental structural and func- on SNPs (http://snp.cshl.org/; http://www.humgen.nl/
tional units across a broad range of eukaryotic organisms, SNP_databases.html, and many more) and CNVs (http://
from the fruit fly to higher primates like Homo sapiens. cnv.gene-quantification.info/ and many more) that are
Such understanding has already led to the discovery of sev- correlated with disease states, including human malfor-
eral gene families and gene-sequence polymorphisms that mations. The analysis of such data must take into account
regulate embryonic development and differentiation, and the population structure, since the expression of a particu-
are involved in the pathogenesis of several human malfor- lar allele may depend on the presence or absence of other
mation disorders. genetic variability in the genome. This is a mammoth and
Among the recent initiatives, new relevant genome proj- challenging task, but it is likely to identify which of these
ects include the Human Variome/Phenome Project (www. polymorphisms are associated with abnormal human devel-
hvp.org), the GenPhen (www.Gen2Phen.org) and Sanger’s opment and disease.
the Decipher Developmental Delay (www.DDDUK.org).
The common theme of these projects is to decipher devel-
C O M PA R AT I VE G E N O M I C S A N D
opmental phenotypes in the most comprehensive genomic
HUM A N D EVE L O PM E N T
context. The positive outcomes of these projects could revo-
lutionize the clinical and preventive approach to a number The sequencing of genomes from different organisms now
of human developmental disorders. provides a splendid opportunity to compare the functional
significance of known or predicted genes in one species
with those in another species. A systematic comparison of
G E N O M E S EQ U E N C E
closely related organisms, such as mouse and human, in
P O LY MO R P H I S M S
which most genes are evolutionarily conserved, has revealed
Human beings do not all possess the same genomic the location of homologous genes. This information is made
sequence. Every person, except probably among the mono- available and regularly updated on the Internet (www.infor-
zygotic twins, has a unique sequence that is different from matics.jax.org/; www.geneontology.org).
any other person’s by at least one change in every 500–1000 Similarly, by comparisons between the genomes of dis-
base pairs (bp). This variation or polymorphism presum- tantly related species, such as humans and fruit flies, that
ably does not carry any significance in terms of structural share common developmental processes, a vast amount of
or functional phenotype (however, it could carry some information about the genetic basis of development has
evolutionary biological significance). Nevertheless, the been accumulated. These comparisons have helped in iden-
process of genome sequencing needs to take into account tifying genes belonging to gene/molecular families that
the existence of this variation between individuals, popu- are derived from an ancestral gene in a common branch
lations, and ethnic/racial groups and its possible influence of the evolutionary tree. The genes in such a family are
6 2 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
Free ebooks ==> www.Ebook777.com
called paralogous, and they originate from repeated gene species. Such genes are called orthologous. Using a cluster
duplication events followed by evolutionary divergence of analysis of such proteins, it is possible to determine which
the duplicated genes (Figure 5.1). The proteins encoded of the synthesized proteins are homologous and therefore
by such genes share common sequence characteristics and likely to have similar functions. This type of analysis has
are referred to as belonging to a gene family. The activi- helped in understanding the evolution of gene/ molecular
ties of such proteins may differ slightly, or the proteins families that influence development (Figure 5.2) (Mount,
may perform a different function as a result of mutation 2004).
or natural selection, and they are also likely to have been
produced at different stages of development. The function
of the proteins may also vary in different tissues due to P H Y LO G ENOMICS AND H U MAN
alternative splicing and mRNA editing, causing modifica- DEVELO PMEN T
tion of the mRNA sequence and the subsequent protein
sequence. It is known that the same gene family identified Developmental biologists remain intrigued by the com-
by the same function can occur in two or more different plexities of developmental mechanisms. Among many
Anterior Posterior
B1 B2 B3 B4 B5 B6 B7 B8 B9 B13
HOxB
Human
Anterior Posterior
Evolutionary conservations of genomic organization and expression patterns of Drosophila fruit fly and mammal Hox genes. The human
Figure 5.1
embryo shows anterior-posterior cluster of four Hox genes based on mouse expression studies. The fruit fly shows Drosophila Hox genes aligned with
their mammalian orthologues and corresponding expression patterns mapped onto the body plan. (Robert, 2008)
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 6 3
www.Ebook777.com
Gene
Chromosome
One function A
Gene duplication
A1 A2
Speciation
Two functions A1 A2 A1 A2
Species A Species B
(same function)
Orthologs
Paralogs
(related by gene duplication)
Repeated duplications
Species A Species B
Orthologous pairs
(rest are members of a paralogous family)
Figure 5.2 Orthologous and paralogous origin of gene families through successive gene duplication events. (Reproduced from Mount, 2004, with permission from Oxford
University Press, New York.)
challenging avenues, by far the most complex is molecu- 1) It sets out the direction of evolutionary diversification
lar evolutionary biology. In the genomic context it is also by clarifying historical relationships;
referred to as phylogenomics. The most challenging ques-
2) it provides the necessary genetic and genomic toolkits
tion remains, how can species share similar developmental
for studying the genome expansion and contraction;
genetic or genomic toolkits but still generate diverse life
forms, ranging from an invertebrate worm to a human, 3) it offers logical clarifications to underlying mechanisms
the most advanced vertebrate? Conversely, how can simi- for evolution and developmental functions; and
lar forms develop from different toolkits? To large extent, finally;
advances and unraveling of genomics have offered many
4) it helps in the identification of conserved non-coding
answers. Genomics bridges the gap between evolutionary
elements and their relationship to genome architecture
and developmental biology and thus helps answer several
and development.
questions on the evolution and development (evo-devo)
philosophy of the developmental biology (Cañestro, In order to identify genes that share sequences with sim-
2012). There numerous ways by which phylogenomics has ilar biological function developmental, biologists search the
emerged as a leading sub-field in developmental biology: genomes of different organisms. This search is commonly
6 4 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
referred to as comparative genomics. It is important because 2004), and comparison of organisms with flagella (such
developmental genetic or genomic toolkits in the two dif- as green algae, flies, roundworms, sea squirts, mice, and
ferent organisms are likely to have originated from a com- humans) and organisms that lack flagella (such as plants; Li,
mon ancestral gene in the evolutionary past. It is likely that Gerdes, et al., 2004), identified several hundred candidate
if mutations in one of the genes result in a developmental genes related to cilia or flagella. An exhaustive search led to
malformation, then the other gene presumably has a simi- the detection of more than 80% of ancestral genes that are
lar role in the organism’s development. This inference holds known to be involved in ciliary function. The proteomic
true even for strikingly dissimilar organisms, such as fruit analysis identified a novel family of proteins (OSEG) that
flies and humans. are essential for the development of cilia in Drosophila
Biologists agree that evolutionarily divergent organ- melanogaster (Holland, 2007). Studies in silico, in vitro, and
isms use similar fundamental gene systems and associ- in vivo in Caenorhabditis elegans validated flagella-related
ated encoded proteins during developmental processes. genes, and identified a novel human gene (BBS5) as defec-
Developmental variation is dependent, not solely on the tive in Bardet-Biedl syndrome (Li, Gerdes, et al., 2004). It
structural variations in these genes, but also on gene function is acknowledged that further applications of this genomic
and regulation during evolutionary cycles. This is referred to strategy will facilitate the identification of candidate genes
as the evo-devo concept of development (David, 2004) and that are important for the development and evolution of a
focuses on identifying basic sets of genes in organisms and variety of developmental traits.
studying how they are regulated in developing cells and tis-
sues. One approach to identifying such gene sets is a detailed
functional identification of specific transcription factors and G ENOME ARC H I T EC T U RE AND
cis-acting binding sites for these factors (Davidson, McClay, DEVELO PMEN T
et al., 2003). Another approach is the search for common bio-
logical functions in distantly related organisms, which may Developmental biologists have demonstrated that, by incor-
help in tracing the evolutionary and developmental origins porating comparative genomics, a clear picture of the devel-
of genes. For example, the genes regulating light perception opmentally relevant part of genomes can be built. These
in a microbe are similar to those that regulate the function so-called genomic developmental toolkits include genome
of chloroplasts, and, eventually, the evolution of complex contraction, genome expansion, epigenetic segments of the
functions of the eye (Gehring, 2002). Other evo-devo studies genome, and evolutionarily conserved non-coding elements
support the premise that variations in the genes themselves (Figure 5.4). In simple terms, these elements constitute the
are important for evolution and development. genome “architecture” relevant to the organism’s develop-
Comparison of genomes from different organisms has ment. The net contribution of the genome architecture is
revealed the fascinating phenomenon of genome expansion evidenced by different expression patterns in different spe-
and contraction. Comparative genomics provides a power- cies (orthologues) compared to different expression pat-
ful tool to discover trait-specific genes on the basis of the terns within species (paralogues). Comparative genomics
assumption that most genes that were expressed exclusively has offered an understanding of mechanisms that govern
in a trait are lost if the trait was secondarily lost (Cañestro, difference in expression patterns and the evolutionary sta-
Yokoi, et al., 2007). This hypothesis is supported by the bility of the genome architecture (Cañestro, Yokoi, et al.,
observation in Drosophila melanogaster and C. elegans 2007).
that genes for cilia or flagella in organisms that have them Another strategy for finding evolutionarily relevant
are absent in other organisms that lack these organelles genes is studying the effects of environmental factors, such
(Figure 5.3). as a dietary supplement. For example, a particular strain
The power of comparative genomics is best exemplified of inbred mice can have either a normal or a highly defec-
by studying genes for cilia and flagella, microtubule-based tive skeleton due to the presence of immature cartilage,
organelles that are important for development of left–right simply by altering the diet (Wu, Li, et al., 2008). This is
asymmetries, heart formation, vertebrate photoreceptors, probably related to the suppression of mutant Hox gene
and invertebrate mechano- and chemoreceptors (Pazour expression by a dietary substance, supporting the argument
and Witman, 2003). Comparative genomics of organisms that expression of accumulated mutations in an important
with cilia (such as flies, roundworms, green algae, proto- developmental gene may not occur until accompanied by
zoans, and humans), and organisms that lack cilia (such as an environmental change (Wu, Li, et al., 2008). It is pos-
plants, yeasts, and slime molds; Avidor-Reiss, Maer, et al., sible that some of these evolutionary events may be neutral
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 6 5
Free ebooks ==> www.Ebook777.com
(A)
Trait present (+)
or absent (–) – + + + – + + – +
Trait loss
Trait loss
Trait loss
+
(B)
Genomes lacking the trait – – –
Intersection enriched
in trait-related genes
Figure 5.3
Comparison of genomes with lost traits: (A) identifies candidate trait-related genes that are present in organisms that have the trait but are
absent in organisms that lack the trait; (B) Comparison of different levels of evolutionary distance for trait-specific genes that are considered lost
without the trait (Cañestro, Yokoi, et al., 2007).
Subphylum Cephalochordates Urochordates Vertebrates
Genome diminution
Determinative development, rapid embryogenesis and life cycle
Figure 5.4
Genome contraction and morphology—note reduced the size of urochordates’ genomes, resulting in the loss of temporal collinearity of
Hox-gene expression by breaking up their Hox-gene cluster; loss of the need to use retinoic acid (RA) for anteroposterior axial patterning associated
with the reorganization of their CNS. Larvaceans lack the classic genetic machinery to synthesize, degrade, and detect RA, and they also lack a
complete genetic system for DNA methylation (carried out by DNA methyltransferases; Dnmts). This illustration demonstrates that the genome
contraction builds a complete chordate body plan that is retained throughout life (Cañestro, Yokoi, et al., 2007).
6 6 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
www.Ebook777.com
Eyespot
Sensory
appendages?
Head
Abdomen Tail
Anus/
Genitals
Mouth Gills?
Conserved developmental patterning systems in a hypothetical ancestral creature; Anterior/posterior axis segmental identity by Hox genes;
Figure 5.5
A/P axis patterning by hedgehog genes and through suppression of BMP signaling; dorsal/ventral division by Notch signaling and promotion of
appendage outgrowth by Distalless; and formation of light-sensitive organs by Eyeless/Pax6. (Reproduced from Bier and McGinnis, In ‘Inborn Errors of Development’, Eds.
Epstein, Erickson et al., 2004, with permission from Oxford University Press.)
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 6 7
Table 5.2 TRANSCRIPTION FACTOR FAMILIES AND FUNCTIONS
transcription factor’s activity to be modulated by transcrip- development. In addition to ocular abnormalities, muta-
tion binding protein (TBP), transcription-associated fac- tions in the PAX6 gene can cause severe nervous system and
tors (TAFs), or other transcription factors. pancreatic optic abnormalities (Cvekl and Tamm, 2004).
There are numerous diseases that result from a defi- PAX6-binding sequences have been found in the enhanc-
ciency of transcription factors, often termed “transcription ers of vertebrate lens crystalline genes and in the genes for
factoropathy” (Gilbert, 2004). The first such human dis- insulin, glucagon, and somatostatin, which are expressed in
ease was androgen insensitivity syndrome (AIS), in which, the endothelial cells of the pancreas.
despite normal testosterone production, the affected male There are some basic principles that govern the role
externally develops as a phenotypical female and fails to played by transcription factors. Firstly, transcription fac-
develop secondary sexual characteristics. In AIS, the testos- tors function in combination with other transcription fac-
terone receptor is either absent or deficient, and its DNA tors. Secondly, transcription factors find their way to the
fails to bind the DNA of male-specific genes (Migeon, nucleus either through cell lineage or induction. Thirdly,
Brown, et al., 1981). Conversely, the binding of DNA transcription factors can continue to be synthesized after
and consequent activation of the receptor site can lead the original signal has ceased. Finally, post-translational
to Waardenburg syndrome, type II. In this disorder, the modification is often necessary to ensure adequate func-
affected heterozygotes have a white forelock, are deaf, and tioning of the transcriptional factors. This mechanism
have multicolored irides. They possess a wild-type allele is probably of major importance in differentiation and
of the microphthalmia (MITF) gene. Activation of this morphogenesis.
transcription factor through the protein tyrosine kinase
cascade enables it to dimerize and bind to the regulatory
OT H E R M EC H A N I S M S
regions of particular genes that open a region of DNA for
transcription. Differential transcription of DNA is not solely dependent
Transcription factors work in conjunction with other on transcriptional factors. There are other mechanisms that
transcription factors to activate particular genes. However, influence developmental gene regulation. Even though
the binding of a specific transcription factor to the enhancer a particular RNA transcript may be synthesized, it is not
or promoter of a gene does not always cause transcription of always possible to generate a functional protein. In order
the gene. Some of these transcription factors, called “mas- for the mRNA to become an active protein, a number of
ter regulatory genes,” are important and take the lead in the other steps are important, including processing of mRNA
transcription process. For example, PAX6 (eye) and MYOD by removal of introns, translocation from the nucleus to the
(muscles) work in concert to initiate cellular differentia- cytoplasm, translation by the protein-synthesizing appara-
tion. The use of PAX6 by different organs illustrates the tus, and post-translational modification to make an active
modular nature of transcriptional regulatory units. PAX6 is protein molecule. Regulation can occur at any of these steps
needed for mammalian eye, nervous system, and pancreatic during development.
6 8 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
RE GU LAT ION OF EM B RYO G ENESIS in the eye, Pax6 acts as the inducer for the ectoderm in the
AND OR G ANO G ENESIS optic vesicle, which in turn acts with fibroblast growth fac-
tor-8 (FGF8) and other factors (Sox2, Sox3, and L-Maf )
Soon after fertilization, the rapid cell division in the result- to ensure the production of the lens. Further induction
ing embryo paves the way for the formation of ectoderm, is called instructive, when the responding tissue depends
mesoderm, and endoderm, the three primordial germ lay- on a specific tissue to begin the process. In general, there
ers. Early embryogenesis is regulated by a complex system are three broad defining principles of instructive induction
of proteins that regulate complex processes of differentia- (Souza, Kuliszewski, et al., 1995): (1) tissue A is necessary
tion and morphogenesis (Table 5.1; Figure 5.6). Toward for tissue B to respond in a desired manner; (2) tissue B
the end of embryonic development and differentiation, does not respond in the desired manner in the absence of
specific organ formation is initiated. Organogenesis is tissue A; (3) tissue B may not respond in the desired man-
highly complex, as different organs are composed of tissues ner in the presence of another tissue, but in the absence
derived from different primordial layers. Disruption in one of tissue A. An example of this form of instruction is in
of these tissues will lead to either a structural malformation the optic vesicle, which, when placed in another part of
or a functional abnormality in the organ. For example, in the developing head ectoderm, can form an ectopic lens.
the eye, a precise arrangement of tissues forming the trans- Induction also depends on environmental factors, which is
parent cornea, lens, vitreous, choroid, and neural retina is called permissive induction.
necessary for normal shape and function. Thus, construc-
tion of organs is accomplished by a group of cells chang-
R EG I O NA L S P EC I FI C IT Y O F I N D U C T I O N
ing the behavior of an adjacent set of cells, causing them to
change their shape, mitotic rate, and eventual fate. This is Induction is a dynamic process that governs the early cell and
called interaction. The interaction between closely located tissue differentiation leading to specific organ formation. It
cells or tissues is referred to as proximate interaction. This particularly involves an interaction between various tissues,
process is continued throughout organogenesis. Proximate particularly those that lie adjacent to each other (Gilbert,
interaction consists of two components—inducers, the tis- 2004). For example, the interaction between epithelial cells
sue producing the signal, and responders, the tissue being and mesenchymal cells is probably the most important in the
induced. The ability of the tissue to respond to the induc- development of several organs (Table 5.3); the best example
tion signal is called competence. It is an active process, of which is the skin. Skin comprises epidermis (epithelial)
as the responding tissue undergoes several changes and and dermis (mesodermal), developed from the interactions
interaction with other factors to ensure formation of the of sheets of epithelial cells and mesenchymal cells derived
intended organ. For example, in the formation of the lens from the mesoderm (see also Chapters 39.1 and 39.2). This
Cytoplasm P P
Extracellular
Adapter
protein
Cytoplasm Inacive P P Ras
responding Acive
protein tyrosine GDP ERK Mitf Melanoblast-
Dormant ATP kinase
specific gene
tyrosine
kinase MEK
ADP Ras GTP
domain
p300/CBP
P
Acive P P
responding RAF
protein Transcription
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 6 9
Table 5.3 ORGANS DERIVED FROM EPITHELIAL–MESENCHYMAL INTERACTIONS (GILBERT, 2004)
phenomenon is regionally specific. For example, the devel- factors activate a set of receptor tyrosine kinases, namely the
oping epidermis signals the underlying dermis, probably FGF receptors (FGFRs). Mutations in some of the FGFRs
through Sonic hedgehog and transforming growth factor result in certain skeletal disorders. For example, mutations
β (TGF-β) proteins, and the condensed dermal mesen- in FGFR3 result in sporadic lethal thanotophoric dysplasia
chyme responds by secreting factors that cause the epider- and autosomal dominant achondroplasia. Receptor tyro-
mis to form regionally specific cutaneous structures. These sine kinases are proteins that extend from the cell surface to
structures could be any of the skin appendages, such as hair, the nucleus. The extracellular part binds with FGFs, and the
nails, or sweat glands (see Chapter 45). Several other organs intracellular component activates dormant tyrosine kinase.
develop from such interactions where the mesenchymal FGFs are associated with a number of developmental func-
cells take the lead in instructing different sets of genes in the tions, including angiogenesis (blood vessel formation),
responding epithelial cells. mesoderm formation, and axon extension. FGF2 is particu-
larly important for angiogenesis, and FGF8 is important for
the development of the mid-brain, eyes, and limbs (Gilbert,
PA R AC R I N E FAC TO R S 2004).
Transmission of signals from the inducer to the responder
is a complex process and depends on several factors. The H E D G E H O G P ROT E I NS
interaction is juxtacrine when cell membrane proteins of
The hedgehog proteins are a family of important paracrine
the responding cell are physically in close contact with the
factors that induce particular cell types and create bound-
cell membrane proteins of the inducing cell. In contrast,
aries between tissues. There are at least three homologues
paracrine interaction depends on the diffusion of proteins
known for the Drosophila hedgehog gene—sonic hedgehog
synthesized by the inducing cell over to the cell membrane
(shh), desert hedgehog (dhh), and Indian hedgehog (ihh).
of the responding cell. This process involves several special
Desert hedgehog is expressed in the Sertoli cells of the tes-
kinds of protein families, collectively referred to as para-
tes, and mice homozygous for a null allele of dhh exhibit
crine factors. Essentially, these are growth and differentia-
abnormal spermatogenesis. Indian hedgehog is expressed in
tion factors (GDFs). Paracrine factors differ from endocrine
the gut and cartilage, and is important for postnatal skeletal
factors (hormones), as they do not travel through the
growth.
blood but are secreted into spaces surrounding the tar-
Sonic hedgehog is perhaps the most widely used hedge-
get cells. These proteins are inducers and are biologically
hog protein. It is expressed in the developing notochord,
similar throughout the animal kingdom, from the fruit fly
and is responsible for the patterning of the neural tube in
Drosophila to humans. There are four major classes of pro-
such a manner that the ventral neurons develop motor neu-
tein families that comprise the majority of the paracrine
rons and the sensory neurons are formed from the dorsal
factors.
neurons (Yamada, Pfaff, et al., 1993). It is also responsible
for patterning the somites. Sonic hedgehog is crucial for the
formation of the left–right axis in many vertebrates. It initi-
FI B RO B L A S T G ROW T H FAC TO R S
ates the anterior–posterior axis in limbs, induces regional
Several FGF genes are important for mammalian develop- specificity in the gut, and induces hair formation (Gilbert,
ment. The FGFs code for specific proteins, of which there 2004). Sonic hedgehog works in conjunction with other
are a number of isoforms produced by alternate RNA splic- paracrine factors, for example, Wnt (wingless integrated)
ing or varying initiation codons in different tissues. These and FGF proteins.
7 0 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
Free ebooks ==> www.Ebook777.com
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 7 1
www.Ebook777.com
(A) (B) (C) (D) (E) (F)
RTK-MAPK
Figure 5.7
Five of the major signal-transduction pathways through which signals from the cell surface are sent into the nucleus. (A) The receptor
tyrosine kinase-mitogen-activated protein kinase (RTK-MAPK pathway); (B) the Smad pathway used by transforming growth factor β (TGF-β)
superfamily proteins; (C) the JAK-STAT pathway; (D) the Wnt pathway; (E) the Hedgehog pathway; and (F) one of the apoptosis pathways used
by mammalian neurons. Abbreviations: ERK, extracellular signal-regulated kinase; GNRP, guanine nucleotide-releasing protein; GSK, glycogen synthase kinase; JAK, Janus kinase; MAPK, mitogen-activated
protein kinase pathway; MEK, MAPK/ERK-kinases; STAT, signal transduction and activator of transcription.
role in the assembly of the extracellular matrix, promoting other extracellular matrix molecules. Another major cel-
cell adhesion and growth, changing cell shape, and permit- lular function of the extracellular matrix is the ability to
ting cell migration. regulate differentiation of the chondrocytes to produce the
In addition to the above-mentioned mechanical roles, cartilage for developing vertebrae and limbs, which is also
the extracellular matrix also plays a role in regulating gene achieved by binding to the integrin. In the absence of inte-
function. It helps in inducing specific gene expression in grin or experimentally blocking the binding of integrin, the
developing tissues; for instance, liver, testis, and mammary developing chondrocytes fail to differentiate into cartilage
glands, through binding the cell substrate that is important and bone (Schwab, Kasper, et al., 2000). Lastly, it has also
for specific transcription factors. The extracellular matrix been shown that branching of some parenchymal organs,
also has a role in inhibiting apoptosis through integrin, such as kidney and lungs, depends on the extracellular
which is the cell membrane receptor for fibronectin and matrix (Kumar, 2008).
7 2 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
A P O P TO S I S A N D D EVE L O PM E N T It is now confirmed that CED-3 and CED-4 proteins
act at the center of the apoptosis pathway. These regulate
Apoptosis, or programmed cell death, is a normal part of
initiation of other genes in the pathway, such as BMP-4.
development. Cells in all animals are programmed to die
Homologues for these proteins are important for auto-cell
every day, and approximately equal numbers of cells are
digestion. The CED-4 protein homologue is called Apaf-1
replaced. For example, adult humans lose as many as 1011
(apoptotic protease activating factor-1), which participates
cells every day, and these are regularly replaced by other
in the cytochrome-dependent activation of the mamma-
cells. It is estimated that the total weight lost every year
lian CED-3 homologues caspase-9 and caspase-3 ( Joza,
through programmed cell death could be equal to the adult
Susin, et al., 2001). Activation of these caspases causes cell
body weight. Apoptosis begins immediately after birth. It
auto-digestion, leading to cell death. Mice homozygous
is estimated that the total number of neurons accumulated
for Apaf-1 deletions have severe craniofacial abnormalities,
throughout the gestation period of nine months is approxi-
brain overgrowth, and syndactyly (webbing between toes).
mately three times that in an adult of average intelligence.
It is important to appreciate that apoptosis can follow
Programmed cell death is a continuous process. It is
more than one pathway. For instance, the “death domain,”
essential for creating proper spaces within an organ, as
containing receptors of the tumor necrosis factor (TNF)
well as between organs or body parts. Examples include
family, can induce apoptosis in several cell systems that can
the middle-ear space, the separation of digits to create the
also be triggered by other apoptosis-inducing factors. This is
proper shape and size of fingers and toes, and the lower
accomplished by blocking the anti-apoptosis signals sent by
vaginal space and opening (Newton and Strasser, 1998).
other factors. One of the developmentally important TNF
Clearly, through apoptosis, redundant tissues and struc-
receptors with a death domain is Edar, a protein required
tures are pruned away. Different tissues use varying signals
for the development of hair, teeth, and other cutaneous
for apoptosis. Among vertebrates, the BMP4-mediated
appendages. Mutations of this gene or its ligand, Eda, cause
signals are important. For example, the connective tissues
X-linked hypohidrotic ectodermal hypoplasia (Online
respond to BMP4 to differentiate into bone. Similarly, the
Mendelian Inheritance in Man [OMIM] 305100), a syn-
surface ectoderm responds to this by differentiating into
drome characterized by lack of sweat glands, sparse hairs,
skin. Another good example is the development of tooth
and poorly formed teeth. An identical syndrome results
enamel. After the tooth cusp has grown, the enamel knot
from deficiency of the adapter protein that binds the death
synthesizes BMP4, which, through apoptosis, stops further
domain of this receptor (Headon, Emmal, et al., 2001).
enamel differentiation (Vaahtokari, Aberg, et al., 1996). As
Instead of resulting in cell death, the activation of the recep-
previously described, the erythropoietin-induced red blood
tor enables continued development of skin appendages.
cell population is programmed for apoptosis. In its absence,
the red cells will undergo apoptosis. This works through
I N FLU E N C E O F E N VI RO N M E N TA L
the JAK-STAT signal-transduction pathway (see previous
A N D O P P O RT U N I S T I C FAC TO R S O N
section).
D EVE L O PM E N T
Apoptosis works through several pathways. One of the
pathways is regulated by genes that were discovered from Developmental biologists agree that the environment plays
studies on C. elegans, appropriately designated as ced-3 and a significant part in producing a phenotype. Nutritional fac-
ced-4 genes. The gene product of these two genes initiates tors undoubtedly result in a number of disease phenotypes,
apoptosis. However, the protein product of another gene such as marasmus, kwashiorkor, rickets, diabetes mellitus,
(ced-9) is shown to inhibit the programmed cell death. and coronary heart disease. There are several genetic factors
Mutations in this gene will accentuate apoptosis by with- that are equally important for creating a pathophysiological
drawing the control. This has been confirmed experimen- state, which predispose to morbid effects of dietary factors
tally when inactivated CED-9 protein led to the death of (see Chapter 12). Dietary supplementation and modifica-
an entire embryo. On the contrary, gain-of-function ced-9 tion are known to significantly alter the phenotype. For
mutations can help cells survive that would have other- example, a normal daily dietary intake of vitamin C pre-
wise died. In other words, wild ced-9 acts as a binary switch vents the development of the clinical effects of vitamin C
between life and death at the cellular level. In mammals, deficiency, as human beings lack naturally occurring vita-
members of the Bcl-2 gene family are the CED-9 protein min C (hypoascorbemia, OMIM 240400) due to deficient
homologues. This gene family is important for red blood gulonolactone oxidase as a result of a mutation in the gulo-
cell development and differentiation. nolactone oxidase gene on the short arm of chromosome
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 7 3
8. This can result in severe childhood connective tissue of phenylalanine can be significantly reduced by dietary
disease leading to death. Gulonic acid oxidase enzyme is restriction of phenylalanine.
the final enzyme leading to the synthesis of ascorbic acid The genomes of primate mammals determine their
(vitamin C). In contrast to humans, several other mammals final physical shape, which is also under the direct influ-
have normal gulonolactone oxidase enzyme activity offer- ence of the environment. For example, the facial phenotype
ing natural protection from the clinical effects of vitamin depends on firm and regular chewing, which stimulates the
C deficiency. facial muscle and bone (maxilla and mandible) develop-
Periconceptional folic acid supplementation is now ment (Corruccini, 1990). The increased prevalence of orth-
commonplace for the prevention of recurrent neural tube odontic problems in young children and adults in modern
defects, and probably even for the primary prevention times is attributed to a soft or mid-textured diet. This has
of some other congenital anomalies. Fetuses with muta- been shown in experimental primates who were fed a soft
tions in genes associated with folate metabolism are at an diet and developed lower jaw malocclusion similar to that
increased risk for neural tube defects (De Marco, Calevo, in children requiring orthodontic treatment (Corruccini,
et al., 2003). One such gene is methylene tetrahydrofolate Whitley, et al., 1985).
reductase (MTHFR), which incorporates folic acid in the In brief, the production of a phenotype such as a devel-
methylation of homocysteine to methionine. Mutations or opmental malformation depends on the genotype, which is
polymorphisms in this gene result in increased homocyste- regulated at numerous levels. The cellular phenotype is the
ine levels, resulting in peripheral vascular disease, and are direct consequence of the genome within the cell and the
associated with myocardial infarction. However, the mech- fate of the community of cells in which it resides. It is also
anisms by which folate deficiency or lack of bio-availability argued that probably even the environment can alter gene
result in neural tube defect are not fully understood. expression (Gilbert and Epel, 2009)!
Several developmental malformation syndromes are
associated with defects in cholesterol biosynthesis. One
MO D E L O RG A N I S M S F O R
of the enzymes in this pathway is 7-dehydoxycholesterol
U N D E R S TA N D I N G D EVE L O PM E N T
reductase. Mutations in the gene for this enzyme (7DHCR)
result in lack of downregulation of Sonic hedgehog (SHh), For several years, geneticists have used a number of model
resulting in a number of abnormal phenotypes such as organisms to dissect how genes control metabolism, repro-
Smith-Lemli-Opitz syndrome (Figure 5.8). It is possible to duction, and development. Similarly, using the same model
ameliorate some of the deleterious effects of this downregu- organisms, biochemists and molecular biologists have dis-
lation by dietary cholesterol supplementation. Similarly, covered mechanisms through which the protein products
dietary restriction of the excess metabolite in a number of of these genes regulate biological processes. These model
inherited metabolic conditions can alter the phenotype. An organisms include different microbes, the budding yeast
excellent example is that of phenylketonuria, in which the (S. cerevisiae), the worm C. elegans (nematode), the plant
behavioral and cognitive effects of excessive accumulation Arabidopsis thaliana, and the fruit fly D. melanogaster. For
Smith–Lemli–Opitz (DHCR7)
Cytoplasm
107
7-DHCR Nucleus
76
Pallister–Hall (GLI3)
SHH PTCH GLI acivator
DHCR7 CR
SMO
Figure 5.8 The gene pathway of 7-dehydroxy cholesterol reductase, Sonic Hedgehog, and GLI3 complex (Brunner and van Driel, 2004).
7 4 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
humans, the mouse model, Mus musculus, is ideal, since unicellular organisms have helped us in understanding
the two species are closely related through evolution. With enzymes that are involved in complex energy metabolism
the rapid progress in DNA sequencing, sequences of the defects resulting in neuromuscular disorders (Berardo,
best-understood genes were deposited into sequence data- DiMauro, et al., 2010).
bases such as GeneBank, maintained at the National Center Studies on invertebrate model organisms, for example,
of Biotechnology Information (NCBI; http://www.ncbi. the Drosophila fruit fly and the nematode (C. elegans),
nlm.nih.gov). This information is used to predict an mRNA have contributed to our understanding of several basic
sequence, which can be readily used to predict the amino biological mechanisms, such as the organization of genes
acid sequence of the corresponding protein. These pro- into independently segregating linear chromosomes, the
tein sequences are known as the sequence signature for that creation of the first chromosome “maps,” the one gene–
particular protein in other organisms. Thus, homologous one protein hypothesis, radiation-induced mutagenesis,
genes in other organisms can be found through searching the principles of pattern formation, and the identification
for nucleic acid sequences that, when translated, produce a of genetic pathways implicated in human disease. As both
similar protein sequence. This newly identified gene is then fruit flies and nematodes have closely related gene coun-
predicted to have similar biological function in the second terparts to many human disease genes, the identification
organism. of new genes in these invertebrates will help define new
Complete genome sequences of model organisms and candidate disease genes that are likely to be involved in
the human genome sequence are available on GeneBank, the same disease processes. A useful example is the Notch
which is now used in searching for new genes and the cor- signaling pathway that has been shown to be important
responding proteins. A newly identified gene sequence in invertebrate development. Mutations in the compo-
can be compared to the existing database of sequences nent genes of the Notch signaling pathway have been
by aligning the predicted protein sequences. If a gene of shown to result in notching of the wing margin in fruit
unknown function is similar to another of a known func- flies and defects in vulval development in worms (Gupta,
tion, the newly discovered gene can be predicted to have Wang, et al., 2003). Notched wings were also observed
similar function. However, not all gene functions can be in mutations in the ligand Delta, the Notch receptor, or
predicted in this manner. In some situations, new genes the signal transducer suppressor (Hairless). Vertebrates
of unknown function are found in two organisms. This have several common paralogues of the Notch signaling
problem can be alleviated by building a protein domain pathway components. Reduced function of the related
knowledge base that can be searched for amino acid sig- ligand, Delta3, or the Notch homologue (Notch1) itself
natures indicative of structure or biochemical activity. results in axial skeletal malformations. A good example is
If present, these signatures provide a clue to the gene spondylocostal dysostosis, an autosomal recessive heter-
function. If two predicted protein sequences from two ogenous condition (OMIM 277300, 608681, 609813).
genomes are similar, then their amino acid alignments The characteristic feature of this disease is the spinal seg-
would be the same. In such a situation, the genes can be mental anomaly, which is also associated with anal and
predicted to have similar functions. urogenital anomalies (OMIM 271520).
Biological makeup of the model organism determines One of the major contributions of studies in inverte-
the extent of information applicable to human develop- brate model organisms has been the discovery of homeo-
ment. Unicellular organisms are an excellent model for tic selector genes, now referred to as Hox genes. The term
studying eukaryotic cell function. For example, genetic homeostasis was coined by William Bateson (1894) for
studies in the yeast (S. cerevisiae) and slime molds the phenomenon in which one segment of an organism is
(Dictyostelium discoideum) can be carried out and repeated “transformed in whole or part to another” (Reid, 2004).
several times, as a billion progeny can be produced in a The genetic basis of these transformations is explained by
relatively short time. These studies provide vital informa- mutations in Hox genes. Systematic analyses of Hox gene
tion about intragenic primary and secondary suppressor mutations in the Drosophila fruit fly revealed an extra
loci. Therefore, unicellular organisms are immensely use- pair of wings due to mutations in the Ultrabithorax (Ubx)
ful in establishing the networks of gene action involved in gene and an extra thoracic leg attached to the head result-
basic cell biological processes. However, such studies have ing from dominant mutations in the Antennapedia (Antp)
limited applications in studying complex cellular func- gene (Gehring, Kloter, et al., 2009). Molecular analysis of
tions such as those of the nervous system, which depend the genomes has revealed that humans and other bilateral
on the interaction between cells. Nevertheless, studies on animals have multiple Hox genes (Figure 5.1), which carry
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 7 5
a common DNA sequence motif called the homeobox. The multiple malformations that include iris defects, pigmenta-
homeobox motif encodes a similar 60-amino acid motif in tion abnormalities, deafness, and the inability to produce a
Hox proteins, termed the homeodomain. The Hox proteins normal number of mast cells. Moreover, these abnormali-
belong to the transcription factor family (Table 5.2). These ties are not related to each other and can occur indepen-
exert their function through activation and repression of dently. This occurs because all body parts can use the MITF
multiple target genes. Arrangement of these genes is strik- protein as a transcription factor. This type of pleiotropy is
ingly similar in the fruit fly and humans. These genes are called mosaic pleiotropy, as the relevant organ or body part
arranged in clusters. There is evidence that this clustered is separately affected by the mutant gene. In contrast, some
arrangement of Hox genes has been maintained for more malformations in the related part do not result directly
than 500 million years, because different genes in the clus- from the abnormal gene function, as the mutant protein
ters are controlled by the same cis-acting DNA regulatory is not expressed. For example, the failure of MITF expres-
regions. In general, there are four clusters—HOXA, HOXB, sion results in the pigmented retina’s not being fully differ-
HOXC, and HOXD. Each gene has a role in the ante- entiated. This in turn causes a malformation involving the
rior–posterior axis patterning of various organs and body choroidal tissue, which results in drainage of the vitreous
parts. This is evident as specific human malformation syn- humor fluid. This further leads to failure of ocular develop-
dromes are now recognized to be associated with HOXA ment, causing microphthalmia (small eye). This phenom-
(OMIM 609296; 601536), HOXB (OMIM 249000, enon, in which several developing tissues or organs might
Meckel syndrome, MKS1), and HOXD (OMIM 606708, be sequentially affected even though they do not express
split hand-foot-absent uterus syndrome; OMIM 127300, the mutant gene, is called relational pleiotropy. This concept
Leri-Well dyschondrosteosis; OMIM 113200, brachydac- is important in dealing with complex clinical genetic situ-
tyly type D; OMIM 112500, brachydactyly type A1). In ations, particularly in prenatal diagnosis where the predic-
addition to these malformations, the HOX genes interact tion of phenotype is important in making informed choices.
with several other patterning genes with a crucial role in
development.
G E N ET I C H ET E RO G E N E IT Y
7 6 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
phenotype only occurs in the heterozygous state. In other microfibrils in elastic connective tissue. The presence of
words, the homozygous state never exists, probably because even minute amounts of mutant fibrillin prohibits the asso-
it is lethal to the embryo. There are several possibilities that ciation of wild-type fibrillin into microfibrils (Watt and
can result in a dominant phenotype (Wilkie, 1997). Chung, 2009).
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 7 7
common phenotypical features (Oti and Brunner, 2007). Phenotypical similarities in different syndromes
The syndrome-family approach was first systematically belonging to a so-called syndrome family could be a reli-
applied to skeletal dysplasias; for example, the family of able indicator of shared biological mechanisms. Apparently,
chondrodysplasias includes several distinct skeletal dyspla- single-gene human genetic disease could in fact be associated
sias such as achondrodysplasia, hypochondrodysplasia, and with mutations in different genes, probably contributing to
so on (Spranger, Winterpacht, et al., 1994). The advances in a common molecular pathway. This was first demonstrated
molecular genetics have vindicated this concept. For exam- in familial elliptocytosis, because linkage to the Rhesus (Rh)
ple, mutations in FGFR3 result in three distinct members blood group was not seen in all families (Morton, 1956).
of the chondrodysplasia family. The converse is true for the It is now clear that nonallelic heterogeneity is extensive in
phenotype of the Stickler-Kniest family, which is related human disease. Although genetic heterogeneity could be
to mutations in three different collagen genes (COL2A1, problematic in conducting familial genetic studies, this can
COL11A2, and COL11A1) (Reginato and Olsen, 2002). be viewed positively, as it might reflect interactions at the
Thus, rapid advances in molecular genetics and genomics protein level: for example, ligand–receptor interactions, dif-
could see merging of syndromes, more splitting of syn- ferent subunits of a multiprotein complex, or proteins that
drome families, and even the complete disappearance of function at different steps of a metabolic pathway. Using
syndromes. However, the description and definition of this strategy, the other genes could be found once the first
syndromes and syndrome families is important. Several gene is discovered. It is also possible that unrelated genes
genetic databases (McKusick’s OMIM, the LDDB, and that result in the same phenotype could also be found and
POSSUM) continue to record these syndromes as inde- will ultimately be shown to have a functional relationship.
pendent entities. Clinical classification is paramount, followed by molecular
There are approximately 200 Mendelian syndromes in verification. In other words, defining the phenotype and
man, which appear in OMIM and other databases. Each syndrome identification could become a functional genom-
syndrome is recognized by a specific phenotypical feature ics tool (Brunner and van Driel, 2004).
or pattern. Some syndromes differ only by a few features.
It is acceptable logic that there could be some biological
T R A N S L AT I O NA L R E S E A RC H
relationship in syndrome families that share the same phe-
I N DYS MO R P H O L O GY
notype. It is successfully argued that a systematic analysis
of phenotypical relationships could be applied in the iden- Clinical studies on human multiple malformation syn-
tification of new genes, providing clues to gene interactions, dromes are not only helpful in medical management but
molecular pathways, and functions (Brunner and van Driel, are also a useful resource for research in understanding the
2004). This has been applied in large-scale mutagenesis pro- basic mechanism of human development, and, eventually,
grams that aim to define the function of genes in a genome mammalian development in general. This approach allows
in relation to mutant phenotypes. This has been completed researchers to gain insight into basic mechanisms of devel-
in yeast, and work is underway for other model systems opment and how genes program organisms to achieve per-
such as C. elegans, the mouse, and the zebrafish. Although manent or adult morphological shapes (Martínez‐Frías,
this strategy has been successful, it is not clear how many 2004). A number of malformation syndromes have over-
different mutants will be required for such screens to be lapping manifestations, despite being phenotypically
comprehensive. It is accepted that creating a single knock- and genetically dissimilar. This fact can be used in basic
out mouse model might not be sufficient to probe a specific research; for example, on developing animal model systems,
gene function in development and homeostasis. A compre- such as fruit flies, mice, worms, and other simple organisms.
hensive analysis of the mutant phenotype would require The data thus generated can be applied in the clinical set-
studying the functional effects of several mutations. Starting ting to understand both the problems patients suffer and
from interesting phenotypical differences and then compar- the mechanisms of development. This is referred to as trans-
ing the underlying mutations might be more productive lational research. However, there are limitations to using
(Brunner and van Driel, 2004). Spontaneous mutations are human subjects in translational research. In humans, it is
frequent, and studying the phenotypical effect can contrib- obviously not possible or ethically appropriate to experi-
ute to our understanding of gene function. Thus, it is impor- mentally manipulate genes to test hypotheses. In addition,
tant to adopt a “phenotype-driven” approach that saturates there is inevitable difficulty in obtaining tissues or organs
the genome with mutations, either experimentally shown in for study due to the lack of consent and/or availability. The
animal models or observed in human disease states. final limitation or disadvantage in carrying out studies on
7 8 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
humans is the enormous cost and length of time it can take wing-vein patterning and lethal malformations manifesting
to complete the study. Moreover, as a genetic model system, with larval death (Kinzler, Ruppert, et al., 1988). It trans-
the human has a long generation time; hence, individuals pired that ci belonged to the genetic pathway downstream
and families must be followed over long periods of time. of the hedgehog signaling molecule (Hh). It was shown that
The typical example of human translational research is the ci protein negatively controlled downstream genes, and
to creatively apply a basic laboratory discovery to the clini- the cleaved ci protein turned off the expression of down-
cal management of a patient with the clinical phenotype of stream genes.
a particular dysmorphic syndrome. This approach would When the same information was applied to humans,
allow testing of the hypothesis as well as providing an it soon became apparent that the truncated GLI3 protein
opportunity to validate the basic laboratory finding. Any had a different clinical outcome (PHS) compared to that
further improvement or information can be fed back to the resulting from haploinsufficiency of the gene (GCPS). In
patient, and the system thus works bidirectionally. Such other words, PHS phenotype was due to the qualitative
a system would allow extracting the maximum amount change, and the quantitative change led to the GCPS phe-
of basic information from the patient. This could then be notype. It is now shown that GLI3 protein, like ci protein,
applied back in the clinical setting to explain the cause is proteolytically processed (Liu, Wang, et al., 2005). This
and the likely outcome of the diagnosis. Thus, researchers example illustrates that molecular studies in rare devel-
can develop novel hypotheses about the mechanisms of opmental disorders like PHS and GCPS can illustrate
development, modifying variables, or other disorders with basic mechanisms of mammalian development. The phe-
overlapping manifestations. The aim is to keep the cycle in notype related to GLI3 mutations also includes polydac-
motion, continually collecting and examining new infor- tyly, imperforate anus, and vertebral anomalies (OMIM
mation, and constantly applying it back in improved clini- 174100).
cal care. The genetic pathway involving GLI3 also includes
The above concept has been applied in a number of the DCHR7 gene (Figure 5.6), mutations in which cause
multiple malformation syndromes with overlapping phe- Smith-Lemli-Opitz syndrome (OMIM 270400), one
notypical manifestations, or families of syndromes (see of the malformation syndromes related to disruption in
previous section). One example is the Pallister-Hall syn- cholesterol biosynthesis. The DCHR7 gene codes for an
drome (PHS), a typical syndrome family disorder in which enzyme called 7-dehydroxycholesterol reductase, which reg-
affected individuals can present with one or more of a range ulates sonic hedgehog (SHH) gene function. One of the
of malformations that include hypothalamic hamartoma, downstream effectors of SHH is GLI3. Other SHH protein
imperforate anus, laryngeal anomalies, and central polydac- homologues include patched protein homologue (PTCH)
tyly, with shortened terminal digits (Biesecker and Graham, and smoothed homologue precursor (SMO), both of
1996). The disorder is inherited in an autosomal dominant which play key roles in regulating GLI3 gene function. It
manner with significant inter- and intra-familial variabil- is likely that the phenotypical similarity in other malforma-
ity. Clinical and genetic studies in families affected with tion syndromes like Optiz G (OMIM 145410) and Mohr
PHS revealed mutations in the zinc finger transcription syndrome (OMIM 252100) could be due to mutations in
factor gene, GLI3 (Kang, Graham, et al., 1997). This was other genes in this gene family (Brunner and van Driel,
an interesting finding, as previously this gene was inciden- 2004).
tally found to be causally related to the Greig cephalopoly- Another illustrative example is McKusick-Kauffman
syndactyly syndrome (GCPS) (Vortkamp, Gessler, et al., syndrome (MKS), which was first described among the Old
1991). Patients with GCPS had balanced chromosomal Order Amish of Lancaster County, Pennsylvania (Biesecker,
rearrangements that had apparently disrupted the GLI3 2002). This disorder is inherited in an autosomal recessive
gene. It is now accepted that GCPS is a distinct develop- manner, and the phenotype includes polydactyly (central
mental syndrome caused by haploinsufficiency of GLI3. and post-axial), congenital heart disease, and hydrome-
This was a challenging observation, as etiologically trocolpos due to congenital uterine outflow obstruction.
significant mutations in the same gene were found in two This disorder is rare, with fewer than 100 cases described
distinct genetic syndromes. The research group led by in the literature. It is likely that the disorder, or a pheno-
Dr. Leslie Biesecker at the National Institutes of Health typically similar disorder, probably occurs in other inbred
(NIH) discovered that mutations in Cubitus interruptus population groups. Increased incidence of autosomal reces-
or ci, the Drosophila homologue of GLI3, were linked to a sive disorders, including multiple malformation syndromes,
wide range of phenotypes in fruit flies, including abnormal is recognized among the Amish and other highly inbred
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 7 9
ethnic population groups. These populations groups are primates. The mechanisms are similar by which the individ-
ideal for conducting homozygosity mapping studies. ual genome specifies the physical and functional state of the
Using the whole-genome-wide scan with 385 markers, human body. This chapter, in brief, summarizes a number
the gene for MKS was mapped to chromosome 20. Further of gene families and numerous related protein families that
molecular studies identified two substitution mutations in regulate the complex process of development. The availabil-
a single mutant chromosome in one of the candidate genes. ity of human and mouse genomes and that of several other
However, this was not associated with a known function. small organisms and animals have provided developmental
Since the Amish are closely related, it was not possible to biologists with powerful new tools for identifying genes,
prove that the sequence variants were pathogenic. A search and their mutations, that control development. The inter-
for non-Amish cases was then made. This proved difficult, ested reader is urged to explore other literature in this com-
as the reported cases were either deceased or no longer avail- plex and stimulating field.
able. In some cases the diagnosis was changed. Eventually,
a newborn girl with features of MKS was recruited to the
study, and was found to have a 2-bp deletion on the same F U RT H ER READIN G
allele in the same gene that was identified in the Amish
(Stone, Slavotinek, et al., 2000). This confirmed that muta- Epstein CJ, Erickson RP, Wynshaw-Boris A (eds.) (2004). Inborn Errors
tions in the novel Amish gene (MKKS) caused MKS. of Development. New York: Oxford University Press.
Jones K (2006). Smith’s Recognizable Patterns of Human Malformations
Since that time, many patients who were originally diag- (5th ed.). New York: Elsevier.
nosed with MKS have developed other clinical features and Stevenson R, Hall JH (eds.) (2006). Human Malformations and Related
Anomalies (2nd ed.). New York: Oxford University Press.
have been diagnosed with Bardet-Biedl syndrome (BBS)
(Slavotinek, Stone, et al., 2000). Affected individuals with
BBS have post-axial polydactyly, mental retardation, pro-
gressive pigmentary retinopathy, and obesity complicated REFERENCES
with diabetes mellitus. It soon became obvious that both
MKS and BBS share common phenotypical features. BBS Adams M, Smith UM, et al. (2008). Recent advances in the molecular
pathology, cell biology and genetics of ciliopathies. Journal of medi-
is caused by mutations in different genes. It is an illustrative cal genetics. 45(5):257–267.
example of a digenic or multigenic disorder, wherein a com- Avidor-Reiss T, Maer AM, et al. (2004). Decoding cilia function: defin-
bination of mutations in different genes results in the BBS ing specialized genes required for compartmentalized cilia biogen-
esis. Cell. 117(4):527–539.
phenotype. Some patients who were not known to have Beldade P, Brakefield PM (2002). The genetics and evo-devo of butterfly
mutations in one of the BBS genes have now been reported wing patterns. Nature reviews genetics. 3(6):442–452.
to have mutations in the MKKS gene (Slavotinek, Searby, Berardo A, DiMauro S, et al. (2010). A diagnostic algorithm for meta-
bolic myopathies. Current neurology and neuroscience reports.
et al., 2002; Katsanis, 2004). These mutations are different 10(2):118–126.
from that found in the Amish. These findings show that it Biesecker LG (2002). Coupling genomics and human genetics to
delineate basic mechanisms of development. Genetics in Medicine.
is important to review all patients with a clinical diagno- 4:39S–42S.
sis of BBS or MKS and apply the molecular techniques to Biesecker LG, Graham J (1996). Pallister-Hall syndrome. Journal of med-
establish diagnosis. This is important in clinical care and ical genetics. 33(7):585–589.
Brunner HG, van Driel MA (2004). From syndrome families to func-
for offering accurate genetic counseling to the family. This tional genomics. Nature reviews genetics. 5(7):545–551.
is another example where the basic scientific data could be Cañestro C (2012). Two rounds of whole-genome duplication: evidence
taken back to the clinical setting for the benefit of patients, and impact on the evolution of vertebrate innovations. In Polyploidy
and Genome Evolution. Springer: 309–339.
parents, and other family members. Cañestro C, Yokoi H, et al. (2007). Evolutionary developmental biology
and genomics. Nature reviews genetics. 8(12):932–942.
Chassaing N, Bourthoumieu S, et al. (2006). Mutations in EDAR
account for one-quarter of non-ED1-related hypohidrotic ectoder-
CONCLU SION mal dysplasia. Human Mutation. 27(3):255–259.
Cohen Jr M (1989). Syndromology: an updated conceptual overview.
In conclusion, understanding normal human develop- III. Syndrome delineation. International journal of oral and maxil-
lofacial surgery. 18(5):281–285.
ment is a prerequisite for analyzing the pathophysiology of Corruccini RS (1990). Australian aboriginal tooth succession, interprox-
developmental malformations occurring either singly or as imal attrition, and Begg’s theory. American Journal of Orthodontics
part of a multiple-malformation syndrome. Development and Dentofacial Orthopedics. 97(4):349–357.
Corruccini, RS, Whitley LD, et al. (1985). Facial height and breadth
in humans is strikingly similar to that in small creatures, relative to dietary consistency and oral breathing in two populations
as well as in several other mammals, including our fellow (North India and US). Human Biology. 57(2):151–161.
8 0 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
Cvekl A, Tamm ER (2004). Anterior eye development and ocular mes- Li, JB, Gerdes JM, et al. (2004). Comparative genomics identifies a fla-
enchyme: new insights from mouse models and human diseases. gellar and basal body proteome that includes the BBS5 human dis-
Bioessays. 26(4):374–386. ease gene. Cell. 117(4):541–552.
David, WM (2004). 4′Consequences of the Genome Project for under- Li L, Stoeckert CJ, et al. (2003). OrthoMCL: identification of ortholog
standing development. The Molecular Basis of Clinical Disorders of groups for eukaryotic genomes. Genome research. 13(9):2178–2189.
Morphogenesis. In ‘Inborn Errors of Development;’ Eds. Epstein, Liao J, Kochilas L, et al. (2004). Full spectrum of malformations
Erickson et al. Oxford University Press, NY, 2004. in velo-cardio-facial syndrome/DiGeorge syndrome mouse
Davidson, EH, McClay DR, et al. (2003). Regulatory gene networks models by altering Tbx1 dosage. Human molecular genetics.
and the properties of the developmental process. Proceedings of the 13(15):1577–1585.
National Academy of Sciences. 100(4):1475–1480. Liu A, Wang B, et al. (2005). Mouse intraflagellar transport proteins reg-
De Marco P, Calevo MG, et al. (2003). Reduced folate carrier poly- ulate both the activator and repressor functions of Gli transcription
morphism (80A→ G) and neural tube defects. European journal of factors. Development. 132(13):3103–3111.
human genetics. 11(3):245–252. Martínez-Frías M (2004). Segmentation anomalies of the vertebras and
Epstein, CJ, Erickson RP, et al. (2004). Inborn Errors of Development: The ribs: one expression of the primary developmental field. American
Molecular Basis of Clinical Disorders of Morphogenesis. Journal of Medical Genetics, Part A. 128(2):127–131.
New York: Oxford University Press. McKusick VA (1969). On lumpers and splitters, or the nosology of
Gehring WJ (2002). The genetic control of eye development and its genetic disease. Birth Defects. 5(1):23–32.
implications for the evolution of the various eye-types. International Migeon, BR, Brown TR, et al. (1981). Studies of the locus for androgen
Journal of Developmental Biology. 46(1):65–74. receptor: localization on the human X chromosome and evidence
Gehring, WJ, Kloter U, et al. (2009). Evolution of the Hox gene complex for homology with the Tfm locus in the mouse. Proceedings of the
from an evolutionary ground state. Current topics in developmental National Academy of Sciences. 78(10):6339–6343.
biology. 88:35–61. Morton NE (1956). The detection and estimation of linkage between
Gilbert SF (2004). General Principles of Differentiation and Morphogenesis. the genes for elliptocytosis and the Rh blood type. American journal
New York: Oxford University Press. 10–24. of human genetics. 8(2):80.
Gilbert SF, Epel D (2009). Ecological Developmental Biology: Integrating Mount DW (2004). Bioinformatics: Sequence and genome analysis. Cold
Epigenetics, Medicine, and Evolution. Sinauer Associates Sunderland, Spring Harbour, NY: Cold Spring Harbour Laboratory Press.
Mass. Nadeau JH (2001). Modifier genes in mice and humans. Nature reviews
Goldstein LS, Gunawardena S (2000). Flying through the Drosophila genetics. 2(3):165–174.
cytoskeletal genome. Journal of cell biology. 150(2):F63–F68. Newton K, Strasser A (1998). The Bcl-2 family and cell death regulation.
Gupta, BP, Wang M, et al. (2003). The C. elegans LIM homeobox gene Current opinion in genetics & development. 8(1):68–75.
lin-11 specifies multiple cell fates during vulval development. Oti M, Brunner H (2007). The modular nature of genetic diseases.
Development. 130(12):2589–2601. Clinical Genetics. 71(1):1–11.
Headon, DJ, Emmal SA, et al. (2001). Gene defect in ectodermal dys- Pazour GJ, Witman GB (2003). The vertebrate primary cilium is a sen-
plasia implicates a death domain adapter in development. Nature. sory organelle. Current opinion in cell biology. 15(1):105–110.
414(6866):913–916. Reginato AM, Olsen BR (2002). The role of structural genes in
Hogan B (1996). Bone morphogenetic proteins: multifunctional the pathogenesis of osteoarthritic disorders. Arthritis research.
regulators of vertebrate development. Genes and development. 4(6):337–345.
10(13):1580–1594. Reid R (2004). Epigenetics and Environment: The Historical Matrix of
Holland LZ (2007). Developmental biology: a chordate with a differ- Matsuda’s Pan-Environmentalism. Cambridge, MA: The MIT Press.
ence. Nature. 447(7141):153–155. Robert JS (2008). Taking old ideas seriously: evolution, development,
Istrail S, Sutton GG, et al. (2004). Whole-genome shotgun assem- and human behavior. New Ideas in Psychology. 26(3):387–404.
bly and comparison of human genome assemblies. Proceedings of Schwab W, Kasper M, et al. (2000). Characterization of caveolins from
the National Academy of Sciences of the United States of America. human knee joint cartilage: expression of caveolin-1, -2, and -3 in
101(7):1916–1921. chondrocytes and association with integrin β1. Histochemistry and
Jones KL, Smith DM (2006). Recognizable Patterns of Human cell biology. 113(3):221–225.
Malformation. New York: Elsevier. Slavotinek A, Searby C, et al. (2002). Mutation analysis of the MKKS
Joza N, Susin SA, et al. (2001). Essential role of the mitochondrial gene in McKusick-Kaufman syndrome and selected Bardet-Biedl
apoptosis-inducing factor in programmed cell death. Nature. syndrome patients. Human genetics. 110(6):561–567.
410(6828):549–554. Slavotinek AM, Stone EM, et al. (2000). Mutations in MKKS cause
Kang S, Graham JM, et al. (1997). GLI3 frameshift mutations cause Bardet-Biedl syndrome. Nature genetics. 26(1):15–16.
autosomal dominant Pallister-Hall syndrome. Nature genetics. Souza P, Kuliszewski M, et al. (1995). PDGF-AA and its receptor influ-
15(3):266–268. ence early lung branching via an epithelial-mesenchymal interaction.
Katsanis N (2004). The oligogenic properties of Bardet-Biedl syndrome. Development. 121(8):2559–2567.
Human molecular genetics. 13(suppl 1):R65–R71. Spranger J, Winterpacht A, et al. (1994). The type II collagenopathies: a
Katsanis N, Lupski JR, et al. (2001). Exploring the molecular spectrum of chondrodysplasias. European journal of pediatrics.
basis of Bardet-Biedl syndrome. Human molecular genetics. 153(2):56–65.
10(20):2293–2299. Stern HM, Brown A, et al. (1995). Myogenesis in paraxial meso-
Kinzler, KW, Ruppert JM, et al. (1988). The GLI gene is a member of the derm: preferential induction by dorsal neural tube and by cells
Kruppel family of zinc finger proteins. Nature. 332(6162):371–374. expressing Wnt-1. Development. 121(11):3675–3686.
Krogan, NJ, Cagney G, et al. (2006). Global landscape of pro- Stone DL, Slavotinek A, et al. (2000). Mutation of a gene encoding a
tein complexes in the yeast Saccharomyces cerevisiae. Nature. putative chaperonin causes McKusick-Kaufman syndrome. Nature
440(7084):637–643. genetics. 25(1):79–82.
Kumar D (2008). Genomic perspectives of human development. In Trehan KS, Gill KS (2002). Epigenetics of dominance for enzyme activ-
Genomics and Clinical Medicine' Eds. Kumar, D & Weatherall ity. Journal of Biosciences. 27(2):127–134.
DW, Oxford University Press, NY. Vaahtokari A, Aberg T, et al. (1996). Apoptosis in the developing
Lander ES, Linton LM, et al. (2001). Initial sequencing and analysis of tooth: association with an embryonic signaling center and suppres-
the human genome. Nature. 409(6822):860–921. sion by EGF and FGF-4. Development. 122(1):121–129.
G e n e s , G e n o m e , a n d D e ve l o pm e n ta l M a l f o r m at i o n s • 8 1
Verloes A (1995). Numerical syndromology: a mathematical approach to Wilkie AO (1997). Craniosynostosis: genes and mechanisms. Human
the nosology of complex phenotypes. American journal of medical molecular genetics. 6(10):1647–1656.
genetics. 55(4):433–443. Wu M, Li J, et al. (2008). Persistent expression of Pax3 in the neural crest
Vortkamp A, Gessler M, et al. (1991). GLI3 zinc-finger gene inter- causes cleft palate and defective osteogenesis in mice. Journal of clini-
rupted by translocations in Greig syndrome families. Nature. cal investigation. 118(6):2076.
352(6335):539–540. Yamada T, Pfaff SL, et al. (1993). Control of cell pattern in the neural
Watt AJ, Chung KC (2009). Generalized skeletal abnormalities. Hand tube: motor neuron induction by diffusible factors from notochord
Clinics. 25(2):265–276. and floor plate. Cell. 73(4):673–686.
8 2 • P r i n c i p l e s o f G e n o m i c M e d i c i n e
6.
BIOINFORMATICS, SYSTEMS BIOLOGY,
AND SYSTEMS MEDICINE
Binay Panda and Neeraja M. Krishnan
83
a focus on data generated by the newer generations of DNA genome medicine, we first provide a short description of the
sequencers. Lastly, biological structure and function along standard analyses of workflow and post raw data retrieval
with its utility in clinical effectiveness will also be discussed. from the sequencing instruments. Various sequencing tech-
nology, the chemistry of producing sequencing libraries,
and study design considerations are reviewed extensively
D N A SE Q U E N C I N G I N T H E elsewhere.18–24
P O S T–H U M A N G E N O ME
P R O J E C T P H A SE
A N A LY T I C A L WO R K FL OW
Post–human genome sequencing (which was achieved pri-
marily by using the automated Sanger capillary sequencing
S EQ U E N C E R E A D S A N D P ROTO C O L S
instruments), the second generation of DNA sequencers have
contributed the most to the data deluge in biology. The first The Illumina platform offers two kinds of protocols for
instrument in the second generation of DNA sequencers was sequencing, the short-insert and the long-insert protocols.
commercially introduced by a Connecticut-based American The former protocol allows users to sequence libraries with
company called 454 Life Sciences in 2005, which was later a shorter insert size of up to 600bp, whereas the latter can
bought by Roche Diagnostics. Although many other com- sequence long insert libraries ranging from 1.5kb to 20kb,
panies introduced their technology in the market post-2005, making it useful for understanding structural variations in the
the one from Solexa (which was later acquired by Illumina) genome. In both protocols, users have an option to produce
employing a massive parallel sequencing-by-synthesis (SBS) either single- or paired-end reads. The basic format for stor-
method using reversible terminator chemistry14 remains the ing the sequenced short reads is the fastq format. The fastq
dominant player, both in terms of the total amount of data format comprises both the sequence and the ASCII-coded
produced so far globally, and of the market penetration in quality scores for each nucleotide sequenced (Figure 6.1A).
large genome centers and individual laboratories. Hence, this
chapter shall cover data analysis mainly related to the sequenc-
ing reads produced by sequencing-by-synthesis technology Q UA L I T Y C H E C K S A N D R E A D
from Illumina, as it is the technology most widely used by P R E -P R O C ESS I N G
researchers. Additionally, due to accessibility constraints, this
review will focus on open-source tools or tools available as Before the short sequenced reads are used further in any
free downloads from commercial providers. analytical workflow, it is important to check the individ-
The second or new generation of DNA sequencers ual read quality and quality of each nucleotide in a read.
(popularly called next-generation sequencers or NGS) has Different sequencing platforms estimate quality differently,
brought about a major transformation in biology, medicine, and there are several read quality control (QC) tools that
and agriculture in the last few years by aiding the discov- access the quality. FastQC25 is a standard modular tool used
ery of genetic, transcriptomic, and epigenetic markers in a to perform quality checks on high-throughput sequence
genome-wide scale. With sequencing becoming ubiqui- data derived from Illumina instruments by running a bat-
tous, transcriptome sequencing is increasingly preferred to tery of tests on the raw data, and it sets a flag when certain
the hybridization-based experiments using whole-genome quality conditions are not met. Trimmomatic26 is a read
microarrays, since sequencing is annotation-independent, QC and trimming tool, which trims the reads based on
digital, and quantitative. RNA and cDNA sequencing also known Illumina-like adapter sequences or a sliding window
overcome disadvantages typically inherent in the microar- of trailing base qualities. Once the standard QC checks are
ray experiments: like probe cross-hybridization, “noise,” and done, researchers studying archival or metagenomic sam-
limited dynamic range of detection.15 Chip-Seq, also a next- ples often run another set of tools, like DeconSeq27 that are
generation technology, assays the entire genome/transcrip- used as pre-processing filters to eliminate contamination
tome for protein-bound DNA/RNA regions, and therefore from genomic or metagenomic sequenced reads. Cancer
has advantages over the array-based ChIp-chip technique, researchers often use a tool called ContEst28 for estimat-
which relies on probes16,17 and suffers from the same draw- ing the amount of cross-sample contamination in sequenc-
backs as array-based gene expression experiments. ing data. ContEst uses a Bayesian framework to estimate
As the focus of this chapter is to elaborate the bioin- contamination levels from array-based genotypes and
formatics and systems biology aspects of high-throughput sequencing reads.
8 4 • P rincip l e s o f G e no m ic M e dicin e
(A)
(B)
Figure 6.1 Sequenced reads in Fastq (A) and SAM (B) formats. Various attributes of the file are shown.
A L I G N I N G R AW R E A D S TO T H E population.44–55 SVDetect,56 Breakway,57 Breakdancer,58
R EFE R E N C E SE Q U E N C E and CREST59 are a few open-source tools that can detect
structural variations and genome breakpoints. Copy num-
Once checked for quality and pre-processed, reads are ber detection tools either measure absolute copy number
aligned to a reference genome using a mapper or read aligner. changes in a single sample or somatic copy number altera-
There are numerous open-source or freely downloadable tions (SCNA) in cancer and other disease samples where
aligners available, each tailored to perform optimally under both control and diseased sample data are used. Both
specific criteria that fall in two major classes. The first class these sets of tools can use whole genome or exome or both
of aligners is built on hash table–based approach of the types of data. RDXplorer,60 CNVNator,61 and Freec62 are
reference genome that includes Bfast,29 Ssaha,30 Smalt,31 CNV detection tools meant for whole genomes of single
Stampy,32 and Novoalign.33 The second class of aligners sample analyses, and COPS,63 SVDetect,56 and CNV-Seq64
relies on creating an efficient index of the reference genome, are whole-genome SCNA detection tools that use disease
and in this category fall BWA34 and Bowtie.35,36 BWA,34 an and matched normal paired samples. In a rigorous bench-
aligner based on the Burrows Wheeler transform, is the most marking of these tools, we found COPS to perform well,
popular short-read aligner used by the research community. in terms of both sensitivity and specificity, in detecting a
In a comparison study37 involving five such short-read align- wide size range of CNAs, using libraries with varying read
ers, we found Novoalign to perform best in read alignment lengths and sequencing coverage using the whole-genome
in terms of sensitivity, perhaps owing to its post-alignment dataset.63 The fragmented spread of exonic regions makes it
base quality recalibration functionality. Following the read difficult to determine CNA boundaries, and hence, break-
alignment to the reference genome, reads are usually stored points accurately. Control-Freec,65 ExomeCNV,66 and
in Sequence Alignment/Map (SAM) format, which is ExomeDepth67 are paired CNA callers, specially designed
evolving as a consensus, flexible, and generic format among for exome dataset.
the research community. Figure 6.1B provides a snippet of
the SAM format.
G E N E A N D VA R I A N T
A N N OTAT I O N
C A LL I N G VA R I A N T S
Once the variants are called, several tools are used for gene
After alignment of reads, single nucleotide polymorphisms and variant annotation. GPAT68 is a rapid gene-annotation
or variants (SNPs or SNVs), insertion and deletion vari- tool based on the genomic positions. It operates currently
ants (indels), and copy number alterations or variations on various genome versions of popular organisms such as
(CNAs or CNVs) are identified in the sequenced genomes human (H. sapiens), mouse (M. musculus), Arabidopsis
and exomes. There are several tools designed to call SNVs (A. thaliana), fruit fly (D. melanogaster), and zebrafish
using aligned sequence data. The most popular one is (D. rerio). The purpose of GPAT is to provide an easy-to-
the Genome Analyses Took Kit (GATK)38 developed by use and convenient tool for rapidly annotating reasonably
the Broad Institute in Cambridge, Massachusetts. In a large sets of genomic positions. In addition to gene anno-
comparative study among the other variant calling algo- tation, GPAT allows the retrieval of some expression data
rithms,37 GATK performed best in calling SNVs, most and chromosome position data. Several in silico tools have
likely due to the base call recalibration and local realign- been developed to predict the effect of single nucleotide
ment step that it incorporates for variant calling. In addi- variants on a protein. Some of the more popular ones are
tion to GATK, tools like Samtools,39 Bambino,40 and SIFT,69 Polyphen,70 VEP,71,72 and PROVEAN.73 Most of
Freebayes41 are some of the other open-source variant call- these tools use predictions based on the degree of con-
ers that are widely used. For indels, Samtools,39 Pindel,42 servation of amino acid residues in sequence alignments
and Dindel43 are widely used to detect indels in the short, derived from closely related sequences, collected through
medium, and long range, in addition to tandem duplica- PSI-BLAST.74,75 PROVEAN73 (Protein Variation Effect
tions and inversions. Copy number alterations (CNAs) Analyzer) predicts the impact of SNP and indel on protein
are an important category of structural aberrations in function, and can thus identify non-synonymous SNPs and
human diseases. CNAs range from one kilobase (kb) to functionally relevant indels. It compares well in its speed
several megabases (mb) and have been implicated in sev- and performance to SIFT69 and PolyPhen-2, VEP,71,72 an
eral human diseases, including cancer, and in the normal Ensembl tool, provides additional annotation of variants
8 6 • P rincip l e s o f G e no m ic M e dicin e
by determining their effects on genes, transcripts, protein as it relies on a expert-curated database storing biological
sequences, and even regulatory regions. interactions and functional annotations from published
literature. Performing pathway analyses after discovery of
variants has many implications for personalized therapy,
VI S UA L I Z I N G G E N O ME DATA especially for pathway-based drug repositioning, where
existing drugs can be tried for new therapeutic uses based
Several visualization platforms have been developed to on pathway information.
aid genome and variant visualization. The common and
most popular ones are the UCSC76 and Ensembl77 genome
browsers. Both genome browsers can import user data in SE Q U E N C E A SSEM B LY
different formats for visualization alongside the publicly
available data and data from multiple types of experiments In the absence of a valid reference genome, de novo sequence
(Figure 6.2). Although genome browsers, like that from assembly is performed, involving self-alignment of the reads
UCSC, are immensely popular, easy to browse, can be cus- and merging of the longest stretch of self-aligned reads
tomized to include external data, and provide data visual- to reconstruct the longest contiguous sequence (contig).
ization for most community and consortium projects (like Contigs are then assembled into scaffolds using gaps between
HapMap and Encode), they lack the ability to visualize data paired sequence reads. Generation of short sequence reads
from two or more faraway regions of the genome at the same of any genome, followed by de novo assembly, can be com-
resolution in a single window. Recently, a nonlinear repre- pared to shredding a paper into several pieces and trying to
sentation of data that are megabases away, where multiple assemble them back to recreate the original shape and form.
regions of the genome can be visualized together in a single This process is a complex one and is often difficult and error-
window, termed the Elastic Genome Browser, has been pro- prone.97 There is a possibility of misplacing a piece of paper
posed.78 For easy and simultaneous visualization of all types from its original location. Particularly, identical parts of the
of variants in individual genomes, Circos plots79 provide us paper may get exchanged, and in this manner, get misas-
with the best option (Figure 6.3). As represented in Figure sembled. The best methods available today towards produc-
6.3, the chromosomes, positions of multiple types of vari- ing complete assemblies of large genomes are methods that
ants, allelic imbalances, genes bearing the SNVs, indels, and take sequence reads from multiple technologies (Sanger,
CNVs can be visualized simultaneously using Circos plots, second- and third-generation), read lengths (using both
aiding visualization at the whole-genome level. long- and short-insert libraries) and chemistry (sequencing-
by-synthesis, sequencing-by-ligation and single molecule
long reads). Several de novo genome assemblers have been
PAT H WAY A N A LYSES developed in the past.98–111 These tools use short sequence
reads, many from both short- and long-insert libraries, to
At the end of next-generation sequencing data analyses, biol- assemble them into different contigs and further assemble
ogists often face the problem of finding meaning in the list the contigs into “scaffolds,” using appropriate gaps spanning
of genes discovered in a particular disease study. This entails those contigs. Long insert mate-pair libraries are typically
extensive functional analyses and validation that require used in the process of scaffolding, and provide the informa-
skills and substantial time to complete. The pathway-based tion on the distance between two contiguous stretches of
approach reduces complexity, increases the explanatory DNA elements. This reduces fragmentation of scaffolds and
power of high-throughput analyses, and simplifies the task pieces together the multitude of contigs to constitute fewer
of looking at many variants throughout the genome. There scaffolds while creating several gaps; e.g., stretches of Ns
are various databases for performing knowledge-based path- or non-ATGC characters within the scaffolds. Assemblers
way analyses, like Kegg,80–83 BioCarta,84 Reactome,85–89 and are also tailored to a particular sequencing platform.
Panther.90–93 DAVID94 offers an application interface that Velvet,110 ABySS,108 and SOAPdenovo111 are commonly
links all the above knowledge databases. Besides these, sev- used genome assemblers for Illumina and SOLiD reads.
eral commercial tools can perform pathway analyses, such AMOS112 can be used to assemble Sanger and 454 reads.
as AvadisNGS,95 which uses a Natural Language Processing CLC Genomics Workbench,113 a proprietary tool, is used
(NLP)–based tool to look up biological interactions from to assemble Sanger, 454, Illumina, and reads. Newbler114 is
literature and Ingenuity Pathway Analysis (IPA).96 IPA is exclusively used to assemble 454 reads. MIRA115 is used for
the most popular pathway analysis tool among biologists, assembling Sanger, 454, Ion Torrent and Illumina PacBio
(B)
Figure 6.2
Visualization using UCSC (A) and Ensembl (B) genome browsers. Genomic coordinate for the gene TP53 is shown with self-explanatory
attributes from the public databases.
(A)
1
X
22
21
20
2
19
18
17 3
16
4
15
CTX
SV
14 CNS ELs
5
-INS IND
13
L SNV
-DE s
12
6 CNV
Het-A
-B
Het
11
7
10
8
6
1 22
(B)
2 21
3 20
4 19
5 18
Chromosome
6 17
7 16
8 15
9 14
10 13
11 12
Figure 6.3
Variant visualization using circular representation of the genome using Circos (A) and linear representation of the genome using the
statistical program R (B). Various variants depicted are: SNVs—single nucleotide variants, CTX—chromosomal translocations; SV—large
insertions/deletions and inversions, detected using the program BreakDancer; CNV-INS and CNV-DEL—insertion and deletion copy number
variants, respectively; Het-A, Het-B—variants found using whole-genome microarrays in A and B allele, respectively; and INDEL—short
insertion/deletion variants found using the program Pindel.
reads. ALLPATHS-LG116 assembles Illumina and SOLiD involve construction of de Bruijn graphs using k-mers (for
reads, and requires overlapping read libraries. While the short reads) or use an overlap-layout-consensus (OLC)
tools mentioned above can produce both contigs and scaf- approach (for longer reads).128 Trinity,129 Trans-ABySS,127
folds, there are tools that perform only scaffolding using Oases,130 and SOAPdenovo-Trans131 are commonly used de
pre-assembled contigs, such as SSPACE,117 Bambus,118 novo transcriptome assemblers. Genome-guided Trinity,132
and Opera.101 Reads from the Pacific BioSciences (PacBio) Scripture,133 and Tophat1-Cufflinks134 are genome-guided
instrument (which are typically of varying lengths but are transcriptome assemblers. Genome-guided Trinity,132
often long) can be used along with the short reads from although it requires a genome, only uses the genome to
Illumina to perform hybrid assembly using the overlap- cluster reads locally and then proceeds to assemble each of
layout-consensus (OLC) approach. PacBiotoCA119 is an these clusters using the same approach as Trinity, which is
error-correction pipeline that corrects the long and error- different from the others in the same group. The former
prone reads by mapping accurate shorter reads from other is used mostly for short reads, while the latter is used for
sequencing instruments to those from PacBio. Another long reads from 454 or PacBio. A detailed account of
algorithm developed by Koren and his co-workers, which various transcriptome assembly tools has been reviewed
is implemented as part of the Celera assembler,120 improves elsewhere.135 Both the de novo and the genome-guided tran-
read accuracy for assembly using long reads from PacBio.121 scriptome assembly processes have their own advantages
The algorithm does this by calculating an accurate hybrid and disadvantages. De novo transcriptome assembly tools
consensus sequence by mapping higher-accuracy short can recover transcript fragments from regions missing in
reads from Illumina to individual lower-accuracy long reads the genome assembly, while the genome-guided assembly
from PacBio. ones can filter out contamination and sequencing artifacts,
Even with the availability of multiple sequencing tech- can recover low-abundance transcripts, and can fill gaps
nology and chemistry, some parts of the genome are inher- using the genome sequence, resulting in full-length tran-
ently hard to sequence. This is primarily due to either the scripts. Recently, from a comparison study of the existing
sequence context (repetitive elements) or the failure of de novo and genome-guided assembler tools, we proposed
a particular technology to read through a stretch of the an augmented method for the transcriptome assembly
same nucleotide (homopolymer issues in pyrosequenc- using the de novo tool with a genome-guided approach has
ing). In order to be practically useful, the completeness of been suggested.136
the sequenced genomes is a prerequisite and imperative for
our understanding the underlying biology. Several methods
have been proposed to fill the gaps in the assembly process F U N C T I O N A L A N N OTAT I O N
with the existing short reads.104,111,122,123 We are currently
working on a proposed method called FILAM (FILling the The process of assembly is usually followed by gene predic-
gAp by re-Mapping) that takes into account the partially tion and annotation. Gene prediction tools utilize gene
mapped reads to the assembly gaps and recreates scaffolds models and training sets, based on previously annotated
post gap filling. In the proposed method, once the partially genes from closely related organisms. In addition, the
mapped reads are identified, the exact gap sizes in the scaf- assembled genome or transcriptome can also be subjected to
folds are determined followed by filling the assembly gaps similarity-based annotation by performing BLAST against
iteratively by extending the reads.124 pre-existing databases. The AutoFACT137 pipeline system-
Transcriptome and EST assembly are more complex atically annotates assemblies based on performing BLAST
than genome assembly. This is due to the presence of against the nr74 (nucleotide and protein), uniref90,138 uni-
spliced site junctions, representation of low copy number ref100,138 kegg,80,82,139,140 and cog141,142 databases. KEGG
transcripts, genes with varying expression levels, and the orthologies using the Kegg Automatic Annotation Server
presence of homologues. Seed length for alignment125 and (KAAS),143 annotate sequences similar to known ortholo-
k-mer size126,127 are important parameters for transcrip- gous genes from a selected range of organisms, across a wide
tome assembly. De novo transcriptome assembly can either spectrum of pathways. However, BLAST-based analyses
make use of the available genome sequence of the indi- can only recover known annotations. Un-annotated genes
vidual organism (genome-assisted) or be performed inde- specific to a given organism need to be validated experi-
pendently. The tools that do the latter use the seed length mentally to confirm their functional role. A potential
parameter while aligning reads to a reference genome prior trick to reduce complexity is to identify known domains
to assembling transcriptome. De novo assembly tools either within the un-annotated genes using domain databases
9 0 • P rincip l e s o f G e no m ic M e dicin e
such as PFam,144 and increase knowledge of these predic- SE C O N DA RY S T RU C T U R E
tions/transcripts. Furthermore, the assembled genome can
be characterized with respect to its repeat content. This is Base pairing within regions of nucleic acids, in single-
typically done using known organism-specific repeat librar- stranded DNA or RNA resulting in a stem-loop second-
ies by tools like RepeatMasker145 and RepeatModeler,146 or ary structure, plays an important role in biological func-
using de novo repeat predictions by tools like LTRFinder,147 tion. Mfold provides secondary structure estimates for
MITE-hunter,148 and RepeatScout149 using certain consen- DNA and RNA, within a certain percent of sub-optimality
sus repeat motifs. of the most optimal (minimum free energy) structure.160
Sfold provides a more accurate representation of the ther-
modynamic spread of all secondary structures by sampling
SE Q U E N C E , S T RU C T U R E , A N D from the Boltzmann-weighted structure ensemble.161 The
F U N C T I O N O F N U C LE I C AC I D S tRNAScan-SE server162 uses the clover-leaf secondary
A N D P R OT E I N S structural fold of a tRNA for detection in an un-annotated
sequence format. Nucleic acid tertiary structures like the
Studying the sequence, structure, and function of DNA, double helix, G-quadruplexes, A-minor motifs, ribose
RNA, and protein forms the foundation of all biological zipper, pseudoknot, riboswitches, and others have been
questions. The ultimate aim of all high-throughput technol- implicated to carry specific functions in vivo. Molecular
ogy, biochemical and analytical, which has led to increasing dynamics simulations have become crucial tools in areas
amounts of sequences and structures of biological macro- such as biomolecular engineering, synthetic biology, and
molecules in the databases, is to understand biological func- drug design.163 Tools can render, model, and compare
tion. Below, a few key methods that apply to sequence and three-dimensional structures of nucleic acids, based on
structure analyses, both for nucleic acids and for protein, are molecular dynamics of their primary and secondary struc-
outlined. tures. NAST164 and SETTER165 are two such tools.
For proteins, alpha helices and beta-sheets are highly
local sub-structures of a protein, defined by hydrogen
MU LT I P LE SE Q U E N C E A L I G N ME N T bonding. The meta-server for protein sequence analysis
(MESSA) is one of the most widely used protein struc-
A multiple sequence alignment (MSA) of DNA, RNA, or ture and function prediction servers,166 which uses mul-
protein helps identify conserved regions of high similarity tiple tools to predict protein properties. YASPIN167 uses
across the sequences that might have arisen out of a com- Hidden Neural Networks and PSI-BLAST75,168 to produce
mon evolutionary origin to support different functions, Position-Specific Scoring Matrix (PSSM169), used for sec-
or alternatively, with some shared and some unique func- ondary structure prediction.
tional domains of proteins. As (proposed by Francois Jacob) While majority of high-resolution protein structures
Nature is a tinkerer and not an inventor,150 the existing have been solved using X-ray crystallography, nuclear mag-
information is adapted and reused rather than new infor- netic resonance (NMR) can provide a time-dependent pro-
mation being created. Pfam151 is a database of all known tein conformational change despite compromising on the
protein domains. ClustalW152 is commonly used in pro- resolution of the predicted structures. Protein misfolding
ducing a multiple sequence alignment. The MSA is used to has been linked to human diseases170 such as Alzheimer’s,
produce a phylogeny, a graphical, treelike representation of Huntington’s, osteogenesis imperfecta, and even some can-
the relatedness between sequences. There are various evo- cers. Folding@home,171 a distributed computing project
lutionary models153 to produce phylogenies. The simplest housed in the Stanford University, is used to analyze the
non-model approach is the principle of parsimony, which tertiary and quaternary structures of proteins. Abalone172
assumes the minimal change required to go from one node and Ascalaph Designer173 are two software suites from
to another. Methods that use Maximum Likelihood and Agile Molecule that perform biomolecular simulations
Bayesian approaches rely on an optimized substitution- using molecular dynamics and protein folding and molecu-
rate matrix to reconstruct ancestral states of the phylogeny, lar building and dynamics using DNA, proteins, hydrocar-
and are known to perform better with minimal composi- bons, and nanotubes respectively. The DisoveryStudio174 is
tional bias. There are various tools available for reconstruct- an entire suite dedicated to discovery of drugs, and it allows
ing phylogeny, such as MrBayes,154 PAUP*,155 RAxML,156 modeling and simulation of small drug molecules docking
Phyml,157 MEGA,158 and PHYLIP.159 to macromolecules.
LUNG CANCER
ARRAY GENOTYPING
ESOPHAGEAL CANCER
BREAST CANCER
9 2 • P rincip l e s o f G e no m ic M e dicin e
6. Lobo D, Malone TJ, Levin M. Towards a bioinformatics of pattern- 3 3. Novocraft. Novoalign. 2011. Accessed 25 November 2011.
ing: a computational approach to understanding regulative mor- 34. Li H, Durbin R. Fast and accurate short read alignment with
phogenesis. Biol open. 2013;2(2):156–169. Burrows-Wheeler transform. Bioinformatics. 2009;25(14):
7. Schneider HC, Klabunde T. Understanding drugs and diseases by 1754–1760.
systems biology? Bioorg med c l. 2013;23(5):1168–1176. 35. Langmead B. Aligning short sequencing reads with Bowtie.
8. Laszlo A, Kripner S. Systems theories: their origins, foundations Current protocols in bioinformatics / editoral board, Andreas D.
and development. In: Systems Theories and A Priori Aspects of Baxevanis . . . [et al.]. Dec 2010;Chapter 11:Unit 11.7.
Perception. Amsterdam: Elsevier Science; 1998:47–74. 36. Langmead B, Salzberg SL. Fast gapped-read alignment with
9. Kritikou E, Pulverer B, Heinrichs A. All systems go! Nat Rev Mol Bowtie 2. Nat Methods. 2012;9(4):357–359.
Cell Biol. 2006;7:801. 37. Pattnaik S, Vaidyanathan S, Pooja DG, Deepak S, Panda B.
10. Auffray C, Chen Z, Hood L. Systems medicine: the future of medi- Customisation of the exome data analysis pipeline using a combina-
cal genomics and healthcare. Genome med. 2009;1(1):2. torial approach. PloS one. 2012;7(1):e30080.
11. Auffray C, Hood L. Editorial: Systems biology and personalized 38. McKenna A, Hanna M, Banks E, et al. The Genome Analysis
medicine—the future is now. Biotech j. 2012;7(8):938–939. Toolkit: a MapReduce framework for analyzing next-generation
12. Auffray C, Imbeaud S, Roux-Rouquie M, Hood L. From functional DNA sequencing data. Genome Res. 2010;20(9):1297–1303.
genomics to systems biology: concepts and practices. Comptes ren- 39. Li H, Handsaker B, Wysoker A, et al. The Sequence Alignment/Map
dus biologies. 2003;326(10–11):879–892. format and SAMtools. Bioinformatics. 2009;25(16):2078–2079.
13. Auffray C, Noble D. Origins of systems biology in William Harvey’s 40. Edmonson MN, Zhang J, Yan C, Finney RP, Meerzaman DM,
masterpiece on the movement of the heart and the blood in animals. Buetow KH. Bambino: a variant detector and alignment viewer
Int J Mol Sci. 2009;10(4):1658–1669. for next-generation sequencing data in the SAM/BAM format.
14. Bentley DR, Balasubramanian S, Swerdlow HP, et al. Accurate Bioinformatics. 2011;27(6):865–866.
whole human genome sequencing using reversible terminator chem- 41. Garrison E, Marth M. Haploytype-based variant detection from
istry. Nature. 2008;456(7218):53–59. short-read sequencing. arXiv:1207.3907v2 [q-bio.GN] 2012.
15. Wang Z, Gerstein M, Snyder M. RNA-Seq: a revolutionary tool for 42. Ye K, Schulz MH, Long Q, Apweiler R, Ning Z. Pindel: a pat-
transcriptomics. Nature Rev Genet. 2009;10(1):57–63. tern growth approach to detect break points of large dele-
16. Barski A, Cuddapah S, Cui K, et al. High-resolution pro-
tions and medium sized insertions from paired-end short reads.
filing of histone methylations in the human genome. Cell. Bioinformatics. 2009;25(21):2865–2871.
2007;129(4):823–837. 43. Albers CA, Lunter G, MacArthur DG, McVean G, Ouwehand
17. Johnson DS, Mortazavi A, Myers RM, Wold B. Genome-wide WH, Durbin R. Dindel: accurate indel calls from short-read data.
mapping of in vivo protein-DNA interactions. Science. Genome Res. 2011;21(6):961–973.
2007;316(5830):1497–1502. 44. Seeger RC, Brodeur GM, Sather H, et al. Association of multiple
18. Mwenifumbo JC, Marra MA. Cancer genome-sequencing study copies of the N-myc oncogene with rapid progression of neuroblas-
design. Nat Rev Genet. 2013;14(5):321–332. tomas. N Engl J Med. 1985;313(18):1111–1116.
19. Mardis ER. Next-generation DNA sequencing methods. Annu Rev 45. Robbins CM, Tembe WA, Baker A, et al. Copy number and tar-
Genom Hum Genet. 2008;9:387–402. geted mutational analysis reveals novel somatic events in metastatic
20. Mardis ER. New strategies and emerging technologies for massively prostate tumors. Genome Res. 2011;21(1):47–55.
parallel sequencing: applications in medical research. Genome Med. 46. Prasad SE, Howley S, Murphy KC. Candidate genes and the behav-
2009;1(4):40. ioral phenotype in 22q11.2 deletion syndrome. Dev Disabil Res
21. Shendure J. Next-generation human genetics. Genome Biol.
Rev. 2008;14(1):26–34.
2011;12(9):408. 47. Newkirk HL, Bittel DC, Butler MG. Analysis of the Prader-Willi
22. Shendure J, Ji H. Next-generation DNA sequencing. Nat Biotech. syndrome chromosome region using quantitative micro-
2008;26(10):1135–1145. sphere hybridization (QMH) array. Am J Med Genet A.
23. Metzker ML. Emerging technologies in DNA sequencing. Genome 2008;146A(18):2346–2354.
Res. Dec 2005;15(12):1767–1776. 48. Elsea SH, Girirajan S. Smith-Magenis syndrome. Eur J Hum
24. Metzker ML. Sequencing technologies—the next generation. Nat Genet: EJHG. 2008;16(4):412–421.
Rev Genet. 2010;11(1):31–46. 49. Sebat J, Lakshmi B, Malhotra D, et al. Strong association
25. Fast QC. http://www.bioinformatics.babraham.ac.uk/projects/ of de novo copy number mutations with autism. Science.
fastqc/. 2007;316(5823):445–449.
26. Lohse M, Bolger AM, Nagel A, et al. RobiNA: a user-friendly, 50. Sebat J, Lakshmi B, Troge J, et al. Large-scale copy number polymor-
integrated software solution for RNA-Seq-based transcriptomics. phism in the human genome. Science. 2004;305(5683):525–528.
Nucleic Acids Res. Jul 2012;40(Web Server issue):W622–W627. 51. Kusenda M, Sebat J. The role of rare structural variants in the
27. Schmieder R, Edwards R. Fast identification and removal of
genetics of autism spectrum disorders. Cytogenet Genome Res.
sequence contamination from genomic and metagenomic datasets. 2008;123(1–4):36–43.
PloS one. 2011;6(3):e17288. 52. Lucito R, Suresh S, Walter K, et al. Copy-number variants in patients
28. Cibulskis K, McKenna A, Fennell T, Banks E, DePristo M,
with a strong family history of pancreatic cancer. Cancer Biol Ther.
Getz G. ContEst: estimating cross-contamination of human 2007;6(10):1592–1599.
samples in next-generation sequencing data. Bioinformatics. 53. Mills RE, Walter K, Stewart C, et al. Mapping copy num-
2011;27(18):2601–2602. ber variation by population-scale genome sequencing. Nature.
29. Homer N, Merriman B, Nelson SF. BFAST: an alignment tool for 2011;470(7332):59–65.
large scale genome resequencing. PloS one. 2009;4(11):e7767. 54. Conrad DF, Pinto D, Redon R, et al. Origins and functional
30. Ning Z, Cox AJ, Mullikin JC. SSAHA: a fast search method for impact of copy number variation in the human genome. Nature.
large DNA databases. Genome Res. 2001;11(10):1725–1729. 2010;464(7289):704–712.
31. Ponstingl H. SMALT. 2011; http://www.sanger.ac.uk/resources/ 55. Redon R, Ishikawa S, Fitch KR, et al. Global variation in copy num-
software/smalt/. ber in the human genome. Nature. 2006;444(7118):444–454.
32. Lunter G, Goodson M. Stampy: a statistical algorithm for sensi- 56. Zeitouni B, Boeva V, Janoueix-Lerosey I, et al. SVDetect: a tool to
tive and fast mapping of Illumina sequence reads. Genome Res. identify genomic structural variations from paired-end and mate-pair
2011;21(6):936–939. sequencing data. Bioinformatics. 2010;26(15):1895–1896.
9 4 • P rincip l e s o f G e no m ic M e dicin e
104. Li R, Fan W, Tian G, et al. The sequence and de novo assembly of 130. Schulz MH, Zerbino DR, Vingron M, Birney E. Oases: robust de
the giant panda genome. Nature. 2010;463(7279):311–317. novo RNA-seq assembly across the dynamic range of expression
105. Narzisi G, Mishra B. Comparing de novo genome assembly: the levels. Bioinformatics. 2012;28(8):1086–1092.
long and short of it. PloS one. 2011;6(4):e19175. 131. Xie Y, Wu W, Tang J, et al. SOAPdenovo-Trans: De novo transcrip-
106. Narzisi G, Mishra B. Scoring-and-unfolding trimmed tree assem- tome assembly with short RNA-Seq reads. arXiv:1305.6760v2
bler: concepts, constructs and comparisons. Bioinformatics. [q-bio.GN]. 2013.
2011;27(2):153–160. 132. TrinityTeam. Genome-guided Trinity. Vol 20132013.
107. Schmidt B, Sinha R, Beresford-Smith B, Puglisi SJ. A fast 133. Guttman M, Garber M, Levin JZ, et al. Ab initio reconstruc-
hybrid short read fragment assembly algorithm. Bioinformatics. tion of cell type-specific transcriptomes in mouse reveals the
2009;25(17):2279–2280. conserved multi-exonic structure of lincRNAs. Nat Biotech.
108. Simpson JT, Wong K, Jackman SD, Schein JE, Jones SJ, Birol I. 2010;28(5):503–510.
ABySS: a parallel assembler for short read sequence data. Genome 134. Trapnell C, Roberts A, Goff L, et al. Differential gene and tran-
Res. 2009;19(6):1117–1123. script expression analysis of RNA-seq experiments with TopHat
109. Warren RL, Sutton GG, Jones SJ, Holt RA. Assembling mil- and Cufflinks. Nat Protoc. 2012;7(3):562–578.
lions of short DNA sequences using SSAKE. Bioinformatics. 135. Martin JA, Wang Z. Next-generation transcriptome assembly. Nat
2007;23(4):500–501. Rev Genet. 2011;12(10):671–682.
110. Zerbino DR, Birney E. Velvet: algorithms for de novo short read 136. Jain P, Krishnan NM, Panda B. Augmenting transriptome assem-
assembly using de Bruijn graphs. Genome Res. 2008;18(5):821–829. bly combinatorially. arXiv:1305.6259v22013.
111. Luo R, Liu B, Xie Y, et al. SOAPdenovo2: an empirically improved 137. Koski LB, Gray MW, Lang BF, Burger G. AutoFACT: an auto-
memory-efficient short-read de novo assembler. GigaScience. matic functional annotation and classification tool. BMC bioin-
2012;1(1):18. formatics. 2005;6:151.
112. Arbuckle JL. Amos 7.0 User’s Guide. 2007. 138. Suzek BE, Huang H, McGarvey P, Mazumder R, Wu CH.
113. CLCBio. CLC Workbench. 2012; www.clcbio.com. UniRef: comprehensive and non-redundant UniProt reference
114. Margulies M, Egholm M, Altman WE, et al. Genome sequenc- clusters. Bioinformatics. 15 2007;23(10):1282–1288.
ing in microfabricated high-density picolitre reactors. Nature. 139. Kanehisa M, Goto S, Kawashima S, Okuno Y, Hattori M. The
2005;437(7057):376–380. KEGG resource for deciphering the genome. Nucleic Acids Res.
115. Chevreux B. Using the miraEST assembler for reliable and auto- Jan 1 2004;32(Database issue):D277–D280.
mated mRNA transcript assembly and SNP detection in sequenced 140. Kanehisa M, Goto S, Sato Y, Furumichi M, Tanabe M. KEGG for
ESTs. Genome Res. 2004;14:13. integration and interpretation of large-scale molecular data sets.
116. Gnerre S, Maccallum I, Przybylski D, et al. High-quality draft Nucleic Acids Res. 2012;40(1):D109–D114.
assemblies of mammalian genomes from massively parallel sequence 141. Tatusov RL, Galperin MY, Natale DA, Koonin EV. The COG
data. Proc Natl Acad Sci U S A. 2011;108(4):1513–1518. database: a tool for genome-scale analysis of protein functions and
117. Boetzer M, Henkel CV, Jansen HJ, Butler D, Pirovano W. evolution. Nucleic Acids Res. 2000;28(1):33–36.
Scaffolding pre-assembled contigs using SSPACE. Bioinformatics. 142. Tatusov RL, Fedorova ND, Jackson JD, et al. The COG data-
2011;27(4):578–579. base: an updated version includes eukaryotes. BMC bioinformat-
118. Pop M. Hierarchical scaffolding with Bambus. Genome Res. ics. 2003;4:41.
2004;14(1):11. 143. Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M. KAAS: an
119. Au KF, Underwood JG, Lee L, Wong WH. Improving PacBio automatic genome annotation and pathway reconstruction server.
long read accuracy by short read alignment. PloS one. 2012;7(10): Nucleic Acids Res. 2007;35(Web Server issue):W182–W185.
e46679. 144. Punta M, Coggill PC, Eberhardt RY, et al. The Pfam protein
120. Assembler C. Whole-genome Shotgun DNA Sequence Assembler. families database. Nucleic Acids Res. Jan 2012;40(Database
2012; http://sourceforge.net/apps/mediawiki/wgs-assembler/ issue):D290–D301.
index.php?title=Main_Page. 145. Smit AFA, Hubley R. Repeat Masker. 1989; http://www.repeat-
121. Koren S, Schatz MC, Walenz BP, et al. Hybrid error correction masker.org. Accessed October 12, 2011.
and de novo assembly of single-molecule sequencing reads. Nat 146. Smit AFA, Hubley R. Repeat Modeller. 1989; http://www.repeat-
Biotech. 2012;30(7):693–700. masker.org. Accessed December 12, 2011.
122. Boetzer M, Pirovano W. Toward almost closed genomes with 147. Xu Z, Wang H. LTR_FINDER: an efficient tool for the predic-
GapFiller. Genome Biol. 2012;13(6):R56. tion of full-length LTR retrotransposons. Nucleic Acids Res. Jul
123. Tsai IJ, Otto TD, Berriman M. Improving draft assemblies by 2007;35(Web Server issue):W265–W268.
iterative mapping and assembly of short reads to eliminate gaps. 148. Han Y, Wessler SR. MITE-Hunter: a program for discovering
Genome Biol. 2010;11(4):R41. miniature inverted-repeat transposable elements from genomic
124. A gap filling and re-scaffolding model for improved genome assem- sequences. Nucleic Acids Res. 2010;38(22):e199.
bly. Ganit Labs, Bio-IT Centre, Institute of Bioinformatics and 149. Price AL, Jones NC, Pevzner PA. De novo identification of
Applied Biotechnology; Unpublished Data. repeat families in large genomes. Bioinformatics. 2005;21(Suppl
125. Schatz MC, Delcher AL, Salzberg SL. Assembly of large genomes 1):i351–i358.
using second-generation sequencing. Genome Res. 2010;20(9): 150. Jacob F. Evolution and tinkering. Science. 1977;196(4295):
1165–1173. 1161–1166.
126. Gruenheit N, Deusch O, Esser C, Becker M, Voelckel C, Lockhart 151. Bateman A, Birney E, Cerruti L, et al. The Pfam protein families
P. Cutoffs and k-mers: implications from a transcriptome study in database. Nucleic Acids Res. 2002;30(1):276–280.
allopolyploid plants. BMC genomics. 2012;13:92. 152. Thompson JD, Gibson TJ, Higgins DG. Multiple sequence align-
127. Robertson G, Schein J, Chiu R, et al. De novo assembly and analy- ment using ClustalW and ClustalX. Current protocols in bioin-
sis of RNA-seq data. Nat Methods. 2010;7(11):909–912. formatics / editoral board, Andreas D. Baxevanis. . . [et al.]. Aug
128. Nagarajan N, Pop M. Sequence assembly demystified. Nat Rev 2002;Chapter 2:Unit 2.3.
Genet. 2013;14(3):157–167. 153. Lio P, Goldman N. Models of molecular evolution and phylogeny.
129. Grabherr MG, Haas BJ, Yassour M, et al. Full-length transcriptome Genome Res. Dec 1998;8(12):1233–1244.
assembly from RNA-Seq data without a reference genome. Nat 154. Huelsenbeck JP, Ronquist F. MRBAYES: Bayesian inference of
Biotech. 2011;29(7):644–652. phylogenetic trees. Bioinformatics. 2001;17(8):754–755.
9 6 • P rincip l e s o f G e no m ic M e dicin e
7.
PHARMACOGENOMICS—CRITICAL COMPONENT
OF GENOMIC MEDICINE
Wolfgang Sadee
97
fact be avoided or at least reduced if we were to better under- pharmacogenomics was initially employed in the pharma-
stand genetic differences between individuals. ceutical industry to accelerate drug target discovery. With
With the emergence of genomics, our view of genetic genome sequencing on the horizon, researchers expected
factors in disease and therapy broadened to account for to find the genes and their protein products critical to the
biological complexity on an ever-increasing scale. Beyond disease process, and hence countless suitable drug targets.
monogenic Mendelian diseases, common disorders such as However, the complexity of human diseases has foiled
cardiovascular and central nervous system (CNS) disorders, rapid discovery of valid targets and drugs, leading to a
diabetes, and cancer are considered to have polygenic origins. reemergence of multi-target and multi-drug therapies—
Moreover, gene–environment interactions are critical deter- surreptitiously confounding attempts to understand drug
minants of well-being and disease, with epigenetic modula- response at the genetic level. On the other hand, upon
tion of cellular regulation playing a key role. In contrast to growing adoption of genomics concepts, pharmacogen-
a less complex genetic architecture of ADRs, we therefore etics morphed into pharmacogenomics with the expressed
expect drug efficacy to depend on complex processes so that goal of optimizing personalized drug therapy, including
variability in drug response is multifaceted—with genetic multi-component therapies. Therefore, pharmacogenomics
factors contributing to a different extent in each specific has become an integral part of the burgeoning “personal-
drug therapy. With functionally relevant genetic variants ized medicine,” or “personalized health care”—the latter
largely uncertain,6 known genetic factors in drug response emphasizing the importance of early intervention and pre-
often account for only a small portion of variability, reduc- vention. A deeper understanding of the causes underlying
ing the value of predictive clinical biomarker tests. Since variable drug response has the potential to minimize ADRs
drug efficacy in the treatment of complex disorders ranges and maximize efficacy; therefore, the use of biomarker tests
from 20–80%, a substantial portion of the patient popula- with drug therapy, or for that matter any other intervention
tion fails to benefit while nevertheless being exposed to risk or preventive measure, emerges as a main goal, predicting
of ADRs. Understanding the causes underlying the failure an individual’s disease risk and response, and guiding op-
of a patient to respond to a specific drug therapy would yield timal therapeutic strategies. Table 7.1 summarizes the main
a quantum leap in optimal care and prevention. goals of pharmacogenomics, leading to biomarkers based
In view of the complex biology impinging on drug re- on DNA sequence variants, but also other -omics panels,
sponse, it became compelling to embrace the emerging such as transcriptomics, proteomics, and metabolomics
genomics concepts and technologies: most prominently, (Figure 7.1). It is noted that such -omics biomarkers rep-
large-scale sequencing, leading to a transformation of resent phenotypes rather than genotypes but are neverthe-
pharmacogenetics into pharmacogenomics. Also, concepts less included with the term “pharmacogenomics”. Because
guiding the study of the genomic DNA sequence, its epigen- all these areas cannot be adequately addressed here, this
etic modifications and chromatin regulation, and the tran- chapter will focus on genetic factors predicting drug re-
scriptome, proteome, and metabolome, merged and indeed sponse and toxicity, arguably one central focus of current
became inseparably intertwined. Therefore, pharmacogen- research. Epigenetic factors and gene–environment interac-
omics now is understood to encompass all these areas, inte- tions, and also phenotypical markers derived from RNA,
grating complex patterns in both genotype and phenotype, protein, and metabolite profiles, will be addressed where
in an attempt to optimize and individualize drug therapy. appropriate.
In the context of genomic medicine, pharmacogenomics has
assumed a critical role in implementing genomic principles
P H A R M AC O G E N O M I C S —BA S I C
in clinical practice. In his article “Genomics, Health Care,
R E S E A RC H
and Society,” Hudson states: “One of the most promising
areas of genomic medicine is the ability to match an individ- Distinct from medical genetics/genomics, pharmacogen-
ual’s genetic profile to the likely effect of particular drugs.”7 omics largely employs drug-related phenotypes, including
pharmacokinetic parameters (PK; metabolism, transport,
clearance, half-life, protein binding, etc.) and pharmaco-
R E S E A R C H A N D D E VE L O PM E N T dynamics (PD; drug response and toxicity). On the other
I N P H A R M AC O G E N O M I C S hand, differential diagnosis of a disease profile can also
guide selection of the drug class, showing that medical
Early pharmacogenetics studies focused on indi- issues and pharmacogenomics cannot be neatly separated.
vidual variability in drug response and toxicity, while Moreover, recorded family histories typically cover disease
Pharmacogenomics biomarkers
phenotype versus genotype
almost exclusively, with drug-effect GWAS results emerg-
ing only recently.
Pharmacogenetics It is perhaps instructive to consider the nature of genetic
Transcriptomics
Sequence variations
mRNA profiles variability associated with disease risk and drug response. On
(genotype)
the basis of traditional genetic studies and GWAS results, a
common assumption is that highly penetrant disease-risk
variants are rare—being under negative evolutionary selec-
Proteomics Metabolomics tion pressure—while numerous gene variants exerting low
protein profiles metabolite profiles
penetrance can be frequent and associate with complex dis-
orders (Figure 7.2). The latter have been revealed by GWAS,
Figure 7.1
Main types of pharmacogenomics biomarkers. Note that only pointing to the pathways involved in the disease process;
the genomics/genetics panel represents a genotype; the other panels are however, assuming additive effects of all known genetic vari-
phenotypes. Biomarkers derived from these areas vary in their clinical
applications. ants associated with a complex disorder, one accounts only
for a small portion of the estimated high degree of herit-
ability in a population. This gap in our current knowledge
incidence but not heritable response to drug therapy. The is called the “missing heritability,”11 also applicable to drug
strong heritability often associated with drug response response phenotypes—we often do not know as yet the
therefore is not readily obvious from family histories or degree of heritability of a drug response trait, and which
linkage studies, but an assessment has to rely on newly genetic factors contribute, even when environmental inter-
designed studies of drug therapies in target populations, or actions are factored in. We can look for two possible paths
involving twin studies.8–10 Also, genome-wide association to clarifying the “missing heritability.” First, using mathem-
studies (GWAS) initially targeted disease phenotypes atical modeling, Lander and colleagues12 have proposed that
??
Frequent,
low Altered protein Altered mRNA
Altered
impact sequence and processing and
transcription
function translation
low
low high
Allele frequency Figure 7.3
The nature of genetic variation with phenotypical
consequences. “SNPs” is used here to represent all sequence variations,
Figure 7.2
Frequency and penetrance of genetic variants in complex traits. including SNPs, insertion/deletions, copy number variants, and
We postulate the existence of frequent variants (mostly regulatory) that chromosomal rearrangements, such as inversions and translocations. For
have accumulated as a result of genetic drift or under positive selection more information, see Sadee et al.6
pressure, and therefore are not primary risk factors per se—only under
specific environmental challenges and with amplified epistatic effects.
Early genetic studies have focused on non-synonymous
SNPs (cSNPs) that alter a protein’s function directly.
epistatic gene–gene interactions have the potential to fill the
However, regulatory variants appear to be more frequent,
gap, thereby moving away from the common additive mod-
affecting gene expression (rSNPs) and RNA functions,
els adopted from GWAS results considering many candidate
such as elongation, splicing, turnover, editing, cellular traf-
single nucleotide polymorphisms [SNPs] each with low
ficking, and translation. We have termed these latter vari-
penetrance. The authors further predict that search for add-
ants “structural RNA SNPs” (srSNPs) (Figure 7.3).7,13 Of
itional variants will account only for a rather small fraction
course, sequence variations also include repeats, insertions/
of the missing heritability under additive model assumption.
deletions, inversions, and translocations, etc., and the term
Second, I propose the existence of frequent polymorphisms
“SNP” is used here for the sake of simplicity (unless speci-
with relatively high penetrance13; however, the affected phe-
fied otherwise). The effect of regulatory variants is subject
notypes may not be directly related to disease risk, except
to the cellular, tissue, whole body, and external environ-
when an individual is exposed to specific environmental
ment, opening many pathways for evolutionary pressures
conditions, including drug exposure, that reveal the genetic
to select optimal conditions for physiological target tissues
effect. Such variants can be frequent because of population
while avoiding deleterious effects in other tissues; hence
substructures, such as bottleneck populations with founder
we can expect some regulatory variants to be frequent
mutations, or they could have been under positive evolu-
without showing obvious disease-risk phenotypes, with
tionary selection because of some reproductive advantage;
effects that are restricted to target organs. I will discuss how
clearly, these genetic variants are unlikely to confer strong
cSNPs and regulatory variants affect drug response and
disease risk and therefore typically are missed by GWAS.
ADRs. Some mutations may be rare with high penetrance,
Multiple studies support the notion that countless regions
leading to idiosyncratic drug reactions; others are quite fre-
of the human genome bear evidence of positive selection,14
quent, affecting ADRs and PK/PD parameters (Table 7.1).
indicating selectable features with phenotypical pene-
Each of these drug phenotypes requires different experi-
trance. Yet, such genetic variants could act as a double-edged
mental approaches to understand the underlying molecular
sword: given certain environmental conditions, the genetic
and genetic mechanisms, and clinical implications (the
variants can turn deleterious, particularly when effects are
Pharmacogenomics Knowledge Base PharmGKB [https://
amplified epistatically (non-linear gene–gene interactions;
www.pharmgkb.org/index.jsp] provides detailed informa-
examples will be provided further below). Drug therapy is
tion on drug–gene interactions).
one such environmental condition that humans have only
recently been exposed to on a massive scale, and therefore,
can reveal hidden variants of strong effect. A case in point is
Discovery of Variants in Genes Relevant to Drug
the frequent mutations in drug-metabolizing enzymes that
Response and Toxicity
can dramatically increase drug exposure in poor metabo-
lizers, or affect their response to environmental toxins and We must tailor the approach to identifying variants with clin-
carcinogens. ical pharmacological relevance to the problem at hand. First,
mRNA
B mRNA (cDNA)
34 35 36 37 38 39 40
DNA
DNA
Figure 7.5
Detection of allelic mRNA ratios, in comparison to allelic gDNA ratios. mRNA and DNA are extracted from a human tissue (e.g., liver),
PCR amplified, and the alleles labeled with fluorescent deoxyribose trinucleotides. The fluorescent amplicons are then analyzed by capillary
electrophoresis. In this example, the allelic gDNA and mRNA ratios differ twofold, revealing a robust allelic expression imbalance (AEI), indicative
of the presence of regulatory variants in this gene locus (cis-eQTL) in this tissue.
Similarly, non-coding RNAs such as microRNAs and lin- debate. Most prominent genes encoding drug-metabolizing
cRNAs are now recognized to play critical roles in all types enzymes and transporters have been intensely studied,
of diseases and drug therapies but have yet to advance to the while regulatory variants still are insufficiently explored.
stage of serving as routine biomarkers. Currently, SNP genotyping panels are commercially avail-
The following sections focus on genetic sequence varia- able, containing hundreds to thousands of ADME vari-
tions, and implications for individualized drug therapy. ants in numerous genes, but one suspects that these panels
I will survey genetic findings related to pharmacokinetics only account for a portion of the genetic variability; more-
and pharmacodynamics (Table 7.2), the latter divided into over, for the vast majority of the genotyped SNPs, a clear
drug efficacy and adverse effects (including IDRs). As many understanding of their function and clinical impact is still
reviews and monographs cover the genes relevant to these pending. While clinical association studies require rigorous
topics, the discussion will focus on overarching principles, replication, molecular genetics studies are often taken at
with only a few select examples. face value before the underlying mechanisms fully under-
stood. Moreover, the prevailing use of surrogate markers ra-
ther than validated causative variants in clinical biomarker
G E N ET I C FAC TO R S I N
tests introduces additional error, lowering any predictive
P H A R M AC O K I N ET I C S
value, even though we do have the research tools to fully
Drug absorption, distribution, metabolism, and excretion understand the underlying biology—an effort that should
(ADME) varies with health status, sex, age, body weight, be of high importance.
diurnal and seasonal rhythms, and environmental condi-
tion (for example, xenobiotics inducing drug-metabolizing
Drug Metabolizing Enzymes (DMEs)
enzymes). In the introduction of this chapter, I have dis-
cussed how genetic factors can have dramatic effects on drug Several metabolizing enzymes are critical for drug elimin-
response; however, to what extent genetics (and epigen- ation from the body, as most drugs are lipophilic and there-
etics) contributes to overall drug response remains a topic of fore not readily excreted into the urine, because of tubular
IIF IIIF F1 F2 R2 R1
Schematics of the mRNA isoforms generated from the NAT1 gene locus. Several transcription start sites, alternative splicing sites, and
Figure 7.6
polyadenylation sites generate multiple mRNA isoforms with distinct properties, while the protein coding region is largely unchanged. NAT1 *10
is located in the 3′UTR, enhancing translation, while the *11 haplotype comprises multiple SNPs across the 3′ end of the coding region and the
3′UTR, altering usage of the poly-adenylation sites, thereby also enhancing translation.33
irinotecan (its active metabolite is eliminated by UGT1A1) co-substrates (e.g., sodium ions and protons) that can drive a
in 9-repeat carriers32; however, this effect may only be sig- drug against its concentration gradient. Typically, SLCs are
nificant if high irinotecan doses are given, so that clinical needed to enhance drug entry into the cell, whereas ABC
utility is still undecided for UGT1A1-irinitecan. Another transporters are extrusion pumps and therefore can func-
common DME reaction is acetylation, catalyzed by tion as drug resistance factors, including MDR1 (ABCB1),
N-acetyltransferases 1 and 2 (NAT1 and NAT2). Showing the first such chemoresistance factor found upregulated
a bimodal distribution, low acetylator status enhances effi- in certain cancer tissues. Because MDR1 is highly preva-
cacy of isonizide in the treatment of tuberculosis, but it can lent in multiple tissues, including the blood–brain barrier,
also enhance toxicity. Rapid NAT2 acetylators may require much research has been devoted to its role and genetic
more frequent dosing to be effective in tuberculosis therapy. factors that might alter drug response. We have reported
Genetic variation in NAT1 was less certain, while the *10 that a synonymous SNP (C3435T) affects mRNA turn-
and *11 allele had been reported to alter expression of the over, thereby lowering expression of the mature protein.34
enzyme by unknown mechanisms, with associations of risk However, the effect is rather small, and countless clinical
of certain cancers—acetylation can lead to both xenobiotic studies have yielded conflicting results. A robust effect may
inactivation or activation to more proximate carcinogens. occur in the treatment of HIV/AIDs with agents that are
Relevant to local effects, NAT1 is expressed in many tissues substrates of MDR1, as the antivirals must enter into lym-
(NAT2 mostly in the liver) and metabolizes amines, includ- phocytes where MDR1 is expressed. Low expression associ-
ing sulfonamides such as sulfamethoxazole, substrates that ated with the 3435T allele appears to enhance drug efficacy
also interact with NAT2. We have determined that the in some but not all studies.35 One therefore needs to distin-
NAT1*10 and *11 alleles enhance protein expression by guish between general PK effects of transporters on overall
regulating translation, thereby protecting against sulfa- clearance and distribution in the body, and specific effects
methoxazole-induced skin rash in HIV/AIDS patients— at target sites, such as lymphocytes and the blood–brain
an effect only observed in poor NAT2 metabolizers, a rare barrier. MDR1 contributes to a formidable barrier for sub-
example of a gene–gene–drug interaction in the pharmaco- strates to enter the brain in appreciable quantities, a char-
genomics literature.33 NAT1 also serves as an example of the acteristic exploited for peripheral use of the potent opioid
diverse mRNA isoforms typically generated from a single loperamide for treatment of diarrhea. Substantial genetic
gene locus, as a result of alternative transcription start sites, differences in MDR1 activity in the blood–brain barrier
alternative splicing events, and alternative polyadenylation (BBB) would readily become apparent by unwanted ad-
sites (Figure 7.6); indeed, we found that *11 alters polyad- verse opioid effects such as respiratory depression. The lack
enylation site usage, there by favoring the mRNA transcript of these events,36 even though loperamide is administered
with enhanced translation capacity.33 to countless patients, suggests that low MDR1 transporter
status is rare, or alternatively, that secondary mechanisms
can substitute for MDR1-mediated efflux.
Transporter Genes
Because of the sheer number of membrane trans-
Numerous transporter genes are involved in pharmaco- porter genes, often with low-affinity promiscuous sub-
kinetics. These are divided into two main classes, solute strate interactions, one would not expect polymorphisms
transporters (SLCs) and active ABC transporters (ATP in single genes to have significant impact on drug distribu-
binding cassette). The former allow facilitated diffusion tion and effect. Nevertheless, a number of examples have
along concentration gradients, or are secondarily driven by emerged where low-activity variants dramatically alter PK
Table 7.3A SELECT ENTRIES IN THE FOOD AND DRUG ADMINISTRATION’S TABLE OF PHARMACOGENOMIC
BIOMARKERS IN DRUG LABELS
URL:
DRUG THERAPEUTIC BIOMARKER LABEL SECTIONS WITH PHARMACOGENOMIC
AREA INFORMATION
Abacavir Antiviral HLA-B*5701 Boxed warning contraindications Warnings and
precautions
Patient counseling
Carbamazepine Neurology HLA-B*1502 Boxed warning
Cetuximab Oncology EGFR Indications and usage warnings and precautions
KRAS Indications and usage
Clopidogrel Cardiovascular CYP2C19 Boxed warning
Dasatinib Oncology Ph Chromosome/BCR-ABL Indications and usage
Doxepin Psychiatry CYP2D6 Precautions
Imatinib C-KIT, BCR-ABL, PDGFR, FIP1L1-PDGFRa Indications and usage
Maraviroc Antivirals CCR5 Warnings and precautions
Tamoxifen Oncology ER Indications and usage
Trastuzumab Oncology Her-NEU Indications and usage
THE LINK IS PROVIDED IN THE TABLE OF PHARMACOGENOMIC BIOMARKERS (TABLE 7.3A). THE TEXT IS ABBREVIATED.
SERIOUS DERMATOLOGIC REACTIONS AND HLA-B*1502 ALLELE
SERIOUS AND SOMETIMES FATAL DERMATOLOGIC REACTIONS, INCLUDING TOXIC EPIDERMAL NECROLYSIS
(TEN) AND STEVENS-JOHNSON SYNDROME (SJS),. . . . ESTIMATED TO OCCUR IN 1 TO 6 PER 10,000 NEW USERS
IN. . . . CAUCASIAN POPULATIONS, BUT THE RISK IN SOME ASIAN COUNTRIES IS ESTIMATED TO BE ABOUT 10
TIMES HIGHER. . . . STRONG ASSOCIATION BETWEEN THE RISK OF DEVELOPING SJS/TEN AND THE PRESENCE OF
HLA-B*1502, AN INHERITED ALLELIC VARIANT OF THE HLA-B GENE. HLA-B*1502 IS FOUND ALMOST EXCLUSIVELY
IN PATIENTS WITH ANCESTRY ACROSS BROAD AREAS OF ASIA. . . . PATIENTS WITH ANCESTRY IN GENETICALLY
AT-RISK POPULATIONS SHOULD BE SCREENED FOR THE PRESENCE OF HLA-B*1502 PRIOR TO INITIATING
TREATMENT WITH TEGRETOL. PATIENTS TESTING POSITIVE FOR THE ALLELE SHOULD NOT BE TREATED WITH
TEGRETOL. . . .
Table 7.4 COMPANION BIOMARKER TESTS FOR TARGETED ANTICANCER AGENTS DIRECTED AGAINST
“DRIVER” ONCOGENE KINASES
INTRODUCTION I N T E G R AT I O N O F G E N O M I C S
I N TO T H E D RU G D I S C O VE RY
In its broadest sense, pharmacogenomics can be defined as A N D D E VE L O PM E N T
the investigation of variations of DNA and RNA character-
istics as related to drug response. The last decade has seen Pharmaceutical companies have historically focused their
a large increase in the amount of genomics data generated drug discovery and development programs on finding ther-
and, with it, increased the expectations of how improved apies for broad use in large disease populations, the “block-
understanding of disease will lead to the development buster business model.” A blockbuster drug is usually defined
of more effective therapies and personalized medicines. as one with peak annual sales of greater than $1 billion and
Despite the regular reports of novel genes being identified is generally developed for long-term use to treat common
in a range of disorders, by 2012 the much-heralded promise complex chronic disorders in the general population. The
of the human genome project has only started to materi- strategy to identify and develop blockbuster drugs has been
alize. However, it is important to realize that this does not the response to the high costs of drug discovery and devel-
represent a failure of the science to deliver as there are mul- opment. A survey of the drug development costs of 68 new
tiple clear examples of the predictability and clinical utility compounds from 10 pharmaceutical companies estimated
of pharmacogenetics but is a reflection of the length of time that the cost to develop a new drug in 2000 was $802 mil-
it takes to develop new drugs and implement changes in lion (DiMasi et al., 2003). The high costs of developing
healthcare. The 1980s and 1990s saw a boom time for the drugs can be attributed to two main factors: the large size
pharmaceuticals industry producing many highly effective and duration of the clinical trials required to provide the
new classes of drugs from statins to proton-pump inhibi- data to show safety and efficacy of the compound, and the
tors and quinolone antibiotics. All were novel therapeutic high rate of attrition of compounds in clinical development;
approaches offering significant benefit to individuals and fewer than 10% of compounds entering phase I clinical de-
society. The science that drove many of these advances was velopment reach the market, the majority failing in clinical
based on the greater understanding of biochemistry and development due to lack of efficacy in phase II. The lack of
pharmacology that emerged during the 1970s and early recent research and development (R&D) success in finding
1980s. This 10- to 15-year time-lag from gaining scien- blockbuster drugs, combined with financial pressure due to
tific knowledge to developing therapies is typical for the patent expiry and downward pressure on pricing, has led to
pharmaceutical industry, and reflects the complexity of a shift in strategy for many companies in the biopharma-
drug discovery and the time required for preclinical and ceutical industry. Companies are shifting towards the dis-
clinical testing to ensure safety and efficacy. This chapter covery and development of stratified medicines. A stratified
will introduce the major concepts of drug discovery and medicine is one that is targeted at a subgroup of a tradition-
development and give a broad overview of how genetics ally classified disease, such as Herceptin for the treatment
and genomics is used across the whole drug discovery and of Her2-overexpressing breast cancer. Stratified medicines
development pipeline, from pre-target identification to offer a significant opportunity: to the industry, as they have
post-marketing surveillance to help discover and develop an increased probability of success and the potential of
improved medicines. It will describe some of the examples smaller programs; to the regulators, as the benefit–risk pro-
of how pharmacogenetics has impacted the lives of patients. files of these medications are greater than with unselected
114
medications; to the payers, as they are more cost effective; The Drug Discovery and Development
and most importantly, to patients, as they are more effective Process
and safer therapies. Genomics has a large role to play in the
The generation of an idea that a particular protein might be
development of stratified medicines, as many of the tools
a suitable therapeutic target for the treatment of a disease
used to stratify the patient populations are genomic, such as
sets in motion what is often depicted as a linear process
selective epidermal growth factor receptor (EGFR) muta-
known as the “drug discovery and development pipeline,”
tion status and Gefitinib, K ras mutation status and Erbitux
in which new medicines follow a set route from early dis-
and Vectibix, Alk4 mutation status and Crizotinib.
covery and preclinical stages through a set of clinical devel-
Pharmacogenomics—the investigation of variations
opment processes to the marketplace (Figure 8.1). In reality,
of DNA and RNA characteristics (germline or tumor) as
the process is generally far from linear, but for the purposes
related to drug response in individual patients or groups of
of describing the component parts, we will consider it a se-
patients—is one of a number of methods employed by the
quential process.
pharmaceutical industry to stratify patient populations.
A major cause of the attrition of drugs for lack of effi-
Candidate-Seeking
cacy is the heterogeneity of the diseases we currently classify
as single entities. Most would be better referred to as syn- The ultimate aim of the drug discovery process is to find a
dromes rather than single diseases. The disease classification chemical (e.g., small molecule) or biological reagent, such
currently used is based on phenotypical consequences of as an antibody, that has the potential to be a drug that can
disease processes rather than on the underlying pathological be moved into preclinical and then clinical testing. In order
mechanisms. This has led to the clustering of heterogeneous to start the process of identifying a potential drug, a bio-
disease syndromes based on symptoms rather than based on logical assay testing interactions with the drug target must
molecular pathology. Genomics will be an important tool be developed. This assay is often based on a cloned and
in reclassifying diseases into a new molecular taxonomy of expressed form of the drug target and will be converted into
human disease. Oncology is one therapeutic area where this a format that will allow high-throughput testing, as millions
is most advanced, as the scientific evidence base for tumor of chemicals may need to be screened in the assay. The need
etiology is more advanced than in other areas. The majority to screen millions of chemicals means that it is usually only
of drug development programs in oncology are now strati- feasible to screen one protein variant of the target in the
fying patient populations based on molecular changes in high-throughput screen. It is therefore vital to screen the
the tumor. During the period from 2005 to 2012, over 5 “right” variant. In the situation where there may be more
stratified medicines in oncology were approved (Table 8.1). than one form of the protein that can be included in the
Most of the current drug development programs in on- screen, it is important to know that the most biologically
cology are using a stratified medicine approach linking the relevant and/or the most common variant is being screened,
target to the dysregulated disease pathways on the tumors and it may be necessary to screen the chemical matter
and only being used when the right pathway is driving against more than one form of the protein. This is not al-
tumorigenesis. It is widely expected that this approach will ways the most common form of the protein—Verumafenib,
expand across other therapeutic areas as our understanding a novel drug for the treatment of malignant melanoma,
of disease biology improves. was identified by specifically screening against the V600E
Candidate Filing
seeking
mutated form of the BRAF protein to ensure that it only toxic effects, it can be insensitive to subtle changes and can
blocked signaling of the pathogenic form. identify species-specific effects that can be difficult to in-
The high-throughput screens generally identify several terpret. A greater understanding of the molecular changes
potential “hits,” which need to be tested in more rigorous following drug administration could identify more subtle
biological assays to determine the type of interaction and effects and species-specific effects. Similarly, the applic-
the effects and then refined using medicinal chemistry. ability of animal models of a disease could be assessed by
Promising “leads” are then developed by a series of minor evaluating molecular changes rather than phenotypical
chemical changes to the original lead, and the final candi- similarities that can be misleading. Greater emphasis is now
date is chosen based on the selectivity and potency criteria being placed on molecular and biomarker changes that re-
required for the drug candidate as well as the physicochemi- sult in organ damage, where they are available (e.g. nephro-
cal properties of the molecule to ensure druglike properties. toxicity biomarkers).
This candidate is then taken forward into preclinical testing. Adverse events can be due to unexpected consequences
The final testing phase is usually based on in vivo test- of the primary pharmacology or to unexpected interactions
ing of the compound in animal models that have been dem- with off-target proteins. Understanding the mechanism of
onstrated to have some translatability to the target human the toxicological effects is important, as this allows a more
disease, or in a range of ex vivo models of human tissue that quantitative evaluation of the risk of the event happening
recapitulate components of the disease. The predictability in humans. Genomics can be used to identify interactions
and translatability of these models to humans varies with with off-target proteins as transcription changes induced
different diseases and is the focus of biomedical research in in the organ damaged by the compound can point to the
many therapeutic areas. mechanism of the toxicity. This is often referred to as toxi-
cogenomics. Multiple consortiums (e.g., the Predictive Safety
Testing Consortium [PSTC] and the Safety in Science in
Preclinical Testing
Medicines Education & Training [SAFESCIMET]) are
Once a drug candidate has been made, it goes into a pre- currently working to identify genomic biomarkers that are
clinical toxicology testing that includes in vitro screening more sensitive than current histopathological scores, allow-
tests to identify potential pharmacological effects at other ing early detection of toxicology and the demonstration of
receptors that could lead to adverse events, and genetic toxi- species-specific toxic effects. Similarly, where specific organ
cology testing, which evaluates mutagenicity and clastogen- toxicity is expected due to the mechanism of action of the
icity. Only if these are satisfactory does animal testing begin. compound or known off-target effects, then transcription
The animal testing is done in two mammalian species and is changes can offer a more sensitive assay to detect early organ
staged to ensure that as few animals as possible are used and damage.
that major problems are picked up early. Toxicology studies
to evaluate long-term exposure, reproductive toxicological
Clinical Development
effects, juvenile toxicity, and carcinogenicity are generally
only performed once the data have been obtained from Once the initial in vitro testing and acute animal toxicology
shorter-term human studies that support safety and efficacy. studies (which generally take 14 days) have been performed,
To date, toxicology induced by new chemicals are identified then it is possible to start testing the candidate in humans.
and classified by standard phenotypical and histological The human studies have traditionally been split into four
changes. While this picks up the majority of potentially phases (phases I–IV), each with specific aims (Box 8.1).
Effect of Genetic Variation on Compound not intended to replace any of the clinical studies required
Screening for exploratory drug development or predict response in
patient populations. The preclinical strategy will produce
Regardless of the original source of the target, genetic analy-
data to inform the pharmacogenomic plan for compounds
ses are important in understanding how to move forward
in exploratory and full development. The challenge facing
in the drug discovery process. Undertaking a comprehen-
pharmacogenomics specialists in the pharmaceutical in-
sive analysis of the genetic variation that exists in putative
dustry is to use the available genomic data to improve the
drug targets will provide information that could have a
efficiency of clinical trials.
powerful impact on drug-discovery processes downstream.
In an internal study within Pfizer, Inc., comparing coding
SNP (cSNP) frequency, a selection of 111 genes encoding A P P LY I N G G E N O M I C S TO D RU G
potential druggable targets and 160 genes considered as D EVE L O PM E N T
“non-druggable” targets found that 15% (26/111) of the pu-
Pharmacogenetics
tative targets were not polymorphic at the amino acid level,
while 40% (45/111) had one or two cSNPs. There are also There are several definitions of pharmacogenetics in the lit-
well-documented differences in the frequencies of specific erature, but the term was originally used in 1959 by Vogel
polymorphisms between ethnic groups. Prior knowledge to describe the inter-individual differences in drug response
of any polymorphisms in a target can be incorporated into due to variations in DNA (Vogel, 1959). Although this
target validation, lead optimization, and inform preclinical is the origin of the term, the concept of inherited differ-
projects supporting the development of the compound. ences in biochemical attributes dates back much further,
The effect of genetic variation can be assessed through in with Garrod describing the inheritance of alcaptonuria and
vitro assays that incorporate a comparison of polymorphic phenylketonuria in 1902, and Snyder in 1932 describing
targets by using either cells or biological reagents obtained the inherited ability to taste (or not) phenylthiocarbamide
from donors of known genotypes (where available), or by (Garrod, 1902; Snyder, 1932). The article by Motulsky in
site-directed mutagenesis. This will facilitate early assess- 1957 was the first serious attempt to understand the basis
ment of the potential impact of genetic variation on the of inherited inter-individual response to drug therapies,
activity of compounds and offer the potential to choose with descriptions of the effects of glucose-6-phosphate
candidates that are the least likely to be influenced by the dehydrogenase (G6PD) deficiency and primaquine in
target polymorphism (Penny and McHale, 2005). African-American soldiers (Motulsky, 1957). During World
Gaining an early understanding of the impact of gen- War II, scientists from the University of Chicago observed
etic variation can increase confidence in chemistry (CIC). that approximately 10% of black American soldiers and
For example, CCR5 has been shown to be the second core- (rarely) some of the white soldiers developed hemolytic an-
ceptor required for primary HIV infection. As such, it was emia of varying severity when given conventional doses of a
a very attractive drug target for the treatment of HIV, as then-new antimalarial drug, primaquine. Further investiga-
blockade of CCR5 should reduce HIV entry into cells and tion revealed that this was due to the lack of the G6PD en-
hence lower viral turnover. There have been multiple poly- zyme in red cells, which was the same genetic defect that had
morphisms reported in the CCR5 gene, and some of these been shown to be responsible for the development of hemo-
have been associated with effects on HIV infection rates lytic anemia in susceptible individuals following the inges-
and/or progression from infection to AIDS. A key question tion of fava beans. This was one of the first descriptions of a
that had to be asked was, What were the functional effects Mendelian (X-linked) pharmacogenetic trait. Also, in 1957,
of these polymorphisms, and would they would impact the Kalow and Genest described an autosomal recessive pharma-
effectiveness of the therapy? Preclinically, it was possible to cogenetic trait (Kalow and Genest, 1957). Approximately 1
demonstrate that the predominant effect of the functional in 2000 subjects undergoing anesthesia develop a prolonged
polymorphisms was to alter receptor expression rather than pharmacodynamic effect of succinyl choline due to a defi-
structure; hence, the variability could be managed by iden- ciency in the enzyme pseudocholinesterase. This autosomal
tifying a dose that could effectively inhibit viral entry across recessive trait has since been recognized in a wide variety of
a wide range of receptor expression levels. ethnic populations, and although the enzyme deficiency was
The pharmacogenomic studies included in the preclin- identified in 1957, it was a further 30 years before the causa-
ical phase of drug discovery that provide CIR and CIC and tive genetic mutations responsible for these reactions were
support nomination of a candidate for development are identified (McGuire et al., 1989).
www.Ebook777.com
Pharmacogenetics remained a relatively small field until the polymorphic metabolism of debrisoquin that significant
the 1990s, due to the fact that although it was well recog- interest grew in the genetic contribution. The cytochrome
nized that all drugs exhibited significant inter-individual P450 (CYP) enzyme family protects the body from xeno-
variability in response, the genetic tools to examine this biotic agents and is the major route of metabolism of many
variability were not available. Apart from a few standard drugs (Danielson, 2002). Several of these enzymes (e.g.,
approaches (e.g., renal impairment studies and gender cytochrome P450 2D6, 2C9, and 2C19) are known to have
differences), there was limited investigation of this phe- functional genetic polymorphisms that result in significant
nomenon during drug development. The approach of the reductions or increases in function (Lee et al., 2002; Shimizu
drug companies and regulators alike was to ensure that all et al., 2003). Genetic variation in cytochrome P450 2D6
compounds had a sufficiently good therapeutic index that (CYP2D6) is well characterized, and approximately 10% of
the average benefit significantly outweighed the potential Caucasians make no CYP2D6 enzyme. Experiments with
risk. This has led to the withdrawal or termination of de- the antihypertensive agent debrisoquin yielded the first
velopment of a number of compounds with good efficacy proven examples of a pharmacogenetic effect. Debrisoquin
but an insufficient population-based safety profile, which is metabolized by the CYP2D6 enzyme. An individual who
can often be driven by a small number of potentially serious makes no CYP2D6 and takes a standard dose of debriso-
adverse events. These events can be categorized into those quin will suffer a profound hypotensive event resulting
that are expected based on an understanding of the pharma- from high plasma exposure levels due to an inability to me-
cological action of the drug (type A), and those that cor- tabolize the drug (Idle et al., 1978). Approximately 20% of
relate with plasma exposure levels or idiosyncratic (type B) all drugs are metabolized by CYP2D6, and subjects who are
(Rawlins and Thompson, 1991). The mechanisms of idio- unable to make this enzyme are at increased risk of devel-
syncratic reactions are generally unknown and do not have oping adverse events when taking one of these compounds
a clear dose–response relationship. (Cascorbi, 2003) (Figure 8.2).
The incorporation of genetic testing for CYP2D6 or
related enzymes in clinical trials has the potential to iden-
Pharmacokinetic Variability
tify, prospectively, subjects who are likely to have adverse
Inter-individual variation in drug metabolism is now events due to poor metabolism, or those who may have
a well-documented phenomenon, but it was not until limited response through inadequate exposure because of
Mahgoub et al, (1977 Lancet 2[8038]:854–856) described ultra-rapid metabolism.
xxx x
xxx xxxx x x
x xx x x xxx xxx
Plasma drug level
8%
10% 80%
2%
No CYP2D6 activity
Below normal CYP2D6 activity
Normal CYP2D6 activity
Excess CYP2D6 activity
Effect (%)
50 50 wt/m
30 variation in genes affecting PD response approaches 0.5,
m/m
0
0 24 hr
0
0 50 100 the predictive power of a test solely looking at drug me-
tabolism decreases. Similarly, the predictive power of a
(B) 100
wt/m
100
wt/wt test evaluating variation in genes impacting PD response
Drug conc.
Effect (%)
50 50 wt/m will vary depending upon PK variability. Most pharma-
65 cogenetic studies that have been published to date have
m/m
0 0
0 24 hr 0 50 100 concentrated on single genes or small numbers of can-
didate genes, which are likely to affect either PK or PD
100
m/m
100
wt/wt variability. It is unsurprising that these studies fail to dem-
Drug conc.
Effect (%)
Adapted from Evans WE, McLeod HL. Pharmacogenomics—drug disposition, drug targets, and side effects. N Engl J Med. 2003 Feb 6;348(6):538–549.
9.
MITOCHONDRIAL GENETICS AND GENOMICS
IN CLINICAL MEDICINE
Agnès Rötig and Dhavendra Kumar
131
www.Ebook777.com
Eye
Heart Optic neuropathy
Conduction disorder Ophthalmoplegia
Wolft-Parkinson-White Retinopathy
Skeletal muscle
Weakness Cardiomyopathy
Fatigue
Myopathy
Neuropathy Liver
Hepatopathy
Brain
Seizures Kidney
Myoclonus Fanconi's syndrome
Ataxia Glomerulopathy
Stroke
Dementia
Migraine
ATP
Nuclear DNA
Pancreas
subunits OX PHOS
mt DNA Diabetes mellitus
O2 H O
2
Defects in intergenomic
communication
Multiple mtDNA deletions and
mt DNA depletion
Blood
Pearson's syndrome
Inner ear
Colon
Sensorineural
Pseudo-obstruction
hearing loss
OX PHOS = OXIDATIVE PHOSPHORYLATION
Outer membrane
H+ H+ H+ H+
c
Q
Inner
membrane Q
Matrix
NADH Succinate O2
Electron flux
Proton flux
ATP ADP
D-loop
NA Phe 16S
7S D Val
Thr 23S
CYB
Pro Leu
H
STRAND PL NDI
Glu IIe
ND6 f-Met
Gln
ND2
ND5 Ala
OL.... Asn
Cys Trp
Tyr
Leu
Ser
His CO1
Ser
ND4
Asp
L ND4L
STRAND CO2
ND3
CO3 Lys
Arg ATPase8
Gly
ATPase6
Figure 9.3
The heavy (H) and light (L) strands of the circular 16,659 bp mtDNA double helix are shown. Protein-coding genes are shaded; transfer
RNA genes are shown as short lines with the name of the amino acid. There are no introns; the heavy arrows indicate the origins and directions
of replication of the two strands; the light arrows show the promoters and direction of transcription of the two multicistronic transcripts that are
subsequently cleaved into individual mRNAs. (Adapted with permission from Figure 9.1 of Strachan & Read, Human Molecular Genetics, 3rd edition, London and New York: Garland; 2004).
genes have been found in patients. Interestingly, most of the molecularly characterized, despite systematic study of the
mutations encountered have been in CI genes; whereas, four genes encoding the CII subunits, suggesting that these
despite the concerted efforts of various teams of researchers, deficiencies could be due to mutations in assembly proteins.
very few mutations in other genes encoding other complex Nevertheless, mutations of the three other genes encoding
subunits have been found. subunits B, C, and D of CII have been reported in heredi-
The first mutation in a gene encoding an RC subunit tary paraganglioma and phaeochromocytoma, suggesting
was reported in 1995 in two sisters with Leigh syndrome that such “housekeeping” genes may be involved in carci-
and CII deficiency. The pathogenic mutation was in the nogenesis.37 It is hypothesized that succinate dehydrogenase
SDHA gene encoding the flavoprotein of CII.34 Mutations (SDH) mutations cause an accumulation of succinate and
in the same gene were subsequently reported in another reactive oxygen species (ROS), which could act as down-
patient who also presented with Leigh syndrome.35,36 stream signaling molecules to activate hypoxia-inducing
However, only a few patients with CII deficiency have been pathways.37
NDUFV1, NDUFS8
NDUFS7, NDUFS1 mtCOXI-III
NDUFA8, NDUFB6 HUMQPC COX4AI2
NDUFS4, NDUFV2 cytb COX6B1 ATP6
mtND1-6
SDH-FP
CI
CIII CIV CV
CII
Figure 9.5
Gene mutations resulting in specific respiratory chain complexes. Genes encoding for the subunits of complexes and assembly factors are
shown in the upper and lower registers, respectively. Genes in boldface are mutations found in several patients or families.
POLG
MPV17
TFAM TFAM CI CIII CIV CV
TOPO
CII
tRNALeu
aminoacyl–tRNA
tRNA synthetases
tRNA processing
EFTu
MRPS22 EFG1
mtDNA rRNA
MRPS16 TSFM
MTFMT MRPL3 C12ORF65
Figure 9.6
Mitochondrial DNA replication and translation of proteins encoded by mtDNA. The gene mutations resulting in mitochondrial diseases
and multiple respiratory chain deficiencies are shown in the black ovals.
1 2 3 4 5 6 7
S U MM A RY
III
Mitochondrial disorders are due to respiratory chain defi-
ciency. Oxidative phosphorylation—ATP synthesis by the
MITOCHONDRIAL oxygen-consuming respiratory chain (RC)—supplies most
MYOPATHY DEAFNESS organs and tissues with a readily usable energy source and is
already fully functioning at birth. This means that, in theory,
RC deficiency can give rise to any symptom in any organ
MYOCLONIC EPILEPSY, DEMENTIA
ABNORMAL EEG or tissue at any age and with any mode of inheritance, due
to the twofold genetic origin of RC components (nuclear
BLACK = SEVERE RED = MILD DNA and mitochondrial DNA). It has long been errone-
ously believed that RC disorders originate from mutations
Figure 9.7
Segregation of variable clinical phenotypes in a family with
multi-system mitochondrial disorder similar to myoclonic epilepsy of mtDNA, as, for some time, only mutations or deletions
lactic acidosis with strokelike symptoms (MELAS). of mtDNA could be identified. However, the number of
148
Performance of Genomic Technologies Across Size Scales
100 101 102 103 104 105 106 107 108 109 bp
PCR-based Methods
MLPA
Southern Blotting
CGH/SNP Arrays
Spectral Karyotyping
Cytogenetics
Figure 10.1
Depicted are the major genomic analysis technologies discussed in this chapter with the approximate genomic size scales to which they are
best suited for detection of anomalies. It should be noted that certain anomalies, such as balanced translocations, require separate consideration. For
example, such translocations may be detected by cytogenics but not by CGH/SNP arrays.
1 5 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
chromosomal bands.14 By this method, late-replicating, anomalies, and chromosomes prepared from these tumors
transcriptionally quiet, and A/T-rich DNA staining is more are generally more condensed and thus have lower banding
intense (G-positive) and early-replicating, transcriptionally resolution, interfering with interpretation. Though modern
active, and relatively G/C-rich DNA is stained more lightly genomics technologies are now beginning to be applied to
(G-negative) (Figure 10.3). The highest quality G-banded these tumors, diagnostic analyses for many large rearrange-
preparations are able to yield up to 850 bands across the ments have typically been performed by FISH.
genome.15 Microscopic analysis of banded chromosomes
assists with their individual identification and can reveal loss
Fluorescence In Situ Hybridization (FISH)
or gain of material down to a few million base pairs (mega-
bases or Mb) in size, though exactly how small depends on In situ hybridization techniques, in particular FISH, have
the location of the defect and the resolution of the bands. It allowed study of the structure of the genome at a level of
can also reveal events such as translocations and inversions, detail greater than that seen by conventional banding tech-
even when no net gain or loss of material occurs (compare niques.25,26,27 However, unlike cytogenetics, FISH is a tar-
with comparative genomic hybridization [CGH] and single geted assay requiring foreknowledge of the expected genetic
nucleotide polymorphisms [SNP] arrays, see below). Other lesion. The method depends on the specific hybridization
banding methods are available that can reveal complemen- of a probe DNA sequence to its complementary sequence
tary information about the structure and organization of the in the genome. Labeling the probe with a fluorescent dye
genome at the chromosomal level. One particular method, allows its location to be revealed by fluorescence micros-
Q-banding, involves a fluorescent dye such as quinacrine, copy. The probes used in FISH experiments are typically
DAPI, or Hoechst 33258, that binds preferentially to A/T- derived from human sequences cloned into bacterial arti-
rich sequences, producing a banding pattern comparable ficial chromosomes (BACs), with sizes of approximately
to G-banding that can be used during fluorescence in situ 100kb. The creation of an extensive human BAC library
hybridization (FISH) experiments (see below).16 (containing approximately 32,000 BACs tiled across the
Though developed many decades ago, cytogenetics genome) was actually the first critical step of the Human
is still widely used in clinical practice today for both con- Genome Project itself, effectively breaking the genome into
stitutional and oncology diagnostics. It is a first-line test “bite-sized” pieces that could be individually sequenced,
for children with developmental abnormalities and for with subsequent assembly to create the final sequence.28
fetal samples obtained via amniocentesis or chorionic vil- This library now serves as a main source of DNA for FISH
lus sampling, when there is reason to suspect a structural probes, thus there are only very few regions of the genome
or numerical chromosomal abnormality.17,18 There is also not amenable to FISH experiments.
a long and rich history of cytogenetic analysis of hemato- FISH can be performed either on metaphase chromo-
logical malignancies, going back to the discovery of the some spreads or on preparations of cells in interphase.29
Philadelphia chromosome in chronic myelogenous leuke- When applied to metaphase spreads, specific signal from
mia in 1960 (the first observation of a recurrent genetic sites of probe binding can be evaluated in the context of
anomaly in cancer),19 followed soon after by the elucidation Q-banding data. Thus, metaphase FISH allows the count-
of the t(9;22) translocation from which it arises.20 Today, ing of probe binding sites, determination of the identity
cytogenetics is still heavily relied upon in the diagnostic of chromosomes showing probe staining, as well as the
work-up of these diseases, particularly the leukemias and sub-chromosomal location of binding events (Figure 10.4).
myelodysplastic syndromes (MDS). There are many cytoge- This is a wealth of useful information, but as with cytoge-
netic signatures for these diseases, including some that are netics, this technique relies on the growth of cells in culture,
disease-defining or major diagnostic criteria, and others that which is expensive and can take days to weeks. Not all speci-
provide therapy-related or prognostic information.21,22,23 men types may grow reliably, and in certain scenarios, this
For instance, the 5q minus syndrome is a specific subtype of process cannot be completed in a clinically relevant time
MDS showing deletions of the long arm of chromosome 5 frame. For example, in cases of suspected acute promyelo-
(often involving 5q31-5q32) that is associated with an over- cytic leukemia, rapid identification of the t(15;17) trans-
all favorable prognosis and a high likelihood of response to location producing the PML-RARa fusion oncogene is
lenalidomide therapy.24 important in order to inform decisions regarding treatment
In contrast to the case in heme malignancies, there is with all trans retinoic acid (ATRA).30 In practice, this is typ-
essentially no routine clinical utility of cytogenetics for solid ically performed either via reverse transcription of RNA to
tumors. Solid tumors tend to have many more cytogenetic DNA followed by PCR (RT-PCR) or via interphase FISH.
www.Ebook777.com
Interphase FISH can be readily performed on essen- the state of the chromatin of non-metaphase cells, which
tially any sample that contains nucleated cells, with no is uncondensed and dispersed throughout the nucleus.
requirement for culture. This includes smears of any cel- However, though the probe lengths may be on the order of
lular bodily fluid, tissue touch preparations and frozen sec- 100kb, this is short enough to produce staining in discrete
tions, and even formalin-fixed, paraffin embedded (FFPE) spots within the nucleus. Of course, by this technique, no
tissue sections. Aside from some differences in the prepa- contextual chromosomal information is produced. This is
ration of cells and target DNA, the underlying concept a limiting factor, yet the process still has many important
is the same as for metaphase FISH. The difference lies in applications:
(A)
Case:
A 2 y old female with developmental delay, short stature and microcephaly, otherwise not dysmorphic. Normal MRI brain scan.
Found to have an apparently balanced translocation between chromosomes 7 and 13 (46,XX,t(7;13)(q21.2;q12.3)de novo), as shown by the two
arrows on the karyotype.
Subsequent array CGH analysis, using a whole genome array of 0.1 Mb BAC clones spaced ~1 Mb apart, has shown that in the regions of
the breakpoints on chromosomes 7 and 13 that there has actually been loss of DNA (indicated in red on the diagram of the chromosomes;
Fig 7.2), and the translocation is therefore not balanced. Loss of 0.2 Mb of chromosome 13 and 8.42 Mb of chromosome 7 is consistent with the
phenotype and its severity.
(Courtesy of Sian Morgan, Cytogenetics Laboratory, Institute of Medical Genetics, Cardiff, UK)
(B)
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X Y
Figure 10.3
Conventional cytogenetic analysis by G-banding. Depicted is a chromosomal spread from a two-year-old female with developmental delay,
short stature, and microcephaly. Cytogenetic analysis revealed a translocation between chromosomes 7 and 13 (46XX, t(7:13)(q21.2;q12.3)), as
shown by the two arrows on the karyotype. A deletion of chromosomal material at the breakpoint region might be expected based on the clinical
scenario, yet the translocation appears to be balanced. However, such is the sensitivity of cytogenetics that even megabase-scale anomalies may not
be readily detectable. (Courtesy of Sian Morgan, Cytogenetics Laboratory, Institute of Medical Genetics, Cardiff, UK)
1 5 2 • P rincip l e s o f G e no m ic M e dicin e
(C)
0
16
32
48
63
i) Chromosome 7 79
95
111
127
143
159
–2.00 –1.60 –1.20 –0.80 –0.40 0.00 0.40 0.80 1.20 1.60 2.00
Chromosome 7 Log2 Ratio Ch1/Ch2
01100200_top - 01100200_bottom.bsn - 14/06/2006
0
11
23
34
46
57
ii) Chromosome 68
13 80
91
103
114
–2.00 –1.60 –1.20 –0.80 –0.40 0.00 0.40 0.80 1.20 1.60 2.00
Chromosome 13 Log2 Ratio Ch1/Ch2
01100200_top - 01100200_bottom.bsn - 14/06/2006
Figure 10.3 Continued
1 5 4 • P rincip l e s o f G e no m ic M e dicin e
(A)
(B)
Normal (overlapping)
Figure 10.5
Interphase FISH. A) For rapid prenatal diagnosis of common aneuploidies, FISH can be performed on cells in interphase obtained at
amniocentesis. In this example, staining with a probe for the centromeric region of chromosome 21 is notable for three signals in every cell, which
is diagnostic for Down syndrome. Unlike conventional cytogenetic analysis, this technique only indicates the number of copies of the probe region,
which does not necessarily equate to the number of copies of whole chromosomes. B) Translocations and other rearrangements with specific
breakpoints can be detected using multicolor “break-apart” FISH. Shown are interphase cells from a 36-year-old patient with newly diagnosed acute
myeloid leukemia (AML). The cells are stained with two probes that target the CBFB gene: a red probe that binds just upstream (5′) of the gene,
and a green probe that binds just downstream (3′). In a normal cell (left), the probes produce essentially overlapping spots, which can appear orange.
In a cell harboring an inversion of chromosome 16 (right) which produces the CBFB-MYH11 fusion gene, one copy of the gene shows physical
separation of the probe spots, indicating a chromosomal breakage between the probe binding sites.(Courtesy of Dr. Peter Thompson, Cytogenetics
Laboratory, Institute of Medical Genetics, Cardiff, UK)(Courtesy of Dr. Gordana Raca, Cancer Cytogenetics Laboratory, Department of Medicine,
University of Chicago, Illinois)
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X Y
Figure 10.6
FISH probes covering entire chromosomes can be tagged with chromosome-specific fluorescent dye signatures and hybridized to
metaphase spreads to produce M-FISH or spectral karyotyping (SKY) images. The example shown is an analysis of the colon cancer cell line
SW480. This allows for the in-depth characterization of complex “marker” chromosomes, including one that contains material from chromosomes
3, 10, and 12 (seen in the chromosome 10 box). This type of determination would be impossible with conventional cytogenetic banding techniques.
(Courtesy of George Poulogiannis, Department of Pathology, University of Cambridge, England)
density (Figure 10.7). It should be noted that array-based number at each probed locus based on binding affinity,
techniques such as CGH can only detect unbalanced chro- while establishing the zygosity of the SNP call at each loca-
mosomal rearrangements where material is lost or gained. tion. Thus, they offer the same type of locus-specific dosage
Balanced inversions and translocations are not detectable data as CGH arrays, but provide the added benefit of up
because the total amount of DNA at each locus remains to millions of SNP genotyping calls genome-wide, which is
unchanged. valuable for a number of different applications.
Such arrays have been the workhorses of genome-wide
association studies (GWAS), which are efforts to uncover
S N P C Y TO G E N O M I C S A R R AYS
genetic underpinnings of complex, multifactorial human
In recent years, BAC-based CGH arrays have largely been diseases or phenotypes. Assaying for many SNPs across the
replaced by newer array systems that use short oligonucle- genome of many affected and normal individuals allows
otide probes, which are easier to mass produce and can be investigators to statistically link the phenotype with par-
made with higher probe density. Array probe counts have ticular SNP markers that may lie in close proximity to the
reached the millions, offering as low as 1kb resolution. responsible genetic factor. This strategy takes advantage
Additionally, many platforms are now based either entirely of the concept of linkage disequilibrium (LD), whereby
or in part on probes that interrogate genomic SNPs.44 One closely adjacent genetic markers will tend to co-segregate
complicating factor when discussing more recent genomics in families rather than being inherited independently, as
technologies is that, for each technology, platforms may be would distantly separated loci; for example, those on sepa-
available from a variety of different companies, each with rate chromosomes. These types of studies have been used to
different chemistries and workflows designed to avoid intel- identify SNPs that predispose to some amount of elevated
lectual property entanglements. Such is the case with SNP risk for a variety of conditions (macular degeneration,
arrays as well as next-generation sequencing platforms (dis- heart disease, diabetes, etc.).45,46 Though there are disagree-
cussed below), among others. However, suffice it to say that ments regarding the clinical utility of this information and
each of the major SNP array platforms can monitor copy the applicability of the data across ethnic groups, certain
1 5 6 • P rincip l e s o f G e no m ic M e dicin e
companies now offer clinical testing using these array plat- Technical Overview
forms, providing individualized risk assessments based on
One of the reasons NGS is so exciting is that it is the first
the results provided by GWA studies. We stress that these
technology with the prospect of detecting genetic varia-
results should be considered and interpreted cautiously, and
tion of every possible size scale. As with array platforms, a
only under the guidance of appropriate experts trained in
variety of NGS platforms exist that share many of the same
clinical genetics and genetic counseling.
important underlying features. As a group, these technolo-
For traditional diagnostic testing, however, these arrays
gies circumvent the signal-to-noise problem of genomic
are proving valuable as a means of surveying the whole
DNA (or any other heterogeneous DNA sample type) in
genome for deletions, duplications, and other anomalies
a fundamentally different way than PCR: by performing
that are too small to be visible by cytogenetics (Figure
independent analyses of many individual DNA molecules
10.8). While the absence of signal from one individual
from a large pool in a massively parallel manner. This is
probe spot may represent noise, a cluster of absent signals
made possible by a critical first step whereby molecules are
can provide statistical confidence of a local genomic dele-
spatially separated from each other, essentially transform-
tion. Thus, depending on overall and local probe density,
ing a complex pool into a collection of isolated single mol-
these arrays can confidently call deletions approximately
ecules. The details of the separation are not critical for this
50kb or less, up to 100 times smaller than those visible
discussion, but (for example) may involve separating DNA
by cytogenetics. The SNP genotyping data can add confi-
molecules into individual microbubbles (emulsion PCR) or
dence to such copy number identification, because a het-
spreading fragments across a surface covered with a “lawn”
erozygous deletion will show all “homozygous” SNP calls
of complementary oligonucleotides. In order to produce
in that region. Additionally, SNP results can detect other
enough signal from each molecule, local PCR amplification
copy-neutral anomalies such as uniparental disomy (which
is performed, creating a population of local amplified clus-
shows a normal number of DNA copies but the absence
ters. Once the individual clonal groups are generated, enor-
of heterozygous SNP calls) or mosaicism (which also has
mous numbers of individual starting DNA molecules can
normal copy number but can show a variety of bizarre SNP
be sequenced simultaneously, with separate data produced
allelic ratio patterns).47 Today, these arrays are widely used
for each.52
in the clinical setting for children with dysmorphic fea-
Nearly any sample of DNA is compatible with NGS
tures and other developmental abnormalities, particularly
analysis. The DNA does not even have to be particularly
when cytogenetics results are normal and when other tar-
intact, because NGS systems require that input DNA
geted testing options (using FISH or PCR, e.g.) have been
take the form of small fragments, typically a few hundred
exhausted.48
base pairs or less. Typically, this is achieved by ultrasonic
fragmentation, or alternative methods based on mul-
N E X T- G E N E R AT I O N S EQ U E N C I N G tiplex PCR, etc. Fragments also typically require that
specific oligonucleotide adapter sequences be incorpo-
Perhaps the most consequential development in biol-
rated onto their ends to make them compatible with the
ogy since the introduction of PCR is the advent of
sequencer. This process is called library preparation, and
next-generation sequencing (NGS).49 NGS represents a
only sequencer-compatible libraries may be applied to the
truly transformational technological shift, because it offers
instruments. After focal amplification of individual input
for the first time the prospect of fast and inexpensive full
library DNA molecules, sequencing primers bind to the
genomic sequence analysis. The implications for research
adapter sequences, and sequencing proceeds inwards into
have been enormous, not just in clinical research but in
the cloned fragment. Some platforms produce only one
virtually every field of biomedicine and basic biology. As
sequence read (single-end sequencing), while others allow
of this writing, the full genome sequences of hundreds
for a second read from the opposite end of each fragment
of animal species have been completed (and much larger
(paired-end sequencing). The details of the sequencing
numbers of bacterial and viral genomes), each offering
reactions are also platform-specific.
new possibilities for discoveries within and across spe-
cies.50 The technique can be used to probe any aspect of
biology related to nucleic acids, from chromatin structure,
Sequencing Strategies
genetics, and epigenetics, to transcription, RNA process-
ing, and more, and can be applied to any species, whether Because of the flexibility of NGS, a myriad of different
well-characterized or novel.51 wet-lab library preparation techniques have been developed
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X Y
(B)
0
16
32
48
i) Chromosome 7 63
79
95
111
127
143
159
–2.00 –1.60 –1.20 –0.80 –0.40 0.00 0.40 0.80 1.20 1.60 2.00
Chromosome 7 Log2 Ratio Ch1/Ch2
01100200_top - 01100200_bottom.bsn - 14/06/2006
0
11
23
34
48
57
ii) Chromosome 13 68
80
91
103
114
–2.00 –1.60 –1.20 –0.80 –0.40 0.00 0.40 0.80 1.20 1.60 2.00
Chromosome 13 Log2 Ratio Ch1/Ch2
Figure 10.7
Low-density CGH array analysis of the two-year-old patient discussed in Figure 10.3. Patient DNA (dark) and control DNA (grey) are
labeled and hybridized together onto an array spotted with BAC clones. Chromosomes 7 and 13 are shown (the chromosomes involved in the
translocation). Most array positions show equivalent binding (normal). However, at the positions corresponding to the breakpoint regions of
both chromosomes are regions of reduced patient signal, indicating loss of chromosomal material. Thus, despite its appearance by conventional
cytogenetics, the translocation is not balanced. It can be seen from the result that approximately 0.2 Mb of chromosome 13 and 8.42 Mb of
chromosome 7 are lost, which is consistent with the patient’s presentation. (Courtesy of Sian Morgan, Cytogenetics Laboratory, Institute of Medical
Genetics, Cardiff, UK)
Copy Number
Genotype
Figure 10.8
SNP array analysis of copy number and genotype. Newer cytogenomics array platforms may contain millions of individual probes, and
can assay the genome at a resolution above one probe per 1000 bp. Many commercial platforms are available, and many options exist with respect
to array composition. Unlike traditional CGH, most modern arrays do not rely on comparative hybridization against a control sample. Instead,
only the test sample is applied to the array, and the data are informatically compared against historical normal samples to derive copy number
and genotype data. Arrays can be used to capture genotype calls genome-wide for GWA studies, etc., or viewed in plots such as the one shown to
evaluate for chromosomal abnormalities. In this example, tumor DNA from a pediatric patient with neuroblastoma is analyzed. Only data from
chromosome 11 are shown. At the end of the p-arm (left), the copy number assessment is normal, as is the genotype pattern (three signal ratios
indicating AA, AB, and BB genotypes). Around the centromere (middle), the copy number is elevated, indicating duplication. The genotype pattern
is complicated by the addition of an extra copy of this chromosomal segment. On the right, a deletion of most of the q-arm can be seen. The copy
number plot shows decreased signal, and the associated genotype plot is consistent with reduced copy number, with each SNP positive for only the
A or B genotype. (Courtesy of Dr. Gordana Raca, Cancer Cytogenetics Laboratory, Department of Medicine, University of Chicago, Illinois)
to focus sequencing efforts on different aspects of genomic continue to decrease, it should see expanded or even routine
biology. Among other applications, ingenious methods clinical use.
have been developed to elucidate the complex domain However, because of the current expense associated
structure of unwound chromatin in interphase nuclei, to with WGS, various targeted sequencing approaches are
reveal genome-wide patterns of transcription factor bind- gaining traction in research and clinical laboratories. These
ing and epigenetic modifications, and to study the spectrum utilize a variety of wet-lab chemistry approaches to create
of RNA/RNA binding protein interactions.53–57 However, sequencing libraries enriched for DNA sequences of inter-
a full description of these techniques is beyond the scope est. Targeted sequencing can be readily performed, either
of this chapter. From the standpoint of clinical medicine, by fishing out fragments of interest from a whole genome
today, most assay designs are geared towards identification library using capture hybridization, or by preparing targeted
of patient-specific genetic variants, either constitutional or sequencing libraries directly from genomic DNA via multi-
somatic, via either a whole-genome or (most frequently) a plex PCR or other methods (Figure 10.9). Library prepa-
targeted sequencing approach. As with any other diagnos- ration strategies can be devised to address essentially any
tic technique, careful planning is required to ensure that an question related to nucleic acid biology, giving laboratories
NGS assay will have the power to detect the full spectrum a great deal of freedom to design assays in order to meet the
of sizes and locations of desired genetic features. needs of their clinician and patient populations.
Despite its bioinformatic complexity, whole genome One important targeted sequencing approach is whole-
sequencing (WGS) is perhaps the simplest sequencing exome sequencing (WES), in which capture baits are used
application to perform, and the creation of a whole genome to select protein-coding sequence fragments from whole-
library is actually a first step for a variety of applications. As genome libraries.59 The resulting sequence data are heav-
described earlier, after the basic steps of fragmentation and ily weighted towards this exonic coding sequence, which
adaptor ligation, genomic DNA fragments are essentially represents only about 2% of the genome, and data from
ready for sequencing. The newest NGS platforms enable the these loci can be attained at a fraction of the cost of
complete sequencing of an entire human genome in only whole-genome sequencing. It has been estimated that as
one day, at a cost of a few thousand dollars, a remarkable much as 80% of Mendelian disease can be explained by
advance reducing cost by several orders of magnitude over mutations in protein-coding DNA. This makes for a very
just the past decade. WGS has been an invaluable research favorable cost–benefit calculation for WES and explains
tool and is beginning to have a clinical impact.58 As costs why it is growing in popularity as a diagnostic choice for
1 6 0 • P rincip l e s o f G e no m ic M e dicin e
Genomic DNA Fragment/repair
Alternate Targeted
Library Prep
Target Capture
Whole Genome Library
NGS
Instrument
PCR Amplification
Targeted Library
Figure 10.9
NGS library preparation. There are many ways to produce sequencer-compatible libraries from genomic DNA. The choice of method
is heavily dependent on the types and sizes of genomic features being investigated, as well as other factors such as cost, sequencing platform,
etc. Typically, genomic DNA sequencing proceeds along a few lines. For WGS and some targeted sequencing approaches, DNA is first sheared
by ultrasonic fragmentation into small pieces (<1kb). After enzymatic end-repair, adapter sequences are ligated to the ends to produce a
sequencer-ready whole-genome library. If this library is applied to the sequencer, the result will be WGS. If instead certain sequences of interest are
purged from this library prior to sequencing, the sequencing will be targeted (i.e., either exome or other targeted capture sequencing). Alternatively,
genomic DNA may be processed in a variety of different ways to produce targeted libraries. For example, targeted libraries can be produced via
multiplex PCR or restriction digest–based methods to produce short fragment libraries with appropriate barcoded adaptors. Every targeted method
can be easily customized to concentrate the sequencing effort on the genomic territory of interest.
contribute extra copies of that chromosome to the mother’s that produce fusion genes.73 Such gene fusion events typi-
blood, which can be detected simply by counting and ana- cally take place via breakpoints within intronic sequences.
lyzing the number of NGS reads that map to each individual As introns are frequently orders of magnitude larger than
chromosome.70 This technique is already having an impact exons, hunting for fusion breakpoints using genomic DNA
in prenatal care, helping to detect these anomalies with requires covering large areas. Breakpoint introns also often
high sensitivity and specificity, thus reducing unnecessary contain repetitive sequences that can create mapping prob-
amniocentesis and chorionic villus sampling procedures.71 lems. The process of transcription and mRNA splicing
It should be noted that RNA is also quite amenable to removes intronic sequences, making the mature mRNA
NGS analysis following reverse transcription into cDNA. a more straightforward target for analysis. Traditional
RNA-seq, as it is termed, can reveal which genes are actively PCR-based translocation assays also routinely utilize RNA
transcribed in a sample, providing important contributory for the same reason.
data to other genomics studies.72 Gene expression levels can
also be compared between samples via analysis of depth
Clinical Integration of NGS
of coverage, and evaluation of read mapping can uncover
information about the expression of specific transcripts The computational processing associated with NGS anal-
and alternative mRNA splicing. Additionally, RNA analy- ysis can be quite demanding. Processing the data from
sis can be used to uncover particular cancer translocations just one instrument involves the production of multiple
70
• Atypical clinical scenario: A mutation of high expected
60
50 MAF > 10%
severity may be difficult to interpret if the patient’s pre-
40 sentation is atypical for the disease. It should be noted
MAF 5–10%
30 that “wellness” is an atypical clinical scenario for every
MAF < 5
20 known disease. This raises the substantial problem of
10 how to interpret and act upon genomic data in the con-
0 text of well patients.
0 100 200 300 400 500 600 700
Sample median exon coverage • Indeterminate mutational effects: Even in a patient with
Figure 10.10
Increased NGS read depth is needed to call low allelic a classical presentation, mutations with unknown effects
percentage mutations with high confidence. In cancer specimens, which on protein function may be difficult to interpret, even if
often show significant sub-clonality and admixture with normal cells, they are on the “correct” gene. This is particularly true of
clinically relevant mutations can be present at very low mutant allelic
frequencies (MAF). This may be true in other clinical scenarios as substitution mutations. Many effects-prediction software
well (e.g., mitochondrial disease). In the cancer setting, it is clinically algorithms are available to help predict protein-function
desirable to be able to reliably detect mutations at 5% MAF. Here, impact, but this is an imperfect science.
samples with MAF between 5–10% required at least 250x read depth
for high-sensitivity detection (compare with 30x read depth, which • Indeterminate gene effects: Similarly, interpretation of
is the gold standard for typical inherited genetics). Below 5% MAF,
extremely high read depths are required. Specificity suffers as well at mutations in genes with little or no known association
very low MAF, because it becomes difficult to discriminate very low with the presumed disease process may be difficult or
percentage mutations from very low percentage sequencing errors. impossible to interpret as clinically pathogenic, even
(Courtesy of Foundation Medicine, Cambridge, Massachusetts)80
if they are predicted to severely affect the gene. Often
this requires perusal of the scientific literature to help
intermediary files potentially equaling hundreds of giga- postulate mechanistic links, though this is most fre-
bytes, and can take days to process on even some of the quently insufficient to produce a clear determination of
fastest computers. The era of big data has arrived, and pathogenicity.
we face large hurdles in the coming years as this technol-
ogy continues to move into the clinic.74,75 For each labo- In some ways, the application of NGS to clinical medi-
ratory to install and maintain sufficient computational cine is a double-edged sword, because as a direct by-product
infrastructure to support NGS applications would be of its power and scope, it raises the likelihood of all of the
prohibitively expensive, and not enough bioinformat- above issues and thus the chance of producing indetermi-
ics expertise currently exists to allow every laboratory to nate results. Fortunately, many tools are emerging to help
participate. For this reason, many groups are developing us sort through the data. Many public and private databases
cloud-computing resources to support NGS operations, are available that contain disease–gene associations and pre-
with the hope that economies of scale and resource shar- viously documented disease mutations, both for constitu-
ing (informatics pipelines, etc.) will allow smaller clinics tional genetic disease (e.g., Online Mendelian Inheritance
and laboratories to utilize NGS testing in the care of their in Man, ClinVar, Human Gene Mutation Database, etc.)
patients.76 and cancer (The Cancer Genome Atlas, Catalogue of
Clinical interpretation of NGS tests is fundamen- Somatic Mutations in Cancer, etc.). Similarly, to help avoid
tally no different from that of more traditional assays. As confusion between common inherited SNPs and potential
always, proper controls, documentation, and adherence to disease mutations, other critical sources provide compen-
strict workflows are required in order to ensure high con- diums of both common and rare inherited variants (e.g.,
fidence in the primary data before it may be interpreted Exome Variant Server, The 1000 Genomes Project, dbSNP,
in the context of the clinical scenario. For findings that HapMap Project, etc.).
are well documented and understood within the clinical While these resources are tremendously valuable, in
context, interpretation is quite straightforward. For exam- practice there is a frequent need for additional informa-
ple, a truncating mutation in the dystrophin gene is eas- tion, which is often unavailable. When presented with a
ily interpretable in the context of a patient with findings variant of uncertain significance (VUS), whether somatic
characteristic of Duchenne muscular dystrophy (DMD). or constitutional, the only ways to develop confidence in
1 6 2 • P rincip l e s o f G e no m ic M e dicin e
its clinical importance are either to document its biological laboratories are adopting the strategy of even wider analysis,
effect in a laboratory, or to produce corroborating evidence performing exome and even genome sequencing in order
from another patient or affected family. What is currently to uncover treatable targets. Some are convinced that this
missing in our health systems is a way to share this type of should be the ultimate strategic goal for all cancer patients,
information broadly between laboratories to help identify while others believe that data from whole genome cancer
similar cases. With the appropriate exchange of knowledge, sequencing studies should be distilled so that clinical diag-
the same VUS identified at two different laboratories could nostics can focus on the relevant (recurrently mutated) tar-
instead become two successful diagnoses. Today, a num- gets for each tumor.
ber of groups are working to create programs and systems With respect to large anomalies, particularly in can-
to enable such sharing, and it is hoped that cloud comput- cer, certain types of variation are more amenable to NGS
ing and other communal resources can help support such analysis than others. For example, NGS is well suited for
endeavors.75,77 Clearly, there are important privacy concerns targeted translocation detection because the identification
surrounding this type of data, and different rules govern- of any cancer-specific fusion sequence is diagnostic. In con-
ing such communications in each country, and any reforms trast, copy number analyses are more difficult to perform in
or agreements must be made in a careful and responsible a superior manner via NGS or SNP arrays compared with
fashion.78,79 FISH or cytogenetics. This is because FISH and cytoge-
netics operate on a single-cell basis, providing data from
each individual cell analyzed. Thus, FISH analysis for Her2
CURRENT AND FUTURE amplification in a breast cancer biopsy would be able to
D I R E C T I O N S I N C L I N I C AL detect significant amplification in a small number of tumor
GENOMICS cells, even if the biopsy was heavily contaminated with nor-
mal cells. In contrast, NGS and SNP arrays only provide
Due to ongoing technological upheaval, the field of clini- an average result of all sampled cells. In order to match
cal genomics is in an exciting state of flux. In many cases, the sensitivity of traditional methods for heterogeneous
there is still minimal consensus regarding what is the most samples (particularly when tumor cell proportion is low),
appropriate way to apply this new technology, particularly single-cell analysis or specific microdissection/sorting may
NGS, to individualized patient care. Technical capabili- be required. Single-cell sequencing is currently performed
ties are changing so quickly that best practices are in many in the research setting, but as of today it is still too expensive
ways a moving target. One clear trend is the current rapid and error-prone to be used as a clinical diagnostic. Thus, for
replacement of traditional Sanger-based gene-sequencing the foreseeable future, FISH and cytogenetics will continue
tests with NGS assays. Full gene-sequencing by Sanger, par- to retain an important role in the work-up of certain tumors.
ticularly for large genes implicated in constitutional genetic Looking to the future, there is vast potential for novel
disease, is a laborious process requiring separate PCRs and and expanded uses of NGS technology in clinical medi-
sequencing reactions for each exon for each patient. In con- cine. While clinical laboratories are beginning to get a
trast, NGS enables the rapid and inexpensive sequencing handle on DNA sequence analysis, many of the sequencing
of panels of many genes, increasing the likelihood of suc- approaches that are routinely used in research laboratories
cessful diagnoses and helping prevent diagnostic odysseys. are still uncharted territory for diagnostics. For example,
Similarly, nearly any assay relying on PCR to uncover small there are essentially no clinically certified assays performed
anomalies can be performed better and for less cost via today that are based on RNA-seq. In clinical laboratories,
NGS, though some smaller mainstay assays continue to rely RNA is typically analyzed either to investigate expression
on PCR. or to look for clinically relevant fusion (translocation)
In cancer diagnostics laboratories, most traditional transcripts. With respect to gene expression, most clinical
assays for detecting small-scale mutational events (includ- laboratories’ needs are met by oligonucleotide microarrays,
ing both oncogene and tumor-suppressor mutations) are which have become quite standardized and inexpensive.81
now being rapidly retired in favor of NGS profiling assays. In contrast, gene-expression analysis by RNA-seq is still
In many academic and commercial laboratories, routine regarded as informatically complex. Therefore, it is most
targeted examination of tens to hundreds of different likely that RNA-seq will first find widespread adoption in
cancer-related genes is now routine, with the goal of pro- clinical laboratories as a tool for translocation detection,
viding individualized treatment recommendations based as oligonucleotide arrays are incapable of detecting these
on each patient’s mutational spectrum.80 Other clinical anomalies.
1 6 4 • P rincip l e s o f G e no m ic M e dicin e
27. van Ommen GJ, et al. FISH in genome research and molecular diag- 48. Miller DT, et al. Consensus statement: chromosomal microar-
nostics. Curr Opin Genet Dev. 1995;5(3):304–308. ray is a first-tier clinical diagnostic test for individuals with devel-
28. Osoegawa K, et al. A bacterial artificial chromosome library
opmental disabilities or congenital anomalies. Am J Hum Genet.
for sequencing the complete human genome. Genome Res. 2010;86(5):749–764.
2001;11(3):483–496. 49. Brenner S, et al. Gene expression analysis by massively parallel sig-
29. Trask BJ, et al. Fluorescence in situ hybridization to interphase cell nature sequencing (MPSS) on microbead arrays. Nature Biotech.
nuclei in suspension allows flow cytometric analysis of chromosome 2000;18(6):630–634.
content and microscopic analysis of nuclear organization. Hum 50. Alföldi J, et al. Comparative genomics as a tool to understand evolu-
Genet. 1988;78(3):251–259. tion and disease. Genome Res. 2013;23(7):1063–1068.
30. Miller WH Jr., et al. Reverse transcription polymerase chain reac- 51. Koboldt DC, et al. The next-generation sequencing revolution and
tion for the rearranged retinoic acid receptor alpha clarifies diag- its impact on genomics. Cell. 2013;155(1):27–38.
nosis and detects minimal residual disease in acute promyelocytic 52. Shendure J, et al. Next-generation DNA sequencing. Nat Biotechnol.
leukemia. Proc Natl Acad Sci U S A. 1992;89(7):2694–2698. 2008;26(10):1135–1145.
31. Wolff AC, et al. Recommendations for human epidermal growth 53. Dixon JR, et al. Topological domains in mammalian genomes identi-
factor receptor 2 testing in breast cancer: American Society of fied by analysis of chromatin interactions. Nature. 2012;485(7398):
Clinical Oncology/College of American Pathologists clinical prac- 376–380.
tice guideline update. J Clin Onc. 2013;31(31):3997–4013. 54. ENCODE Project Consortium. An integrated encyclopedia of DNA
32. Schrock E, et al. Multicolor spectral karyotyping of human chromo- elements in the human genome. Nature. 2012;489(7414):57–74.
somes. Science. 1996;273(5274):494–497. 55. Lister R, et al. Human DNA methylomes at base resolution show wide-
33. Schrock E, et al. Spectral karyotyping refines cytogenetic diagnos- spread epigenomic differences. Nature. 2009;462(7271): 315–322.
tics of constitutional chromosomal abnormalities. Hum Genet. 56. Robertson G, et al. Genome-wide profiles of STAT1 DNA associa-
1997;101(3):255–262. tion using chromatin immunoprecipitation and massively parallel
34. Karpova MB, et al. Combined spectral karyotyping, compara- sequencing. Nat Methods. 2007;4(8):651–657.
tive genomic hybridization, and in vitro apoptyping of a panel 57. Licatalosi DD, et al. HITS-CLIP yields genome-wide insights into
of Burkitt’s lymphoma-derived B cell lines reveals an unexpected brain alternative RNA processing. Nature. 2008;456(7221): 464–469.
complexity of chromosomal aberrations and a recurrence of 58. Bainbridge NM, et al. Whole-genome sequencing for optimized
specific abnormalities in chemoresistant cell lines. Int J Oncol. patient management. Sci Transl Med. 2011;3(87): 87re3.
2006;28(3):605–617. 59. Ng SB, et al. Targeted capture and massively parallel sequencing of
35. Veldman T, et al. Hidden chromosome abnormalities in haemato- 12 human exomes. Nature. 2009;461(7261):272–276.
logical malignancies detected by multicolour spectral karyotyping. 60. Worthey EA, et al. Making a definitive diagnosis: successful clinical
Nat Genet. 1997;15(4):406–410. application of whole exome sequencing in a child with intractable
36. Southern EM. Detection of specific sequences among DNA frag- inflammatory bowel disease. Genet Med. 2011;13(3):255–262.
ments separated by gel electrophoresis. J Mol Biol. 1975;98(3): 61. Bamshad MJ, et al. Exome sequencing as a tool for Mendelian dis-
503–517. ease gene discovery. Nat Rev Genet. 2011;12(11):745–755.
37. Monaghan KG, et al. ACMG standards and guidelines for frag- 62. Langmead B, et al. Ultrafast and memory-efficient alignment
ile X testing: a revision to the disease-specific supplements to the of short DNA sequences to the human genome. Genome Biol.
standards and guidelines for clinical genetics laboratories of the 2009;10(3):R25.
American College of Medical Genetics and Genomics. Genet Med. 63. Li H, et al. Fast and accurate short read alignment with Burrows–
2013;15(7):575–586. Wheeler transform. Bioinformatics. 2009;25(14):1754–1760.
38. Schouten JP, et al. Relative quantification of 40 nucleic acid
64. Ulahannan D, et al. Technical and implementation issues in using
sequences by multiplex ligation-dependent probe amplification. next-generation sequencing of cancers in clinical practice. Br J
Nucleic Acids Res. 2002;30(12):e57. Cancer. 2013;109(4):827–835.
39. Hardenbol P, et al. Multiplexed genotyping with sequence-tagged 65. Ruibin X, et al. Detecting structural variations in the human
molecular inversion probes. Nat Biotechnol. 2003;21(6):673–678. genome using next generation sequencing. Brief Funct Genomics.
40. Willis AS, et al. Multiplex ligation-dependent probe ampli-
2010;9(5–6):405–415.
fication (MLPA) and prenatal diagnosis. Prenat Diagn. 66. Jiang Y, et al. PRISM: pair-read informed split-read mapping for
2012;32(4):315–320. base-pair level detection of insertion, deletion and structural vari-
41. Hömig-Hölzel C, et al. Multiplex ligation-dependent probe ampli- ants. Bioinformatics. 2012;28(20):2576–2583.
fication (MLPA) in tumor diagnostics and prognostics. Diagn Mol 67. Wang C, et al. High-throughput, high-fidelity HLA genotyping with
Pathol. 2012;21(4):189–206. deep sequencing. Proc Natl Acad Sci U S A. 2012;109(22): 8676–8681.
42. Pinkel D, et al. High resolution analysis of DNA copy number varia- 68. Baslan T, et al. Genome-wide copy number analysis of single cells.
tion using comparative genomic hybridization to microarrays. Nat Nature Protocols. 2012;7(6):1024–1041.
Genet. 1998;20(2):207–211. 69. Lo YM, et al. Presence of fetal DNA in maternal plasma and serum.
43. Solinas-Toldo S, et al. Matrix-based comparative genomic hybridiza- Lancet. 1997;350(9076):485–487.
tion: biochips to screen for genomic imbalances. Genes Chromosomes 70. Chiu RW, et al. Noninvasive prenatal diagnosis of fetal chromo-
Cancer. 1997;20(4):399–407. somal aneuploidy by massively parallel genomic sequencing of
44. Sapolsky RJ, et al. High-throughput polymorphism screening and DNA in maternal plasma. Proc Natl Acad Sci U S A. 2008;105(51):
genotyping with high-density oligonucleotide arrays. Genet Anal. 20458–20463.
1999;14(5–6):187–192. 71. Mersy E, et al. Noninvasive detection of fetal trisomy 21: systematic
45. Klein RJ, et al. Complement factor H polymorphism in age-related review and report of quality and outcomes of diagnostic accuracy
macular degeneration. Science. 2005;308(5720):385–389. studies performed between 1997 and 2012. Hum Reprod Update.
46. The Wellcome Trust Case Control Consortium. Genome-wide 2013;19(4):318–329.
association study of 14,000 cases of seven common diseases and 72. Jongeneel CV, et al. An atlas of human gene expression from
3,000 shared controls. Nature. 2007;447(7145):661–678. massively parallel signature sequencing (MPSS). Genome Res.
47. Conlin LK, et al. Mechanisms of mosaicism, chimerism and unipa- 2005;15(7):1007–1014.
rental disomy identified by single nucleotide polymorphism array 73. Edgren H, et al. Identification of fusion genes in breast cancer by
analysis. Hum Mol Genet. 2010;19(7):1263–1275. paired-end RNA-sequencing. Genome Biol. 2011;12(1):R6.
1 6 6 • P rincip l e s o f G e no m ic M e dicin e
11.
MICROBIAL GENOMICS: TARGETED ANTIMICROBIAL
THERAPY AND GENOME VACCINES
Immaculada Margarit and Rino Rappuoli
167
Table 11.1 LICENSED VACCINES FOR HUMAN USE
Personalized Therapeutics
medicine
Figure 11.1
Impact of the recent advances in microbial and human genomics on the development of novel tools for the diagnosis, monitoring,
prevention, and treatment of human infectious diseases.
Pizza and coworkers on Neisseria meningitidis serogroup its 2,158 open reading frames (ORFs). Based on predic-
B (MenB)(5,6). MenB, causing 50% of the meningococ- tion algorithms, 570 ORFs were expected to encode sur-
cal meningitis worldwide, had been refractory to vaccine face-exposed or secreted proteins that might be accessible
development due to the identity of its capsular polysaccha- to the immune system. Further steps towards the selection
ride to a human self-antigen and to the extreme variability of the best vaccine candidates comprised expression of the
of its major outer membrane proteins. The MenB genom- predicted antigens as recombinant proteins in Escherichia
ics vaccine project started with full DNA sequencing of coli (350 ORFs), assessment of their high exposure on the
the virulent strain MC58 and bioinformatics analysis of meningococcal surface (91 candidates), and testing of their
ability to elicit antibodies mediating MenB killing by serum
bactericidal assays and/or protection against lethal challenge
in an animal infection model (28 selected candidates). After
screening these 28 antigens on a panel of diverse isolates to
Epidemiology determine whether their sequence was well conserved, a
multi-component vaccine was finally selected for develop-
ment. This genome-based MenB vaccine consists of three
Production of
recombinant proteins recombinant proteins representing five MC58 antigens,
plus outer membrane vesicles derived from another MenB
isolate(7). Notably, the main vaccine antigens were identified
Genome sequence, as important virulence factors(8). Factor H binding protein
antigen prediction (fHbp) binds a key inhibitor of the complement alternative
pathway enabling the meningococcus to evade killing by the
innate immune system; the Neisserial heparin binding anti-
In vitro/in vivo
testing
gen (NHBA) also plays a role in serum resistance; and the
40
Neisserial adhesin A (NadA) mediates bacterial adhesion to
20
host cells.
0 0
10 102 104
Further molecular epidemiological investigations
Vaccine testing revealed a certain degree of sequence variability in the vac-
in humans
cine antigens expressed by MenB isolates obtained from dif-
ferent patients. Antibody cross-reactivity was demonstrated
between the NHBA variants and also between the three
main NadA variants identified in hyper-virulent strains.
Little cross-protection was instead observed between
Figure 11.2
Reverse vaccinology approach for the identification of vaccine the fHbp variant present in the MenB vaccine and the
candidates starting from the genome sequences of the pathogen. other two fHbp variant groups. Additionally, the level of
expression of all MenB antigens was shown to vary between is serotype-specific and that non-typeable isolates not
strains. For these reasons, a new typing system (the menin- expressing any capsule cannot be protected against by
gococcal antigen typing system, or MATS) was developed polysaccharide-based vaccines. Analysis of the full genome
to predict potential vaccine coverage among infective iso- of eight different GBS strains by Tettelin et al. revealed
lates in different geographical settings(9). MATS is a sand- novel genes to be added to the species gene pool after each
wich enzyme-linked immunosorbent assay (ELISA) that strain was sequenced. This observation highlighted GBS
measures the amount of each target antigen expressed by a intra-species diversity and introduced the concept of the
strain and its immunological cross-reactivity with the pro- “pan-genome,” which comprises “core” genes shared by all
tein variant present in the MenB vaccine. The data obtained strains, and “dispensable” genes present only in one or a few
with MATS correlate with the killing of strains in a serum strains(14). Maione and colleagues applied the pan-genome
bactericidal activity assay, and allowed the prediction of notion to design a universal vaccine against GBS(15). By
78% coverage of the European MenB isolates. Results from computational analysis of the eight sequenced genomes,
clinical trial studies have shown safety and robust immune they predicted 589 surface-exposed proteins, 396 of which
responses in both adults and infants(10), forming the basis encoded in core genes and 193 in dispensable genes. Of
for the recent licensure of this vaccine by the European these 589, 312 were successfully expressed as recombinant
Medicines Agency. proteins in E. coli and evaluated for their ability to mediate
protection in a mouse ‘maternal immunization–neonatal
pup challenge’ model. A four-antigen combination proved
ADDRESSING ANTIGEN
protective against a large panel of strains. Three of these
VA R I A B I L IT Y BY M I C RO B I A L
protective antigens were encoded in dispensable genes, and
C O M PA R AT I VE G E N O M I C S
would not have been identified if only a single genome had
Emerging technologies have greatly accelerated genome been screened. Interestingly, these three proteins were seen
sequencing during the past decade, leading to a further evo- to assemble into previously undescribed long filamentous
lution of reverse vaccinology to incorporate comparative in pilus-like structures extending outside the bacterial surface
silico analysis of multiple genomes from different strains of that were shown to play an important role during bacterial
the same species. This approach allows the selection of vac- infection. Subsequent genomic analysis using a wider col-
cine antigen candidates, while taking into account the intra- lection of strains revealed three different pilus islands, PI-1,
species antigen diversification stratagem adopted by many PI-2a, and PI-2b, at least one of which was present in a wide
pathogenic species to escape the immune system. Intra- panel of strains. More interestingly, a vaccine incorporating
species diversity is generated by a variety of mechanisms, one component of each pilus variant was shown to provide
including mutation, horizontal gene transfer through a high level of mouse protection against virulent isolates
mobile genetic elements and recombination(11). representing all GBS serotypes(16).
Multigenome reverse vaccinology was first applied to A similar comparative genomics approach was success-
Streptococcus agalactiae (Group B streptococcus, or GBS), a fully applied to Streptococcus pneumoniae (pneumococcus),
Gram-positive microorganism that colonizes the ano-geni- another major human pathogen causing sepsis, meningitis,
tal tract of 20–30% of healthy women and is a major cause pneumonia, otitis media, and sinusitis, which accounts for
of neonatal sepsis and meningitis. The pathogen can also more than 10% of the mortality worldwide in children under
cause severe invasive infections in the elderly, in pregnant five years old(17). Pneumococcus can be classified into more
women, and in patients with underlying disease(12). There than 90 capsular serotypes, and the recently introduced
are ten GBS serotypes distinguished by their capsular poly- polysaccharide-conjugate vaccines have proven highly effec-
saccharide, and the amount of maternal antibodies directed tive in preventing pneumococcal infections against their
against each polysaccharide type is inversely proportional represented serotypes(18). Pneumococcal protein antigens
to the risk of neonatal infection with strains of that spe- have additionally been evaluated for their use in universal
cific serotype. This observation established the basis for the serotype-independent vaccines to face variable regional dis-
development of vaccines based on capsular polysaccharides tributions of serotypes, the occurrence of serotype replace-
conjugated to carrier proteins, which induce long-lasting ment after vaccination, as well as the complexity and cost
immune responses(13). of conjugate vaccines(19). The availability of multiple pneu-
Parallel efforts to identify protective protein antigens mococcal genome sequences, combined with an increased
capable of conferring wide coverage were also under- understanding of pili in GBS and in other Gram-positive
taken, given that protection by GBS polysaccharides pathogens, led to the discovery of pneumococcal pilus
M icrobial G enomics • 1 7 1
www.Ebook777.com
proteins eliciting high protection in mouse infection mod- followed by mass spectrometry (MS) analysis of gener-
els as potential components of a broad-coverage vaccine ated peptides, and identified proteins highly expressed on
combination(20). the bacterial surface and thus accessible to antibodies(23).
A further step in multi-genome reverse vaccinology Similarly, antigens on the surface of Gram-negative bacte-
introduced an additional criterion for the selection of ria were identified by MS analysis of membrane fragments
antigens specific to pathogenic strains and absent in com- released by the bacteria upon genetic modifications to
mensal strains of the same species, with the aim of reducing weaken their outer membrane(24).
the potential impact of a vaccine on the commensal flora. An alternative approach aimed at reducing the number
Comparative analysis of the genomes of two E. coli strains of antigens in the pre-selection stage consists of interrogat-
causing meningitis, five uro-pathogenic strains, one avian ing the entire antigenic repertoire of a particular pathogen
strain, and the non-pathogenic K12, identified 230 surface by using representative libraries of recombinant peptides
antigens present in the extra-intestinal pathogenic E. coli that can be displayed on the surface of bacteria or bacte-
but absent (or poorly conserved) in the non-pathogenic rial phages, or spotted onto microarrays. These libraries can
isolate. Nine potential vaccine antigens were able to induce then be screened with sera from infected individuals who
protection in a mouse-challenge sepsis model, some of recovered from infection for the presence of specific anti-
which also present in intestinal pathogenic E. coli, show- bodies, leading to the identification of a discrete number of
ing promise for a broad-coverage vaccine against different antigen targets(25).
pathogenic E. coli(21). In a recent study, Bensi et al. devised a strategy that incor-
Genomic reverse vaccinology approaches have now porates quantification of bacterial surface proteins using
been applied to the discovery of new vaccines against antibodies raised against recombinant surface-predicted
many other pathogens, including Chlamydia pneumoniae, antigens, MS proteomic analysis, and high-throughput
Bacillus anthracis, Porphyromonas gingivalis, Mycobacterium screening of human sera, for the rapid selection of a limited
tuberculosis, Helicobacter pylori, hepatitis C virus, the coro- number of vaccine candidates prior to biological testing.
navirus responsible for severe acute respiratory syndrome By applying this combined approach to GAS, highly selec-
(SARS), and the malaria parasite Plasmodium falciparum(8). tive identification of few protective antigens was achieved,
Promising results towards defeating malaria have recently which allowed the definition of a multi-protein formula-
been obtained by vaccinating infants and children with tion conferring consistent protection against multiple GAS
the circumsporozoite protein 1 (CSP-1) fused with the serotypes in mouse models of infection(26).
hepatitis-B surface antigen(22). The new information derived from the growing list
of protective antigens from different microbial species is
expected to bring an additional improvement to the predic-
I N T EG R AT I N G G E N O M I C S ,
tion of vaccine candidates by bioinformatics tools. Indeed,
P ROT EO M I C S , A N D I M MU N O M I C S
several curated databases have been established, based on
F O R VAC C I N E D I S C OVE RY
the information obtained from experimentally validated
A common drawback of genome-based approaches for antigens. Ultimately, improved algorithms are expected to
vaccine discovery is the need to screen a large number of be developed that will allow, not only better prediction of
candidates by laborious and time-consuming in vivo and/ surface localization, but also the identification of common
or in vitro assays, in order to select a limited number of signatures among protective antigens, that will guide the
antigens conferring effective protection. Therefore, several identification of novel vaccine candidates.
pre-screening strategies aimed at reducing the number of
antigens for further biological testing have been attempted.
R AT I O NA L D E S I G N O F E FFEC T I VE
Based on the observation that bacterial vaccines induc-
VAC C I N E S BY S T RU C T U R E -BA S E D
ing protective antibodies are mainly constituted by highly
A P P ROAC H E S
expressed surface-exposed antigens and/or secreted tox-
ins, proteomic-based methods have been used to selec- Recent advances in X-ray crystallography and nuclear
tively identify these categories of proteins. In a pioneering magnetic resonance (NMR) spectroscopy have greatly
approach, Rodriguez-Ortega et al. analyzed the surface of accelerated structural studies on vaccine antigens and their
Streptococcus pyogenes (Group A streptococcus, or GAS), a epitopes, opening the path for the structural design of
severe human pathogen for which a vaccine is not yet avail- novel and improved vaccines(27). This type of approach can
able, by digestion of live bacteria with different proteases, be applied to address the high variability of many antigen
M icrobial G enomics • 1 7 3
the vaccine necessary to cope with unceasingly arising process often takes several days, or even weeks in the case
flu variants. This is particularly true in the case of pan- of slow-growing microorganisms. The “omics” revolution is
demics, as vaccine prevention tools are required well expected to deeply transform routine clinical microbiology
before the appearance of the peak of disease incidence. laboratory practices in the coming years by the progressive
Genomic-based approaches have come to our aid in this substitution of many of these complex multifaceted proce-
case also, in the form of “synthetic vaccinology.” In a sim- dures with genome-based technologies(36,37).
ulated response to a pandemic with a H7N9 “bird flu,” High-density pan-microbial microarrays containing
Dormitzer et al. utilized the viral DNA sequence informa- nucleic acid probes specific for various pathogen sequences
tion to synthesize the genes coding for the necessary HA now allow the rapid screening of a large number of patho-
and neuraminidase (NA) antigen variants(35). The genes gens: nucleic acids from a clinical specimen can be ampli-
were built enzymatically in cell-free reactions that included fied randomly and then hybridized to the chips for species
a critical error correction step. Co-transfection of canine identification(38).
kidney cells with the synthetic genes cloned into optimized A proteomics-based approach that has further accel-
expression vectors and plasmid DNA containing improved erated species identification consists of comparing mass-
viral backbone genes produced a combination resulting in spectrometry profiles of pure microbial suspensions with
high yields of vaccine antigens. The authors could demon- available databases(39). Polymerase chain reaction (PCR)
strate the potential of this new procedure to save weeks assays can also assist in the rapid identification and typing
off the time needed for vaccine manufacture to respond of bacterial and viral pathogens, and in some cases they
quickly to a sudden pandemic. The study also offered proof have become an integral part of the standard of care. For
of concept for the potential of synthetic vaccinology for instance, DNA-sequencing analysis of the HIV genotype
the generation of tailor-made microorganisms optimized for the presence of mutations conferring resistance to anti-
for the expression of newly designed vaccines. retroviral drugs, and PCR-based measurement of viral
loads, can be used to choose medication and help predict
the responses to therapy(40), and similar approaches have
M O N I TO R I N G I N F E C T I O N been applied to other viral diseases like hepatitis B and C
E M E R G E N C E , T R A N S M I S S I O N, A N D and influenza. As far as testing of antimicrobial susceptibil-
PAT H O G E N E VO LU T I O N BY N E X T- ity in concerned, checking for the presence or absence of
G E N E R AT I O N S E Q U E N C I N G antimicrobial resistance genes by PCR can accelerate effec-
tive treatment of infected patients.
In this section, we will discuss how genomic-based tech- Application of next-generation full-genome sequencing
nologies can be used in clinical microbiology both for the will soon be sufficiently fast, accurate, and cheap to be rou-
diagnosis and the management of individual infections, as tinely used by the clinical laboratory for improved patient
well as for monitoring the emergence and epidemiology of care. The major advantage of whole-genome sequencing is that
infectious diseases. comprehensive DNA information can be obtained in a single
rapid step, providing all necessary data for diagnostic and typ-
ing needs. Nevertheless, substantial challenges will need to be
GENOMICS IN THE CLINICAL
overcome before microbial full-genome sequencing can be
M I C RO B I O L O GY L A B O R ATO RY
fully embraced in a clinical laboratory setting and gradually
The main task of the clinical microbiology laboratory is replace present-day methodologies. Success will depend on the
the rapid management of individual infections by isolat- development and implementation of reliable and user-friendly
ing pathogens from clinical samples, identifying species bioinformatics tools to rapidly extract the most meaningful
for diagnostic purposes, and testing for antimicrobial sus- genetic information from the generated complex data sets.
ceptibility. Many basic microbiology practices to accom-
plish these tasks were developed over decades, and are
time-consuming and labor-intensive. Most often the
T R AC I N G I N FEC T I O N O U T B R E A K S
pathogen has to be cultured before isolation, and complex
A N D T R A N S M I S S I O N EVE N TS
selective media are required to treat samples contaminated
with colonizing flora. Moreover, diagnostic characteriza- Microbial genomes are much smaller than eukaryotic
tion depends on a wide range of biochemical testing path- genomes but much more diverse, as up to 40% of their
ways that are often species-specific. This multiple-step DNA may consist of dispensable sequences that are not
M icrobial G enomics • 1 7 5
were found to possess an excess of mutations that truncated G E N O M I C S A N D TA RG ET-BA S E D
proteins, including a transcriptional regulator implicated A N T I M I C RO B I A L D I S C O VE RY
in pathogenicity. Another study, of a 16-year outbreak of
Genome-based approaches applied at the early stage of the
chronic Burkholderia dolosa infection, revealed evidence
drug discovery process have generated a valuable inventory
for parallel adaptive evolution of 17 genes across 14 cystic
of genes and cellular processes from which to further test
fibrosis patients.
and validate novel antibacterial targets(47). Most of these
targets are selected for their essential role during in vitro
M ETAG E N O M I C S A N D T H E HUM A N growth, usually by means of genetic manipulation (e.g., gene
M I C RO B I O M E knockout) of the relevant bacteria. An alternative approach
for target selection focuses on virulence factors required by
The different sites of the human body are populated by
specific bacteria to cause disease, like toxin delivery or cell
complex microbial communities, which have become the
adhesion, and has the potential advantages of better pre-
subject of a new field in microbiology aimed at defining
serving the host microbiome and of decreasing the proba-
the “microbiome” composition, its interactions with the
bility of antibiotic resistance(48). Novel antimicrobial targets
human host, and its role in human health(46). Given the
can also be selected by “metabolomics” approaches—that is,
impossibility of cultivating most of the bacteria compos-
analysis of metabolite production by host cells by NMR or
ing the human microbiome, broadly applicable techniques
Mass Spectrometry(49). For instance, changes in the meta-
for analyzing massive amounts of DNA sequence data
bolic flux of human cytomegalovirus–infected cells were
have been developed, which have contributed significantly
used to identify metabolic pathways upregulated by viral
to the growing field of metagenomics. These studies have
infection, and potential targets for novel antivirals aimed at
demonstrated great variations between host individuals
blocking viral replication.
and confirmed that substantial alterations in the human
By comparative genomics, target genes can be selected
microbiome are important for a variety of disease states,
for narrow- or large-spectrum therapeutic solutions on the
including psoriasis, sexually transmitted infections, Crohn
basis of their conservation profiles across species. Once vali-
disease, gastroesophageal reflux disease, and others. Recent
dated, these target genes can be cloned and sequenced, and
studies have also addressed the role of the gut microbiome
their protein products expressed in an optimized expression
in the development of the immune system. The established
system (e.g., Pichia pastoris, Baculovirus, E. coli). Targets
links between microbial communities and the etiology of
are often screened by high-throughput methods against
human disease will inform future design of better vaccines
large libraries of combinatorial chemistry-derived com-
and therapeutics.
pounds(50,51). Each specific molecular target or pathway of
interest is combined systematically with each possible drug
compound by automated platforms, and positive results,
C O N T R O L L I N G B AC T E R I A L or “hits,” are subsequently characterized with respect to
I N F E C T I O N BY G E N O M E -B A S E D potency, mechanism of inhibition, spectrum, and selectiv-
ANTIMICROBIALS ity(36,52). An example of an antimicrobial target identified
by a genomics-driven approach is the product of the def
Most of the antibiotics in use today originated many decades gene, which is present in all pathogenic bacteria and does
ago as natural products isolated from bacteria and fungi. not share a functionally equivalent gene in mammalian
The antimicrobial industry has excelled at fine-tuning these cells. The gene encodes a peptide deformylase belonging to
natural molecules to improve their spectrums, efficacy, the matrix metallo-protease family of enzymes, and a selec-
and safety, especially by means of semi-synthetic chemis- tive inhibitor could be identified by screening a library of
try approaches. Nonetheless, the growing emergence of metallo-enzyme inhibitors.
antibiotic-resistant bacterial strains and the public health An important condition that dictates the efficiency of
threat of pandemic viral infections have recently raised antimicrobial compounds is their need to cross the micro-
a renewed interest in the discovery and development of bial membrane barrier and to avoid subsequent extrusion by
novel non-toxic and fast-acting antimicrobial drugs. The multidrug-resistance efflux pumps. One strategy that takes
Infectious Disease Society of America estimates that 70% of into account permeability and efflux issues consists of com-
hospital-acquired infections in the United States are resis- bining genomics with classic whole-cell screening methods
tant to one or more antibiotics. by using genetically modified microorganisms that can
M icrobial G enomics • 1 7 7
Large-scale profiling of the full RNA of peripheral blood 5. M. Pizza et al. Identification of vaccine candidates against serogroup
B meningococcus by whole-genome sequencing. Science 287, 1816
mononuclear cells (PBMCs) by microarrays or deep RNA (Mar 10, 2000).
sequencing approaches has been applied to the identifica- 6. H. Tettelin et al. Complete genome sequence of Neisseria menin-
tion of patterns of gene expression characteristic of cer- gitidis serogroup B strain MC58. Science 287, 1809 (Mar 10, 2000).
7. M. M. Giuliani et al. A universal vaccine for serogroup B meningo-
tain human infections like tuberculosis, dengue, influenza, coccus. Proceedings of the National Academy of Sciences of the United
S. aureus and salmonellosis, among others. Data derived States of America 103, 10834 ( Jul 18, 2006).
from these approaches will assist the diagnosis and progno- 8. K. L. Seib, X. Zhao, R. Rappuoli. Developing vaccines in the era of
genomics: a decade of reverse vaccinology. Clinical Microbiology and
sis of disease, the design of antiviral and antibiotic therapies, Infection: the Official Publication of the European Society of Clinical
and the development of genetic tests to predict adverse reac- Microbiology and Infectious Diseases 18 Suppl 5, 109 (Oct, 2012).
tions to antimicrobial drugs. 9. J. Donnelly et al. Qualitative and quantitative assessment of
meningococcal antigens to evaluate the potential strain coverage
In the vaccines field, novel tools of systems biology of protein-based vaccines. Proceedings of the National Academy of
can be applied to the analysis of human immunological Sciences of the United States of America 107, 19490 (Nov 9, 2010).
response patterns to vaccines in order to uncover molecu- 10. A. R. Gorringe, R. Pajon. Bexsero: a multicomponent vaccine
for prevention of meningococcal disease. Human Vaccines and
lar signatures of vaccine efficacy and guide the design and Immunotherapeutics 8, 174 (Feb, 2012).
evaluation of new vaccines(61). This “systems vaccinology” 11. T. T. Binnewies et al. Ten years of bacterial genome sequenc-
strategy has been applied to examine the initial molecular ing: comparative-genomics-based discoveries. Functional and
Integrative Genomics 6, 165 ( Jul, 2006).
signatures in individuals vaccinated against yellow fever, or 12. A. Schuchat. Group B streptococcus. Lancet 353, 51 ( Jan 2, 1999).
after administration of the trivalent inactivated influenza 13. C. J. Baker, M. S. Edwards. Group B streptococcal conjugate vac-
cines. Archives of Disease in Childhood 88, 375 (May, 2003).
virus vaccine. Similar approaches have been used to study 14. H. Tettelin et al. Genome analysis of multiple pathogenic iso-
immune responses to Brucella melitensis and fungal infec- lates of Streptococcus agalactiae: implications for the microbial
tions. The obtained data will ideally lead to the design of “pan-genome.” Proceedings of the National Academy of Sciences of the
United States of America 102, 13950 (Sep 27, 2005).
vaccines capable of inducing optimal immune responses 15. D. Maione et al. Identification of a universal Group B streptococcus
without toxic effects, thus improving vaccine safety profiles. vaccine by multiple genome screen. Science 309, 148 ( Jul 1, 2005).
Integration of increasingly complex high-throughput 16. I. Margarit et al. Preventing bacterial infections with pilus-based vac-
cines: the group B streptococcus paradigm. The Journal of Infectious
data into descriptive and predictive equations for immune Diseases 199, 108 ( Jan 1, 2009).
responses to vaccines is expected to drive faster and 17. K. L. O’Brien et al. Burden of disease caused by Streptococcus pneu-
more accurate ways of screening vaccine candidates for moniae in children younger than 5 years: global estimates. Lancet
374, 893 (Sep 12, 2009).
their effectiveness. Pulendran et al.(62) have predicted the 18. K. L. Moffitt, R. Malley. Next generation pneumococcal vaccines.
development of a vaccine chip microarray, similar to the Current Opinion in Immunology 23, 407 ( Jun, 2011).
MammaPrint prognostic chip that was developed for 19. T. M. Wizemann et al. Use of a whole genome approach to identify
vaccine molecules affording protection against Streptococcus pneu-
breast cancer, which will be able to predict the immunoge- moniae infection. Infection and Immunity 69, 1593 (Mar, 2001).
nicity of any vaccine. 20. J. L. Telford, M. A. Barocchi, I. Margarit, R. Rappuoli, G. Grandi.
Like in other genomic fields, the successful clinical Pili in Gram-positive pathogens. Nature Reviews. Microbiology 4,
509 ( Jul, 2006).
application of these novel technologies will depend on the 21. D. G. Moriel et al. Identification of protective and broadly con-
development of translational research approaches capable served vaccine antigens from the genome of extraintestinal patho-
of dealing with the enormous amount and different types genic Escherichia coli. Proceedings of the National Academy of Sciences
of the United States of America 107, 9072 (May 18, 2010).
of generated information and with the uncertainty that is 22. S. T. Agnandji et al. First results of phase 3 trial of RTS,S/AS01
typical of common clinical scenarios. malaria vaccine in African children. The New England Journal of
Medicine 365, 1863 (Nov 17, 2011).
23. M. J. Rodriguez-Ortega et al. Characterization and identification of
vaccine candidate proteins through analysis of the group A strepto-
coccus surface proteome. Nature Biotechnology 24, 191 (Feb, 2006).
REFERENCES 24. F. Berlanda Scorza et al. Proteomics characterization of outer
membrane vesicles from the extraintestinal pathogenic Escherichia
1. R. Rappuoli, H. I. Miller, S. Falkow. Medicine. The intangible value coli DeltatolR IHE3034 mutant. Molecular and Cellular
of vaccination. Science 297, 937 (Aug 9, 2002). Proteomics: MCP 7, 473 (Mar, 2008).
2. D. K. Kaushik, D. Sehgal. Developing antibacterial vaccines in 25. C. Giefing et al. Discovery of a novel class of highly conserved vac-
genomics and proteomics era. Scandinavian Journal of Immunology cine antigens using genomic scale antigenic fingerprinting of pneu-
67, 544 ( Jun, 2008). mococcus with human antibodies. The Journal of Experimental
3. S. A. Plotkin. Vaccines: past, present and future. Nature Medicine Medicine 205, 117 ( Jan 21, 2008).
11, S5 (Apr, 2005). 26. G. Bensi et al. Multi high-throughput approach for highly selective
4. R. D. Fleischmann et al. Whole-genome random sequencing and identification of vaccine candidates: the Group A streptococcus
assembly of Haemophilus influenzae Rd. Science 269, 496 ( Jul 28, case. Molecular and Cellular Proteomics: MCP 11, M111 015693
1995). ( Jun, 2012).
M icrobial G enomics • 1 7 9
12.
NUTRITIONAL GENOMICS
Zhenglong Gu, Kaixiong Ye, and Patrick J. Stover
180
Free ebooks ==> www.Ebook777.com
Food Intolerances Dietary Requirements 3) Nutritional systems biology, the application of systems
biology approaches in nutritional studies. It integrates
global data at multiple levels (genome, transcriptome,
proteome, metabolome, interactome, etc.) and utilizes
Dietary systemic analysis and model-based computational
components Human Genome
simulation to study biological networks (e.g.,
regulatory, signaling, and metabolic networks) under
various physiological and nutritional states.
genomics, like most “-omics” fields, is focused on the biol- Human populations and individuals within populations dif-
ogy of the individual, but is distinguished by its unique fer in their sensitivities to nutrient deficiencies and excesses.
potential to advance our understanding of disease preven- Human genetic variation contributes to differences in physi-
tion and healthy aging through manipulation of gene–diet ological responses to diet. The United Nations Educational,
interactions. In addition, nutritional genomics has thera- Scientific and Cultural Organization (UNESCO) recog-
peutic applications through the rational design of dietary nized both the influence of human genetic variation on
interventions to manage chronic disease. human nutrition and the concept that nutrient utilization
Advances in nutritional genomics research are antici- and efficacy are unique characteristics of individuals, in the
pated to illuminate the mechanisms underlying the acute Universal Declaration on the Human Genome and Human
and long-lasting diet/nutrition–genome interactions that Rights, Section A, Article 3, which states:
promote health and revolutionize both clinical and public
health nutrition practice, and culminate in: 1) genetically The human genome, which by its nature evolves, is
informed nutrient- and food-based dietary guidelines for subject to mutations. It contains potentialities that
disease prevention and healthy aging, 2) improved and/or are expressed differently according to each individ-
individualized nutritional therapeutic regimes for disease ual’s natural and social environment including the
management, and 3) better targeted public health nutri- individual’s state of health, living conditions, nutri-
tion interventions (e.g., micronutrient fortification and tion and education.
supplementation) that maximize benefit and minimize
—UNESCO Document 27 V / 45, adopted by the Thirty-First
adverse outcomes within human populations. These objec-
General Assembly of UNESCO, Paris, November 11, 1997
tives will be met only with further development of the basic
science that underpins nutritional genomics, and effective
Indeed, human evolution is a continuous, albeit irregular,
translation of this knowledge into nutrition practice in the
process made manifest through the generation and expan-
following areas:
sion of DNA mutations that permit survival amidst erratic
and unpredictable environmental exposures. Changes in
1) Nutritional genetics, the identification, classification,
DNA primary sequences enable human evolution through
and characterization of human genetic variation that
the generation of adaptive genetic variants that alter an
modifies nutritional requirements and food tolerances;
organism’s response to environmental challenges and hence
2) Nutritional epigenetics, the modification of chromatin to its fitness. Human DNA primary sequence was origi-
structure (and hence gene expression) by diet, through nally estimated to differ by approximately 0.2–0.4% among
post-replication and post-translational modification humans, mostly due to single-nucleotide polymorphisms.6,7
of DNA and protein, respectively, which serves to It was recently estimated that 1–3% of human genomes are
program or reprogram biological networks, with different considering various types of structural variations.8
multigenerational consequences; These observed variations are partly products of historical
N utrition a l G e no m ic s • 1 8 1
www.Ebook777.com
interactions among humans and their environment, includ- sequence similarities among family members, Alu insertions
ing dietary patterns. can serve as nucleation sites for unequal intrachromosomal
and interchromosomal homologous recombination events
that result in chromosomal aberrations such as deletion and
T Y P E S O F G E N ET I C VA R I AT I O N
translocation events.
Differences in DNA primary sequence constitute a primary Viral elements can confer new types of regulation to
molecular basis for human phenotypical variation, includ- existing genes, including regulation by essential nutrients,
ing metabolic efficiency and disease susceptibility. Genomic but they may also interrupt genes or create new genes. Alu
polymorphisms emerge through the sequential processes of elements can function as transcriptional silencers or activa-
DNA mutation and expansion of the mutation within a tors that are responsive to nutrient status. Some Alu ele-
population. Environmental exposures can accelerate both ments have retinoic acid (vitamin A) response elements
processes. Polymorphisms are classified according to the that bind nuclear receptor transcription factors and there-
origin and nature of the genomic mutation, and include fore confer retinoic acid responsiveness to genes that neigh-
single-nucleotide polymorphisms (SNPs), micro- and bor the insertion site. Alu elements also contain multiple
macrosatellite repeat sequences and structural variations, CpG “islands,” a nucleotide sequence that attracts the DNA
such as repetitive element insertions and copy number methylation of the cytosine (C) base. DNA methylation
variations.9,10 SNPs are common nucleotide base-pair dif- typically serves to suppress transcription at that locus, and
ferences in the primary sequence of DNA. As of 2011, the degree of methylation can be sensitive to dietary intake
there were more than 50 million SNPs submitted to the of vitamins involved in one-carbon metabolism, including
National Center for Biotechnology Information (NCBI) folic acid.21
dbSNP database (dbSNP build 137), about 38 million of Alu insertions are also associated with metabolic dis-
which had been validated.11 SNPs can be single base-pair ease. New Alu insertions may account for up to 0.1% of
insertions, deletions, or substitutions of one base pair for human genetic disorders, including Apert syndrome, cho-
another. Nucleotide substitutions are the most common linesterase deficiency, and breast cancer.22 Alu-mediated
polymorphism, whereas insertion/deletion mutations unequal homologous recombination events are respon-
occur at one-tenth that frequency.12 SNPs, like other poly- sible for 0.3% of inherited human genetic diseases, such as
morphisms, are usually defined as genetic variants that have type 2 insulin-resistant diabetes and familial hypercholes-
a frequency of at least 1% in human populations.13 terolemia.14,17 Such recombination events are rare because
Genetic variations can also result from the integration Alu-mediated unequal homologous recombination events
and/or transposition of retroviral DNA.14 Approximately are usually inhibited by CpG methylation of the element.
50% of noncoding human DNA originates from transpos- Copy number variations (CNV) involve changes of copy
able elements that are highly mobile and contain repetitive number for DNA segments that are ~1 kb or longer.23,24
sequences.15–17 Retrotransposons are classified by size and Usually, the insertion or deletion of transposable elements is
include long interspersed nuclear elements (LINEs) and not classified as a CNV.24 The estimated genome-wide muta-
short interspersed nuclear elements (SINEs). About 10% of tion rate of CNV ranges from 1.7 x 10–6 to 1.0 x 10–4 per
the human genomic sequence consists of 280 base-pair Alu locus per generation, which is 100 to 10,000 times higher
SINE elements. Over 1,200 Alu elements integrated into than nucleotide substitution rates.25 Due to the length of
the human genome following early human migrations out DNA involved, CNV usually affects more nucleotides per
of Africa.18 Today, there are an estimated 1.4 million such genome than SNPs.23,25 Because CNVs can modify gene
elements in the human genome, with a new Alu insertion dosage, interfere with proper splicing, disrupt the coding
event occurring every 200 births.17 Alu elements are believed region, and alter regulation of a nearby gene, they can have
to be catalysts for organismal evolution.17,19 They display a significant functional impact on a genome, and therefore
promoter activity, but their transcripts lack an open read- are subject to selection.23,26 CNVs can be subject to puri-
ing frame and therefore are not translated into protein. Alu fying (negative) selection by disrupting proper gene func-
insertions can alter genome stability and/or the function tions, and to directional (positive) selection by contributing
of a single gene function near their site of integration. Alu to regional adaptation. Association studies have identified
elements contain splice acceptor sites; therefore, their inte- CNVs that contribute to phenotypical diversity, such as dis-
gration within an intron can lead to the generation of new ease pathology, drug sensitivity, steroid hormone, xenobiotic
proteins through alternative splicing of gene transcripts.20 metabolism, prostate cancer, nicotine metabolism, regula-
Because of their repetitive sequence and the high degree of tion of food intake and body weight, neurodevelopment
1 8 2 • P rincip l e s o f G e no m ic M e dicin e
and neurological disorders, colonic Crohn’s disease, toxin average for autosomes in regions of the genome presumed
resistance, coronary heart disease risk, Alzheimer’s disease, to be nonfunctional, including intronic and intergenic
HIV infection, and AIDS progression.10,24,27–29 Variations in regions).12,40
the copy number of the amylase gene resulted from dietary
adaptation during human evolution.30
D I ET A N D MU TAT I O N E X PA N S I O NS
I N HUM A N P O P U L AT I O NS
D NA MU TAT I O N R AT E S A N D D I ET
There is increasing evidence that dietary challenges have
The human genome is assembled from deoxynucleotide increased human genetic variation by driving the expansion
monomers that are synthesized de novo from metabolic pre- of rare gene variants in isolated human populations in iso-
cursors derived from energy, amino acid, and one-carbon lated geographic regions. Mutations that promote survival
metabolism. DNA mutations arise as a consequence of the in challenging dietary environments expand within popula-
inherent chemical instability of DNA bases, errors associ- tions through the generation of environmentally adaptive
ated with the fidelity of DNA replication and recombina- gene alleles. Mutations that expand and reach a specific fre-
tion, exposure to chemical oxygen radicals that are generated quency within a population contribute to genetic variation
during oxidative metabolism, as well as by numerous as polymorphisms, and this expansion within human popu-
genotoxic xenobiotics that are present in the food supply. lations is the molecular basis for the evolution of genomes.
Therefore, some DNA mutations are unavoidable; however, Germline mutation alone is necessary, but not sufficient, for
their deleterious impact is minimized by DNA repair sys- establishing genetic variation.
tems that detect and correct most mutation events. Not all genes “evolve” or change at the same rate. The
Environmental exposures such as nutrient deficiencies neutral theory of evolution (DNA mutation in the absence
or excesses, factors that increase cellular oxidative stress, or of selection) does not account for the extent of amino
genetic variations that modify the metabolism of dietary acid substitutions observed in mammalian genomes.41–44
components can accelerate DNA mutation rates by inducing Natural selection, which is the differential contribution of
DNA modification reactions and/or by accelerating DNA genetic variants to future generations, is the only evolu-
polymerase error rates. Many of the essential B-vitamins tionary force that has adaptive consequences.45 Darwinian
(e.g., folate, niacin, flavin, vitamin B6, vitamin B12) and selection favors the conservation and expansion of favor-
minerals (e.g., zinc, iron) are required for nucleotide biosyn- able mutations (by positive or balancing selection), and the
thesis; nutritional deficiencies can impair the synthesis of elimination of mutations that are deleterious (referred to
nucleotide precursors and lower DNA synthesis rates (rates as “negative” or “purifying” selection). Not all genes evolve
of mitosis) and/or the fidelity of DNA replication (DNA at the same rate, because positive selection only acceler-
mutation rates). For example, folate deficiency inhibits ates mutation fixation at defined loci within the genome.
thymidylate (dTMP) synthesis, which increases incorpora- Mutations that confer reproductive and/or survival advan-
tion of uridylate (dUTP) into DNA, resulting in increased tage within a single environmental context expand within
frequencies of DNA strand breaks.31–35 Furthermore, defi- populations at higher rates than neutral mutations and
ciency of dietary antioxidants that scavenge chemical radi- replace a population’s preexisting variation.
cals, or excesses of pro-oxidant nutrients such as iron, may Mutations that alter physiological processes are under
increase mutation rates.36–38 Other dietary components constraint and subject to positive, balancing, or negative
affect DNA mutation rates by altering cellular redox states selection. Patterns of genetic variation across the human
or functioning as genotoxic radicals that chemically modify genome are affected by demographic history, mutation,
purine and pyrimidine bases. Certain aflatoxins, a com- recombination, and, in some cases, selection.14 Although
mon class of natural xenobiotics found in soil molds that protein-coding sequences are conserved among mammals
contaminate certain foods, increase DNA mutation rates, in general, rates of amino acid substitution vary markedly
leading to the transformation of somatic cells and localized among proteins compared to rates of synonymous substitu-
cancer epidemics.39 However, only mutations that occur in tion among genes (changes in the coding region of genes
the germline contribute to a species’ heritable genetic varia- that do not affect protein sequence).42 The proportion of
tion. Mutations that have no functional consequences are amino acid substitutions that result from positive selection
anticipated to be phenotypically silent and therefore selec- is estimated to be 35–45%.42 Mutations that alter amino
tively neutral. The level of nucleotide diversity is a function acid sequence, which affects protein structure and function,
of the DNA mutation rate (estimated to be 2.5 x 10–8 on can have physiological consequences that may be beneficial,
N utrition a l G e no m ic s • 1 8 3
deleterious, or neutral, and thereby influence an organ- particular disease state, or could affect biomarkers associ-
ism’s fitness in specific environmental and dietary contexts. ated with that disease. Model organisms, including yeast,
Likewise, mutations that affect protein expression can alter Drosophila, Caenorhabditis elegans, and mice, have been
biological network outputs, leading to altered physiology. excellent resources to identify potential candidate genes
Mutations can expand in the absence of selection and and to confirm their contribution to a metabolic pheno-
contribute to metabolic disease. The rate of mutation fixa- type. Other advancements, including the availability of
tion is a function of the effective population size, popula- high-density SNP maps of the human genome have accel-
tion demographic history, and the effect of the mutation erated the identification of human disease alleles, including
on an organism’s fitness.14 Polymorphisms can expand and low-penetrant alleles that may make relatively small con-
become fixed within a population through the processes of tributions to the initiation and/or progression of complex
genetic drift or natural selection. Drift is a stochastic process disease.14 Furthermore, haplotype maps of human genetic
resulting from the random assortment of chromosomes at variation offer advantages for disease-associational studies
meiosis. Because only a fraction of all possible zygotes are because they are simpler than SNP maps,46 but their util-
generated and survive to reproduce, mutations can expand ity may be limited because of the variability in haplotype
through many generations by the random sampling of gam- diversity across candidate genes.47 The candidate gene
etes in the absence of selection.14 Drift is expected to have a approach, while successful in identifying alleles underlying
greater influence on genetic variation in small populations monogenic traits,5,48 is limited by incomplete knowledge
expanding rapidly; drift in large, static populations is not of gene function, incomplete knowledge of transcriptional
usually as significant. Genetic drift becomes relevant in large and metabolic networks that suggest candidate genes for
populations undergoing “bottlenecks” (massive reductions analyses, and the multifactorial nature of most complex
in population) or in founding events that have occurred dur- human metabolic chronic diseases (most are polygenic
ing human migrations; for instance, population groups that traits with multiple environmental components, which,
include the Old Order Amish, Hutterites, and Ashkenazi in isolation, make relatively minor contributions to the
Jews.14 In such populations, rare disease alleles can expand disease phenotype). This is witnessed by the many incon-
rapidly and increase the incidence of diseases, including sistent findings that have emerged within the nutritional
breast cancer, Tay-Sachs, Gaucher, Niemann-Pick, and and genetic epidemiological literature, especially for the
familial hypercholesterolemia.14 Although most human involvement of low-penetrant genetic alleles in chronic
genetic variations arose as a result of the neutral processes metabolic disease.49
of mutation and genetic drift, variation resulting from drift
rarely has physiological consequences in static environ- Linkage analysis and association studies
ments. However, environmental shifts like alterations in the
food supply can challenge biological systems and convert Linkage analysis and association studies are two primary
otherwise physiologically “silent” genetic variations into methods to map causal alleles for human traits, including
functional gene variants. Relevant examples are discussed diseases. Linkage analysis compares genetic markers across
below. the genome in normal and affected individuals from the
same family and determines whether certain markers are
inherited along with the trait. The tool has limited power
I D E N T I FI C AT I O N O F NU T R IT I O NA L LY for research in complex diseases because sample sizes are
A N D E N VI RO N M E N TA L LY S E NS I T I VE usually small, due to the limited number of meiosis events
HUM A N A L L E L E S within families.17,74,75 The genome-wide association studies
(GWAS) genotype a set of genetic markers, usually SNPs,
Candidate gene approach
in a group of affected individuals and unaffected control
The vast majority of known functional polymorphisms that individuals to detect an association between a particular
contribute to food intolerances and metabolic disorders genomic region and the specific traits of interest. Facilitated
were first identified as highly penetrant disease alleles from by the HapMap project and the availability of large-scale
epidemiological or clinical studies (Table 12.1). Candidate genotyping platforms, this approach has been widely used
genes were analyzed for genetic variation; their selection to examine genome-wide genetic markers in samples with
as candidate genes was based on existing knowledge of thousands of or even tens of thousands of subjects. The list
metabolic pathways and inference that their impairment of traits under GWAS study has been categorized in a pub-
could result in metabolic phenotypes associated with a lic database, and the new candidate genes generated from
1 8 4 • P rincip l e s o f G e no m ic M e dicin e
Table 12.1 CANDIDATE HUMAN DISEASE ALLELES THAT AFFECT THE
UPTAKE OR METABOLISM OF DIETARY COMPONENTS
Vitamin B12
MTR N919G 258
Minerals
Iron HFE C282Y 154,263
Lipids
APOB many 266,267
APOE many 68
Alcohol
ADH/ALDH2 many 80,82
Carbohydrate
Lactose LCT promoter 64
Detoxification/Oxidative Stress
NAT1/NAT2 many 269,270
these studies provide novel hypotheses for disease initiation common disease–common variant hypothesis states that
and progression. disease-susceptibility alleles arose before humans migrated
out of Africa and therefore exists at high frequency across
all human populations.52,53 However, both “single-gene”
Evolutionary analysis
disorders, including cystic fibrosis and hemochromato-
Genes that have undergone accelerated changes or evo- sis, as well as complex diseases, can be associated with
lution display genomic signatures that can be identified geographically restricted populations because the alleles
computationally. The identifiable genomic signatures arose after migrations out of Africa.7,19,51,54–56 Therefore,
include the presence of an excess of rare variants within although 85–90% of all human genetic variation is found
a population (which can be indicative of a selective within populations and presumably arose prior to human
sweep), large allele frequency differences among popula- migrations, some of the 10–15% of variation among pop-
tions, and a common haplotype that remains intact over ulations probably arose from recent selective pressures
long distances.19,45,50,51 The identification of polymorphic that contributed to both simple and complex disease.57,58
alleles that have arisen as a result of historical selection Comparison of genomic sequence divergence among
resulting from nutritional challenges offers the oppor- mammalian species enables the identification of ancient
tunity to identify genes that contribute to monogenic selection throughout the process of speciation and genetic
metabolic disorders, as well as low-penetrant alleles divergence (Table 12.2). These approaches can identify
that contribute to complex metabolic disease.14,51 The single genes or pathways within biological networks that
N utrition a l G e no m ic s • 1 8 5
Table 12.2 DIET-RELATED GENES THAT DISPLAY Bitter-taste receptor
GENOMIC SIGNATURES OF ADAPTIVE EVOLUTION
Recognition of bitter taste may have conferred a selective
GENE SPECIES/FUNCTION REFERENCE NO.
advantage by deterring the consumption of plant toxins
lysozyme langur monkey 42,274,275
that often elicit the sensation of bitterness.61 The TAS2R16
ribonuclease langur monkey 42,275
gene encodes a G protein-coupled receptor that is activated
Cox4 Primates 276
by salicin in fruit, amygdalin in almonds, and many com-
LCT human lactose metabolism 63
mon β-glucopyranosides that elicit cyanogenic toxicity.
ADH1B human ethanol metabolism 81 The K172V polymorphism increases the receptor’s sensi-
ALDH2 human ethanol metabolism 83 tivity to cyanogenic glycosides and displays signatures of
HFE human iron homeostasis 77 positive selection. The adaptive allele arose in the Middle
PPARγ human nuclear receptor 45 Pleistocene era prior to human migrations out of Africa.61
PTC human bitter-taste receptor 99
1 8 6 • P rincip l e s o f G e no m ic M e dicin e
and glucose synthesis, resulting in severe hypoglycemia the MTHFR (A222V) is protective against colon cancer
following ingestion of fructose. Individuals carrying poly- in folate replete subjects.74 The MTHFR A222V variant
morphic variants of aldolase B are asymptomatic in the protein has reduced affinity for riboflavin cofactors and is
absence of fructose or sucrose consumption and can avoid thermolabile, resulting in reduced cellular MTHFR activ-
the recurrence of symptoms by remaining on a fructose- and ity; its stability is increased when folate is bound.75 The
sucrose-free diet. Chronic fructose ingestion in infants ulti- prevalence of the MTHFR allelic variant varies markedly
mately leads to hepatic and/or renal failure and death. The among human population and occurs with an allelic fre-
prevalence of these variants differs throughout Europe; the quency of nearly 40% in some Hispanic populations, but
L288 delta C frameshift mutation is restricted to Sicilian it is mostly absent in African populations.76 However, it has
subjects. These aldolase B variants probably emerged in not been reported to display signatures of positive selection.
populations through random drift; fructolysis is not an Although the biochemical role (if any) of these polymor-
essential metabolic pathway for humans, and fructose has phisms in the etiology of neural tube defects and cancer is
not been an abundant dietary component throughout most unknown, it is demonstrated that some carriers of MTHFR
of human history. However, the incidence of HFI intoler- variants require higher folate intakes than others to: 1) sta-
ance has increased since the widespread use of sucrose and bilize the MTHFR protein, 2) lower the concentration of
fructose as nutrients and sweeteners, providing an excel- the metabolic intermediate homocysteine, and 3) decrease
lent example of an environmental shift that has resulted in a women’s risk of bearing children with developmental
the apparent conversion of normally nonpenetrant “silent” anomalies, including neural tube defects.76 The fortification
aldolase B alleles into HFI disease alleles.67 of the food supply with folic acid that occurs in many coun-
tries targets women of childbearing age for birth-defect
prevention, with genetically at-risk subgroups receiving the
Lipid metabolism
most benefit.
Apolipoprotein E (apoE) is a polymorphic protein that
functions in lipid metabolism and cholesterol transport.68
Iron metabolism
All human populations display apoE polymorphism.
There are the three common allelic variants, ε2, ε3, and Hereditary hemochromatosis is a recessive iron-storage dis-
ε4, whose relative distribution varies among populations; ease that is prevalent in populations of European descent,
the frequency of the ε4 allele declines from northern to with an incidence of one in 300 persons. The HFE gene
southern Europe. These variant alleles encode proteins that is polymorphic and encodes a protein that regulates iron
differ in their affinity both for lipoprotein particles and for homeostasis. A common polymorphism, HFE C282Y,
low-density lipoprotein receptors. The ε4 allele increases emerged approximately 138 generations ago.19,77,78 This
the risk for late-onset Alzheimer’s disease and arterioscle- SNP is associated with the disease phenotype in 60–100%
rosis with low penetrance. Carriers of the ε2 allele tend to of Europeans. The HFE C282Y allele is not present in
display lower levels of total plasma cholesterol, whereas car- Asian and African populations, despite the presence of
riers of the ε4 allele, which may be ancestral, display higher iron-storage diseases in those populations indicating that
cholesterol levels. Therefore, serum cholesterol levels are other genes are associated with hereditary hemochroma-
likely to be more responsive to low-fat and low-cholesterol tosis. Furthermore, the penetrance of the C282Y HFE
diets in carriers of the ε4 allele.69,70 allele for the iron-overload phenotype varies widely among
homozygotes, with some individuals being asymptomatic,
indicating the presence of modifying alleles. The recent
One-Carbon metabolism
expansion of this polymorphism may have conferred selec-
Folate-mediated one-carbon metabolism is required for tive advantages in iron-poor environments77,78 or resistance
purine, thymidylate, and methionine biosynthesis, and it to microbial infection.79
affects genome synthesis, stability, and gene expression.71
Several polymorphic alleles have been identified as associ-
Alcohol metabolism
ated with metabolic perturbations that can confer both pro-
tection and risk for specific pathologies and developmental Ethanol metabolism efficiency varies considerably among
anomalies.72 SNPs in MTHFR (A222V) and MTHFD1 human populations.80 Ethanol is oxidized to acetaldehyde
(R653Q),73 which encode folate-dependent enzymes, by the enzyme alcohol dehydrogenase, encoded by the ADH
are associated with increased risk for neural tube defects; genes. Acetaldehyde, a toxic metabolite, is subsequently
N utrition a l G e no m ic s • 1 8 7
oxidized to acetic acid by the enzyme aldehyde dehydroge- in mitochondria. This variant has been proposed to confer
nase, which is encoded by ALDH2. Seven ADH genes that advantage to meat-eating populations but is detrimental to
are clustered on chromosome 4 encode proteins with dis- vegetarians. It displays evidence for positive selection.85 The
tinct catalytic properties and tissue-specific expression pat- allelic frequency varies among human populations and cor-
terns. Two of the genes encoding class I enzymes (ADH1B relates with historical dietary patterns; it is present at a fre-
and ADH1C) are expressed in liver, function in systemic quency of 28% in Saami and 2.3% and 3% in Chinese and
ethanol clearance, and display functional polymorphism. Indian Hindus respectively.
A variant ADH1B* 47His allele predominates in Japanese
and Chinese populations but is rare in European and
Vitamin C transport
northern-African populations.81 The variant allele encodes
an enzyme with elevated enzyme activity, leading to more The individuality of vitamin C needs was first recog-
rapid formation of acetaldehyde. The ADH1C*349Ile vari- nized in 1967.86 Vitamin C is required for the function
ant is found in Europeans while the ADH1B*369Arg vari- of at least 8 mammalian enzymes and can scavenge reac-
ant is mostly restricted to individuals of African descent. tive oxygen species.87 There are two genes that encode
ALDH2 is also highly polymorphic, and Asian populations sodium-dependent vitamin C transporters, SLC23A1 and
carry a common dominant null allelic variant (E487K) and SLC23A2. These genes resulted from an early gene dupli-
develop a characteristic “flush” reaction when consuming cation event but appear to have acquired distinct func-
alcohol, resulting from acetaldehyde accumulation.82 ADH tions. SLC23A1 is responsible for intestinal and renal
and ALDH alleles that predominate in east Asian popula- absorption of vitamin C and therefore has the potential
tions display signatures of positive selection, and the expres- to affect whole-body vitamin C accumulation.87 The over-
sion of these variant alleles results in elevated acetaldehyde all mutation rate of these genes is similar, as evidenced by
concentrations following alcohol consumption, which may the similarity in their non-synonymous substitution rates.
have conferred advantage by protecting against parasite However, four population-specific non-synonymous sub-
infection.83 The high frequency of ADH1B *47His in East stitutions are seen in SLC23A1 that arose after the migra-
Asians may have resulted from adaptation to rice domesti- tions out of Africa, indicating a potential role for selection
cation and the consumption of fermented beverages.84 in the expansion of these variant alleles.87 Only synony-
mous substitutions are observed in SLC23A2, and deletion
of the orthologous slc23a2 in mice is lethal, indicating the
Glyoxylate metabolism and kidney stone disease
critical, non-redundant function of this transporter that
Kidney stone disease resulting from calcium oxalate seems to be under selective constraint. A common SNP in
(CaOx) formation is common in Western populations SLC23A2, rs1279386, has been linked to plasma vitamin
and results from multiple etiologies.85 Polymorphism in C concentrations, with carriers of the GG genotype exhib-
the AGXT gene that encodes the enzyme alanine: gly- iting lower plasma vitamin C concentrations than do car-
oxylate aminotransferase (AGT) results in an accumula- riers of other genotypes.88 The effects of non-synonymous
tion of glyoxylate, a toxic intermediary metabolite that is SLC23A1on vitamin C transport or physiology have not
converted to oxalate. Oxalate does not undergo further been investigated.87
metabolism and accumulates as insoluble CaOx precipi-
tates that accumulate in the kidney and urinary tract. There
Energy metabolism
are two major precursors of glyoxylate: glycolate, which is
present in plant-based foods, and hydroxyproline, which is The “thrifty gene” hypothesis was first proposed over
present in meat collagen. Glycolate metabolism occurs in 40 years ago to account for the epidemic of type 2 dia-
the peroxisomes, while hydroxyproline is metabolized in betes observed in non-Western cultures that adopted
mitochondria. Among mammals, the intracellular localiza- Western-style diets and lifestyles.89,90 The hypothesis states
tion varies among carnivores and omnivores; it is primar- that exposure to frequent famine selected for gene variants
ily peroxisomal in herbivores, mitochondrial in carnivores, that enabled the more efficient conversion of food into
and both mitochondrial and peroxisomal in omnivores. In energy and fat disposition during times of unpredictable
humans, AGT is usually localized to peroxisomes. A com- and sometimes scant food supplies. The putative adapta-
mon AGT variant, Pro11Leu, decreases enzyme activity by tions also may have resulted in more efficient adaptations
about 70% but also results in the formation of a mitochon- to fasting conditions (e.g., more rapid decreases in basal
drial leader sequence and 5% of AGT protein localization metabolism) and/or physiological responses that facilitate
1 8 8 • P rincip l e s o f G e no m ic M e dicin e
excessive intakes in times of plenty. Conclusive genomic Adaptive alleles may become recessive-disease alleles, or
data have not yet supported this hypothesis.90,91 disease alleles even in heterozygote individuals, when the
environmental conditions change profoundly, such as those
brought about by the advent of civilization and agriculture,
Starch Digestion
including alterations in the nature and abundance of the
A well-known example for the adaptive role of CNV food supply.41,42,45,50,95–99 Adaptive alleles may be respon-
is involved in starch digestion. The salivary amylase sible for the generation of metabolic disease alleles both
gene (AMY1) shows extensive variation in copy num- within and across ethnically diverse human populations,
ber among individuals and between populations. It and therefore are strong, nonbiased candidate genes for
was also demonstrated that the gene copy number is disease association studies: the interacting and modifying
positively correlated with the protein level. Populations environmental factors can be inferred from the nutrients
consuming a high-starch diet, such as agricultural popu- and/or metabolites that are known to interact with the gene
lations of European Americans and Japanese, and Hadza product.14
hunter-gatherers, who rely extensively on starch-rich
roots and tubers, have higher copy number of AMY1
than populations consuming a low-starch diet, such as N U T R I T I O N AL E P I G E N ET I C S
hunter-gatherers in the rainforests and near the Arctic
Circle. Comparison with other great apes, chimpanzees, Traits can be inherited from one generation to the
bonobos, New World monkeys, and Old World monkeys next through both genetic and epigenetic mechanisms.
indicate that the increased copy number of AMY1 origi- Classical genetic inheritance refers to the transmission of a
nated in the human lineage. The low amount of nucleo- DNA primary sequence from one generation to the next.
tide divergence among different gene copies indicates a Concordance among monozygotic twins illustrates the pre-
recent origin that may be within the timeframe of mod- dominant yet non-exclusive contribution of DNA primary
ern human origins (~200,000 years ago). Taken together, sequence to human phenotypes; other modes of inheritance
the copy number variations of the AMY1 gene among dif- must also be operative.100,101 Epigenetics refers to the inheri-
ferent populations might represent regional adaptation to tance of traits through mechanisms that are independent
diets with varying starch content, an interesting example of DNA primary sequence. There are now many examples
demonstrating the role of diet in modulating the human demonstrating the inheritance of gene expression patterns
genome.30 and/or levels independent of DNA primary sequence, and
such differences can elicit phenotypical differences among
individuals, including monozygotic twins through multiple
Oxidative metabolism
generations.100
Variations that have an impact on human nutrition and Interest in the relationships among epigenetic events
metabolism may have arisen independently of direct and human nutrition was ignited by the “fetal origins of
nutritional challenges. The enzyme glucose-6-phosphate adult disease” hypothesis, originally put forward by David
dehydrogenase is solely responsible for the generation of Barker and colleagues.102 Barker proposed that nutrition
reduced nicotinamide adenine dinucleotide phosphate acts very early in life to program risk for adverse outcomes
(NADPH) in red blood cells and therefore is required in adult life (Figure 12.2). The notion that phenotypical
to prevent oxidative damage. Variants with low activ- plasticity was associated with in utero environmental expo-
ity resulting from amino acid substitutions, including sures had been validated in the toxicology literature, but
the G6PD-202A allele, are enriched in sub-Saharan not well considered in the nutrition literature.103,104 Until
African populations and arose 2,500 to 6,500 years recently, Barker’s hypothesis was supported only by epide-
ago.92 Presumably, this allelic variant became enriched in miological associations among early nutritional exposures
populations as a result of balancing selection because it and increased risk in adulthood for obesity, hypertension,
conferred resistance to malarial disease in heterozygous and insulin resistance, which are the antecedents of adult
females and hemizygous males.93,94 chronic disease, diseases that include cardiovascular disease
These examples illustrate the role of environmental (CVD), diabetes, and metabolic syndrome.102,105 But now
exposures, including pathogens and dietary components, as there is an emerging, basic science literature that supports
selective forces that facilitated the expansion of alleles that the concept that fetal environment can, in fact, “program or
alter the utilization and metabolism of dietary components. reprogram” the fetal genome with lifelong consequences.103
N utrition a l G e no m ic s • 1 8 9
The human genome has evolved in the context of a nutrient and metabolic networks.113 DNA is modified by meth-
environment that was often scanty and always unpredict- ylation of cytosine deoxyribonucleotides present in the
able. Hence, organisms developed the capacity to “sense” sequence CpG. Cytosine methylation is usually associ-
and “adapt” to the food supply. These genomics adapta- ated with gene silencing; methylcytosine is bound by
tions have been referred to as “metabolic imprinting” or methylcytosine-binding proteins that heterochromatize
“metabolic programming.” Such adaptations occur within DNA and hence silence the gene. Histone proteins are
critical windows in development, are seemingly irreversible, essential components of chromatin and are subject to mod-
and permit in utero survival in the context of a suboptimal ification by methylation, phosphorylation, ubiquitination,
nutrient environment, but they may predispose the affected ribosylation, and acetylation, all of which modify gene
individual to metabolic disease in adulthood.106–111 transcription efficiency and can influence DNA stability.114
Waterland and Garza described specific criteria to dif- Alterations in DNA and histone methylation constitute
ferentiate adaptive metabolic imprinting phenomena from the epigenetic signatures that enable genome programming
toxicological responses that resulted in permanent genomic because of their potential connections to metabolic net-
alterations during the affected individual’s lifetime. These works, chromatin structure, and transcriptional networks.
criteria include: 1) a susceptibility window limited to a Furthermore, DNA and histone methylations are meta-
critical ontogenic window in development, 2) a persistent stable, heritable, and alter genome expression and stabil-
effect lasting through adulthood, 3) a specific and measur- ity. Methylation is a higher order genomic signal that can
able outcome, and 4) a dose–response or threshold rela- override transient metabolic or hormonal signals such as
tionship between a specific exposure and an outcome.111 the regulation of transcriptional networks through nuclear
Mechanisms for metabolic imprinting are beginning to be receptors (e.g., vitamin A, vitamin D, steroid hormones).
understood and modeled in animal systems; some exam- The molecular mechanisms that describe the interac-
ples are illustrated below. Metabolic programming mecha- tions among nutrients, metabolism, and gene expression/
nisms are more complex than those associated with toxic genome programming are mostly unknown. Following
or deficiency states. Whereas many teratogens are exog- are two illustrative examples of genome programming by
enous agents that disrupt biological processes, metabolic nutrition.
imprinting is a conserved and adaptive response that opti-
mizes biological function in one life stage through genomic
M AT E R NA L F O L AT E , O N E - C A R B O N
mechanisms that result in permanent functional character-
M ETA B O L I S M , FETA L G E N O M E
istics. These imprints, however, may prevent or limit the
P RO G R A M M I N G, A N D IN U TE RO S U RVI VA L
range of other adaptive mechanisms that are protective in
subsequent life stages when environments change, such as Folate is a B-vitamin and a family of metabolic cofactors that
in a transition from “dearth” to “surplus.” Thus, once the carry and chemically activate one-carbon units for the de novo
program is established, a system’s buffering capacity is lim- synthesis of purine nucleotides and thymidylate (dTMP),
ited. The novelty of sustained surplus food present in most and for the remethylation of homocysteine to methionine, a
Western cultures may explain some of the limited response metabolic network known as “folate-mediated one-carbon
capability apparently at the core of the present obesity/type metabolism” (Figure 12.3). Methionine can in turn be ade-
2 diabetes epidemic.112 nosylated to form S-adenosylmethionine (AdoMet), which
Genome programming results from chemical modifi- is a cofactor for numerous cellular methylation reactions,
cations of chromatin, either DNA or histone proteins, at a including histone and DNA methylation.72 Impairments in
specific locus that leads to programming of transcriptional this metabolic network by nutritional deficiencies or highly
“Program”
“Imprint”
Figure 12.2
The fetal origins of disease hypothesis. Fetal environmental exposures, especially nutritional, act in early life to program risk for adult
health outcomes.
1 9 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
penetrant SNPs increase risk for pathologies such as cancer folate-dependent dTMP synthesis, catalyzed by thymi-
and cardiovascular disease, and developmental anomalies dylate synthase (TS) and 5-methylTHF synthesis (leading
such as spontaneous abortion (SA) and neural tube defects to Adomet synthesis) catalyzed by methylenetetrahydrofo-
(NTDs).72 Folate supplementation can reduce the risk for late reductase (MTHFR), are competitive pathways within
these disorders; maximal benefit is achieved in genetically the one-carbon network.122 They compete for a limiting
susceptible individuals and populations. pool of the cofactor methylenetetrahydrofolate (methy-
Methionine and dTMP synthesis are the most vul- leneTHF) (Figure 12.3).72,117,118 This metabolic competition
nerable pathways within the network; their impairments is unbalanced by the previously described MTHFR A222V
compromise the fidelity of DNA synthesis and cellular meth- polymorphism. This functional SNP reduces MTHFR
ylation reactions.72,115;31 AdoMet-dependent methyltrans- activity and has two effects on the network. It impairs the
ferases, including histone and DNA methyltransferases, are remethylation of homocysteine to methionine, thereby
subject to product inhibition by S-adenosylhomocysteine reducing global DNA methylation and thus also influences
(AdoHyc), which accumulates during folate deficiency gene expression.123,124 It also increases the conversion of
(Figure 12.3).116–118 Hence, methylation of chromatin is sen- deoxyuridine monophosphate (dUMP) to deoxythymidine
sitive to the efficiency of the one-carbon metabolic network; monophosphate (dTMP).125 These changes in the network
methyltransferases sense the efficiency of the folate metabolic are associated with increased risk for spontaneous abortions
network because their activity is determined by the cellular (SA) and NTDs, but decreased risk for adult colon cancer,76
AdoMet/AdoHyc ratio, otherwise known as the “methyla- illustrating that optimal network function or outputs differs
tion potential” of the cell.117,118 Global genomic methylcyto- between the fetal and adult environments. Identifying the
sine content is highly sensitive to the AdoMet/AdoHyc ratio, precise mechanism for folate-related pathologies in experi-
which can affect both gene expression and DNA stability.119 mental systems is challenging because any factor, genetic or
The mechanisms underlying folate-associated patholo- environmental, that influences the metabolic competition
gies, including NTDs, SA, and cancers are assumed to be for methyleneTHF may simultaneously alter the efficiency
the result of insufficient flux through the dTMP and/or of both dTMP and AdoMet synthesis (Figure 12.3).
AdoMet synthesis pathways.120,121 Therefore, the etiology Alterations in one-carbon metabolism, and the AdoMet
involves either impairments in genome synthesis (mitotic) cycle in particular, can have dramatic effects on genome
rates, genome stability and/or methylation-sensitive methylation. Both genome-wide and allele-specific DNA
gene expression.72 Numerous studies have indicated that methylation are influenced by alterations in folate metab-
olism.126 DNA hypomethylation induced by folate defi-
ciency alters transcription of genes regulated by promoter
10-formylTHF PURINES methylation, including tumor suppressor genes,126,127 and
enables interchromosomal recombination events through
common retroviral repeat sequences whose activity nor-
TS
methyleneTHF dTMP mally is silenced by methylation.128 Patients with hyper-
homocysteinemia, a clinical state that results from the
MTHFR
inability to effectively metabolize homocysteine, accu-
5-methylTHF mulate cellular AdoHyc and exhibit alterations in gene
expression. Patients exhibit DNA hypomethylation and a
homocysteine-dependent shift from monoallelic to biallelic
homocysteine
Methionine expression of genetically imprinted genes, including H19.
AdoHyc Folate supplementation in these patients restores homo-
AdoMet cysteine levels to baseline, reverses global DNA hypometh-
ylation, and restores monoallelic expression of imprinted
Methylation Reactions genes.129 Interestingly, SNPs in the H19 gene are associated
DNA with cord blood Isulin Growth Factor II (IGF-II) levels and
Histones
birth size.130
Figure 12.3
Folate-mediated one-carbon metabolism. Tetrahydrofolate Folate-mediated alterations in genome methylation can
(THF)-mediated one-carbon metabolism is required for the be set irreversibly or “imprinted” during early development.
synthesis of purines, thymidylate, and methionine. MTHFR,
methylenetetrahydrofolate reductase; TS, thymidylate synthase; In the viable yellow agouti (Avy) mouse model, maternal diet
AdoMet, S-adenosylmethionine; AdoHcy, S-adenosylhomocysteine. determines the coat color of offspring.131 This mouse strain
N utrition a l G e no m ic s • 1 9 1
www.Ebook777.com
contains an Intracisternal A-particle (IAP) element, which maternal and fetal endocrine, nutritional, or hormonal
is a retrotransposon, that integrated into a 5′ exon of the imbalances; maternal and fetal infections; and endome-
agouti gene, resulting in cryptic and constitutive expression triosis (Table 12.3).139 Few specific environmental risk fac-
of the agouti gene. This aberrant expression of the agouti tors for SA have been identified, but known factors include
results in a “yellow” coat color and an obesity phenotype.132 low maternal folate status, diabetes (type 1), and elevated
The IAP retroviral element also attracts DNA methylation homocysteine (which most often results secondarily to pri-
to that locus, and the degree of methylation determines the mary or conditioned folate deficiency).142–148 For example,
expression level of the agouti gene. variants in the transcobalamin II (TC-II) gene are associ-
The Avy mouse is sensitive to maternal folate and ated with sporadic and recurrent miscarriage.149 TC-II plays
one-carbon status during gestation. Within a critical win- a role in the delivery of vitamin B12 to peripheral tissues,
dow in development, maternal diet determines the density indicating a role for maternal vitamin B12 status, and the
of cytosine methylation at the agouti locus and hence the genes responsible for its processing and utilization, in the
level of agouti gene transcription, coat color, and propen- etiology of miscarriage.
sity for obesity.133 The methylation patterns and subsequent Developmental anomalies (DA) and SA have both inde-
effects on coat color and, presumably, associated metabolic pendent and shared etiologies (Table 12.3).135 The under-
characteristics are maintained throughout the lifetime of lying mechanisms for over 75% of DA are unknown and
experimental animals21 and are heritable.132 The identifi- assumed to be multifactorial; only 15% of those whose etiol-
cation of other genes that are influenced by alterations in ogies have been identified are solely genetic, including chro-
the AdoMet/AdoHyc ratio through chromatin modifica- mosomal abnormalities or autosomal/sex-linked genetic
tions, and the critical developmental windows that enable disease. Ten percent of DA are attributed to environmental
genome programming, are essential to elucidate the mech- factors, 4% of which are attributed to disruptions in mater-
anisms of folate-related pathologies and developmental nal/fetal metabolism or suboptimal nutrition that includes
anomalies. Equally important, this mouse model illustrates micronutrient under-nutrition, starvation, Phenylketonuria
the concept that epigenetic modifications in the developing (PKU), diabetes, alcoholism, etc. Infectious agents account
embryo induced by maternal diet can “rescue” deleterious for another 4% of DA; mechanical disruption accounts
genetic insults, such as retroviral insertions in gene promot- for 2%; and known chemical/prescription toxins account
ers, and restore the “normal” phenotype. for less than 1% of DA.135 Polymorphic variants of two
The ability of maternal folate and one-carbon sources genes that encode folate-dependent enzymes, MTHFS
to compensate or “rescue” genetic deficiencies may not be A222V and MTHFD1 R653Q, are associated with risk for
limited to the Avy mouse model, and therefore may have developmental anomalies, including neural tube defects.73
implications for women taking nutritional supplements Interestingly, human alleles associated with developmen-
during pregnancy at intake levels that exceed dietary recom- tal anomalies that encode folate-dependent metabolic
mendations. Fetal genotypes that cannot support basic bio- enzymes are not in Hardy-Weinberg equilibrium in some
logical processes in the embryonic and fetal stages usually studies (alleles are not inherited at the expected frequency),
are eliminated. In primates, this is achieved by spontaneous consistent with evidence that elevated homocysteine is a
abortion (miscarriage). Humans may be unique compared risk factor for spontaneous miscarriage and decreased fetal
to other mammalian species in their high rates of fetal viability.71,73,147,150,151
loss134; high SA rates may be a selective pressure that accel- The concept that embryos can be rescued by maternal
erates the expansion of polymorphic alleles within human nutritional status, or that “good diet hides genetic muta-
populations. Approximately 75% of human conceptions are tions,”152 is suggested by numerous examples of nutritional
lost spontaneously before term; 80% of all SA occur within rescue or compensation (viability or phenotype) of gene dis-
the first trimester.135–137 It is estimated that half of SA occur ruptions through diet in mice and yeast.71,72,153–157 Individual
before the first three weeks of gestation and generally are nutrients can rescue severe genetic lesions in mice when
unnoticed; many embryos fail to implant in the uterus.138 administered in supra-physiological levels during critical
Risk for SA increases and fertility decreases in women developmental windows. Maternal retinoic acid adminis-
over the age of 30 years.139,140 Although the etiologies of tration between 7.5 and 9.5 days post-conception rescued
SA are generally not established, the etiologies of most SA deafness and inner ear development in Hoxa1–/– mice,153
are likely to be multifactorial. Many SA fetuses have struc- and folic acid can rescue skeletal defects associated with
tural and/or genetic anomalies.135,141 Potential inducers of deletion of a Hox gene, as well as neural tube defects in mice
SA include maternal immune responses; fetal genotypes; that have no evidence of disrupted folate metabolism.152
1 9 2 • P rincip l e s o f G e no m ic M e dicin e
Table 12.3 MATERNAL RISK GENOTYPES AND REPRODUCTIVE OUTCOMES
NTD
Down syndrome
Adult CVD
MTFD1 one-carbon metabolism NTD 73
NTD
IL6 (-174 G--> C) Cytokine SA 279
The most comprehensive studies have been performed Clearly, maternal folate and other methyl donor supple-
in yeast. Gene-deletion studies indicate that 80% of yeast mentation alters the methylation status of targeted alleles
genes are nonessential for survival under laboratory condi- in the mouse embryo, and methylation patterns and subse-
tions. A recent examination of yeast metabolic networks quent effects on gene expression persist throughout adult-
using an in silico model revealed that culture conditions, hood.21 Epigenetic phenomena may provide mechanistic
especially the use of nutrient-rich culture media, can com- insight into the many observational studies that associate
pensate for the disruption of 37% to 68% of the organism’s risk for adult chronic disease with maternal nutrition and
genes. In microbial systems, the maintenance of enzymatic embryonic nutrient exposures (as proposed by Barker and
flux under highly diverse environmental conditions appears colleagues).111
to be a primary selective pressure that maintains gene
sequence, starvation being among the most common envi-
G LU C O C O RT I C O I D S A N D M ETA B O L I C
ronmental stresses.158 The concept of nutritional rescue of
D I S E A S E : MO D I FI C AT I O N O F T H E
genetic mutations is exceptionally salient to human moder-
P L AC E N TA L BA R R I E R BY M AT E R NA L
nity because of the unprecedented degree to which we
NU T R IT I O N
can manipulate our nutritional environments. The inborn
error of metabolism phenylketonuria provides the classical The consequences of fetal glucocorticoid (GC) exposure
example for the effectiveness of dietary manipulations in on adult chronic disease provide some of the best support-
modifying deleterious phenotypes resulting from genetic ing evidence for the fetal origins of disease hypothesis.161–164
mutations that alter metabolism. Restriction of phenylala- Complementary human clinical and animal studies have
nine from the diet can prevent severe cognitive deficits in revealed the long-term consequences and associated mech-
children with mutations in the phenylalanine hydroxylase anisms of fetal GC exposure (Figure 12.4).165–167 Fetal GC
gene.159 Likewise, maternal folic acid supplementation and levels are maintained at low concentrations relative to mater-
fortification of the food supply with folic acid reduces the nal concentrations, primarily through the action of placen-
occurrence and recurrence of neural tube defect–affected tal 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2),
pregnancies in genetically susceptible populations.160 which catalyzes the oxidative inactivation of cortisol and
N utrition a l G e no m ic s • 1 9 3
Initiating signal Programming
GC Induced events
–– Small placenta
–– CNS defects
–– Attenuated HPA axis feedback sensitivity
–– Altered GR promoter methylation/expression declines
GR knock-out in mice –– Altered dopaminergic programming
–– Increased PEPCK expression
(in adulthood/2nd generation)
Outcomes
PEPCK over expressing mice
–– inhibited insulin suppression of –– Low Birth weight (IUGR)
gluconeogenesis
–– increased insulin –– Elevated plasma GC in adulthood
–– glucose intolerance –– Hypertension
–– Hyperglycemia
–– Insulin resistence
–– Hyperinsulinaemia
–– Anxiety
corticosterone.168 Elevated fetal exposures to GC dur- axis (HPA) activity. These disorders persist not only into
ing late gestation169 (which can result from 11β-HSD2 adulthood, but also into the next generation.174
inhibitors, rare mutations in the human 11β-HSD2 gene, GCs are steroid hormones that serve as ligands for the
or large existing variation in placental 11β-HSD2 activity glucocorticoid receptor (GR), a member of the nuclear
among humans) have lifelong consequences for the fetus, receptor superfamily.175 GRs are complexed with heat
including low birth weight (IUGR), elevated plasma GC, shock proteins and do not affect transcription in the
hypertension, hyperglycemia, insulin resistance, hyperin- absence of bound ligand. With GC bound, GR remod-
sulinemia, and anxiety.164 IUGR and preeclampsia also are els chromatin through transient interactions that recruit
associated with elevated cortisol and low birth weight.170,171 remodeling proteins to a defined locus and “open” chro-
Furthermore, elevated GC and the metabolic syndrome matin thereby enabling transcription factor binding.175
(i.e., the combination of type 2 diabetes/insulin resistance, GR also may recruit transcription factors directly to the
dyslipidemia, and hypertension) are also characteristic of transcriptional preinitiation complex,175 and may target
Cushing’s syndrome.162 promoter demethylation.176 GCs are required for normal
Interestingly, low maternal dietary protein intake dur- central nervous system (CNS) development, apoptosis,
ing gestation causes a specific loss of placental 11β-HSD2 and synapse formation.177 Both GR and 11β-HSD2 are
expression as well as fetal outcomes similar to those present in the developing brain; 11β-HSD2 expression
observed from elevated fetal GC exposure.172 11β-HSD2 begins to dissipate and brain GC accumulates from week
activity in placenta correlates with birth weight,173 and dis- 19 to 26 during end stages of neurogenesis.173,178 Lifelong
ruption of the murine gene encoding 11β-HSD2 reduces consequences associated with fetal GC exposure may
birth weight.164 Similarly, obstetrical GC therapy to acceler- result from premature GR-mediated chromatin remodel-
ate lung development prior to anticipated preterm deliveries ing in the hippocampus.164 Interestingly, GR programming
increases risk for reduced fetal birth weight, and long-term resulting from fetal GC exposure, and its deleterious reper-
susceptibility to hypertension, hyperglycemia, cardiovascu- cussions, can be erased in the adult animal by treatment
lar disease, and increased hypothalamic-pituitary-adrenal with histone deacetylase inhibitors.164
1 9 4 • P rincip l e s o f G e no m ic M e dicin e
GC homeostasis is maintained by the HPA axis, which manifestations of the functioning or malfunctioning of bio-
is imprinted or programmed by fetal GC exposures dur- logical systems. Therefore, in contrast to traditional reduc-
ing gestation. Plasma GC concentrations are normally tionist approaches, a systems framework is fundamental to
regulated by a feedback loop that involves GR in the hip- understanding the concept of life and disease, and to pro-
pocampus. Persistent prenatal GC exposure decreases fetal vide evidence-based guidance for effective disease preven-
GR expression, and these expression levels can be “set” or tion and management. This systems perspective has been
“memorized,” with lifelong consequences. Low hippocam- recognized for almost one century,189 but it was not until the
pal GR levels increase adult corticosterone plasma levels beginning of the twenty-first century that systems biology
and thereby may reinforce the decreased GR expression lev- emerged as a new discipline.190,191 The remarkable growth
els. Maternal undernutrition can elicit the same effect, pre- of systems biology during the past decade was enabled by
sumably by decreasing placental 11β-HSD2, and thereby the rapid development of technologies that permit global
program hypertension and hyperglycemia in the fetus. Fetal cell measurements and computational analyses. Global cell
GC exposure also affects the dopaminergic system179 and measurement technologies at the levels of the genome, tran-
alters amygdala function, including the regulation of fear scriptome, proteome, metabolome, and interactome have
and anxiety.180 In addition, maternal GC exposure affects opened the possibility of enumerating all players within
glucose and insulin homeostasis181 by programming hepatic the biological networks and all types of interactions among
Phosphoenolpyruvate carboxykinase (PEPCK) levels with them. The availability of comprehensive high-throughput
effects that persist into adulthood182,183 and are carried to data sets, in the context of long-accumulated biochemical
the next generation.174 knowledge and advanced mathematical and computational
Interestingly, GR levels also can be imprinted epigeneti- modeling approaches, permits comprehensive construction
cally and postnatally in rodents through maternal behavior, of biological networks. Network models serve as scaffolds
including handling, licking, and grooming.161 At day 20 to assist the analysis and interpretation of high-throughput
post-conception, the GR exon 17 promoter is unmethyl- data190,192,193 and elucidate the structure and dynamics of
ated. One week after birth, offspring from low licking/ biological systems and their roles in physiological and path-
grooming-arched-back nursing mothers uniformly exhib- ological states. Nutritional systems biology is an integral part
ited methylated CpG islands in the NGFI-A cis-element of systems biology, with a specific focus on the impact of
located within the exon 17 promoter and elevated GR nutrients and other environmental exposures on biological
expression in the hippocampus, whereas offspring from networks, the interrogation of perturbed networks underly-
high licking/grooming-arched-back nursing mothers rarely ing metabolic diseases using nutrients and other bioactive
methylated this sequence.184,185 These elevations in GR food components, the discovery of metabolic patterns or
expression lower plasma GC levels and the stress response metabolites as biomarkers, and the network-assisted devel-
through the HPA axis throughout the animal’s lifetime. In opment of nutritional interventions to prevent or manage
humans, elevated maternal choline intake, which increases metabolic diseases.188
the cellular methylation capacity, results in elevated placen-
tal promoter methylation of the cortisol-regulating genes
T H E N ET WO R K BA S I S O F M ETA B O L I C
corticotropin-releasing hormone and glucocorticoid recep-
DISEASES
tor, and 33% lower concentration of cortisol cord plasma.186
A systems approach is essential to understanding molecular
processes underlying metabolic diseases. From the point of
N U T R I T I O N AL SYS T EMS B I O L O GY view of systems biology, metabolic diseases are manifesta-
tions of the dysregulation of biological networks, which
All biological processes, including gene transcription, emphasizes the impairment of multiple processes, includ-
protein synthesis, as well as signaling and metabolic path- ing signaling, regulation, catalysis, and transport, or altered
ways, are interrelated, interdependent, and function as a interactions among them. The proper functioning of a
dynamic and complex cellular network. Food and nutrient metabolic network is represented by the coordinated flows
intakes, as well as other environmental exposures, modu- of metabolites through the network. The flow of metabo-
late life processes by interacting with biological networks. lites is measured by metabolic flux, which is defined as the
The impact of such external inputs is rarely limited to their rate at which every metabolite is produced or consumed by
primary targets, due to the interconnectedness of pathways each reaction.187,194 Although metabolic diseases are gener-
within the networks.187,188 Human health and disease are ally characterized as a series of impaired metabolic fluxes,
N utrition a l G e no m ic s • 1 9 5
these can originate from a single, causal primary defect accounts for comorbidity among related diseases.187,199–201
as occurs in most inborn errors of metabolism (IEM), or Clustering diseases based on their shared phenotypes may
from the multiple subtly altered metabolic fluxes observed assist us in the identification of their common molecular
in most complex diseases. Complex diseases illustrate the basis, but this approach is challenged by variable penetrance
importance and complexity of interacting genes and path- and by a lack of consistent diagnoses of phenotypes.198
ways in pathophysiological processes, whereas monogenic Interestingly, this concept has instead been supported
IEM illustrate the principles by which impaired networks through a reverse approach: disease etiology with common
cause diseases. Although most IEM are caused by single molecular bases tends to occur simultaneously in the same
gene defects, the effect of this impairment can spread individuals.199,202 By integrating available databases of meta-
across the metabolic network, as accumulated metabo- bolic networks203,204 and compilations of disease–gene asso-
lites can be toxic and impair developmental or neurologi- ciations,205 it is possible to cluster diseases based on their
cal processes and/or be shunted to secondary pathways shared molecular basis. Diseases can be clustered by their
and influence their metabolic fluxes. Moreover, upstream common relatedness to genes, expression patterns, partici-
and downstream alterations in metabolite concentrations pation in multiprotein complexes, or adjacency within a
can inhibit or activate other pathways. For example, gly- metabolic network.199–201 A metabolism-based disease net-
cogen storage disease type Ia (GSD-Ia), an IEM caused work (MDN) was constructed by clustering diseases that
by defects in glucose-6-phosphatase alpha (G6PC), leads are associated with specific genes that function in common
to the accumulation of glucose-6-phosphate and reduced pathways.199 In this MDN, connected disease pairs are three
supply of glucose; hypoglycemia following a short fast is times more likely to occur together in the same individu-
a hallmark of GSD-Ia. G6PD catalyzes the hydrolysis of als than the average of all disease pairs. The more a disease
glucose-6-phosphate (G6P) into glucose and inorganic is connected to other diseases in the MDN, the higher its
phosphate, the last step of glycogenolysis and gluconeo- prevalence in the population and the higher its mortal-
genesis. In this disorder, G6P is shunted to other pathways, ity rate.199 Similar results were found in disease networks
including glycogenesis, glycolysis, and lipogenesis, leading built based on protein–protein interactions, co-expression
to complications such as hepatomegaly, nephromegaly, and disease-gene sharing.201 These disease networks are the
hyperlipidemia, and lactic academia.195,196 Furthermore, the phenotypical representations of the underlying cellular
accumulated G6P may also play a regulatory role in the acti- networks.
vation of transcription of lipogenic genes, further disturb-
ing lipid metabolism.197
G L O BA L R EC O N S T RU C T I O N O F
Similar clinical presentations can result from muta-
HUM A N M ETA B O L I C N ET WO R K S
tions in different genes when the encoded genes are closely
linked within the network architecture, such as serving as The global reconstruction of biological networks is a pre-
components of a multiprotein complex, a pathway, or a requisite for the application of a systems approach to study
cellular organelle, and contribute to the same downstream human physiology and pathology.193,204 It provides a scaf-
metabolic fluxes.198 For example, glycogen-storage disease fold and a computable model for analyzing and interpreting
type I (GSD-I), caused by impaired hydrolysis of G6P into large-scale omics data to reveal perturbed pathways under-
glucose and inorganic phosphate that originates by several lying different pathophysiological states. Moreover, the
mechanisms, including mutations in G6PC, as explained reconstruction could be transformed into an in silico model
above for GSD-Ia. Related phenotypes result from the that is amenable to computational analysis and math-
impairment of the G6P translocase (G6PT), which is ematical modeling. Computational analysis of the in silico
responsible for the transport of G6P from cytosol into the model may unravel human-specific network properties
lumen of the endoplasmic reticulum, where the hydrolysis or principles regarding network structure and dynamics.
takes place. GSD-I caused by G6PT deficiency is referred Mathematical modeling enables simulations and predic-
to as “GSD-Ib,” which has clinical presentation essentially tions of network responses to genetic and environmental
identical to that of GSD-Ia.195,197 perturbations, including drug treatments and nutritional
Network architecture defines both the clinical pre- interventions, and it can assist in the discovery of biomark-
sentation and the underlying molecular basis of diseases, ers and the development of disease management strategies.
as well as variations in symptoms and clinical biomarkers The comprehensive reconstruction of biological net-
that define disease, in disorders that include obesity, dia- works is a daunting undertaking. The various types of cellu-
betes, Gaucher disease, and Parkinson disease.199 It also lar networks, usually classified as signaling, gene regulatory,
1 9 6 • P rincip l e s o f G e no m ic M e dicin e
and metabolic, must be integrated. Each of these networks associations.204 Transcriptomic data are currently the most
exhibits different network properties and requires different readily available high-throughput data for this type of anal-
strategies of reconstruction and modeling.206 Each network ysis, due to the well-developed microarray and RNA-seq
usually contains thousands of components and a myriad of technologies.213,214 Recon 1 can contextualize transcrip-
interactions among them. A complete reconstruction of cel- tomic data within the human metabolic network model.204
lular network requires integrations of these different types The application of the human metabolic network model in
of networks, and must take cellular compartmentalization contextualization of transcriptomic data was demonstrated
into account. Cell-type specificity also calls for the recon- along with the release of Recon 1. In this demonstration,
struction of cell-specific models. Furthermore, models at gene expression data derived from skeletal muscle isolated
the level of tissues have to consider the integration of mul- from morbidly obese patients before and after gastric-bypass
tiple cell-specific networks and the communication among surgery were reanalyzed and mapped onto Recon 1 in order
them.193,206 to identify metabolic changes following gastric-bypass
In spite of these challenges, considerable progress has surgery. Signature patterns in anaerobic metabolism,
been made in network reconstructions. The release of the including downregulated oxidative phosphorylation and
global reconstruction of the human metabolic network, mitochondrial bioenergetics, were observed following sur-
named “Recon 1,” was a milestone event.204 This reconstruc- gery.204 Others have used Recon 1 to investigate the effect
tion was built based on the human genome sequence and of dietary interventions on transcriptome profiles at differ-
the accumulated knowledge of human metabolism going ent stages of the intervention. Studies of human subjects
back more than 50 years, encompassing 1,496 genes, 2,004 challenged with an energy-restriction phase followed by a
proteins, 2,712 metabolites, and 3,311 metabolic reactions. weight-stabilization phase209 revealed interactions among
Recon 1 is mass- and charge-balanced and accounts for metabolic and inflammatory pathways in adipose tissue
seven intracellular compartments (cytoplasm, mitochon- and their impact on insulin sensitivity. Gene expression
dria, nucleus, endoplasmic reticulum, Golgi apparatus, patterns for both adipocytes and macrophages were char-
lysosome, and peroxisome).193,204 Although Recon 1 is still acterized within an unbiased context provided by Recon
being perfected and there are ongoing efforts to continu- 1. Interestingly, adipocyte genes involved in metabolism
ously update and fill gaps and missing information, it has and macrophage genes participating in immune pathways
enabled development and applications in multiple areas. exhibited an opposite pattern of responses to the dietary
Firstly, the current reconstruction is fueling a wave of intervention. Adipocytes metabolic genes were downreg-
hypothesis-driven studies that will advance our understand- ulated during energy restriction and upregulated during
ing of human metabolism.207,208 Also, Recon 1 serves as the weight stabilization, whereas macrophage immune-related
foundation for generating cell-specific, tissue-specific, and genes were not changed or upregulated during energy
condition-specific models, and as the scaffold for contex- restriction and down-regulated in weight-stabilization
tualizing high-throughput data to unravel the mechanistic phase.209 Comprehensive reconstructions of the whole-body
processes of disease and drug-treated states. Another attrac- metabolic network are required to simultaneously analyze
tive application of Recon 1 is mathematical modeling and high-throughput data from different cell types and body
computational simulation to identify biomarkers and to fluids.215,216 For example, to determine differential meta-
develop disease-management strategies.193,204 bolic activity between obese and diabetic obese individuals,
a multi-tissue type genome-wide metabolic network was
built by integrating three cell-specific networks (hepato-
N ET WO R K-A S S I S T E D SYS T E M I C
cyte, myocyte, and adipocyte) representing three tissues
I N T E R RO G AT I O N O F T H E MO L ECU L A R
(liver, skeletal muscle, and adipose tissue, respectively) and
BA S I S O F M ETA B O L I C D I S E A S E S
one blood compartment connecting the three cell types.216
A N D N U T R IT I O NA L I N T E RVE N T I O N
Various multicellular models of brain energy metabolism
Network-assisted systems approaches are fundamental to have been reconstructed to study Alzheimer’s disease.215
elucidating the molecular basis of diseases and develop- Facilitated by the rapid technological advances in mass
ing nutritional and/or pharmaceutical interventions using spectrometry (MS) and nuclear magnetic resonance (NMR)
high-throughput data,193 including transcriptomic,204,209,210 spectrometry, and high-resolution separation technologies,
proteomic, and metabolomic data.211,212 The mapping of such as high performance liquid chromatography (HPLC)
genes, transcripts, and proteins onto the metabolic network is and gas chromatography (GC),217 network-assisted
conducted according to the gene-transcript-protein-reaction approaches are starting to be applied in a systemic analysis
N utrition a l G e no m ic s • 1 9 7
of metabolomic data.211,212 Pathways and affected enzymes derivatives alone can provide remarkable insights into the
responsible for Leigh’s syndrome (LS) were identified using mechanistic bases of metabolic disease and health manage-
fibroblasts obtained from normal and LS subjects grown in ment, and generate a large number of hypotheses await-
media with 13C-labeled glucose. Time-course metabolomic ing experimental verifications.220–223 One study utilized
data revealed that fibroblasts from LS patients exhibited Recon 1–based computational simulation and considered
slower metabolism and less adenosine triphosphate (ATP) additional constraints, from enzyme solvent capacity, to
production. The model predicted mutations in succinate investigate the causes of the Warburg effect, which is the
cytochrome c reductase as the underlying cause of LS.212 preferred use of glycolysis over respiration even in the pres-
Unbiased metabolic network modeling approaches can ence of oxygen in cancer cells.223 The model predicts that
also identify gaps in our understanding of the contributions the Warburg effect is the direct consequence of a metabolic
of genes to metabolic disease. A reanalysis of the meta- adaptation of cancer cells to fast proliferation. In addition,
bolic responses to an oral glucose tolerance test (OGTT) the computational modeling also captures several experi-
was conducted in 25 normal subjects and 25 subjects with mentally observed phenotypes during cancer development,
impaired glucose tolerance.211 The initial analysis of the including the preferential uptake of glutamine over other
data that considered only the established insulin-related amino acids.223 There is little doubt that network-based
pathways (glycolysis, lipolysis, ketogenesis, and proteolysis) mathematical modeling and computational simulation will
revealed that responses in all these pathways are blunted in greatly accelerate hypothesis-driven studies and enhance
subjects with impaired glucose tolerance. Further investi- our understanding of human diseases and health.
gation identified 18 plasma metabolites that were respon- The reconstruction of the human metabolic network
sive to glucose ingestion in normal individuals only. These also helps us make large-scale identifications of various
metabolites could not be mapped to established pathways types of interacting relationships among different network
and were not previously linked to glucose homeostasis.218 components. Based on these relationships, new candidate
Reanalysis of these metabolic profiles, assisted with the disease genes or drug targets that closely interact with
human Recon 1, revealed the unexpected involvement of well-established targets can be identified. For example,
solute carriers in the metabolic pathways of OGTT.211 Such “correlated sets of reactions” (Co-Sets) under a specific
studies give us confidence that metabolomic profiling com- condition can be identified from the condition-specific
bined with network-assisted data analysis will significantly reconstruction of the human metabolic network as reac-
enhance our further understanding of the molecular basis tions whose fluxes are perfectly correlated, such that the
of diseases and assist in the development of effective nutri- flux through one reaction indicates equivalent flux through
tional interventions. correlated reactions.204,224,225 Reactions in the same sets can
There are meaningful limitations to the use of be either continuous in a linear pathway or be present in dif-
Recon 1 and its derivatives as scaffolds to contextualize ferent pathways. Although the continuous cases are intui-
high-throughput data. Although these approaches consider tive in understanding, the non-linear cases are less apparent
the interconnectedness of genes, enzymes, and metabolites and require additional analysis based on comprehensive
in the network, the scaffold design of these network recon- networks.224 For example, correlated sets of reactions were
structions ignores reaction stoichiometry, network topol- identified in cells conducting aerobic glucose metabolism.
ogy, the conservation of mass, and the balance of charge.204 One of the largest sets of correlated reactions is related to
Accounting for this unexploited information and consid- cholesterol biosynthesis, and specifically to the reaction
ering these factors as constraints within a mathematically catalyzed by 3-hydroxy-3-methylglutaryl-CoA reductase, a
representative system of reactions enables a higher-level primary metabolic target of the anti-lipidemic drug statin.
use of the reconstructions, referred to as “constraint-based It is therefore anticipated that other enzymes in the same set
computational modeling,” which provides an intriguing will be drug targets for treating the same disorders.224
application for the reconstruction of the human metabolic
network.193 Constraint-based modeling and simulation
S YS T E M I C I D E N T I FI C AT I O N O F
allows for the prediction of the metabolic effects of genetic
M ETA B O L I C PAT T E R N S O R M ETA B O L IT E S
variants, enzyme deficiencies, and environmental expo-
A S B I O M A R K E R S
sures.219–222 The successful application of this simulation
approach for biomarker identification will be further dis- Biomarkers are biological characteristics that are objectively
cussed in the next section. Computational simulation of measured and evaluated as indicators of normal biologi-
the human metabolic network reconstructions or their cal processes, pathological processes, or pharmacological
1 9 8 • P rincip l e s o f G e no m ic M e dicin e
responses to a therapeutic intervention.226 Biomarkers have systematically validated by comparing them with experi-
been widely used in clinical practices to predict, diagnose, mentally confirmed biomarkers for each disease recorded
and classify diseases, or to monitor outcomes of disease in the OMIM database. The biomarker predictions from
management or nutritional intervention.227,228 For instance, this method were 10 times higher than random chance.219
elevated blood-glucose levels are used to diagnose diabe- This study also illuminated the current limitations to
tes, and increased plasma level of low-density lipoprotein network-based computational simulation in biomarker pre-
(LDL) cholesterol is indicative of higher risk for cardio- diction. These limitations lie in both the network model
vascular diseases.227,229 As indicated in the definition,226 a and the simulation methods. The network model used
biomarker is not necessarily a single molecule (metabolite, in the study (Recon 1) is still incomplete and is not inte-
protein, etc.). Rather, it could be a combination of mul- grated with the gene-expression regulatory network. The
tiple molecules or their related features.228,230 The currently prediction of extracellular biomarkers will benefit from
available omics data, especially metabolomic data, provide a whole-body network model that incorporates different
opportunities for identifying metabolic patterns or metab- cells and tissues.219 Moreover, simulation techniques with
olites as biomarkers.227,228 The analysis of large-scale omics human metabolic networks have to be further improved.207
data for biomarker identification also benefits from the use
of network models. For example gene-expression data were
recently used to identify novel biomarker candidates for N U T R I T I O N AL G E N O M I C S
type 2 diabetes.210 Transcriptomic data from skeletal muscle A P P L I C AT I O N A N D P U B L I C H EALT H
of individuals with different phenotypes (insulin resistance,
type 2 diabetes) were compared to reveal differentially The use of genetic information in current nutrition practice
expressed genes, which were further mapped onto the net- and policy is very limited, but such consideration has the
work models. Reporter metabolites were identified based potential to “personalize” nutrition and thereby prevent
on their association with enzyme-coding genes that were chronic disease and promote healthy aging through diet
enriched in the transcriptional response to diabetes. Some by targeting the molecular antecedents of disease. Cellular
of these reporter metabolites were from pathways known networks are sensitive to internal perturbations that result
to be associated with type 2 diabetes, including tricarbox- from genetic variation and gene mutations; identification
ylic acid cycle (TCA cycle), oxidative phosphorylation, and of the gene variant and/or metabolites that induce network
lipid metabolism, whereas others were novel discoveries, dysfunction can be informative in the diagnoses or predic-
such as NADH and ATP, representing candidate biomark- tions of various health and disease outcomes. For example,
ers awaiting thorough experimental verification.210 as discussed in the section on nutritional systems biology,
In addition to network-assisted mining of omics data, classic monogenic disorders, including the in-born errors of
computational simulation based on reconstructions of metabolism phenylketoneuria and galactosemia, illustrate
human metabolic networks also enables the prediction the severe consequences that can result from catastrophic
of candidate biomarkers. This approach is especially effi- metabolic network disruptions. Perhaps more importantly,
cient for monogenic hereditary diseases, such as IEMs, these early clinical studies also demonstrated that single-gene
because of their relatively simple etiologies. The power of disorders can be managed and/or alleviated through dietary
network-assisted biomarker identifications is exemplified interventions (e.g., the use of phenylalanine-restricted diets
by the pioneering effort to systemically predict biomark- to prevent or mitigate the cognitive impairments resulting
ers for more than 300 metabolic disorders.223 This novel from mutations in the phenylalanine hydroxylase gene) and
computation approach applied a constraint-based mod- thereby established the principles of nutritional interven-
eling method206 to predict the level of metabolites under tion. Nutrients, like pharmaceuticals, are powerful modi-
both normal and diseased states. For each metabolic disor- fiers of genome and network function and stability, and
der, the morbid state was simulated by deleting the known gene–nutrient interactions can be optimized for disease
causative gene. For 304 metabolic disorders documented in prevention and management (Figure 12.1).
the Online Mendelian Inheritance in Man (OMIM) data-
base231 whose causative genes are also present in the Recon
P E R S O NA L I Z E D N U T R I T I O N
1,204 the computational approach made 3,912 predictions
involving 233 metabolites whose concentrations change in Salient examples of nutritional intervention are pro-
at least one disease state, and 176 diseases that are associ- vided by past experiences that demonstrate that mater-
ated with at least one candidate biomarker. Predictions were nal and perinatal nutrition can improve birth outcomes,
N utrition a l G e no m ic s • 1 9 9
including cognitive development (e.g., ensuring iodine establishment of policies that result in pharmacological
sufficiency), lifelong chronic disease resistance (e.g., pre- intakes of nutrients and other dietary components.
venting small-for-gestational-age births), and increased
longevity (e.g., optimizing dietary fat consumption and
A S S I S T E D R E P RO D U C T I O N
immune function).232 It also is apparent that inappropri-
A N D O P T I M A L C U LT U R E M E D I A
ate uses of nutrition to maximize reproductive outcomes
will present new potential risks.71,157 High-dose vitamin Elevated SA rates have been observed in some, but not all,
therapy has been advocated to rescue impaired metabolic studies of human in vitro fertilization (IVF) pregnancies
reactions that result from mutations and polymorphisms compared to natural conceptions. These findings may be
that decrease the affinity of substrates and cofactors for the result of early harvest and early manipulations of eggs
the encoded enzyme.48 ω-3 fatty acids and tocopherols may and embryos in culture media.139,236 Furthermore, other
promote healthy aging and longevity by modulating the studies have found that human IVF procedures result in
inflammatory response by altering gene transcription.232 higher-than-expected incidences of IUGR.237 Numerous
Genes and their allelic variants that influence longevity (a studies have shown that the composition of the embryo
trait that is nonadaptive) are being identified at accelerated culture medium affects the expression and methylation
rates,233 and their penetrance can be modified by the ratio- status of imprinted genes, including H19, Igf2 and Igf2r
nal design of nutrition-based interventions and therapies.232 in ovine and other mammalian embryos, resulting in large
Repression of energy metabolism, through caloric restric- offspring syndrome.237–241 There appear to be many critical
tion or transcriptional regulation of metabolic enzymes, windows associated with the establishment of environmen-
reduces oxidative stress and promotes longevity in many tally sensitive methylation patterns from early embryogen-
experimental model systems. Manipulation of these tran- esis through the suckling period, all of which are cell-type
scriptional and/or metabolic networks by designer vitamin and/or allele specific.184,185 Some of these networks are
supplements may promote healthy aging. However, caution sensitive to folate-mediated one-carbon metabolism, as
is warranted. Genes encoding virtually all physiological illustrated by the impact of maternal nutrition on genomic
process are not adapted to excessive nutrient intake expo- methylation in the viable yellow agouti (Avy) mouse131;
sures that exceed what has been achieved in historical and other networks may respond to the allelic- or locus-specific
healthy food-based diets. Therefore, new risks and toxicities targeting of methylase/demethylase activity, as seen with
should be anticipated in human populations or population glucocorticoid-receptor programming, described earlier.176
subgroups when nutrients are administered at pharma- The increasing evidence that early nutritional exposures can
cological levels, as illustrated by the introduction of high increase the risk for late-onset metabolic diseases through
levels of fructose into the food supply.67 Some of the unin- epigenetic mechanisms illuminates the major challenges
tended consequences may involve genome programming, that are made more immediate by the increased demand for
including permanent alterations in genome-wide methyla- assisted reproduction.
tion patterns in stem cell populations, as has been observed
in mouse embryos whose mothers received elevated doses
D I ETA RY R EC O M M E N DAT I O N S
of folic acid and one-carbon donors during gestation.21
F O R P O P U L AT I O N S
Methylation patterns that are established in utero, and per-
haps in adult stem cell populations, can be metastable and Food-based and nutrient-based dietary guidelines were
influence gene expression and, potentially, mutation rates established to help individuals and populations achieve
throughout the lifespan.21 Furthermore, although antioxi- adequate dietary patterns to maintain health. The deriva-
dants can decrease mutation rates, they can also function tion and goals of these guidelines evolve as new knowledge
as pro-oxidants in vivo234 and may be cancer-promoting becomes available.242,243 Guidelines for single nutrients and
when consumed at elevated intakes by inhibiting cellular other food components are scientifically and quantitatively
death programs in transformed cells.235 In conclusion, elu- derived, and are usually based on the level of nutrient intake
cidation of robust gene-by-nutrient interactions will inform that prevents a clinical and/or biochemical outcome that is
dietary approaches for individuals and for populations that associated with a particular nutrient deficiency. Numerical
aim to prevent and/or manage complex metabolic disease, standards for nutrients are essential to validate the effi-
as has been accomplished for rare inborn errors of metabo- cacy of food-based guidelines.243 Nutrient requirements
lism. Equally important, these and other examples indicate vary among individuals within all human populations,
that rigorous hazard-identification is essential prior to the and can be modified by age, gender, and life stage, among
2 0 0 • P rincip l e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
other variables. Therefore, recommended nutrient intakes will be selected against in large part through fetal loss and
are often derived separately for population subgroups. the failure to survive to reproduce. Allelic variants that con-
Although genetic variation can modify the efficacy, dos- fer atypical nutrient requirements that cannot be met by the
age, and safety of pharmaceutical agents244 and tolerance or mother or that result in severe metabolic disruptions are
intolerance for certain foods,62 the contribution of genetics expected to be embryonically lethal. Alleles that confer sub-
to nutrient requirements within and among human popula- tler differences in nutrient requirements or food tolerances
tions remains to be evaluated. However, the characterization are expected to be enriched in subgroups or populations,
of gene variants that modify optimal nutrient requirements and contribute to disease in certain environmental contexts
will enable the classification of genetic subgroups for whom (Table 12.1 and Table 12.2).
generalized nutritional requirements may be valid. The concept of generalized nutrient requirements
within and among human populations is nullified only when
a level of nutrient intake that represents minimal nutrient
R EC O M M E N D E D DA I LY A L L OWA N C E
adequacy for one genetic subgroup exceeds a safe intake
A N D U P P E R L I M ITS
level for another group, assuming that nutrient deficiency
The “recommended daily allowance” (RDA) for each nutri- and harm/toxicity-avoidance are the primary criteria for
ent is defined as the level of dietary intake that is sufficient that requirement. For example, it is widely recognized that
to meet the requirements of 97% of healthy individuals optimal folate intakes may differ among identifiable genetic
in a particular life stage and gender group (Figure 12.5). subgroups.245 However, it is not at all certain that the mag-
When there are insufficient data to calculate an RDA for a nitude of the genetic contribution to variations in adequate
nutrient, an “adequate intake” (AI), which is an estimated dietary folate intake warrants genotype-specific recommen-
recommended intake value, is established. Some nutrients dations, especially considering that folic acid intakes up to
demonstrate toxicities at elevated intake levels; therefore, 1 mg/day are not associated with known toxicities.124 Iron
a “tolerable upper intake level” (UL), which represents the is another candidate nutrient for genotype-specific nutrient
highest level of nutrient intake that can be achieved without requirement recommendations.79,246–248 For these and other
incurring risk for adverse health effects for the vast majority cases, the penetrance (contribution of the individual allele
of individuals in the general population, is established. to variation in nutrient requirements) and the prevalence
Human genetic variation is not anticipated to confer of these functional gene variants must be elucidated both
extreme variations in optimal nutrient requirements among within and among human populations to validate the con-
individuals and populations. Nutrition, unlike pharmaceu- cept of generalized nutrient requirements for all human
ticals, is an in utero and lifelong exposure that can serve as populations. Unlike the effects of gender and life cycle, no
a selective pressure to eliminate genomes that are not com- common allelic variant has been shown to be sufficiently
patible with the nutrient environment. Therefore, human penetrant and to warrant genotype-specific numerical stan-
genotypes that do not support basic physiological processes dards for nutrient adequacy or upper levels of intake associ-
ated with harm or toxicity. At this time, genetic variation
is known only to influence nutrient and food intolerance
EAR (Table 12.1 and Table 12.2). However, genetic variation has
AI UL not been characterized in many geographically and cultur-
Risk of inadequacy
Risk of excess
N utrition a l G e no m ic s • 2 0 1
www.Ebook777.com
result of differential processing of the LCT transcript.250 disease. For example, there is evidence that docosahexaenoic
Pyridoxine-5′-beta-D-glucoside activity is necessary for the acid (DHA)256 or choline257 supplementation during prena-
bioavailability of pyridoxine-5′-beta-D-glucoside, the major tal and/or early postnatal life improves CNS function and
form of vitamin B6 in plant derived foods.251 Therefore, cognitive performance throughout life. Continued progress
LCT variation predicts both lactose tolerance as well as pre- in these areas requires the identification and understanding
ferred dietary sources of vitamin B6 in adulthood. of nutritionally modifiable networks that enable successful
adaptations in successive life stages, their associated critical
windows, and validation of the efficacy, effectiveness, and
G E N O M I C C R IT E R I A F O R S ET T I N G
safety of the nutritional interventions.
R EQ U I R E M E N TS A N D TOX I C IT I E S
2 0 2 • P rincip l e s o f G e no m ic M e dicin e
6. Venter JC, Adams MD, Myers EW, et al. The sequence of the human 33. Fenech M. Genome health nutrigenomics: nutrition and the
genome. Science. 16 Feb 2001;291(5507):1304–1351. science of optimal genome maintenance. Asia Pac J Clin Nutr.
7. Tishkoff SA, Kidd KK. Implications of biogeography of human 2004;13(Suppl):S15.
populations for “race” and medicine. Nat Genet. Nov 2004;36(11 34. Fenech M. Chromosomal biomarkers of genomic instability rele-
Suppl):S21–S27. vant to cancer. Drug Discov Today. 15 Nov 2002;7(22):1128–1137.
8. Venter JC. Multiple personal genomes await. Nature. 1 Apr 35. Fenech M. Recommended dietary allowances (RDAs) for genomic
2010;464(7289):676–677. stability. Mutat Res. 1 Sep 2001;480:51–54.
9. Zhang F, Gu W, Hurles ME, Lupski JR. Copy number variation in 36. Ames BN. DNA damage from micronutrient deficiencies is likely to
human health, disease, and evolution. Annu Rev Genomics Hum be a major cause of cancer. Mutat Res. 18 Apr 2001;475(1–2):7–20.
Genet. 2009;10:451–481. 37. Goswami T, Rolfs A, Hediger MA. Iron transport: emerging roles in
10. Redon R, Ishikawa S, Fitch KR, et al. Global variation in copy health and disease. Biochem Cell Biol. 2002;80(5):679–689.
number in the human genome. Nature. 23 Nov 2006;444(7118): 38. Hediger MA, Rolfs A, Goswami T. Iron transport and hemochro-
444–454. matosis. J Investig Med. Sep 2002;50(5):239S–246S.
11. Sherry ST, Ward MH, Kholodov M, et al. dbSNP: the NCBI 39. Mishra HN, Das C. A review on biological control and metabolism
database of genetic variation. Nucleic Acids Res. 1 Jan 2001;29(1): of aflatoxin. Crit Rev Food Sci Nutr. 2003;43(3):245–264.
308–311. 40. Roach JC, Glusman G, Smit AF, et al. Analysis of genetic inheri-
12. Nachman MW, Crowell SL. Estimate of the mutation rate per tance in a family quartet by whole-genome sequencing. Science. 30
nucleotide in humans. Genetics. Sep 2000;156(1):297–304. Apr 2010;328(5978):636–639.
13. Frazer KA, Murray SS, Schork NJ, Topol EJ. Human genetic varia- 41. Clark AG, Glanowski S, Nielsen R, et al. Inferring nonneutral evo-
tion and its contribution to complex traits. Nat Rev Genet. Apr lution from human-chimp-mouse orthologous gene trios. Science.
2009;10(4):241–251. 12 Dec 2003;302(5652):1960–1963.
14. Tishkoff SA, Verrelli BC. Patterns of human genetic diver-
42. Wolfe KH, Li WH. Molecular evolution meets the genomics revo-
sity: implications for human evolutionary history and disease. Annu lution. Nat Genet. Mar 2003;33 Suppl:255–265.
Rev Genomics Hum Genet. 2003;4:293–340. 43. Kimura M. Evolutionary rate at the molecular level. Nature. 17 Feb
15. Van Blerkom LM. Role of viruses in human evolution. Am J Phys 1968;217(129):624–626.
Anthropol. 2003;Suppl 37:14–46. 44. King JL, Jukes TH. Non-Darwinian evolution. Science. 16 May
16. Brookfield JF. Mobile DNAs: the poacher turned gamekeeper. Curr 1969;164(881):788–798.
Biol. 28 Oct 2003;13(21):R846–R847. 45. Akey JM, Zhang G, Zhang K, Jin L, Shriver MD. Interrogating a
17. Batzer MA, Deininger PL. Alu repeats and human genomic diver- high-density SNP map for signatures of natural selection. Genome
sity. Nat Rev Genet. May 2002;3(5):370–379. Res. Dec 2002;12(12):1805–1814.
19. Bamshad M, Wooding SP. Signatures of natural selection in the 46. Gabriel SB, Schaffner SF, Nguyen H, et al. The structure
human genome. Nat Rev Genet. Feb 2003;4(2):99–111. of haplotype blocks in the human genome. Science. 21 Jun
20. Gal-Mark N, Schwartz S, Ast G. Alternative splicing of Alu
2002;296(5576):2225–2229.
exons—two arms are better than one. Nucleic Acids Res. Apr 47. Crawford DC, Nickerson DA. Definition and clinical importance
2008;36(6):2012–2023. of haplotypes. Annu Rev Med. 2005;56:303–320.
21. Waterland RA, Jirtle RL. Transposable elements: targets for early 48. Ames BN, Elson-Schwab I, Silver EA. High-dose vitamin therapy
nutritional effects on epigenetic gene regulation. Mol Cell Biol. Aug stimulates variant enzymes with decreased coenzyme binding affin-
2003;23(15):5293–5300. ity (increased K(m)): relevance to genetic disease and polymor-
22. Deininger PL, Batzer MA. Alu repeats and human disease. Mol phisms. Am J Clin Nutr. Apr 2002;75(4):616–658.
Genet Metab. Jul 1999;67(3):183–193. 49. Bracken MB. Genomic epidemiology of complex disease: the need
23. Conrad DF, Pinto D, Redon R, et al. Origins and functional impact for an electronic evidence-based approach to research synthesis. Am
of copy number variation in the human genome. Nature Apr 2010; J Epidemiol. 15 Aug 2005;162(4):297–301.
1;464(7289):704–712. 50. Wooding S. Natural selection: sign, sign, everywhere a sign. Curr
24. Freeman JL, Perry GH, Feuk L, et al. Copy number variation: new Biol. 7 Sep 2004;14(17):R700–R701.
insights in genome diversity. Genome Res. 2006;16(8):949–961. 51. Akey JM, Eberle MA, Rieder MJ, et al. Population history and natu-
25. Zhang F, Gu W, Hurles ME, Lupski JR. Copy number variation in ral selection shape patterns of genetic variation in 132 genes. PLoS
human health, disease, and evolution. Annu Rev Genomics Hum Biol. Oct 2004;2(10):e286.
Genet 2009;10:451–481. 52. Chakravarti A. To a future of genetic medicine. Nature. 15 Feb
26. Eichler EE, Nickerson DA, Altshuler D, et al. Completing the map 2001;409(6822):822–823.
of human genetic variation. Nature. May 10, 2007;447(7141): 53. Reich DE, Lander ES. On the allelic spectrum of human disease.
161–165. Trends Genet. Sep 2001;17(9):502–510.
27. Xue Y, Sun D, Daly A, et al. Adaptive evolution of UGT2B17 54. Pritchard JK. Are rare variants responsible for susceptibility to com-
copy-number variation. Am J Hum Genet. 2008;83(3):337–346. plex diseases? Am J Hum Genet. Jul 2001;69(1):124–137.
28. Tuzun E, Sharp AJ, Bailey JA, et al. Fine-scale structural variation of 55. Pritchard JK, Cox NJ. The allelic architecture of human disease
the human genome. Nat Genet. 2005;37(7):727–732. genes: common disease-common variant . . . or not? Hum Mol Genet.
29. Sebat J, Lakshmi B, Troge J, et al. Large-scale copy number polymor- 1 Oct 2002;11(20):2417–2423.
phism in the human genome. Science 2004;305(5683):525–528. 56. Jorde LB, Watkins WS, Bamshad MJ. Population genomics: a bridge
30. Perry GH, Dominy NJ, Claw KG, et al. Diet and the evolution from evolutionary history to genetic medicine. Hum Mol Genet.
of human amylase gene copy number variation. Nat Genet. Oct 1 Oct 2001;10(20):2199–2207.
2007;39(10):1256–1260. 57. Jorde LB, Watkins WS, Bamshad MJ, et al. The distribution
31. Blount BC, Mack MM, Wehr CM, et al. Folate deficiency causes of human genetic diversity: a comparison of mitochondrial,
uracil misincorporation into human DNA and chromosome break- autosomal, and Y-chromosome data. Am J Hum Genet. Mar
age: implications for cancer and neuronal damage. Proc Natl Acad 2000;66(3):979–988.
Sci U S A. 1 Apr 1997;94(7):3290–3295. 58. Jorde LB, Wooding SP. Genetic variation, classification and “race.”
32. Fenech M. Micronutrients and genomic stability: a new paradigm Nat Genet. Nov 2004;36(11 Suppl):S28–S33.
for recommended dietary allowances (RDAs). Food Chem Toxicol. 59. Charlesworth D. Balancing selection and its effects on sequences in
Aug 2002;40(8):1113–1117. nearby genome regions. PLoS Genet. Apr 2006;2(4):e64.
N utrition a l G e no m ic s • 2 0 3
60. Collins FS. What we do and don’t know about “race,” “ethnicity,” 81. Osier MV, Pakstis AJ, Soodyall H, et al. A global perspective
genetics and health at the dawn of the genome era. Nat Genet. Nov on genetic variation at the ADH genes reveals unusual patterns
2004;36(11 Suppl):S13–S15. of linkage disequilibrium and diversity. Am J Hum Genet. Jul
61. Bufe B, Hofmann T, Krautwurst D, Raguse JD, Meyerhof W. The 2002;71(1):84–99.
human TAS2R16 receptor mediates bitter taste in response to 82. Loew M, Boeing H, Sturmer T, Brenner H. Relation among alcohol
beta-glucopyranosides. Nat Genet. Nov 2002;32(3):397–401. dehydrogenase 2 polymorphism, alcohol consumption, and levels
62. Enattah NS, Sahi T, Savilahti E, Terwilliger JD, Peltonen L, Jarvela of gamma-glutamyltransferase. Alcohol. Apr 2003;29(3):131–135.
I. Identification of a variant associated with adult-type hypolactasia. 83. Oota H, Pakstis AJ, Bonne-Tamir B, et al. The evolution and
Nat Genet. Feb 2002;30(2):233–237. population genetics of the ALDH2 locus: random genetic drift,
63. Bersaglieri T, Sabeti PC, Patterson N, et al. Genetic signatures of selection, and low levels of recombination. Ann Hum Genet. Mar
strong recent positive selection at the lactase gene. Am J Hum Genet. 2004;68(Pt 2):93–109.
Jun 2004;74(6):1111–1120. 84. Peng Y, Shi H, Qi XB, et al. The ADH1B Arg47His polymorphism
64. Poulter M, Hollox E, Harvey CB, et al. The causal element for the in East Asian populations and expansion of rice domestication in
lactase persistence/non-persistence polymorphism is located in a 1 history. BMC Evol Biol. 2010;10:15.
Mb region of linkage disequilibrium in Europeans. Ann Hum Genet. 85. Danpure CJ. Molecular etiology of primary hyperoxaluria
Jul 2003;67(Pt 4):298–311. type 1: new directions for treatment. Am J Nephrol. May-Jun
65. Wooding S. PopHist: inferring population history from the spec- 2005;25(3):303–310.
trum of allele frequencies. Bioinformatics. 1 Mar 2003;19(4): 86. Williams RJ, Deason G. Individuality in vitamin C needs. Proc
539–540. Natl Acad Sci U S A. Jun 1967;57(6):1638–1641.
66. Esposito G, Vitagliano L, Santamaria R, Viola A, Zagari A,
87. Eck P, Erichsen HC, Taylor JG, et al. Comparison of the genomic
Salvatore F. Structural and functional analysis of aldolase B mutants structure and variation in the two human sodium-dependent vita-
related to hereditary fructose intolerance. FEBS Lett. 6 Nov min C transporters, SLC23A1 and SLC23A2. Hum Genet. Sep
2002;531(2):152–156. 2004;115(4):285–294.
67. Cox TM. The genetic consequences of our sweet tooth. Nat Rev 88. Zanon-Moreno V, Ciancotti-Olivares L, Asencio J, et al.
Genet. Jun 2002;3(6):481–487. Association between a SLC23A2 gene variation, plasma vitamin
68. Fullerton SM, Clark AG, Weiss KM, et al. Apolipoprotein E varia- C levels, and risk of glaucoma in a Mediterranean population. Mol
tion at the sequence haplotype level: implications for the origin and Vis. 2011;17:2997–3004.
maintenance of a major human polymorphism. Am J Hum Genet. 89. Neel JV. Diabetes mellitus: a “thrifty” genotype rendered detri-
Oct 2000;67(4):881–900. mental by “progress?” Am J Hum Genet. Dec 1962;14:353–362.
69. Ordovas JM, Corella D. Nutritional genomics. Annu Rev Genomics 90. Diamond J. The double puzzle of diabetes. Nature. 5 Jun
Hum Genet. 2004;5:71–118. 2003;423(6940):599–602.
70. Ordovas JM. The quest for cardiovascular health in the genomic 91. Rockman MV, Wray GA. Abundant raw material for
era: nutrigenetics and plasma lipoproteins. Proc Nutr Soc. Feb cis-regulatory evolution in humans. Mol Biol Evol. Nov 2002;19
2004;63(1):145–152. (11):1991–2004.
71. Stover PJ. Nutritional genomics. Physiol Genomics. 15 Jan
92. Verrelli BC, McDonald JH, Argyropoulos G, et al. Evidence for
2004;16(2):161–165. balancing selection from nucleotide sequence analyses of human
72. Stover PJ. Physiology of folate and vitamin B12 in health and dis- G6PD. Am J Hum Genet. Nov 2002;71(5):1112–1128.
ease. Nutr Rev. Jun 2004;62(6 Pt 2):S3–S12; discussion S13. 93. Watkins WS, Rogers AR, Ostler CT, et al. Genetic variation
73. Brody LC, Conley M, Cox C, et al. A polymorphism, R653Q, in among world populations: inferences from 100 Alu insertion poly-
the trifunctional enzyme methylenetetrahydrofolate dehydroge- morphisms. Genome Res. Jul 2003;13(7):1607–1618.
nase/methenyltetrahydrofolate cyclohydrolase/formyltetrahydro- 94. Tishkoff SA, Varkonyi R, Cahinhinan N, et al. Haplotype
folate synthetase is a maternal genetic risk factor for neural tube diversity and linkage disequilibrium at human G6PD: recent
defects: Report of the Birth Defects Research Group. Am J Hum origin of alleles that confer malarial resistance. Science. 20 Jul
Genet. Nov 2002;71(5):1207–1215. 2001;293(5529):455–462.
74. Ma J, Stampfer MJ, Giovannucci E, et al. Methylenetetrahydrofolate 95. Baier LJ, Permana PA, Yang X, et al. A calpain-10 gene polymor-
reductase polymorphism, dietary interactions, and risk of colorectal phism is associated with reduced muscle mRNA levels and insulin
cancer. Cancer Res. 15 Mar 1997;57(6):1098–1102. resistance. J Clin Invest. Oct 2000;106(7):R69–R73.
75. Guenther BD, Sheppard CA, Tran P, Rozen R, Matthews RG, 96. Inoue I, Nakajima T, Williams CS, et al. A nucleotide substitu-
Ludwig ML. The structure and properties of methylenetetra- tion in the promoter of human angiotensinogen is associated with
hydrofolate reductase from Escherichia coli suggest how folate essential hypertension and affects basal transcription in vitro. J
ameliorates human hyperhomocysteinemia. Nat Struct Biol. Apr Clin Invest. 1 Apr 1997;99(7):1786–1797.
1999;6(4):359–365. 97. Wray GA, Hahn MW, Abouheif E, et al. The evolution of transcrip-
76. Bailey LB. Folate, methyl-related nutrients, alcohol, and the
tional regulation in eukaryotes. Mol Biol Evol. Sep 2003;20(9):
MTHFR 677C-->T polymorphism affect cancer risk: intake rec- 1377–1419.
ommendations. J Nutr. Nov 2003;133(11 Suppl 1):3748S-3753S. 98. Wright AF, Carothers AD, Pirastu M. Population choice in map-
77. Toomajian C, Ajioka RS, Jorde LB, Kushner JP, Kreitman M. A ping genes for complex diseases. Nat Genet. Dec 1999;23(4):
method for detecting recent selection in the human genome from 397–404.
allele age estimates. Genetics. Sep 2003;165(1):287–297. 99. Wooding S, Kim UK, Bamshad MJ, Larsen J, Jorde LB, Drayna D.
78. Toomajian C, Kreitman M. Sequence variation and haplotype Natural selection and molecular evolution in PTC, a bitter-taste
structure at the human HFE locus. Genetics. Aug 2002;161(4): receptor gene. Am J Hum Genet. Apr 2004;74(4):637–646.
1609–1623. 100. Dennis C. Epigenetics and disease: altered states. Nature. 13 Feb
79. Beutler E. Iron absorption in carriers of the C282Y hemochromato- 2003;421(6924):686–688.
sis mutation. Am J Clin Nutr. Oct 2004;80(4):799–800. 101. Meda F, Folci M, Baccarelli A, Selmi C. The epigenetics of autoim-
80. Bosron WF, Li TK. Genetic polymorphism of human liver alco- munity. Cell Mol Immunol. May 2011;8(3):226–236.
hol and aldehyde dehydrogenases, and their relationship to alco- 102. Barker DJ. Intrauterine programming of coronary heart disease
hol metabolism and alcoholism. Hepatology. May-Jun 1986;6(3): and stroke. Acta Paediatr Suppl. Nov 1997;423:178–182; discus-
502–510. sion 183.
2 0 4 • P rincip l e s o f G e no m ic M e dicin e
103. Wu G, Imhoff-Kunsch B, Girard AW. Biological mechanisms for one-carbon incorporation into human DNA deoxynucleosides. J
nutritional regulation of maternal health and fetal development. Nutr. Mar 2005;135(3):389–396.
Paediatr Perinat Epidemiol. Jul 2012;26 Suppl 1:4–26. 126. Zingg JM, Jones PA. Genetic and epigenetic aspects of DNA meth-
104. Blake MJ, Castro L, Leeder JS, Kearns GL. Ontogeny of drug ylation on genome expression, evolution, mutation and carcino-
metabolizing enzymes in the neonate. Semin Fetal Neonatal Med. genesis. Carcinogenesis. May 1997;18(5):869–882.
Apr 2005;10(2):123–138. 127. Mason JB, Kim Y. Nutritional strategies in the prevention of
105. Rasmussen KM. The “fetal origins” hypothesis: challenges and colorectal cancer. Curr Gastroenterol Rep. Aug 1999;1(4):341–353.
opportunities for maternal and child nutrition. Annu Rev Nutr. 128. Kim M, Trinh BN, Long TI, Oghamian S, Laird PW. Dnmt1 defi-
2001;21:73–95. ciency leads to enhanced microsatellite instability in mouse embry-
106. Waterland RA, Garza C. Potential mechanisms of metabolic onic stem cells. Nucleic Acids Res. 2004;32(19):5742–5749.
imprinting that lead to chronic disease. Am J Clin Nutr. Feb 129. Ingrosso D, Cimmino A, Perna AF, et al. Folate treatment and
1999;69(2):179–197. unbalanced methylation and changes of allelic expression induced
107. Waterland RA. Is epigenetics an important link between early life by hyperhomocysteinaemia in patients with uraemia. Lancet. 17
events and adult disease? Horm Res. Jan 2009;71 Suppl 1:13–16. May 2003;361(9370):1693–1699.
108. Waterland RA. Early environmental effects on epigenetic regula- 130. Petry CJ, Ong KK, Barratt BJ, et al. Common polymorphism in
tion in humans. Epigenetics. 16 Nov 2009;4(8):523–525. H19 associated with birthweight and cord blood IGF-II levels in
109. Hochberg Z, Feil R, Constancia M, et al. Child health, develop- humans. BMC Genet. 10 May 2005;6(1):22.
mental plasticity, and epigenetic programming. Endocr Rev. Apr 131. Waterland RA, Jirtle RL. Early nutrition, epigenetic changes at
2011;32(2):159–224. transposons and imprinted genes, and enhanced susceptibility to
110. Waterland RA, Kellermayer R, Laritsky E, et al. Season of concep- adult chronic diseases. Nutrition. Jan 2004;20(1):63–68.
tion in rural gambia affects DNA methylation at putative human 132. Morgan HD, Sutherland HG, Martin DI, Whitelaw E. Epigenetic
metastable epialleles. PLoS Genet. 2010;6(12):e1001252. inheritance at the agouti locus in the mouse. Nat Genet. Nov
111. Waterland RA, Garza C. Potential mechanisms of metabolic 1999;23(3):314–318.
imprinting that lead to chronic disease. Am J Clin Nutr. Feb 133. Waterland RA, Travisano M, Tahiliani KG. Diet-induced
1999;69(2):179–197. hypermethylation at agouti viable yellow is not inherited trans-
112. Waterland RA. Epigenetic epidemiology of obesity: applica- generationally through the female. FASEB J. Oct 2007;21(12):
tion of epigenomic technology. Nutr Rev. Aug 2008;66 Suppl 3380–3385.
1:S21–S23. 134. Delhanty JD. Preimplantation genetics: an explanation for poor
113. Cedar H, Bergman Y. Programming of DNA methylation pat- human fertility? Ann Hum Genet. Jul 2001;65(Pt 4):331–338.
terns. Annu Rev Biochem. 2012;81:97–117. 135. Brent RL, Beckman DA. The contribution of environmental
114. Rando OJ. Combinatorial complexity in chromatin structure and teratogens to embryonic and fetal loss. Clin Obstet Gynecol. Sep
function: revisiting the histone code. Curr Opin Genet Dev. Apr 1994;37(3):646–670.
2012;22(2):148–155. 136. Edmonds DK, Lindsay KS, Miller JF, Williamson E, Wood
115. Suh JR, Herbig AK, Stover PJ. New perspectives on folate catabo- PJ. Early embryonic mortality in women. Fertil Steril. Oct
lism. Annu Rev Nutr. 2001;21:255–282. 1982;38(4):447–453.
116. Clarke S, Banfield K. S-adenosylmethionine-dependent meth- 137. Edwards RG. Recent scientific and medical advances in assisted
yltransferases. In: Carmel R, Jacobson DW, eds. Homocysteine in human conception. Int J Dev Biol. Apr 1997;41(2):255–262.
Health and Disease. Cambridge, UK: Cambridge University Press; 138. Wilcox AJ, Weinberg CR, O’Connor JF, et al. Incidence of early
2001. 65–80. loss of pregnancy. N Engl J Med. 28 Jul 1988;319(4):189–194.
117. Finkelstein JD. Homocysteine: a history in progress. Nutr Rev. Jul 139. Bulletti C, Flamigni C, Giacomucci E. Reproductive failure due
2000;58(7):193–204. to spontaneous abortion and recurrent miscarriage. Hum Reprod
118. Finkelstein JD. Pathways and regulation of homocysteine metabo- Update. Mar-Apr 1996;2(2):118–136.
lism in mammals. Semin Thromb Hemost. 2000;26(3):219–225. 140. Brock DJ, Holloway S. Fertility of older women. Lancet. 16 Jun
119. Choi SW, Mason JB. Folate and carcinogenesis: an integrated 1990;335(8703):1470.
scheme. J Nutr. Feb 2000;130(2):129–132. 141. Cowchock FS, Gibas Z, Jackson LG. Chromosome errors as a
120. Beaudin AE, Stover PJ. Insights into metabolic mechanisms under- cause of spontaneous abortion: the relative importance of mater-
lying folate-responsive neural tube defects: a minireview. Birth nal age and obstetric history. Fertil Steril. May 1993;59(5):
Defects Res A Clin Mol Teratol. Apr 2009;85(4):274–284. 1011–1014.
121. Beaudin AE, Abarinov EV, Malysheva O, Perry CA, Caudill 142. Gris JC, Perneger TV, Quere I, et al. Antiphospholipid/antiprotein
M, Stover PJ. Dietary folate, but not choline, modifies neural antibodies, hemostasis-related autoantibodies, and plasma homo-
tube defect risk in Shmt1 knockout mice. Am J Clin Nutr. Jan cysteine as risk factors for a first early pregnancy loss: a matched
2012;95(1):109–114. case-control study. Blood. 15 Nov 2003;102(10):3504–3513.
122. Herbig K, Chiang EP, Lee LR, Hills J, Shane B, Stover PJ. 143. Zetterberg H. Methylenetetrahydrofolate reductase and trans-
Cytoplasmic serine hydroxymethyltransferase mediates com- cobalamin genetic polymorphisms in human spontaneous abor-
petition between folate-dependent deoxyribonucleotide and tion: biological and clinical implications. Reprod Biol Endocrinol.
S-adenosylmethionine biosyntheses. J Biol Chem. 11 Oct 17 Feb 2004;2(1):7.
2002;277(41):38381–38389. 144. Zetterberg H, Regland B, Palmer M, et al. Increased frequency
123. Oyama K, Kawakami K, Maeda K, Ishiguro K, Watanabe G. The of combined methylenetetrahydrofolate reductase C677T and
association between methylenetetrahydrofolate reductase poly- A1298C mutated alleles in spontaneously aborted embryos. Eur J
morphism and promoter methylation in proximal colon cancer. Hum Genet. Feb 2002;10(2):113–118.
Anticancer Res. Mar-Apr 2004;24(2B):649–654. 145. Zetterberg H, Regland B, Palmer M, et al. The transcobalamin
124. Shelnutt KP, Kauwell GP, Gregory JF, 3rd, et al. codon 259 polymorphism influences the risk of human spontane-
Methylenetetrahydrofolate reductase 677C-->T polymorphism ous abortion. Hum Reprod. Dec 2002;17(12):3033–3036.
affects DNA methylation in response to controlled folate intake in 146. Zetterberg H, Zafiropoulos A, Spandidos DA, Rymo L, Blennow
young women. J Nutr Biochem. Sep 2004;15(9):554–560. K. Gene–gene interaction between fetal MTHFR 677C>T and
125. Quinlivan EP, Davis SR, Shelnutt KP, et al. Methylenetetrahydro transcobalamin 776C>G polymorphisms in human spontaneous
folate reductase 677C-->T polymorphism and folate status affect abortion. Hum Reprod. Sep 2003;18(9):1948–1950.
N utrition a l G e no m ic s • 2 0 5
147. Nelen WL, Blom HJ, Steegers EA, den Heijer M, Eskes TK. 169. Nyirenda MJ, Lindsay RS, Kenyon CJ, Burchell A, Seckl JR.
Hyperhomocysteinemia and recurrent early pregnancy loss: a Glucocorticoid exposure in late gestation permanently programs
meta-analysis. Fertil Steril. Dec 2000;74(6):1196–1199. rat hepatic phosphoenolpyruvate carboxykinase and glucocorti-
148. Nelen WL. Hyperhomocysteinaemia and human reproduction. coid receptor expression and causes glucose intolerance in adult
Clin Chem Lab Med. Aug 2001;39(8):758–763. offspring. J Clin Invest. 15 May 1998;101(10):2174–2181.
149. Yamada H, Sata F, Saijo Y, Kishi R, Minakami H. Genetic factors 170. Goland RS, Jozak S, Warren WB, Conwell IM, Stark RI, Tropper
in fetal growth restriction and miscarriage. Semin Thromb Hemost. PJ. Elevated levels of umbilical cord plasma corticotropin-releasing
Jun 2005;31(3):334–345. hormone in growth-retarded fetuses. J Clin Endocrinol Metab. Nov
150. Nelen WL, Blom HJ, Steegers EA, den Heijer M, Thomas CM, 1993;77(5):1174–1179.
Eskes TK. Homocysteine and folate levels as risk factors for recur- 171. Goland RS, Tropper PJ, Warren WB, Stark RI, Jozak SM, Conwell
rent early pregnancy loss. Obstet Gynecol. Apr 2000;95(4):519–524. IM. Concentrations of corticotrophin-releasing hormone in the
151. Nelen WL, Blom HJ, Thomas CM, Steegers EA, Boers GH, umbilical-cord blood of pregnancies complicated by pre-eclampsia.
Eskes TK. Methylenetetrahydrofolate reductase polymor- Reprod Fertil Dev. 1995;7(5):1227–1230.
phism affects the change in homocysteine and folate concen- 172. Bertram C, Trowern AR, Copin N, Jackson AA, Whorwood CB.
trations resulting from low dose folic acid supplementation in The maternal diet during pregnancy programs altered expression
women with unexplained recurrent miscarriages. J Nutr. Aug of the glucocorticoid receptor and type 2 11beta-hydroxysteroid
1998;128(8):1336–1341. dehydrogenase: potential molecular mechanisms underlying
152. Pennisi E. Evolution of developmental diversity. Evo-devo the programming of hypertension in utero. Endocrinology. Jul
devotees eye ocular origins and more. Science. 10 May 2001;142(7):2841–2853.
2002;296(5570):1010–1011. 173. Stewart PM, Rogerson FM, Mason JI. Type 2 11 beta-hydroxysteroid
153. Pasqualetti M, Neun R, Davenne M, Rijli FM. Retinoic acid res- dehydrogenase messenger ribonucleic acid and activity in human
cues inner ear defects in Hoxa1 deficient mice. Nat Genet. Sep placenta and fetal membranes: its relationship to birth weight and
2001;29(1):34–39. putative role in fetal adrenal steroidogenesis. J Clin Endocrinol
154. Zhao R, Russell RG, Wang Y, et al. Rescue of embryonic lethal- Metab. Mar 1995;80(3):885–890.
ity in reduced folate carrier-deficient mice by maternal folic acid 174. Drake AJ, Walker BR, Seckl JR. Intergenerational consequences of
supplementation reveals early neonatal failure of hematopoietic fetal programming by in utero exposure to glucocorticoids in rats.
organs. J Biol Chem. 30 Mar 2001;276(13):10224–10228. Am J Physiol Regul Integr Comp Physiol. Jan 2005;288(1):R34–R38.
155. Finnell RH, Spiegelstein O, Wlodarczyk B, et al. DNA methyla- 175. Hebbar PB, Archer TK. Chromatin remodeling by nuclear recep-
tion in Folbp1 knockout mice supplemented with folic acid during tors. Chromosoma. May 2003;111(8):495–504.
gestation. J Nutr. Aug 2002;132(8 Suppl):2457S-2461S. 176. Thomassin H, Flavin M, Espinas ML, Grange T.
156. Pal C, Papp B, Hurst LD. Genomic function: Rate of evolution Glucocorticoid-induced DNA demethylation and gene memory
and gene dispensability. Nature. 30 Jan 2003;421(6922):496–497; during development. Embo J. 17 Apr 2001;20(8):1974–1983.
discussion 497–498. 177. Weinstock M. Alterations induced by gestational stress in brain
157. Stover PJ, Garza C. Bringing individuality to public health recom- morphology and behaviour of the offspring. Prog Neurobiol. Dec
mendations. J Nutr. Aug 2002;132(8 Suppl):2476S-2480S. 2001;65(5):427–451.
158. Gasch AP, Werner-Washburne M. The genomics of yeast responses 178. Stewart PM, Murry BA, Mason JI. Type 2 11 beta-hydroxysteroid
to environmental stress and starvation. Funct Integr Genomics. Sep dehydrogenase in human fetal tissues. J Clin Endocrinol Metab. Jun
2002;2(4–5):181–192. 1994;78(6):1529–1532.
159. van Spronsen FJ. Phenylketonuria management from an 179. Diaz R, Fuxe K, Ogren SO. Prenatal corticosterone treat-
European perspective: a commentary. Mol Genet Metab. Jun ment induces long-term changes in spontaneous and
2010;100(2):107–110. apomorphine-mediated motor activity in male and female rats.
160. Beaudin AE, Stover PJ. Folate-mediated one-carbon metabo- Neuroscience. Nov 1997;81(1):129–140.
lism and neural tube defects: balancing genome synthesis 180. Dunn AJ, Berridge CW. Physiological and behavioral responses to
and gene expression. Birth Defects Res C Embryo Today. Sep corticotropin-releasing factor administration: is CRF a mediator
2007;81(3):183–203. of anxiety or stress responses? Brain Res Brain Res Rev. May-Aug
161. Seckl JR, Meaney MJ. Glucocorticoid programming. Ann N Y 1990;15(2):71–100.
Acad Sci. Dec 2004;1032:63–84. 181. Gatford KL, Wintour EM, De Blasio MJ, Owens JA, Dodic M.
162. Seckl JR, Walker BR. Minireview: 11beta-hydroxysteroid dehy- Differential timing for programming of glucose homoeostasis, sen-
drogenase type 1- a tissue-specific amplifier of glucocorticoid sitivity to insulin and blood pressure by in utero exposure to dexa-
action. Endocrinology. Apr 2001;142(4):1371–1376. methasone in sheep. Clin Sci (Lond). May 2000;98(5):553–560.
163. McMillen IC, Robinson JS. Developmental origins of the meta- 182. Valera A, Pujol A, Pelegrin M, Bosch F. Transgenic mice over-
bolic syndrome: prediction, plasticity, and programming. Physiol expressing phosphoenolpyruvate carboxykinase develop
Rev. Apr 2005;85(2):571–633. non-insulin-dependent diabetes mellitus. Proc Natl Acad Sci U S
164. Seckl JR. Prenatal glucocorticoids and long-term programming. A. 13 Sep 1994;91(19):9151–9154.
Eur J Endocrinol. Nov 2004;151 Suppl 3:U49–62. 183. Rosella G, Zajac JD, Kaczmarczyk SJ, Andrikopoulos S, Proietto
165. Reinisch JM, Simon NG, Karow WG, Gandelman R. Prenatal J. Impaired suppression of gluconeogenesis induced by overexpres-
exposure to prednisone in humans and animals retards intrauterine sion of a noninsulin-responsive phosphoenolpyruvate carboxyki-
growth. Science. 27 Oct 1978;202(4366):436–438. nase gene. Mol Endocrinol. Nov 1993;7(11):1456–1462.
166. French NP, Hagan R, Evans SF, Godfrey M, Newnham JP. 184. Weaver IC, Cervoni N, Champagne FA, et al. Epigenetic program-
Repeated antenatal corticosteroids: size at birth and subsequent ming by maternal behavior. Nat Neurosci. Aug 2004;7(8):847–854.
development. Am J Obstet Gynecol. Jan 1999;180(1 Pt 1):114–121. 185. Weaver IC, Diorio J, Seckl JR, Szyf M, Meaney MJ. Early environ-
167. Newnham JP, Evans SF, Godfrey M, Huang W, Ikegami M, Jobe A. mental regulation of hippocampal glucocorticoid receptor gene
Maternal, but not fetal, administration of corticosteroids restricts expression: characterization of intracellular mediators and potential
fetal growth. J Matern Fetal Med. May–Jun 1999;8(3):81–87. genomic target sites. Ann N Y Acad Sci. Jun 2004;1024:182–212.
168. White PC, Mune T, Agarwal AK. 11 beta-Hydroxysteroid dehy- 186. Jiang X, Yan J, West AA, et al. Maternal choline intake alters
drogenase and the syndrome of apparent mineralocorticoid excess. the epigenetic state of fetal cortisol-regulating genes in humans.
Endocr Rev. Feb 1997;18(1):135–156. FASEB J. Aug 2012;26(8):3563–3574.
2 0 6 • P rincip l e s o f G e no m ic M e dicin e
187. Lanpher B, Brunetti-Pierri N, Lee B. Inborn errors of metabo- 212. Vo TD, Paul Lee WN, Palsson BO. Systems analysis of energy
lism: the flux from Mendelian to complex diseases. Nat Rev Genet. metabolism elucidates the affected respiratory chain complex in
Jun 2006;7(6):449–460. Leigh’s syndrome. Mol Genet Metab. May 2007;91(1):15–22.
188. Panagiotou G, Nielsen J. Nutritional systems biology: definitions 213. Ozsolak F, Milos PM. RNA sequencing: advances, challenges and
and approaches. Annu Rev Nutr. 2009;29:329–339. opportunities. Nat Rev Genet. Feb 2011;12(2):87–98.
189. Arkin AP, Schaffer DV. Network news: innovations in 21st cen- 214. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold
tury systems biology. Cell. 18 Mar 2011;144(6):844–849. B. Mapping and quantifying mammalian transcriptomes by
190. Ideker T, Galitski T, Hood L. A new approach to decoding life: sys- RNA-Seq. Nat Methods. Jul 2008;5(7):621–628.
tems biology. Annu Rev Genomics Hum Genet. 2001;2:343–372. 215. Lewis NE, Schramm G, Bordbar A, et al. Large-scale in silico mod-
191. Kitano H. Systems biology: a brief overview. Science. 1 Mar eling of metabolic interactions between cell types in the human
2002;295(5560):1662–1664. brain. Nat Biotechnol. Dec 2010;28(12):1279–1285.
192. Chuang HY, Hofree M, Ideker T. A decade of systems biology. 216. Bordbar A, Feist AM, Usaite-Black R, Woodcock J, Palsson BO,
Annu Rev Cell Dev Biol. 10 Nov 2010;26:721–744. Famili I. A multi-tissue type genome-scale metabolic network
193. Bordbar A, Palsson BO. Using the reconstructed genome-scale for analysis of whole-body systems physiology. BMC Syst Biol.
human metabolic network to study physiology and pathology. J 2011;5:180.
Intern Med. Feb 2012;271(2):131–141. 217. Dunn WB, Bailey NJ, Johnson HE. Measuring the metab-
194. Orth JD, Thiele I, Palsson BO. What is flux balance analysis? Nat olome: current analytical technologies. Analyst. May
Biotechnol. Mar 2010;28(3):245–248. 2005;130(5):606–625.
195. Chou JY, Mansfield BC. Mutations in the glucose-6-phosphatase- 218. Shaham O, Wei R, Wang TJ, et al. Metabolic profiling of the
alpha (G6PC) gene that cause type Ia glycogen storage disease. human response to a glucose challenge reveals distinct axes of insu-
Hum Mutat. Jul 2008;29(7):921–930. lin sensitivity. Mol Syst Biol. 2008;4:214.
196. Froissart R, Piraud M, Boudjemline AM, et al. Glucose-6- 219. Shlomi T, Cabili MN, Ruppin E. Predicting metabolic biomarkers
phosphatase deficiency. Orphanet J Rare Dis. 2011;6:27. of human inborn errors of metabolism. Mol Syst Biol. 2009;5:263.
197. Bandsma RH, Smit GP, Kuipers F. Disturbed lipid metabolism in 220. Gille C, Bolling C, Hoppe A, et al. HepatoNet1: a comprehensive
glycogen storage disease type 1. Eur J Pediatr. Oct 2002;161 Suppl metabolic reconstruction of the human hepatocyte for the analysis
1:S65–S69. of liver physiology. Mol Syst Biol. 7 Sep 2010;6:411.
198. Oti M, Brunner HG. The modular nature of genetic diseases. Clin 221. Folger O, Jerby L, Frezza C, Gottlieb E, Ruppin E, Shlomi T.
Genet. Jan 2007;71(1):1–11. Predicting selective drug targets in cancer through metabolic net-
199. Lee DS, Park J, Kay KA, Christakis NA, Oltvai ZN, Barabasi works. Mol Syst Biol. 2011;7:501.
AL. The implications of human metabolic network topol- 222. Frezza C, Zheng L, Folger O, et al. Haem oxygenase is synthetically
ogy for disease comorbidity. Proc Natl Acad Sci U S A. 22 Jul lethal with the tumour suppressor fumarate hydratase. Nature. 8
2008;105(29):9880–9885. Sep 2011;477(7363):225–228.
200. Goh KI, Cusick ME, Valle D, Childs B, Vidal M, Barabasi AL. 223. Shlomi T, Benyamini T, Gottlieb E, Sharan R, Ruppin E.
The human disease network. Proc Natl Acad Sci U S A. 22 May Genome-scale metabolic modeling elucidates the role of prolifera-
2007;104(21):8685–8690. tive adaptation in causing the Warburg effect. PLoS Comput Biol.
201. Park J, Lee DS, Christakis NA, Barabasi AL. The impact of cellular Mar 2011;7(3):e1002018.
networks on disease comorbidity. Mol Syst Biol. 2009;5:262. 224. Jamshidi N, Palsson BO. Systems biology of SNPs. Mol Syst Biol.
202. An S, Kumar R, Sheets ED, Benkovic SJ. Reversible compartmen- 2006;2:38.
talization of de novo purine biosynthetic complexes in living cells. 225. Burgard AP, Nikolaev EV, Schilling CH, Maranas CD. Flux cou-
Science. 4 Apr 2008;320(5872):103–106. pling analysis of genome-scale metabolic network reconstructions.
203. Kanehisa M, Goto S, Hattori M, et al. From genomics to chemical Genome Res. Feb 2004;14(2):301–312.
genomics: new developments in KEGG. Nucleic Acids Res. 1 Jan 226. Biomarkers Definitions Working Group. Biomarkers and surrogate
2006;34(Database issue):D354–D357. endpoints: preferred definitions and conceptual framework. Clin
204. Duarte NC, Becker SA, Jamshidi N, et al. Global recon- Pharmacol Ther. Mar 2001;69(3):89–95.
struction of the human metabolic network based on 227. Mamas M, Dunn WB, Neyses L, Goodacre R. The role of metabo-
genomic and bibliomic data. Proc Natl Acad Sci U S A. 6 Feb lites and metabolomics in clinically applicable biomarkers of dis-
2007;104(6):1777–1782. ease. Arch Toxicol. Jan 2011;85(1):5–17.
205. Hamosh A, Scott AF, Amberger JS, Bocchini CA, McKusick VA. 228. Kouskoumvekaki I, Panagiotou G. Navigating the human metabo-
Online Mendelian Inheritance in Man (OMIM), a knowledge lome for biomarker identification and design of pharmaceutical
base of human genes and genetic disorders. Nucleic Acids Res. 1 Jan molecules. J Biomed Biotechnol. 2011; Jan-Feb;29(1):94–110.
2005;33(Database issue):D514–D517. 229. Tobert JA. Lovastatin and beyond: the history of the
206. Machado D, Costa RS, Rocha M, Ferreira EC, Tidor B, Rocha I. HMG-CoA reductase inhibitors. Nat Rev Drug Discov. Jul
Modeling formalisms in Systems Biology. AMB Express. 2011;1:45. 2003;2(7):517–526.
207. Orth JD, Palsson BO. Systematizing the generation of missing meta- 230. Weljie AM, Dowlatabadi R, Miller BJ, Vogel HJ, Jirik FR. An
bolic knowledge. Biotechnol Bioeng. 15 Oct 2010;107(3):403–412. inflammatory arthritis-associated metabolite biomarker pat-
208. Heinemann M, Sauer U. Systems biology of microbial metabolism. tern revealed by 1H NMR spectroscopy. J Proteome Res. Sep
Curr Opin Microbiol. Jun 2010;13(3):337–343. 2007;6(9):3456–3464.
209. Capel F, Klimcakova E, Viguerie N, et al. Macrophages and adipo- 231. McKusick VA. Mendelian Inheritance in Man and its online ver-
cytes in human obesity: adipose tissue gene expression and insu- sion, OMIM. Am J Hum Genet. Apr 2007;80(4):588–604.
lin sensitivity during calorie restriction and weight stabilization. 232. Kornman KS, Martha PM, Duff GW. Genetic variations and
Diabetes. Jul 2009;58(7):1558–1567. inflammation: a practical nutrigenomics opportunity. Nutrition.
210. Zelezniak A, Pers TH, Soares S, Patti ME, Patil KR. Metabolic Jan 2004;20(1):44–49.
network topology reveals transcriptional regulatory signatures of 233. Vijg J, Suh Y. Genetics of longevity and aging. Annu Rev Med.
type 2 diabetes. PLoS Comput Biol. Apr 2010;6(4):e1000729. 2005;56:193–212.
211. Deo RC, Hunter L, Lewis GD, et al. Interpreting metabolomic 234. Seifried HE, Anderson DE, Sorkin BC, Costello RB. Free radi-
profiles using unbiased pathway models. PLoS Comput Biol. Feb cals: the pros and cons of antioxidants. Executive summary report.
2010;6(2):e1000692. J Nutr. Nov 2004;134(11):3143S–3163S.
N utrition a l G e no m ic s • 2 0 7
235. Zeisel SH. Antioxidants suppress apoptosis. J Nutr. Nov 258. Jacques PF, Bostom AG, Williams RR, et al. Relation between
2004;134(11):3179S-3180S. folate status, a common mutation in methylenetetrahydrofolate
236. Stillman RJ, Rosenberg MJ, Sachs BP. Smoking and reproduction. reductase, and plasma homocysteine concentrations. Circulation.
Fertil Steril. Oct 1986;46(4):545–566. 1 Jan 1996;93(1):7–9.
237. De Rycke M, Liebaers I, Van Steirteghem A. Epigenetic risks 259. Tsai MY, Yang F, Bignell M, Aras O, Hanson NQ. Relation between
related to assisted reproductive technologies: risk analysis and epi- plasma homocysteine concentration, the 844ins68 variant of the
genetic inheritance. Hum Reprod. Oct 2002;17(10):2487–2494. cystathionine beta-synthase gene, and pyridoxal-5′-phosphate con-
238. Sinclair KD, Dunne LD, Maxfield EK, et al. Fetal growth and centration. Mol Genet Metab. Aug 1999;67(4):352–356.
development following temporary exposure of day 3 ovine 260. Afman LA, Trijbels FJ, Blom HJ. The H475Y polymorphism in
embryos to an advanced uterine environment. Reprod Fertil Dev. the glutamate carboxypeptidase II gene increases plasma folate
1998;10(3):263–269. without affecting the risk for neural tube defects in humans. J Nutr.
239. Young LE, Fernandes K, McEvoy TG, et al. Epigenetic change in Jan 2003;133(1):75–77.
IGF2R is associated with fetal overgrowth after sheep embryo cul- 261. Devlin AM, Ling EH, Peerson JM, et al. Glutamate carboxypep-
ture. Nat Genet. Feb 2001;27(2):153–154. tidase II: a polymorphism associated with lower levels of serum
240. Khosla S, Dean W, Reik W, Feil R. Culture of preimplantation folate and hyperhomocysteinemia. Hum Mol Genet. 22 Nov
embryos and its long-term effects on gene expression and pheno- 2000;9(19):2837–2844.
type. Hum Reprod Update. Jul-Aug 2001;7(4):419–427. 262. Uitterlinden AG, Fang Y, Bergink AP, van Meurs JB, van
241. Gao S, Latham KE. Maternal and environmental factors in Leeuwen HP, Pols HA. The role of vitamin D receptor gene
early cloned embryo development. Cytogenet Genome Res. polymorphisms in bone biology. Mol Cell Endocrinol. 29 Nov
2004;105(2–4):279–284. 2002;197(1–2):15–21.
242. Schneeman BO, Mendelson R. Dietary guidelines: past 263. Griffiths W, Cox T. Haemochromatosis: novel gene discovery and
experience and new approaches. J Am Diet Assoc. Oct the molecular pathophysiology of iron metabolism. Hum Mol
2002;102(10):1498–1500. Genet. Oct 2000;9(16):2377–2382.
243. Schneeman BO. Evolution of dietary guidelines. J Am Diet Assoc. 264. Geller DS. A genetic predisposition to hypertension? Hypertension.
Dec 2003;103(12 Suppl 2):S5–S9. Jul 2004;44(1):27–28.
244. Weinshilboum R. Inheritance and drug response. N Engl J Med. 6 265. Jeck N, Waldegger S, Lampert A, et al. Activating mutation of the
Feb 2003;348(6):529–537. renal epithelial chloride channel ClC-Kb predisposing to hyper-
245. Solis C, Veenema K, Ivanov AA, et al. Folate intake at RDA lev- tension. Hypertension. Jun 2004;43(6):1175–1181.
els is inadequate for Mexican American men with the methy- 266. Bentzen J, Jorgensen T, Fenger M. The effect of six polymorphisms
lenetetrahydrofolate reductase 677TT genotype. J Nutr. Jan in the apolipoprotein B gene on parameters of lipid metabolism in
2008;138(1):67–72. a Danish population. Clin Genet. Feb 2002;61(2):126–134.
246. Hunt JR, Zeng H. Iron absorption by heterozygous carriers of the 267. Hubacek JA, Pistulkova H, Skodova Z, Berg K, Poledne R.
HFE C282Y mutation associated with hemochromatosis. Am J Association between apolipoprotein B promotor haplotypes and
Clin Nutr. Oct 2004;80(4):924–931. cholesterol status. Ann Clin Biochem. Jul 2001;38(Pt 4):399–400.
247. Swanson CA. Iron intake and regulation: implications for iron 268. Brown S, Ordovas JM, Campos H. Interaction between the
deficiency and iron overload. Alcohol. Jun 2003;30(2):99–102. APOC3 gene promoter polymorphisms, saturated fat intake and
248. Moirand R, Guyader D, Mendler MH, et al. HFE based plasma lipoproteins. Atherosclerosis. Oct 2003;170(2):307–313.
re-evaluation of heterozygous hemochromatosis. Am J Med Genet. 269. Hein DW. Molecular genetics and function of NAT1 and
1 Sep 2002;111(4):356–361. NAT2: role in aromatic amine metabolism and carcinogenesis.
249. Baier LJ, Hanson RL. Genetic studies of the etiology of type 2 dia- Mutat Res. 30 Sep 2002;506: 65–77.
betes in Pima Indians: hunting for pieces to a complicated puzzle. 270. Hein DW, Doll MA, Fretland AJ, et al. Molecular genetics and epi-
Diabetes. May 2004;53(5):1181–1186. demiology of the NAT1 and NAT2 acetylation polymorphisms.
250. Tseung CW, McMahon LG, Vazquez J, Pohl J, Gregory JF, 3rd. Cancer Epidemiol Biomarkers Prev. Jan 2000;9(1):29–42.
Partial amino acid sequence and mRNA analysis of cytosolic 271. Ferre N, Camps J, Fernandez-Ballart J, et al. Regulation of serum
pyridoxine-beta-D-glucoside hydrolase from porcine intestinal paraoxonase activity by genetic, nutritional, and lifestyle factors in
mucosa: proposed derivation from the lactase-phlorizin hydrolase the general population. Clin Chem. Sep 2003;49(9):1491–1497.
gene. Biochem J. 15 May 2004;380(Pt 1):211–218. 272. Chistyakov DA, Savost’anov KV, Zotova EV, Nosikov VV.
251. Mackey AD, McMahon RJ, Townsend JH, Gregory JF, 3rd. Uptake, Polymorphisms in the Mn-SOD and EC-SOD genes and their
hydrolysis, and metabolism of pyridoxine-5′-beta-D-glucoside in relationship to diabetic neuropathy in type 1 diabetes mellitus.
Caco-2 cells. J Nutr. Apr 2004;134(4):842–846. BMC Med Genet. 2001;2(1):4.
252. Csete M, Doyle J. Bow ties, metabolism and disease. Trends 273. Van Landeghem GF, Tabatabaie P, Kucinskas V, Saha N, Beckman
Biotechnol. Sep 2004;22(9):446–450. G. Ethnic variation in the mitochondrial targeting sequence poly-
253. Honein MA, Paulozzi LJ, Mathews TJ, Erickson JD, Wong morphism of MnSOD. Hum Hered. Jul 1999;49(4):190–193.
LY. Impact of folic acid fortification of the US food sup- 274. Messier W, Stewart CB. Episodic adaptive evolution of primate
ply on the occurrence of neural tube defects. JAMA. 20 Jun lysozymes. Nature. 9 Jan 1997;385(6612):151–154.
2001;285(23):2981–2986. 275. Zhang J, Zhang YP, Rosenberg HF. Adaptive evolution of a dupli-
254. Delange FM. Control of iodine deficiency in Western and Central cated pancreatic ribonuclease gene in a leaf-eating monkey. Nat
Europe. Cent Eur J Public Health. Sep 2003;11(3):120–123. Genet. Apr 2002;30(4):411–415.
255. Center for Disease Control. Alcohol consumption among women 276. Wu W, Goodman M, Lomax MI, Grossman LI. Molecular evo-
who are pregnant or who might become pregnant—United States, lution of cytochrome c oxidase subunit IV: evidence for positive
2002. Morb Mortal Wkly Rep. 2004;53(50):1178–1181. selection in simian primates. J Mol Evol. May 1997;44(5):477–491.
256. Mellott TJ, Williams CL, Meck WH, Blusztajn JK. Prenatal 277. Wooding SP, Watkins WS, Bamshad MJ, Dunn DM, Weiss RB,
choline supplementation advances hippocampal development Jorde LB. DNA sequence variation in a 3.7-kb noncoding sequence
and enhances MAPK and CREB activation. FASEB J. Mar 5′ of the CYP1A2 gene: implications for human population history
2004;18(3):545–547. and natural selection. Am J Hum Genet. Sep 2002;71(3):528–542.
257. Heird WC, Lapillonne A. The role of essential fatty acids in devel- 278. Bailey LB, Gregory JF, 3rd. Folate metabolism and requirements.
opment. Annu Rev Nutr. 2005;25:549–571. J Nutr. Apr 1999;129(4):779–782.
2 0 8 • P rincip l e s o f G e no m ic M e dicin e
279. von Linsingen R, Bompeixe EP, Bicalho Mda G. A case-control the South Indian population. Hum Reprod. Nov 2004;19(11):
study in IL6 and TGFB1 gene polymorphisms and recurrent 2648–2652.
spontaneous abortion in southern Brazilian patients. Am J Reprod 285. Schweikert A, Rau T, Berkholz A, Allera A, Daufeldt S, Wildt
Immunol. Feb 2005;53(2):94–99. L. Association of progesterone receptor polymorphism with
280. Prigoshin N, Tambutti M, Larriba J, Gogorza S, Testa R. Cytokine recurrent abortions. Eur J Obstet Gynecol Reprod Biol. 15 Mar
gene polymorphisms in recurrent pregnancy loss of unknown 2004;113(1):67–72.
cause. Am J Reprod Immunol. Jul 2004;52(1):36–41. 286. Sata F, Yamada H, Kondo T, et al. Glutathione S-transferase M1
281. Daher S, Shulzhenko N, Morgun A, et al. Associations between and T1 polymorphisms and the risk of recurrent pregnancy loss.
cytokine gene polymorphisms and recurrent pregnancy loss. Mol Hum Reprod. Mar 2003;9(3):165–169.
J Reprod Immunol. Feb 2003;58(1):69–77. 287. Finan RR, Tamim H, Ameen G, Sharida HE, Rashid M, Almawi
282. Perni SC, Vardhana S, Tuttle SL, Kalish RB, Chasen ST, Witkin WY. Prevalence of factor V G1691A (factor V-Leiden) and pro-
SS. Fetal interleukin-1 receptor antagonist gene polymor- thrombin G20210A gene mutations in a recurrent miscarriage
phism, intra-amniotic interleukin-1beta levels, and history of population. Am J Hematol. Dec 2002;71(4):300–305.
spontaneous abortion. Am J Obstet Gynecol. Oct 2004;191(4): 288. Tempfer C, Unfried G, Zeillinger R, Hefler L, Nagele F, Huber
1318–1323. JC. Endothelial nitric oxide synthase gene polymorphism in
283. Sata F, Yamada H, Yamada A, et al. A polymorphism in the CYP17 women with idiopathic recurrent miscarriage. Hum Reprod. Aug
gene relates to the risk of recurrent pregnancy loss. Mol Hum 2001;16(8):1644–1647.
Reprod. Nov 2003;9(11):725–728. 289. Gloria-Bottini F, Lucarini N, Palmarino R, et al. Phosphoglucomu
284. Suryanarayana V, Deenadayal M, Singh L. Association of tase genetic polymorphism of newborns. Am J Hum Biol. Jan–Feb
CYP1A1 gene polymorphism with recurrent pregnancy loss in 2001;13(1):9–14.
N utrition a l G e no m ic s • 2 0 9
13.
GENOMICS IN PUBLIC AND POPULATION HEALTH
Anastasia L. Wise and Teri A. Manolio
210
Table 13.1 PHARMACOGENOMIC VARIANTS IN COLORECTAL CANCER 3
been shown to provide resistance to monoclonal antibod- and for those with one or two copies of the variant, only
ies directed against the upstream epidermal growth factor treating with carbamazepine when the benefits outweigh
receptor (EGFR), as both are components of a cellular the risks of the drug.6
pathway leading to abnormal cell growth and cancer. Thus,
cetuximab and panitumumab (anti-EGFR antibodies) are
P O P U L AT I O N S C I E N C E S
given only to individuals with normally functioning KRAS,
where blocking EGFR can have an effect.3 The population sciences, such as epidemiology, focus
Pharmacogenomics can also be useful in determining on studying whole populations rather than individuals.
drug dosage for colorectal cancer. For example, the Food Through studying environmental, genomic, and social fac-
and Drug Administration (FDA) recommends testing tors that affect human health, population-level interven-
for UGT1A1 variants when administering irinotecan, as tions can be identified.
individuals who are homozygous or heterozygous for the Population variation is an important consideration
UGT1A1*28 allele are at increased risk of developing neu- when studying common complex conditions that are influ-
tropenia and severe infections.5 Individuals with inactivat- enced by multiple genetic, environmental, and social risk
ing UGT1A1 variants are therefore recommended to be factors, such as asthma. Over 300 million individuals of
started at a lower dosage of irinotecan to reduce the risk of all ages worldwide have asthma.7 Prevalence estimates can
neutropenia.5 vary greatly by ethnicity, however, from 2–33%.8 In the
In addition to modifying drug dosing, pharmacoge- United States, prevalence ranges from approximately 8%
nomic information can also be used in drug selection, in European Americans to 12% in African Americans and
to choose agents more likely to give a beneficial response 7% in Hispanic Americans.9 Within admixed populations,
based on a patient’s genetically driven ability to metabolize such as Hispanic Americans, even greater variation can
them. For example, in patients of Asian ancestry who are be seen when populations are further sub-stratified, with
given carbamazepine (used to treat epilepsy and bipolar Mexican American populations around 6%, while Puerto
disorder) the HLA-B*1502 allele has been associated with Rican populations are closer to 19%.9 Genetic studies have
Stevens-Johnson syndrome/toxic epidermal necrolysis, a shown that at least some of this variation is due to differ-
life-threatening skin condition. This allele can be found ences in genetic variants, with 35–80% of the variation in
in over 15% of the population in some regions in Asia, asthma heritability explained by genetic factors.10,11 For
including Hong Kong, Thailand, Malaysia, and parts of the example, variants in ADAM33 have been seen in European,
Philippines, and is very rare in other populations outside African American, and some Hispanic populations, but not
Asia.6 in other European American, Mexican, Puerto Rican, and
Within populations of Asian ancestry, there can also Korean populations, all of which found different variants in
be great variation, such as is seen within China, where the ADAM33 associated with asthma (Figure 13.1).12
HLA-B*1502 allele frequency varies from 0–36% depend- Studying the interplay between environmental,
ing upon ethnicity (Table 13.2). Thus, the FDA recom- genetic, and social risk factors is also critical to under-
mends genetic testing for the HLA-B*1502 variant before standing the etiology of this complex disease. For exam-
prescribing carbamazepine for patients of Asian ancestry, ple, the effect of air pollution on asthma case reports is
270° 300°
30°
0°
−30°
Example of the variation in allele frequency by population for rs2280089 in ADAM33. The A allele has been previously associated with
Figure 13.1
predisposition to asthma and bronchial hyper-responsiveness in populations from the United States, the United Kingdom, and China.48–51
in particular has been associated with increased risk for adding rs10757274 genotyping to the Framingham risk
developing both Alzheimer’s disease and atherosclerosis, score improved its ability to determine the individuals who
along with a protective effect against developing macular would suffer later cardiovascular events, independent of
degeneration.33 their family history.35 Such models can be used to screen
populations to determine individuals at increased risk of
disease and recommend further testing and individualized
CROSS-CUTTING ISSUES genomic medicine.
O F P O P U L AT I O N G E N O M I C S While chronic conditions such as cardiovascular disease
make up the majority of deaths in the developed world,
Though all three of the sciences contributing to popula- infections are still a major health concern within devel-
tion genomics work together, there are also some issues that oping countries and are equally amenable to study using
more broadly span the field of population genomics and its population genomics. Genomics has made possible the
relationship to medicine and public health. As was touched rapid identification of the organisms causing recent pan-
upon in many of the examples above, it is important to demic outbreaks, including H1N1 (swine flu) and severe
consider the broader implications of population genomics acute respiratory syndrome (SARs), as well as identifying
to global health and how population- and individual-level the source of foodborne illnesses. Genomic sequence infor-
views of health can work together to improve health mation on malaria parasites, mosquito vectors, and their
worldwide. human hosts is being leveraged to produce more rapid diag-
nosis and better drugs, vaccines, and intervention strategies
to fight malaria.36,37
G L O BA L H E A LT H
To maximize the benefit of population genomics
Cardiovascular disease (CVD) is a leading cause of death advances to global health, it is also important to include
worldwide, with over 13.5 million deaths from ischemic multiple populations of diverse ages, ethnicities, and gen-
heart disease, stroke, or another form of cerebrovascu- ders in disease research. As evidenced by the example of
lar disease in 2008, and it is highly amenable to study asthma genomics above, the prevalence of disease can be
using population genomics techniques.34 For example, highly variable across ancestral groups, and genetic variants
9 10 11 12
8
7
3 6
4 5
1
2
Y
19 21
20
22
16 18
15 17
14
13
X
APOE (19q13.32)
14 Alzheimer's disease GWAS
8 Cholesterol GWAS
19 3 C-reactive protein GWAS
1 Response to statin therapy GWAS
1 Brain iamging GWAS
The NHGRI GWAS Catalog showing that many genetic loci are associated with multiple phenotypes. APOE on 19q13.32 is highlighted,
Figure 13.2
along with examples of the disease phenotypes associated with the gene.19
often vary in frequency as well. Thus, while a single pathway modification), environmental (such as diet), and social fac-
may be implicated in disease across many populations, the tors (such as exercise), all of which contribute to this com-
most frequent variant in each population may lie in differ- plex disease.38 The prevalence of type 2 diabetes varies by
ent genes or gene regions. country, from approximately 5–29% with risk alleles such
Local environmental and social factors that affect dis- as the C allele in rs11196205, and decreasing in frequency
ease and population health should also be incorporated from sub-Saharan Africa to Asia.39,40 (Figure 13.3). Effects
into studies of population genomics to produce the most of other risk factors also vary across different populations,
complete picture of disease etiology. For example, the with the relative risk of diabetes for each 5-kg/m2 increase
prevalence of type 2 diabetes is increasing globally and in body mass index (BMI), for example, being 2.4 in Asian
has been associated with multiple genetic (more than 60 Americans, 2.2 in Hispanic Americans, 2.0 in European
genes to date), epigenetic (such as methylation or histone Americans, and 1.6 in African Americans.41
270° 300°
30°
0°
−30°
Figure 13.3 Example of the variation in allele frequency by population for rs11196205 in TCF7L2, a SNP previously associated with type 2 diabetes
risk.40,48
218
largely replacing chromosome analysis, and next-generation if they are already affected or at increased risk. The point of
sequencing (NGS). doing this is to reap the benefits of an early (presymptom-
One application of NGS is to analyze in parallel (i.e., atic) diagnosis, where this brings the advantages of a bet-
simultaneously) a set of many genes, mutation in any one ter prognosis, or to gain access to screening for those in a
of which may lead to a similar pattern of disease. Thus, high-risk group. In contrast, “testing” in the narrower sense
all the genes implicated in particular symptom clusters— refers to the examination or investigation of an individual
such as retinal degeneration, early-onset epilepsy, muscu- in the context of their personal or family history of disease;
lar dystrophy, or hypertrophic cardiomyopathy—may be that is, where there is a specific reason to believe that their
analyzed simultaneously; this can accelerate the diagnosis individual chance of carrying a disease-associated genetic
of such conditions and highlight potential interactions variant, or of developing a particular type of genetic illness,
between variants at different loci, which might otherwise may be greater than the population’s average.
have remained unrecognized. Problems of interpreta- An individual who seeks access to over-the-counter
tion can arise but do so less frequently with these selective genetic testing in a pharmacy or through the Internet may
approaches to sequencing. therefore be arranging a genetic test on him- or herself if
As it becomes both simpler and cheaper to be less selec- s/he has a strong family history of a relevant disease, or they
tive about the sequence information to be generated, labora- may be arranging a genetic screening test if there are (as far as
tories are moving towards the use of exome sequencing (ES) they know) no particular features in their personal or fam-
and whole-genome sequencing (WGS) for diagnostic as ily history suggesting an increased risk. It is therefore the
well as research purposes. The limiting factor in laboratory context that distinguishes a specific test applied to one indi-
diagnostics is no longer the generation of sequence infor- vidual from a population screening test applied to others;
mation but is becoming the interpretation of the sequence there may also be differences in the technology employed,
data generated. but they will reflect differences in the scale of the enterprise
rather than the distinction between individual testing and
population screening.
N OT J U S T T E R M I N O L O GY: G E N E This setting illustrates one of the serious problems
A N D G E N O ME , T ES T I N G A N D raised by direct-to-consumer (DTC) genetic screening: a
S C R EE N I N G company may claim that their single nucleotide polymor-
phism (SNP)-based genome-wide assessment of disease risk
Some comments on terminology will be helpful. The term indicates risk categories that people at population risk may
“genetic testing” can be used to refer to a wide range of find helpful. However, those seeking such tests may well be
activities. We must be clear as to the scope intended by this motivated by a family history of disease. They may actually
term, so we must distinguish between two pairs of words— be in a high-risk group that only a thorough risk assessment
genetic and genomic, testing and screening. and the sequencing of relevant genes of major effect could
For the purposes of clinical investigation, a “genetic” clarify, and which the SNP-based risk modifying screen
test will refer to a test performed on one particular gene; available DTC does not address. So they may buy a genetic
performing a test on several different genes would be a series screen of no established utility under the illusion that it is,
of genetic tests. Performing a genome-wide investigation— for them, a highly specific genetic test relevant to their per-
WGS or ES, investigating all genes simultaneously, in par- sonal situation. When this confusion is placed alongside the
allel—would be a genomic investigation. A genome-wide fact that most of these SNP-based tests use genetic variation
association study (GWAS) is another example of a genomic that accounts for only a small fraction of the genetic con-
investigation but would be carried out for research rather tribution to risk of disease, it will be clear why such DTC
than in clinical practice. Sequencing a cluster of phenotypi- screening tests attract much professional hostility.
cally related genes, such as all the genes known to be impli- In addition to genetic testing of an individual in their
cated in causing a particular disorder, would still be a set of family context, and the population genetic screening of
genetic tests for that disorder but might now employ NGS individuals intended to identify those carrying a particu-
rather than Sanger sequencing. lar genetic variant predisposing them to disease, there is
The word “screening” also requires some clarification, another type of genetic test in clinical use—the testing of a
as it is slippery and is used with a range of different mean- tissue sample from a patient with disease, where the genetic
ings. It may be regarded as the testing of a person (or of changes may be present only in some tissues rather than
people) at more-or-less population risk of a disorder, to see constitutionally (i.e., in all tissues). This usually entails the
2 2 0 • P rincipl e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
circumstance, however, there may be several different family disease (HD), the situation is very different and will remain
contexts in which genetic testing of an individual may be so until an effective treatment has been developed for symp-
considered appropriate. In relation to their own health and tomatic disease, or effective, pre-symptomatic disease pre-
health care, genetic testing is most likely to be appropriate vention has become possible. In this context, the genetic
when the person knows they are at risk of a disorder pres- professional—whether clinical geneticist or genetics
ent in other members of the family but does not yet know counselor—will engage the at-risk individual who is seek-
if they will be affected; this amounts to predictive genetic ing predictive genetic testing in an extended conversation.
testing. This process may entail challenging the individual’s state-
Such testing has been available for Huntington’s disease ments and attitudes, to help them make the best decision
for more than two decades, and for several familial cancer for them(4,5); this is designed to help them decide whether
syndromes for nearly as long. There have been numerous or not predictive testing would be helpful and, where the
studies of the impact of testing on individuals and families, client chooses to have the test, to prepare for the full range
and of the circumstances in which testing can be clinically of possible test results.
helpful and emotionally tolerable or even beneficial. One Even in such an apparently simple context, there are
important factor is whether the genetic test result guides complexities to consider before proceeding with the test.
the medical care of the individual, including surveillance Clients will be helped to think through how they, and oth-
for treatable complications of the disease. Where the test ers, would react to a favorable or an unfavorable result,
result does have implications for medical management, or to a result of uncertain significance (an intermediate,
the health professionals involved will need to ensure that “gray-zone” allele), or to a change of heart—a decision not
their client understands this. Indeed, professionals may to go ahead with testing (for the moment). Not everyone
wish frankly to recommend testing if this is the best way wishes, or is able, to engage in discussions about such hypo-
to safeguard their patient’s health and welfare. If the medi- thetical scenarios(6), but it is good practice to encourage
cal benefits are less clear, it may be more difficult for them such reflection(7,8). Having clear reasons for testing, and the
to help the client make the best decision for their personal willingness to engage in advance in reflection about the pos-
circumstances, when it is the family and emotional factors sible consequences of testing, may be associated with better
that will be more relevant. outcomes after testing(9,10).
The usually “nondirective” ethos of genetic services will There may be decisions to make about whom to tell
not be so appropriate where the decision about testing has about the risk, the test, and the result, and there may be
clear implications for medical management(2). An example important potential consequences of such testing for the
of such a context might arise with an adolescent or young person’s insurance, employment, or career choice, as well
adult at risk of the familial adenomatous polyposis coli as personal relationships and emotional equilibrium. The
(FAP) present in other family members. Because the prog- implications of testing for insurance and employment will
nosis in an affected individual is so poor in the absence of differ between countries, according to the legal context
tumor surveillance, and of colectomy in those with bowel and the system of health and social care. There will also be
tumors at high risk of malignant transformation, it would consequences of testing for other individuals, including the
be good practice for any health professional involved to person’s partner, any surviving parents, and their siblings
recommend genetic testing. A similar situation arises with and children; there will be an impact on the whole family
infants at risk of inherited retinoblastoma, where the ben- system(11). The most appropriate clinical approach to adopt
efits of genetic testing as an aid to the coordinated manage- has been refined as experience has accumulated of predic-
ment of the at-risk child are so substantial that the only tive testing in this context of high risk without effective
sensible course of action is to recommend it, so that frequent treatments(12–15). However, the early finding that, whatever
surveillance for tumors can be provided to those who are at the result, there will be costs as well as benefits for those
high risk and can be avoided in those whose risk is low (that who are tested has not been contradicted by subsequent
of the general population). Because of the implications of experience(16).
testing results for the use of other medical resources, a num- One type of difficulty can emerge, even with a favorable
ber of perspectives unfamiliar to clinical geneticists may result, if life-plans have been made and acted upon and the
need to be introduced into decisions about which genetic result now undermines the basis of those decisions. Perhaps
testing services it would be appropriate to fund(3). a decision was made about marriage, children, or career
In the context of an individual at risk of a usually late- that seemed safe in the context of genetic risk, but with the
onset neurodegenerative disorder, such as Huntington’s result known, years of life may now, in hindsight, be viewed
www.Ebook777.com
with regret. Difficulties may also arise if the results, whether genetic test, so issues of family communication (or lack of
“good” or “bad,” conflict with a prior expectation of the per- it) are central. Equally, while many women at risk of breast
son tested or his or her family. Individuals and families can and ovarian cancer find testing helpful in both the practical
have their own ideas about the pattern of inheritance that and emotional domains, there are often emotional barriers
applies to them: perhaps it is the oldest girl in each genera- to their discussing test results with relatives. Indeed, given
tion who is affected, or the youngest boy, or the one with red that the primary motivation of many is to provide informa-
hair. Such ideas, which can be referred to by the term “lay tion for their children, the unanticipated emotional and
beliefs about inheritance,” can lead to serious distortions communication difficulties that arise for those who have
in medical care: for example, women at risk of breast can- undergone testing constitute an important phenomenon(22).
cer but related to affected relatives through their father are Those already affected by cancer may have additional coun-
less likely to seek risk assessment or additional surveillance selling and support needs, in part because of the additional
because many families imagine that susceptibility to breast implications for their own health(23), such as when a woman
cancer has to be transmitted through the female line(17). with breast cancer finds that a test result shows that she is
Particular individuals may also be picked out by others in now also at high risk of ovarian cancer.
the family as destined to develop the family disease—they Even apparently favorable results may have some dif-
are “preselected”—and this phenomenon can contribute to ficult consequences, such as altered family relationships(13)
complex processes of family psychodynamics(13). or an unanticipated reluctance to discontinue screening for
Experience has shown that those given an unfavorable disease(24).
predictive result for HD (and many other conditions) do The value of a test—in the sense of its clinical utility—
often experience distress, but, as with those given a favorable can vary enormously with the details of the family history
result, there is a fall in their distress and anxiety over some and the availability for testing of other family members or of
months, associated with the loss of uncertainty, that returns stored pathological specimens. If a family mutation can be
to baseline levels by one year. Overall, the best predictor identified in a sample from a definitely affected individual,
of a person’s post-test emotional state is the pre-test state, then the interpretation to be made of the results of testing
and those whose motivation for testing is unfocused and an at-risk relative can be defended with greater confidence
non-specific tend to fare worse than others(10,18). Those who than if no family mutation is known and only the at-risk
show greater concern and distress during the counselling individual can be tested. In the former case, a negative test
before the testing may be those who can engage construc- result (the failure to find the family’s mutation in one of
tively with their risk in advance, and they may cope better the two BRCA loci) will give very substantial reassurance,
with an unfavorable result(9). Those who proceed with test- although it will still leave the woman concerned with at least
ing for HD are a self-selected group(16), and results of these the population risk of developing a breast cancer. In the lat-
studies cannot be generalized to others—we could not draw ter setting, finding no mutation in either of the two loci will
any conclusions from these studies about how those at low provide much less reassurance. It is for this reason that some
risk of a serious disorder would respond to unfavorable test services only perform mutation searches in samples from a
results. definitely affected individual and only when the chance of
Using the framework of Burke and colleagues to evalu- finding a BRCA gene mutation exceeds a given (arbitrary)
ate genetic tests(19–21), through examining both the clinical threshold. Testing for the family’s particular mutation in an
validity of testing and the scope for useful medical inter- individual at risk becomes possible only once it has been
ventions, we can turn to contexts where there may be some identified by a mutation search in the affected individual.
health benefits from testing. Testing for familial predisposi- This is one way of attempting to contain the costs of health
tion to breast and ovarian cancer at the BRCA1 and BRCA2 care, although other models of service provision could well
loci is in this category, with the benefits of prophylactic sur- be defended and might permit at-risk women with fewer
gery in the prevention of malignancy now better defined. surviving female relatives to be given access to testing.
Here, the potential benefits of testing need to be weighed One approach, which would maximize the efficiency of
against the possible drawbacks jointly by the at-risk indi- cascade testing within families, would be to make mutation
vidual and their health professionals, and the contextual searching available to those affected individuals found to
details of the individual case will often be crucial. One of have a malignancy at a young age and who have a family his-
the factors to be taken into account is the ability or willing- tory of relevant disease(25). This may be more cost-effective
ness of the concerned individual to approach any surviving, in the long term than a genetics service operating solely
affected relatives to see if they would be willing to have a in response to concerns from unaffected, at-risk relatives.
2 2 2 • P rincipl e s o f G e no m ic M e dicin e
Proactively suggesting to such affected individuals that patient, and perhaps their relatives, that the investigations
molecular testing might be appropriate would substantially may generate results with implications for others in the
alter the emotional context for the patient. This greater family, the fact of these genetic implications—a potential
readiness to test the individual woman affected by breast impact on the immediate or even the extended family—will
cancer needs to be assessed not only for its utility, but also emerge once the diagnosis has been made, whatever tech-
for the associated distress and emotional burden. There may nology is used to make the diagnosis. If a child has cystic
be several reasons for this burden, including both the direct fibrosis (CF) or Duchenne muscular dystrophy (DMD),
prognostic implications for themselves and the need for for example, then there will be a risk of recurrence within
them now to consider the implications for others in their the immediate family, and there may be healthy carri-
families, with the associated burden of having to raise this ers of the disease in the extended family, too. This is true
topic with their loved ones. whether the diagnostic process entails DNA technology or
Where testing requires access to information about, and more traditional methods (assay of sweat electrolytes in CF,
perhaps samples from, other family members, the decision or the pathological and immunohistochemical examination
to seek testing will for many be difficult and, even if they of a muscle biopsy in DMD). Similarly, if a healthy preg-
decide to go ahead, may be frustrated by a lack of coopera- nant woman has renal cysts identified in herself during an
tion by their relatives. Furthermore, the possible adverse ultrasound scan carried out on her fetus, then the possible
consequences of unfavorable test results for insurance cov- implications for her own health, the fetus, and the family
erage, and the possible distress or anger that may be antici- will clearly emerge before any specifically genetic testing has
pated in close family members and those whose risks will be taken place.
modified by the result, may deter an individual from pro- It is the clinical imperative of arriving at the correct diag-
ceeding. This also raises the question of “whose information nosis for a sick individual that drives the process of investi-
is it anyway?” and there are good grounds for weakening gation, whether or not that entails genetic investigations.
the right of the affected individual to prevent the applica- Therefore, it is not only the patient but also other family
tion of their test result to the health care of their relatives. members who may need to come to terms simultaneously
Technology may come to the rescue, however, in that the with both the serious nature of the established diagnosis in
cost of testing the bundle of relevant genes in those at risk is the patient and the fact of the genetic implications for oth-
falling, so that it will soon make good sense to test anyone at ers. It is important to make sure that patients and families
risk for mutations with the same panel of genes and to give caught up in such a situation have ready access to timely and
each individual their own risk level, without the interpreta- supportive clinical genetic assessment and genetic counsel-
tion depending so much upon their relatives’ results. ling, as well as the best clinical care for established disease.
One context that has generated much debate among There are rather different considerations in the context
professionals, and also within families, is that of testing an of a child with developmental difficulties and dysmorphic
individual at one in four (25%) prior risk of HD, when an features. In this setting, diagnostic investigations will usu-
unfavorable result on the individual implies an unfavor- ally include tests for genetic conditions. The problems that
able predictive result for the person’s parent, who was pre- commonly arise in this context are threefold:
sumably at one in two (50%) prior risk but who may have
decided not to have predictive testing and may resent the
1. The failure to make a diagnosis that accounts for the
fact that they have, in effect, been “tested” without giving
child’s difficulties, with resulting uncertainty for the
consent(26).
prognosis, the risk of recurrence, and the implications
As our knowledge of human genetic variation and our
for other family members.
ability to interpret it both improve, testing one individual
in a family will become progressively less dependent upon 2. The distress caused by the specific diagnostic label
information about other members of the family. that is attached, either because it confirms the
syndromic association of dysmorphic features with
the developmental problems, or because the parents
D I AG N O S T I C G E N ET I C T E S T I N G
suddenly come to appreciate that the prognosis for their
When an individual presents with a disease that may have a child is substantially worse than they had previously
genetic basis, whether or not there is a relevant family his- accepted: they may be forced to recognize that the
tory, genetic investigations may very reasonably be initiated. child is unlikely ever to talk, or to walk, or to lead an
While it will often be prudent and helpful to explain to the independent life as an adult.
2 2 4 • P rincipl e s o f G e no m ic M e dicin e
once so many can be examined, except perhaps in commu- fundamental ways, for both the timing and the risks of the
nities with many private recessive disorders if ES would be test. Fetal DNA is present in maternal plasma in adequate
substantially more costly. quantities from seven weeks of gestation onward, and
One way of framing the concerns about social discrimi- there is no risk of miscarriage from the procedure. These
nation on genetic grounds is to consider how to reconcile are major changes to the context of prenatal screening or
the different perspectives of the (affected) individual, the diagnosis, and such changes will substantially alter the
family (at risk of having an affected child), and society ethical considerations surrounding prenatal testing and the
(which is challenged by the expectation to meet the medical termination of pregnancies. Thoughtful responses to these
and social needs of all its members). How can the multiple new challenges will be important to prevent the excessive
and conflicting goals of these parties be met, simultaneously commercialization and/or the trivialization of antenatal
respecting the worth of each affected individual, promoting screening.
the reproductive autonomy of prospective parents, all while If a program of comprehensive carrier-screening is intro-
containing the cost of care of children and adults with all duced alongside antenatal screening by maternal blood
forms of special needs, at a time when families’ expectations sample for fetal aneuploidy and chromosomal copy number
of government support are rising(32)? These goals are framed variants (CNVs), then the incidence of both chromosomal
at different levels of social organization—the individual, and recessive disorders at birth may fall dramatically, with
the family, and society at large—and there are powerful ten- unpredictable consequences for social attitudes to these
sions between them. conditions and perhaps other causes of disability. Is this, or
The other application of NGS to reproduction is in to what extent, or under what circumstances, would this be
the context of antenatal screening. While screening for something to be welcomed?
structural anomalies will continue to use ultrasound scan- Pre-implantation genetic testing (PGT) or screening
ning, the current methods of screening for chromosomal may be employed by those using IVF methods to achieve
anomalies are likely to give way to methods based on deep a conception, whether because of fertility problems or to
sequencing of free DNA in maternal plasma, a proportion avoid transmitting a serious genetic disorder. The applica-
of which will be fetal in origin. The formerly crucial distinc- tion of NGS methods in single-cell PGT raises concerns
tion between prenatal diagnosis and antenatal screening that too much knowledge may be generated about a child,
collapses as the screening method becomes the diagnostic whose full genome sequence may become known from
method, and the level of resolution (the whole chromosome before birth. This will give rise to questions considered
or the nucleotide base pair) depends largely upon the cover- below concerning the genetic testing of children.
age and depth of sequencing; in effect, the cost.
Diagnosing fetal chromosome anomalies is now largely
G E N ET I C I N F O R M AT I O N A B O U T C H I L D R E N
based on invasive prenatal diagnosis, but it is beginning
to employ high-throughput methods, such as array CGH There has been a broad professional consensus to avoid
analysis, in place of conventional cytogenetics(33). By setting generating predictive genetic information about children
diagnostic criteria that minimize the number of VUSs, this unless it is to their direct medical benefit. This consensus
may be appropriate practice, although our experience is too has been in place in the United Kingdom and Europe for
limited to regard aCGH as a complete substitute for cyto- about two decades and is still in place(37–39), although recent
genetics in prenatal diagnosis. Avoiding too much informa- recommendations of the American Academy of Pediatrics
tion—too many VUSs or IFs—remains a higher priority and the American College of Genetics and Genomics(40,41)
in this setting than after birth, and assessing the effects of have somewhat weakened the previously similar U.S. policy.
a CNV will vary greatly with the precise diagnostic route These policies also apply to information of importance in
(the sequence of events) that led to the aCGH being reproduction rather than personal health care, such as car-
performed(34). rier status for recessive disorders, but there may be rather
The other technology to consider here is noninvasive less at stake in these contexts, especially when the condition
prenatal diagnosis through the sequencing of free DNA under consideration is autosomal recessive. There may be
in maternal plasma, as this derives from both the mother more at stake for the child’s future when the condition is
and the fetus. Sequencing of chromosome 21 elements sex-linked or results from a chromosomal rearrangement,
from maternal plasma enables the reliable identification when all the child’s offspring will be at risk, whereas if the
of fetal trisomy 21 (Down syndrome)(35,36). This changes condition is autosomal recessive, then the risk only materi-
the domain of prenatal screening and diagnosis in two alizes if the child’s future partner is also a carrier.
2 2 6 • P rincipl e s o f G e no m ic M e dicin e
unit, although no strictly medical benefits for the child The costs of storage will be substantial, as active man-
(e.g., screening for Duchenne muscular dystrophy)(50). The agement of the data will be required because both soft-
introduction of TMS as the method of biochemical mea- ware and information technology hardware will change.
surement has led practitioners to begin to assess “screening Furthermore, repeat or additional analyses on fresh samples
by TMS” as a potential alternative to earlier assay methods may be required (e.g., to capture markers of epigenetic influ-
that diagnosed many fewer conditions. Instead of making ences and gene expression), and not simply data storage.
a decision “disease-by-disease,” the question of introducing These would be real challenges for such a program, which
TMS could be approached by judging between methods of may render carefully stored data relatively useless. It will be
screening, taking what is revealed by TMS as a single entity. instructive to follow the emerging experience of this pro-
That approach has led to the packaging of a long list of met- gram; one can only hope that the children do not become
abolic disorders diagnosable through TMS as the new stan- its casualties.
dard. The process of deciding which diseases to report may
have included a step in which the potential health benefits
G E N ET I C T E S T I N G A N D T H E R A P EU T I C
of an early diagnosis for each disease were considered, but,
B E N E FIT
beneath the disease-by-disease rhetoric, the decision being
made was in fact about whether TMS should be used or Moving from prediction through to therapy, one can be
whether laboratories should close their eyes to the potential optimistic that genome-based knowledge is set to transform
of the new technology. A decision to use TMS, but practice medicine. The first area in which major benefits can already
as if it had not been developed, would not have been sus- be seen in the clinic is that of oncology. Genomic analysis of
tainable. The Wilson and Jungner approach has been super- a patient’s malignancy is already guiding the choice of treat-
seded by a technology-led decision to make the diagnoses ments. However, it is also becoming clear in the area of rare
that the new technology permits. diseases that a gene-based understanding of disease is open-
Now that the technology of NGS has arrived, will this ing up immense therapeutic possibilities.
simply follow the precedent of TMS to take over labora- In malignancies, rational treatments can be developed
tory practice and to drive the clinical practices of explana- following the model of inhibitors of the Philadelphia fusion
tion, taking consent, and reporting results? If so, will it be protein, the BCR-ABL tyrosine kinase(54), and the generation
applied solely in private health care, or in state-sponsored of monoclonal antibodies against the disease-modifying
programs, too? These questions will have particular reso- protein products of tumor-amplified genes. Understanding
nance in United States, not only because that is where this the cellular role of DNA mismatch repair systems has led to
will be happening first on a large scale, but also because the use of Poly Adenosine diphosphate Ribose Polymerase
of the pattern of healthcare funding in the United States. (PARP) inhibitors in treating breast cancers in those with
Will health risks identified through newborn screening by constitutional BRCA gene mutations(55). The post-genome
WGS be managed as an entitlement through state health- move into proteomics also creates opportunities to classify
care, as would often occur for the continuing dietary needs tumors more sensitively (e.g., with certain lung cancers and
of children with phenylketonuria (PKU)? Or will a state leukemias), which can then lead to the improved targeting
program of newborn early diagnosis leave a generation of of therapies, although this chapter is not the place to review
children with knowledge of their state of risk but without these developments.
the resources to act upon this information? Among the rare (“orphan”) diseases, gene product
In addition, of course, there are some additional ques- replacement can sometimes be helpful, as in the inherited
tions about the management of the information generated forms of leptin deficiency(56) and growth hormone deficiency.
by WGS of healthy newborn infants. Will this be used by The administration of developmental signaling proteins
insurance companies to restrict access to health care? Will at the appropriate stage has effectively treated the mouse
the program store the sequence data indefinitely, as a life- with hypohidrotic ectodermal dysplasia(57), and human tri-
long resource, as suggested by Biesecker(51)? Will periodic als of this approach are proceeding. The gene-based under-
reanalyses be expected, as the ability to interpret genomic standing of disease mechanisms is also proving beneficial in
data improves? How will information be passed to the tuberous sclerosis (TS); thus, the tumors that cause so many
children as they mature? How will the liminal category of of the problems of TS can be stabilized or sometimes made
“patient in waiting”(52) be managed by professionals, parents, to regress with inhibitors of the mTOR pathway(58).
and then by the child? Will such results lead to unhelpful Understanding the precise mutation in a disease-
distress and anxiety(53)? associated gene may also open up therapeutic avenues. One
2 2 8 • P rincipl e s o f G e no m ic M e dicin e
screening for carriers of recessive Mendelian disease such as loci—that are associated with hypercholesterolemia, so that
cystic fibrosis(71), at least in the United Kingdom. This may the molecular diagnostic work involved in cascade testing
be because: (i) a family history of CAD has immediate rele- within families is substantial. Measurement of serum cho-
vance to the well-being of the practitioner’s patients instead lesterol remains the primary screening test, but molecular
of only a long-term relevance to potential future patients, diagnostics nevertheless improves the sensitivity of test-
and (ii) it does not raise the difficult and emotional topics ing within families once the family’s FH-associated muta-
of prenatal diagnosis and the termination of wanted (but tion has been identified(78,79). The question then arises as to
affected) pregnancies, as inevitably occurs in the context whether establishing the diagnosis of FH in an individual
of screening for carriers of recessively inherited disorders. through “traditional” biochemical measures of serum cho-
There are several single-gene disorders that nevertheless lesterol or with molecular methods of mutation detection
raise important issues for decisions about screening for the will have different consequences for patients’ perceptions of
complex, multifactorial, diseases. the disease and for their motivation to comply with medi-
cal recommendations. There is some evidence from experi-
ence with newborn screening that a DNA-based diagnosis
FA M I L I A L H Y P E RC H O L E S T E RO L E M I A ( FH )
of FH can lead to a sense of fatalism(80). This area of research
This autosomal dominant condition affects about one in is currently being pursued vigorously, given its potential
500 of the population in the United Kingdom and many public health importance, but whether these behavioral
developed countries; it predisposes to CAD at an early age. considerations will influence the choice of screening or
The prevention of such excess and early CAD is feasible in cascade-testing laboratory methods, however, is rather
many of those at risk by treatment with diet and medica- doubtful, as the test’s sensitivity, specificity, and cost are
tion (the statin drugs, inhibitors of HMG CoA reductase) likely to be decisive.
to achieve a reduction of serum cholesterol levels to within The general question of whether individuals with an
the normal range. Before such safe and effective treatment increased but readily modifiable risk of disease should
of hypercholesterolemia became available, measurement of be identified through population screening or through
serum cholesterol was often advocated to assess the risk of family-based cascade testing is important, and will have to
CAD, and that context—the awareness of risk without any be kept under review as laboratory methods develop over
highly effective remedy—raised a number of issues. time for each relevant disease. Family-based cascade test-
It has long been recognized that families have their ing for FH is certainly more cost-effective than population
own understanding of how they come to be at risk of heart screening(81,82), but there is debate about the “best” (most
disease—acknowledging the effects of smoking, diet, and effective and most ethical) way to achieve effective and
genetic factors in an intuitive fashion and without necessar- efficient cascading through a family(83). There are good rea-
ily formulating a clear mechanism through which such fac- sons for primary care practitioners to remain alert to those
tors could operate(72). While identifying those at increased at risk of FH—active, opportunistic case ascertainment—
risk of disease can motivate some to comply with medical and to consider serum cholesterol screening in healthy
advice, it can also lead to paradoxical consequences such as young and middle-aged adults. Once one case in a family
inappropriate feelings of fatalism or of invulnerability, per- has been identified biochemically, cascade testing may then
haps encouraging indulgence in harmful behaviors(73) and use molecular or biochemical methods, or both. Clearly,
therefore leading on to unhelpful health consequences(74,75). it is important that cases of FH are identified so that they
Interestingly, CAD is often seen as a disease predominantly can be offered treatment for their susceptibility to CAD;
affecting men, so there can be a readiness to accept that men this remains a Mendelian predisposition to cardiac disease,
are at increased risk but a reluctance to acknowledge that however, rather than a common, complex disorder.
women can also be affected(76). This could, of course, be There are small but important Mendelian subsets con-
highly relevant to decisions about who seeks screening for cealed within many of the common, complex disorders. We
disease risk, and who complies with behavioral recommen- have already referred to some of the familial cancer predis-
dations or takes a prescribed risk-reducing medication(77). positions, including the BRCA1 and BRCA2 loci, in which
Now that effective treatments are available for elevated mutations predispose to breast, ovarian, and other cancers,
serum cholesterol, will the introduction of genetic testing the apc locus associated with familial adenomatous polypo-
improve the management of at-risk individuals and fami- sis coli (FAP), and the mismatch repair loci associated with
lies? There are many mutations recognized—predominantly hereditary nonpolypotic colon cancer (HNPCC). In cur-
in the LDLR gene but also in the apoB gene and some other rent medical practice, the role of genetic testing for cancer
2 3 0 • P rincipl e s o f G e no m ic M e dicin e
Free ebooks ==> www.Ebook777.com
mutation rates—the latter will hitchhike along with the the functional consequences of genetic diversity for disease.
success of the former. These are reasons why the analysis of Other necessary elements of this are the Gluckman and
clinical and genetic data to give insights into the causation Hanson model of predictive adaptive responses, derived
of the common complex diseases (CCDs) will be very com- from Barker’s “thrifty phenotype” hypothesis of physiologi-
plex and difficult. This is not a reason for not setting out to cal adaptation and an understanding of history’s contribu-
perform such research, but it is a reason, in conjunction with tion to genetic variation through drift (e.g., fall in population
the limited effect of genetic information on health-related size during pogroms); selection for physical endurance, such
lifestyles, for caution in making claims about the ready clini- as during hazardous migrations, relating to Neel’s “thrifty
cal applicability of such research. gene” hypothesis(95); and Wilkinson and Pickett’s “Spirit
Indeed, the attempt to focus on genetic risk factors and Level” study on the biological consequences of the distri-
individualized, behavioral responses serves to distract our bution of wealth and power(96). The political consequences
attention from the risk factors that are the result of collec- of the biological understanding of these population effects
tive, societal practices and that may be far more amenable must not be forgotten(97,98). The physical, nutritional envi-
to collective solutions (such as controls on industrial pollu- ronment and the social setting work together to influence
tion, transport policy, vitamin supplementation of essential the development of disease through genetic, epigenetic, and
foods, etc.). To individualize the problems, when the only physiological mechanisms.
effective solutions are likely to be collective, is profoundly
unhelpful and likely to be driven by a specific political
agenda. C O N C LU S I O N
Challenges that remain, and that have to be tackled
afresh in each study, include the possibility of population Developments in the technologies of genomics—the
stratification (which might obscure real effects or generate high-throughput, genome-wide technologies, especially
confounding and misleading associations)(94), the appro- array CGH and NGS—are bringing rapid change to the
priate way to make allowance for the performance of mul- clinical practice of genetic medicine. The proportion of rare
tiple tests when assessing the significance of findings, and diseases for which a definitive diagnosis can be achieved
the difficulty of determining the biological basis of a true is rising, and the treatment of many malignancies, as well
SNP–disease association (identifying the causal factor in as some rare genetic disorders, is being greatly improved
disequilibrium with the SNP, lying in the same haplotype by these laboratory developments in genetic laboratory
block, if the SNP is not it). While methods are being devel- diagnostics.
oped to recognize SNPs that influence local and remote Some difficulties arise, however, from the very same
gene expression, the interactions between polymorphic features of these technologies that make them so attractive.
structural arrangements, gene expression, and disease asso- The sheer volume of sequence generated leads to difficul-
ciation remain challenging. ties in interpreting the findings, especially when novel vari-
Another temptation that some investigators succumb to ants are found whose pathological significance is unclear.
is the use of social categories as if they were discrete, bio- Furthermore, because these are genome-wide investigations,
logical entities. In particular, finding differences (e.g., in they will often generate findings incidental to the original
SNP allele frequencies) between racial or national groups indication for genetic testing, whether or not there are also
does not establish the two populations as discrete; clines in helpful findings pertinent to that original diagnostic ques-
allele frequencies across continents are common, so allele tion. These changes mean that the formerly clear distinction
frequencies will differ between arbitrarily defined groups in between a focused genetic test and genome-wide screening
the absence of biological boundaries. Indeed, the relevance collapses; the laboratory process and interpretive pipeline
of population group to the interpretation of genetic infor- may be the same, but the context determines what counts as
mation should diminish as the genome-wide causal factors a pertinent result from a focused genetic test or as an inci-
involved are understood, so that the inadequate proxy of dental product of genomic screening.
“race” will become irrelevant. The occurrence of these VUSs and IFs requires changes
Finally, we need to learn to bring together our under- in the conversation between patient and professional, espe-
standing of genetics, epigenetics, and selection, and the role cially the talk in which a proposed genetic investigation is
of these factors in shaping the variation in disease within explained, and in the process of documenting appropriate
and between populations. This can be seen as the founda- consent for the investigation. There must be an appropri-
tion stone of the grand synthesis required to understand ate process for deciding what incidental results—especially
www.Ebook777.com
about a child—should be disclosed to the patient (or their 9. DudokdeWit AC, Tibben A, Duivenvoorden HJ, Niermeijer MF,
Passchier J, Trijsburg RW, and the Rotterdam/Leiden Genetics
parents). If samples are stored long-term, or their test results Workgroup. Distress in individuals facing predictive DNA test-
are maintained on file for future reference, then specific ing for autosomal dominant late-onset disorders: comparing
plans should be established about when a reinterpretation questionnaire results with in-depth interviews. Am J Med Genet.
1998;75:62–74.
would be performed and how to contact the patient in the 10. Decruyenaere M, Evers-Kiebooms G, Cloostermans T, et al.
future if additional findings emerge from our developing Psychological distress in the 5-year period after predictive testing for
collective ability to interpret genome data. Huntington’s disease. Eur J Hum Genet. 2003;11:30–38.
11. Sobel S, Cowan DB. Impact of genetic testing for Huntington dis-
When a family is unaware of the potential finding ease on the family system. Am J Med Genet. 2000;90:49–59.
of important, family-specific IFs when their child has a 12. Harper PS. Presymptomatic testing for late-onset genetic disor-
genome assessment, then the considerations that usually ders. Lessons from Huntington’s disease. In: Harper P, Clarke AJ.
Genetics, Society and Clinical Practice. Oxford, UK: Bios Scientific
caution against revealing predictive genetic information Publishers; 1997:31–48.
about a child in relation to a future risk of an adult-onset 13. Kessler S, ed. Resta RC. Psyche and Helix. Psychological Aspects of
disease may no longer apply. The alternative to generating Genetic Counselling. New York and Chichester, England: Wiley-Liss;
2000.
and disclosing such predictive information to an unpre- 14. Soldan J, Street E, Gray J, Binedell J, Harper PS. Psychological
pared family, who may then have the opportunity to pre- model for presymptomatic test interviews: lessons learned from
pare in whatever way is possible, is for all parties to remain Huntington disease. J Genet Counsel. 2000;9(1):15–31.
15. Tibben A. Genetic counselling and presymptomatic test-
ignorant until someone in the family comes to be affected, ing. In: Bates G, Harper PS, Jones L, eds. Huntington’s Disease.
with even less opportunity to prepare. How to manage New York, USA: Oxford University Press; 2002:198–248.
16. Codori AM, Brandt J. Psychological costs and benefits of predictive
these human, interactive consequences of the technological testing for Huntington’s disease. Am J Med Genet (Neuropsychiatric
developments in genomics is being debated among profes- Genet). 1994;54:174–184.
sionals with a passionate intensity. These challenges will 17. Richards MPM. Families, kinship and genetics. In: Marteau T,
Richards M, eds. The Troubled Helix. Cambridge, UK: Cambridge
become simpler over the next decade as laboratory scientists University Press; 1996:249–273.
and clinical practitioners jointly develop their approaches 18. Broadstock M, Michie S, Marteau T. Psychological consequences
to handling genomic uncertainty and negotiate strategies of predictive genetic testing: a systematic review. Eur J Hum Genet.
2000;8:731–738.
for an appropriate narrowing or widening of the focus of 19. Burke W, Pinsky LE, Press NA. Categorizing genetic tests to iden-
investigation. As our collective experience of interpreting tify their ethical, legal and social implications. Am J Med Genet.
genomic investigations accumulates over the years, the cor- 2001;106:233–240.
20. Burke W, Zimmern RL. Ensuring the appropriate use of genetic
responding level of justified confidence in these interpreta- tests. Nature Rev Genet. 2004;5:955–959.
tions will improve. 21. PHG Foundation. 2011. Next steps in the sequence: the implica-
tions of whole genome sequencing for health in the UK.
22. Lim J, Macluran M, Price M, Bennett B, Butow P and the kCon-
Fab Psychosocial Group. Short- and long-term impact of receiving
R EFE R E N C ES genetic mutation results in women at increased risk for hereditary
breast cancer. J Genet Counsel. 2004;13(2):115–133.
1. PHG Foundation. 2013. Managing incidental and pertinent find- 23. Hallowell N, Foster C, Eeles R, Ardern-Jones A, Murday V, Watson
ings from WGS in the 100,000 Genome Project (www.phgfounda- M. Balancing autonomy and responsibility: the ethics of generating
tion.org.uk). and disclosing genetic information. J Med Ethics. 2003;29:74–83.
2. Elwyn G, Gray J, Clarke A. Shared decision making and 24. Michie S, Smith JA, Senior V, Marteau TM. Understanding why
non-directiveness in genetic counselling. J Med Genet. negative genetic test results sometimes fail to reassure. Am J Med
2000;37:135–138. Genet. 2003;119A:340–347.
3. Fulda KG, Lykens K. Ethical issues in predictive genetic testing: a 25. Wevers MR, Ausems MG, Verhoef S, Bleiker EM, Hahn DE,
public health perspective. J Med Ethics. 2006;32:143–147. Hogervorst FB, et al. Behavioral and psychosocial effects of rapid
4. Wolff G, Jung C. Nondirectiveness and genetic counseling. J Gen genetic counseling and testing in newly diagnosed breast cancer
Counsel. 1995;4:3–25. patients: design of a multicenter randomized clinical trial. BMC
5. Clarke A. The process of genetic counselling: beyond nondirective- Cancer. 10 Jan 2011;11:6.
ness. In: Harper P, Clarke AJ. Genetics, Society, and Clinical Practice. 26. Lindblad AN. To test or not to test: an ethical conflict with pres-
Oxford, UK: Bios Scientific Publishers; 1997:179–200. ymptomatic testing of individuals at 25% risk for Huntington’s dis-
6. McAllister M. Predictive genetic testing and beyond: a theory of order. Clin Genet. 2001;60:442–446.
engagement. J Health Psychol. 2002;7(5):491–508. 27. Savulescu J, Kahne G. The moral obligation to create children with
7. Sarangi S, Bennert K, Howell L, Clarke A, Harper P, Gray J. the best chance of the best life. Bioethics. 2009;23:274–290.
Initiation of reflective frames in counselling for Huntington’s dis- 28. Bekker H, Modell M, Denniss G, et al. Uptake of cystic fibro-
ease predictive testing. J Genet Counsel. 2004;13:135–155. sis testing in primary care: supply push or demand pull? BMJ.
8. Sarangi S, Bennert K, Howell L, Clarke A, Harper P, Gray J. (Mis) 1993;306:1584–1586.
alignments in counselling for Huntington’s disease predictive 29. HGC (Human Genetics Commission). Increasing Options,
testing: clients’ responses to reflective frames. J Genet Counsel. Informing Choice: A Report on Preconception Genetic Testing and
2005;14:29–42. Screening. London: Department of Health; 2011.
2 3 2 • P rincipl e s o f G e no m ic M e dicin e
30. Grody WW. Expanded carrier screening and the law of unin- 50. Parsons EP, Clarke AJ, Hood K, Lycett E, Bradley DM. Newborn
tended consequences: from cystic fibrosis to fragile X. Genet Med. screening for Duchenne muscular dystrophy: a psychosocial study.
2011;13(12):996–997. Arch Dis Child Fetal Neonatal Ed. 2002;86:F91–F95.
31. Lazarin GA, Haque IS, Nazareth S, Iori K, Patterson AS, Jacobson 51. Biesecker LG. Opportunities and challenges for the integration of
JL, et al. An empirical estimate of carrier frequencies for 400+ causal massively parallel genomic sequencing into clinical practice: lessons
Mendelian variants: results from an ethnically diverse clinical sam- from the ClinSeq Project. Genet Med. 2012;14(4):393–398.
ple of 23,453 individuals. Genet Med. 2013;15(3):178–186. 52. Timmermans S, Buchbinder M. Patients-in-waiting: living between
32. de Wert GM, Dondorp WJ, Knoppers BM. Preconception care and sickness and health in the genomics era. J Health Soc Behav.
genetic risk: ethical issues. J Community Genet. 2012;3(3):221–228. 2010;51(4):408–423.
33. Faas BH, Feenstra I, Eggink AJ, et al. Non-targeted whole genome 53. Goldenberg AJ, Sharp RR. The ethical hazards and program-
250K SNP array analysis as replacement for karyotyping in fetuses matic challenges of genomic newborn screening. JAMA. 1 Feb
with structural ultrasound anomalies: evaluation of a one-year expe- 2012;307(5):461–462.
rience. Prenat Diagn. 2012;32(4):362–370. 54. Savage DG, Antman KH. Imatinib mesylate—a new oral targeted
34. Benn PA. Prenatal counseling and the detection of copy-number therapy. N Engl J Med. 2002;346(9):683–693.
variants. Genet Med. 2013;15(4):316–317. 55. Lee JM, Ledermann JA, Kohn EC. PARP inhibitors for BRCA1/2
35. Ehrich M, Deciu C, Zwiefelhofer T, Tynan JA, Cagasan L, Tim R, mutation-associated and BRCA-like malignancies. Ann Oncol. Nov
et al. Noninvasive detection of fetal trisomy 21 by sequencing of 12, 2013. Online Access, https://www.acmg.net/docs/Incidental_
DNA in maternal blood: a study in a clinical setting. Am J Obstet Findings_in_Clinical_Genomics_A_Clarification_081413.pdf.
Gynecol. 2011;204(3):205.e1–11. 56. Oral EA, Simha V, Ruiz E, et al. Leptin-replacement therapy for
36. Palomaki GE, Kloza EM, Lambert-Messerlian GM, et al. DNA lipodystrophy. N Engl J Med. 2002;346(8), 5700578.
sequencing of maternal plasma to detect Down syndrome: an 57. Gaide O, Schneider P. Permanent correction of an inherited
international clinical validation study. Genet Med. 2011;13(11): ectodermal dysplasia with recombinant EDA. Nature Med.
913–920. 2003;9:614–618.
37. Borry P, Evers-Kiebooms G, Cornel MC, Clarke A, Dierickx K, 58. Davies DM, de Vries PJ, Johnson SR, et al. Sirolimus therapy for angi-
and the Public and Professional Policy Committee (PPPC) of omyolipoma in tuberous sclerosis and sporadic lymphangioleiomyo-
the European Society of Human Genetics (ESHG). Genetic test- matosis: a phase 2 trial. Clin Cancer Res. 2011;17(12):4071–4081.
ing in asymptomatic minors: background considerations towards 59. Du M, Liu X, Welch EM, Hirawat S, Peltz SW, Bedwell DM.
ESHG Recommendations. Eur J Hum Genet. 2009;17(6):711–719 PTC124 is an orally bioavailable compound that promotes suppres-
(Recommendations, 720–721). sion of the human CFTR-G542X nonsense allele in a CF mouse
38. European Society of Human Genetics. Genetic testing in asymp- model. Proc Natl Acad Sci U S A. 2008;105(6):2064–2069.
tomatic minors: recommendations of the European Society of 60. Aartsma-Rus A, Fokkema I, Verschuuren J, et al. Theoretic applica-
Human Genetics. Eur J Hum Genet. 2009;17(6):720–721. bility of antisense-mediated exon skipping for Duchenne muscular
39. British Society for Human Genetics. 2010. Genetic Testing of
dystrophy mutation. Hum Mutat. 2009;30(3):293–299.
Children. Report of a working party of the British Society for 61. Jih KY, Hwang TC. Vx-770 potentiates CFTR function by promot-
Human Genetics. http://www.bsgm.org.uk/media/678741/gtoc_ ing decoupling between the gating cycle and ATP hydrolysis cycle.
booklet_final_new.pdf Proc Natl Acad Sci U S A. 2013;110(11):4404–4409.
40. American Academy of Pediatrics and the American College of 62. Benkhelifa-Ziyyat S, Besse A, Roda M, et al. Intramuscular
Medical Genetics and Genomics. Ethical and policy issues in genetic scAAV9-SMN injection mediates widespread gene delivery to the
testing and screening of children. Pediatrics. 2013;131:620–622. spinal cord and decreases disease severity in SMA mice. Mol Ther.
41. Ross LF, Saal HM, David KL, Anderson RR; American Academy of Feb 2013;21(2):282–290.
Pediatrics; American College of Medical Genetics and Genomics. 63. Guy J, Gan J, Selfridge J, Cobb S, Bird A. Reversal of neuro-
Technical report: ethical and policy issues in genetic testing and logical defects in a mouse model of Rett syndrome. Science
screening of children. Genet Med. 2013;15(3):234–245. 2007;315:1143–1147.
42. Green RC, Berg JS, Grody WW, et al. ACMG recommendations 64. Abdala AP, Dutschmann M, Bissonnette JM, Paton JF. Correction
for reporting of incidental findings in clinical exome and genome of respiratory disorders in a mouse model of Rett syndrome. Proc
sequencing. Genet Med. 2013;15(7):565–574. Natl Acad Sci U S A. 2010;107:18208–18213.
43. American College of Medical Genetics and Genomics. Incidental 65. Barker DJP, ed. Fetal and Infant Origins of Adult Disease.
Findings in Clinical Genomics: A Clarification. A policy statement London: British Medical Journal; 1992.
of the American College of Medical Genetics and Genomics. 66. Gluckman P, Hanson M. The Fetal Matrix. Evolution, Development,
Bethesda, MD: ACMGG; 2013. and Disease. Cambridge, UK: Cambridge University Press; 2005.
44. Feinberg J. The child’s right to an open future. In: Aiken W, La 67. Pembrey M, Bygren LO, Kaati G, et al., and the ALSPAC Study
Fallette H, eds. Whose Child? Children’s Rights, Parental Authority Team. Sex-specific, male-line transgenerational responses in humans.
and State Power. Totowa, NJ: Littlefield, Adams; 1980:124–153. Eur J Hum Genet. 2006;14:159–166.
45. Davis D. Genetic Dilemmas: Reproductive Technology, Parental 68. Humphries SE, Talmund PJ, Hawe E, Bolla M, Day INM, Miller GJ.
Choices, and Children’s Futures. 2nd ed. Oxford, UK: Oxford Apolipoprotein E4 and coronary heart disease in middle-aged men
University Press; 2010. who smoke: a prospective study. Lancet. 2001;358:115–119.
46. Hendriks KSWH, Grosfeld FJM, van Tintelen JP, et al. Can parents 69. Ulvik A, Evensen ET, Lien EA, et al. Smoking, folate and methy-
adjust to the idea that their child is at risk of sudden death? Am J lenetetrahydrofolate reductase status as interactive determinants
Med Genet. 2005;138A:107–112. of adenomatous and hyperplastic polyps of colorectum. Am J Med
47. Hercher L, Bruenner G. Living with a child at risk for psychotic ill- Genet. 2001;101:246–254.
ness. Am J Med Genet. 2008;146A; 2355–2360. 70. Wang X, Zuckerman B, Pearson C, et al. Maternal cigarette smok-
48. Wilson JMG, Junger G. Principles and Practices of Screening for ing, metabolic gene polymorphism, and infant birth weight. JAMA.
Disease. Geneva: World Health Organisation; 1968. 2002;287:195–202.
49. National Screening Committee. Second Report of the National
71. Payne Y, Williams M, Cheadle J, et al. Carrier screening for cystic
Screening Committee. London: Health Departments of the United fibrosis in primary care: evaluation of a project in South Wales. Clin
Kingdom; 2000. Genet. 1997;51:153–163.
2 3 4 • P rincipl e s o f G e no m ic M e dicin e
15.
BIOBANKING FOR GENOMICS-BASED
TRANSLATIONAL MEDICINE
Steven J. Madore
235
and compile medical histories and lifestyle and genealogi- donors or unaffected controls. Many epidemiological
cal information from the original sample donor. Samples studies are focused on or include the goal of biomarker
stored in large population biobanks have proven to be identification(31). Collection of the various sample types
extremely useful as input material for genome-wide associa- and accompanying relevant medical records from study
tion studies (GWAS) to identify genes that contribute to participants is often coordinated by the biobank. The
human disease(22–25). degree and accuracy of sample-linked data adds tremen-
Critical to the downstream usefulness of biospecimens dous value to the sample when linking genetic findings
is that they be obtained under appropriate ethical and legal with donor phenotype.
criteria(26). Proper consent for biospecimen acquisition Biospecimens are collected at sites away from the biore-
is least problematic when donors are healthy and compe- pository; thus the mode of packaging and transportation is
tent adults. In contrast, consent considerations can arise a critical parameter in assuring biospecimen stability during
when potential donors are sick and/or vulnerable(27–30). transit(32–35). This may include the addition of wet or dry ice
While the consent of participants is usually required before in sufficient quantity and in insulated containers to account
samples can be used in research, the nature of this consent for the anticipated period of transit from collection site
and how it is obtained can vary widely. Given the impor- to biobanking facility. Blood packages shipped overnight
tance of human biospecimens, their collection and pro- by commercial carrier may encounter extreme seasonal
cessing must adhere to standards that maintain biological temperatures, and careful precautions need to be taken to
integrity as well as ensure that relevant patient-related and ensure that the shipping containers include adequate pro-
biospecimen-specific information remains linked to each tection against extreme ambient temperature deviations.
specimen(26–30). While a full discussion of the issue of sample Prolonged storage of blood at room temperature prior to its
consent is extremely important, it is beyond the scope of processing can affect the stability of the macromolecules like
this review; however, researchers requesting biospecimens DNA and proteins, as well as the efficiency of lymphocyte
from established biobanks need to be aware of legal and separation by Ficoll density centrifugation(36–40). Biofluids
ethical implications of obtaining properly consented sam- such as plasma, serum, urine, or cerebrospinal fluid (CSF)
ples for their research programs. are shipped frozen on dry ice. Established cell lines can be
Human biospecimen sample types found in biobanks shipped as live cultures in growth media or cryopreserved
include normal and diseased tissue, whole blood, plasma, under liquid nitrogen vapor.
serum, urine, cerebral spinal fluid, and saliva, while tumor Human tissue biobanking—the procurement of tissue
banks consist of well-annotated tumor tissue (either samples from live donors or cadavers—requires special con-
stored frozen or as formalin-fixed paraffin-embedded siderations related to minimizing pre-analytical variation.
[FFPE] blocks or slides) obtained from cancer patients. Pre-analytical variables, such as the time the tissue remains
Immortalized cell lines derived by transformation of periph- at room temperature, can cause significant variability and
eral blood premature B-cells with Epstein-Barr virus (EBV) bias in downstream molecular analysis, including reactive
remain a rich source for genomic-based research. Other changes that begin with oxidative, hypoxic, and meta-
types of cell lines found in biobanked collections include bolic stress, and culminate in apoptosis(41). Standardized
lines derived from tumor biopsies, as well as those that have methods for tissue procurement to minimize sources of
been genetically manipulated to express mutant or altered pre-analytical variation as well as the development of bio-
versions of certain genes. Biobanks may also contain large markers indicative of the biological state of the tissue prior
collections of animal and plant specimens, as well as various to freezing remain active areas of research in biobanking
bacterial and viral strains. Finally, reagents useful for basic science(41–43).
research, including genomic DNA, RNA, plasmid con-
structs, purified proteins, tissue extracts, and polyclonal and
monoclonal antibodies, can also be important components B I O B A N K E D S A M P LE T Y P ES
of the biorepository. A N D G E N O M I C T E C H N O L O G I ES
So what are the origins of biospecimens stored within
biobanks? Probably the most valuable samples for transla- Upon arrival at the biobank, samples are usually assigned
tional research originate from well-planned, large clinical a unique reference ID to facilitate tracking during labora-
studies designed with specific aims and outcomes. These tory processing and inventory during storage and distribu-
studies are often longitudinal and powered to recruit a tion. While some submitted samples remain unprocessed
sufficient patient population that also includes healthy and are placed in safe storage after accessioning, biobanks
2 3 6 • P rincip l e s o f G e no m ic M e dicin e
may extract or isolate important components or “prod- in GWAS technologies, human genetic association studies
ucts” from submitted samples like whole blood or saliva. have moved from looking at a single SNP or a small set of
Large epidemiological studies focused on genetics include SNPs to very large studies involving international consor-
whole-blood specimens that yield sufficient quantities of tium analyzing genome-wide analysis of tens of thousands
high-quality DNA amendable to several genomic analysis of individuals(55–57). The increase in cohort size and breadth
platforms. Human blood is collected into polypropylene of coverage across the genome lends authenticity and reli-
tubes containing the anticoagulants ethylenediaminetet- ability to GWAS studies(54).
raacetic acid (EDTA), heparin, or acid citrate dextrose. The availability of large collections of archived saliva
If shipped properly, DNA can be extracted directly from from GWA studies represents a valuable resource for
fresh blood or from blood that has been stored frozen advancing the study of the microbiota in the mouth as
at –80 C. Most large biobanks rely on robotic instrumen- well(58–60). The salivary microbiota are a potential diagnos-
tation for DNA extraction, using a number of different tic indicator of several diseases, and because the oral cavity
extraction methodologies that are scalable and appropri- is often the entry point for bacteria, there is the likelihood
ate for automation(44). These include salt precipitation that interactions between the saliva microbiome and other
(“salting out”), and direct capture on silica membranes or microbiomes in the human body (particularly within the
magnetic beads, with typical DNA yields ranging from intestinal tract) play a role in human disease(58–60). Genomic
100 µg to 400 µg per 10 ml of whole blood(45–47). Purified DNA purified from human saliva can be analyzed by
genomic DNA must undergo rigorous quality assessment. high-throughput DNA sequencing to characterize the
For example, DNA purity can be determined using spec- enormous diversity of microbial organisms present in the
trophotometry, and DNA size by agarose gel electropho- human salivary microbiome. This characterization should
resis or lab-on-a-chip techniques using a set of molecular provide insight into what role the oral microbial commu-
size standards for comparison(48). Due to shearing forces nity has in human health and disease(58).
introduced during most extraction methods, average DNA Another rich source of genomic DNA found in bio-
size with typical extraction methods described above aver- banks is that isolated from lymphoblastoid cell lines, or
ages around 100,000 base pairs, which is sufficient for most LCLs, which are EBV-immortalized B-cells derived from
existing genomic technologies(44–47). donor blood. LCLs, which can be grown easily in the labo-
Whole blood collected in PAXgene tubes containing ratory and represent a renewable source of genomic DNA
a stabilization reagent to neutralize intrinsic RNase activ- from the original donor, have been used in a wide variety
ity can be used for the extraction of high-quality total of genomic studies, including GWAS and gene-expression
RNA that can serve as input material for high-throughput profiling, and next-generation sequencing data are now
expression-profiling analysis. Whole blood can also serve as publicly available for hundreds of established LCLs. Large,
raw material for the isolation of live cells such as peripheral well-characterized collections of LCLs that have been used
blood mononuclear cells (PBMCs). Primary fibroblasts can in pharmacogenomic studies include the HapMap and
be isolated from skin-punch biopsies by outgrowth in tissue Human Variation Panel lymphoblastoid cell lines(61,62)—
culture dishes using appropriate growth media. Blood can both of which are publicly available from the Coriell Institute
be further processed into plasma or serum for proteomics for Medical Research (www.coriell.org). Researchers have
studies. Biofluids such as urine and cerebral spinal fluid also used cultured LCLs as a cost-effective, unlimited cell
can serve as substrate for both proteomic and metabolomic resource to identify genes and variants capable of predicting
analyses. both drug response and drug toxicity. The utility of LCLs in
DNA samples of sufficient quantity and quality for cancer pharmacogenomic studies has recently been demon-
GWAS can be isolated from saliva, which can be easily strated by the identification of polymorphisms associated
self-collected in prefabricated vessels that contain a stabili- with clinical phenotypes such as survival and response to
zation solution(49–52). Once collected, the saliva can be stored chemotherapy(63).
for months at room temperature with minimal effects on the The recent development of “next-generation sequenc-
yield or quality of extracted genomic DNA(53). The analysis ing” (NGS) technologies holds great promise in accel-
of saliva-derived genomic DNA by ultra-high-throughput erating genomic studies by increasing efficiency and
genotyping, in which up to one million distinct single-base significantly reducing the cost of generating whole
positions can be interrogated in parallel, has greatly facili- genome sequences. Most new techniques require exten-
tated the ability to assess common variation across the sive biochemical labeling and sample preparation, and
genome(54). Along with the ease of collection and advances do not allow long, single-molecule read lengths to be
2 3 8 • P rincip l e s o f G e no m ic M e dicin e
shown to be involved in a variety of biological processes. technology engine towards key discoveries and advance-
Some miRNAs exhibit differential expression levels in can- ments in translational research.
cer and have been shown to affect cellular transformation,
carcinogenesis, and metastasis, acting either as oncogenes or
as tumor suppressors(89,90). Careful attention must be paid AC K N OW LE D G ME N T S
to RNA extraction methods so that small RNAs, including
miRNAs, are retained in the final RNA preparation. For The author wishes to apologize to all whose published works
blood collected in PAXgene tubes, there are several com- were not referenced in the manuscript due to space limitations.
mercially available kits amenable to automation that use Special thanks to Drs. Joseph Jarvis, Christine Beiswanger,
silica membranes embedded in disposable spin columns and Victoria Kelly for their helpful review of this manuscript.
or magnetic bead technology to capture pure RNA from a
biological sample. Binding and elution conditions must be
optimized so that the final purified RNA sample includes
R EFE R E N C ES
the miRNA fraction.
Fluid biospecimens such as blood, urine, saliva, CSF, 1. Goldblatt EM, Lee WH: From bench to bedside: the growing use
seminal fluid, and bronchial lavage contain cell-free pro- of translational research in cancer medicine. Am J Transl Res 2:1–18,
2000.
teins, nucleic acids, lipids, and metabolites, all of which 2. Woolf SH: The meaning of translational research and why it mat-
can be utilized as biomarkers of disease. The availability of ters. JAMA 299:211–213, 2008.
high-quality biospecimens is a requirement for novel bio- 3. Zerhouni EA: Translational and clinical science—time for a new
vision. NEJM 353:1621–1623, 2005.
marker discovery that can promote disease risk-assessment, 4. Marincola FM: Translational medicine—a two-way road. J Transl
diagnosis, management, and treatment(91). Biospecimen Med 1:1, 2003.
quality depends upon a number of pre-analytical variables 5. International Human Genome Sequencing Consortium: Finishing
the euchromatic sequence of the human genome. Nature 431:931–
such as collection and processing techniques, as well as 945, 2004.
freezing/thawing stability and long-term storage stabil- 6. Venter JC, Adams MD, Myers EW, et al.: The sequence of the
ity(91–93). Millions of fluid biospecimens are currently stored human genome. Science 291:1304–1351, 2001.
7. NCI Dictionary of Cancer Terms, at http://www.cancer.gov/
in biobanks across the world, and with appropriate annota- dictionary?cdrid=446543.
tion and consent, and if properly collected, processed, and 8. Hawkins N, de Vries J, Boddington P, Kaye J, Heeney C: Planning
stored, archived fluid biospecimens are valuable raw mate- for translational research in genomics. Genome Med 1:87–94, 2009.
9. Kumar, D: Genomic medicine: a new frontier of medicine in the
rial for translational research. twenty-first century. Genomic Med 1:3–7, 2007.
10. Loft S, Poulsen HE: Cancer risk and oxidative DNA damage in
man. J Mol Med 74:297–312, 1996.
11. Cambon-Thomsen A, Ducoumau P, Gourraud P-A, Pontille
C O N C LU S I O N S A N D S U MM A RY D: Biobanks for genomics and genomics for biobanks. Comp Funct
Genom 4:628–634, 2003.
The availability of high-quality, well-annotated bio- 12. Department of Health and Human Services (Internet). Personalized
Healthcare, at http://www.hhs.gov/myhealthcare/news/presonalized-
specimens to both basic and clinical research efforts is a healthcare-9-2007.html.
critical component for advancing biomedical research. 13. Compton C: Getting to personalized cancer medicine—taking out
the garbage. Cancer 110:1641–1643, 2007.
Because biospecimen quality is affected by a number of 14. Park A: Ten ideas changing the world right now. Time, 173(11):17–
pre-analytical factors that are introduced through the 22, 2009.
variety of biospecimen collection, processing, and storage 15. Baker, M: Biorepositories: Building better biobanks. Nature
486:141–146, 2012.
procedures, there is a need to establish best practices and 16. Riegman PHJ, Morente MM, Betsou F, de Blasio P, Geary P, the
recognize and further advance the field of “biospecimen Marble Arch International Working Group on Biobanking for
science”—the emerging field of study that is attempting Biomedical Research: Biobanking for better healthcare. Mol Oncol
2:213–222, 2008.
to quantify and control such variability(94). Various efforts 17. Steinberg K, Beck J, Nickerson D, et al.: DNA banking for epidemi-
are now under way around the world to establish research ologic studies: a review of current practices. Epidemiology 13:246–
programs, evidence-based biospecimen protocols, and 254, 2002.
18. Murtagh MJ, Demir I, Harris JR, Burton PR: Realizing the promise
standards to improve the overall quality of biospecimens of population biobanks: a new model for translation. Hum Genet
for research(95). Thus, biobanks will continue to be a valu- 130:333–345, 2011.
able resource for biological raw material for both basic 19. Swede H, Stone CL, Norwood AR: National population-based bio-
banks for genetics research. Genet Med 9:141–149, 2007.
and clinical researchers, and hold great promise as provid- 20. Austin MA, Harding S, McElroy C: A comparison of eight pro-
ers of the raw material necessary for fueling the genomic posed international genetic databases. Comm Genet 6:37–45, 2003.
2 4 0 • P rincip l e s o f G e no m ic M e dicin e
69. Manrao EA, Derrington IM, Laszlo AH, et al.: Reading DNA at 82. Stunnenberg HG, Verneulen M: Towards cracking the epigen-
single-nucleotide resolution with a mutant MspA nanopore and etic code using a combination of high-throughput epigenomics
Phi29 DNA polymerase. Nat Biotechnol 30:349–353, 2012. and quantitative mass spectrometry-based proteomics. Bioessays
70. Zhang M, Zhang Y, Scheuring CF, Wu C-C, Dong JJ, Zhang 33:547–551, 2011.
H-B: Preparation of megabase-sized DNA from a variety of organ- 83. Chi P, Allis CD, Wang GG: Covalent histone modifications—mis-
isms using the nuclei method for advanced genomics research. Nat written, misinterpreted and mis-erased in human cancers. Nat Rev
Protocols 7:467–478, 2012. Cancer 10:457–469, 2010.
71. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D: Global 84. Li XY, Biggin MD: Genome-wide in vivo cross-linking of
cancer statistics. CA Cancer J Clin 61:69–90, 2011. sequence-specific transcription factors. Methods Mol Biol. 809:
72. Cheng KWA, Tanyi J, Mills GB: Genomic technologies in cancer 3–26, 2012.
research, drug discovery and development. CME J Gynecol Oncol 85. Wang Z, Gerstein M, Snyder M: RNA-Seq: a revolutionary tool for
7:16–24, 2002. transcriptomics. Nat Rev Genet 10:57–63, 2009.
74. Bell DW: Our changing view of the genomic landscape of cancer. 86. Ozolak F, Milos PM: RNA sequencing: advances, challenges and
J Pathol 220:231–243, 2010. opportunities. Nat Rev Genet 12:87–98, 2011.
75. Greenman C, Stephens P, Smith R, et al.: Patterns of somatic muta- 87. Faghihi MA, Wahlstedt C: Regulatory roles of natural antisense
tion in human cancer genomes. Nature 446:153–158, 2007. transcripts. Nat Rev Mol Cell Biol 10:637–643, 2009.
76. Jones S, Zhang X, Parsons DW et al.: Core signaling pathways in 88. Esteller M: Non-coding RNAs in human disease. Nat Rev Genet
human pancreatic cancers revealed by global genomic analyses. 12:861–874, 2011.
Science 321:1801–1806, 2008. 89. Ryu S, Joshi N, McDonnell K, et al.: Discovery of novel human
77. Parsons DW, Jones S, Zhang X, et al.: An integrated genomic analy- breast cancer microRNAs from deep sequencing data by analysis of
sis of human glioblastoma multiforme. Science 321:1807–1812, pri-microRNA secondary structures. PLoS ONE 6: e16403, 2011.
2008. 90. Medina PP, Slack FJ: MicroRNAs and cancer: an overview. Cell
79. Taylor BS, Schultz N, Hieronymus H, et al.: Integrative genomic Cycle 7:2485–2492, 2008.
profiling of human prostate cancer. Cancer Cell 18:11–122, 2010. 91. Hubel A, Aksan A, Skubitz APN, Wendt C, Zhong X: State of
78. Kan Z, Jaiswal BS, Stinson J, et al.: Diverse somatic mutation pat- the art in preservation of fluid biospecimens. Biopreserv Biobank,
terns and pathway alterations in human cancers. Nature 466:869– 9:237–244, 2011.
873, 2010. 92. Landi MT, Caporaso N: Sample collection, processing and storage.
79. Gilbert MT, Haselkorn T, Bunce M, et al.: The isolation of nucleic IARC Sci Publ 142:223–236, 1997.
acids from fixed, paraffin-embedded tissues—which methods are 93. Elliot P, Peakman TC, UK Biobank: The UK Biobank sample
useful when? PLoS ONE 2:e537, 2007. handling and storage protocol for the collection, processing and
80. Schweiger MR, Kerick M, Timmermann B, et al.: Genome-wide archiving of human blood and urine. Int J Epidemiol 37:234–244,
massively parallel sequencing of formaldehyde fixed-paraffin embed- 2008.
ded (FFPE) tumor tissues for copy number and mutational analysis. 94. Vaught JB, Henderson MK, Compton CC: Biospecimens and
PLoS ONE 4:e5548, 2009. biorepositories: from afterthought to science. Cancer Epidemiol
81. Kerick M, Isau M, Timmermann B, et al.: Targeted high throughput Biomarkers Prev 21:253–5, 2012.
sequencing in clinical settings: formaldehyde fixed-paraffin embed- 95. Moore HM, Compton CC, Alper J, Vaught JB: International
ded (FFPE) tumor tissues, input amount and tumor heterogeneity. approaches to advancing biospecimen science. Cancer Epidemiol
BMC Med Genomics 4:68–72, 2011. Biomarkers Prev 20:729–32, 2011.
242
Table 16.1 INDIVIDUALS WHO SHOULD BECOME The U.S. Department of Health and Human Services
EDUCATED ABOUT MEDICAL GENETICS identified public education in genetics as one of a number
GENERAL PUBLIC of critical areas that need to be reviewed and addressed as
the healthcare system prepares for genomic medicine and
Non-Health Professionals:
direct-to-consumer genetic testing. For example, when
Educational administrators
patients were surveyed in 2007, most reported receiving no
Teachers
genetic information from their health care provider about a
Legislators and their aides condition that was known to be familial[10]. In response, the
Policymakers National Human Genome Research Institute (NHGRI)
Health Professionals: established an Education and Community Involvement
Physicians generally, especially Branch that hosts a website with a rich variety of informa-
Primary care physicians tion for the public (http://www.genome.gov/Education)
Oncologists and also provides a compendium of information from other
Cardiologists sources[11]. A recent initiative is GeneEd, designed for high
Neurologists school students, teachers, and the general public[12].
Medical geneticists Notable recent initiatives include “DNA Day” on April
Genetic counselors
25 (in honor of the anniversary of the publication of Watson
and Crick’s seminal paper on the double helix) sponsored
Nurses and nurse practitioners
by the NHGRI and the ASHG[13]. One component is a
Physician assistants
national competition among high school students to select
the most outstanding essays submitted on a topic that var-
focus on concepts and investigation (not “what to think,” ies each year[14]. Additionally, the ASHG sponsors a one-day
but “how to think”), but the fact that the program persists workshop for high school science teachers and students
to this day is a testimony to its many successes[5]. More than immediately preceding their annual national meeting[15].
three decades ago, a survey in Canada indicated that high Colleges and universities may partner with local commu-
school students had a strong preference for learning more nity schools to enhance teaching and learning in STEM. For
about genetics, including genetic screening[2]. However, example, Johns Hopkins, supported by an NSF grant, has
deficiencies in the education of our youth are not limited to developed the STEM Achievement in Baltimore Elementary
biology and genetics. Schools, which initially will benefit 1,600 students in grades
Students in the United States continue to fall behind three through five in nine elementary schools[16]. A project
their peers in other developed nations in the quality and at Harvard, the Personal Genetics Education Project (pged.
the quantity of STEM (science, technology, engineering org) has focused on high school students and teachers
and mathematics) subjects from primary through second- through workshops and curriculum development[17].
ary school. One result is that fewer American youth major Planned for 2013 is a major exhibition at the
in these subjects when they do have an opportunity to Smithsonian Institution in Washington developed in col-
pursue higher education. Another result is that they are laboration with the NHGRI[18]. The project is funded
less likely to be interested in or understand reports of new by the National Institutes of Health (NIH) Foundation
scientific advances that appear in all of the various popular (non-governmental funds), the Smithsonian, and private
media. A number of organizations are working to improve sources. After a stay of a year in Washington, the exhibit will
STEM learning, especially the American Association for travel internationally.
the Advancement of Science (AAAS) and the National Providers of genetic services, mainly commercial ones
Science Foundation (NSF)[6–8]. The federal Department thus far, have also developed educational materials for
of Education also identifies improving STEM learning as the general public. While typically slanted toward use
a priority. Too few states and localities, which are the true of the specific genetic and genomic testing being offered,
arbiters of curricula, have prioritized STEM subjects. the material is nonetheless typically sound and approach-
With regard to genetics and genomics, a 2011 survey by able. Some companies also provide mechanisms for poten-
the American Society of Human Genetics (ASHG) found tial customers to pose questions and to have them answered.
that few states have standards for core concepts in basic Thus far these companies reach only a small segment of
genetics[9]. Of the 19 core concepts assessed, 85% of states the population. For example, several groups have shown
were judged to have inadequate standards. that people who avail themselves of personal genetic and
MDs are practical; most genomic advances have been rel- REFERENCES
evant to only a few clinical situations;
1. National Institutes of Health Office of Science Policy. http://
New ideas, even well-tested ones, get neglected unless they oba.od.nih.gov/oba/SACGHS/reports/SACGHS_education_
are relatively easy to understand; report_2011.pdf. Accessed 14 April 2014.
2. Scriver CR, et al. The education of citizens: human genetics. Am
Few recommendations for the use of genomic information Biol Teacher 40:280–284, 1978.
have evidence-based support (either because they apply 3. Fitzsimmons JS. The teaching of human genetics in schools. J Med
Genet 20:244–248, 1983.
to rare diseases, or because research about their appli- 4. McInerney JD. DNA in medicine: school-based education. Am
cation to common disease is lacking), so few practice J Hum Genet 42:635–636, 1988.
guidelines adopt them; 5. McInerney JD. The Human Genome Project relevant to genetics
education in high school. Am J Hum Genet 52:235–238, 1993.
There are lots of other educational reforms and demands. 6. Alberts B. Trivializing science education. Science 335:263, 2012.
7. Alberts B. Teaching real science. Science 335:380, 2012.
Genomics needs to be integrated, but this is difficult 8. Fedoroff NV. The global knowledge society. Science 335:503, 2012.
without wholesale alteration of curricula (see above); 9. Dougherty M, et al. A comprehensive analysis of high school
genetics standards: are states keeping pace with modern genetics?
and CBE-Life Sciences 10:318–327, 2011.
10. Harvey EK, et al. Providers’ knowledge of genetics: a survey of 5915
All of this has been done in the context of the “push individuals and families with genetic conditions. Genet Med 9:
model”; what we need is a “pull model.” 259–267, 2007.
11. NHGRI Online Genetics Education Resources. http://www.
genome.gov/10000464. Accessed 14 April 2014.
This will require support for research that documents 12. GeneEd. http://geneed.nlm.nih.gov/. Accessed 14 April 2014.
the clinical utility of genetic technologies. Evidence for 13. NHGRI DNA Day. http://www.genome.gov/10506367. Accessed
cost-effectiveness/benefit would also help stimulate inter- 14 April 2014.
14. ASHG DNA Day. http://www.ashg.org/education/dnaday_
est in education[63]. The Centers for Disease Control and winners_2012.shtml. Accessed 14 April 2014.
Prevention (CDC) Office of Public Health Genomics 15. ASHG High School Workshop 2012. http://www.ashg.org/
launched the Evaluation of Genomic Applications in 2012meeting/pages/workshops.shtml#2. Accessed 14 April 2014.
16. Johns Hopkins receives $7.4 million grant to boost STEM edu-
Practice and Prevention (EGAPP) program in 2004[64]. cation in Baltimore city. http://releases.jhu.edu/2012/09/25/
Most of the analyses performed thus far found a lack of johns-hopkins-receives-7-4-million-grant-to-boost-stem-education-
firm evidentiary basis for clinical applications of genetic in-baltimore-city/. Accessed 14 April 2014.
17. Kung JT, Gelbart ME. Getting a head start: the importance of per-
tests that assess risk of common diseases or affect drug sonal genetics education in high schools. Yale J Biol Med 85:87–92,
choice or dosing. As important as this activity could be for 2012.
health professionals, insurers, and policymakers, EGAPP 18. NHGRI: Genome: unlocking life’s code. http://www.genome.gov/
Smithsonian/. Accessed 14 April 2014.
is under-resourced and potentially overwhelmed. Given 19. Gollust S, et al. Motivations and perceptions of early adopters of
that EGAPP and others have identified a lack of rigor- personalized genomics: perspectives from research participants.
Publ Health Genomics 15:22–23, 2012.
ous research on clinical utility of genomics, and that the 20. Pyeritz RE. The family history: the first genetic test, and still useful
NHGRI strategic plan clearly prioritizes establishing major after all those years? Genet Med 14:3–9, 2012.
clinical applications in the next decade, it is ironic that little 21. NHGRI: The U.S. Surgeon General’s family history initiative.
http://www.genome.gov/17516481. Accessed 14 April 2014.
funding exists for the requisite research. Ultimately research 22. U.S. Department of Health & Human Services: Surgeon General’s
on the clinical utility of genomics, coupled with due atten- family health history initiative. http://www.hhs.gov/familyhis-
tion to ethical, economic, legal, and social implications, will tory/. Accessed 14 April 2014.
23. CDC Public Health Genomics. http://cdc.gov/genomics. Accessed
determine both the approach to, and the success of, educa- 14 April 2014.
tional efforts for all stakeholders. 24. American Association for the Advancement of Science: Fellowships.
http://fellowships.aaas.org/. Accessed 14 April 2014.
25. Klatt EC, Klatt CA: How much is too much reading for medical
C O DA students? Assigned reading and reading rates at one medical school.
Acad Med 86:1079–1083, 2011.
This brief review of education in genetics and genom- 26. Feero WG, Green RD. Genomics education for health care profes-
sionals in the 21st century. JAMA 306:989–990, 2011.
ics has, in the interests of conserving space, been highly 27. Emanuel EJ, Fuchs VR. Shortening medical training by 30%. JAMA
selective and focused primarily on efforts in the United 307:1143–1144, 2012.
250
Free ebooks ==> www.Ebook777.com
can be traced to the same origin. The perceived power and by genetic health professionals occur when individual
inescapably deterministic feature of Mendelian genetics autonomy conflicts with familial solidarity. Should an
has led to a fear of stigmatization and unfair discrimina- individual be able to consent alone to a test that will reveal
tion, which in turn has led to the introduction of genetic information about a family member who does not want to
non-discrimination legislation and insurance moratoria in know their result?
many countries. The treatment of genetic information in Maintaining patient confidentiality and an individu-
this way, as needing special protection above and beyond al’s right to privacy is important in clinical genetics and, as
other biomedical data—a practice known as “genetic excep- such, genetic diagnoses are generally treated no differently
tionalism”—also derives to some extent from the wide- from other potentially sensitive personal medical informa-
spread misunderstanding that genetic tests deliver certainty. tion. However, unlike the case with most other medical
Although genetic exceptionalism has been widely criticised,5 data, respecting individual privacy and or choice can be
and is based on the false belief that most genetic informa- problematic in the context of “at risk” families in which
tion is deterministic, clinicians must nonetheless address it is possible that individual family members will have dif-
and respond to these preconceptions and worries when ferent values and conflicting opinions. Does an individual
working with patients. have a right not to know their own genetic makeup,6 or
The emotion attached to a diagnosis of many Mendelian to withhold access to it when a family member is in need
diseases may be very significant for both the individual and of the same information7? In such cases, the value of pri-
family. The discipline of genetic counselling has developed vacy needs to be balanced against the rights and freedoms
from the (patient-led) necessity for psychosocial and infor- of others, and in certain circumstances it may be justified
mational support to help individuals and families cope with to break confidence in order to avoid serious harm to a
the impact of a genetic condition. Genetic counselling for relative.8
rare, highly penetrant, serious—often life-threatening— Particular difficulties arise around testing those who
conditions is available from specially trained clinical cannot give consent (minors and those lacking capacity).
geneticists and genetic counsellors. These professionals use Testing for preexisting conditions or those where early pre-
established, evidence-based communication models that ventative actions may be taken is usually advised, and offered
offer time and space to individuals and families to consider with consent from a parent or legal guardian. However,
the emotional and psychological implications of being most professional guidelines in Europe9 recommend that
tested for a family condition. Many recognized ethical, minors should not be tested for adult-onset conditions
legal, and social issues have emerged from genetic counsel- for which no immediate preventative action exists, which
ling practice over the last 50 years, and any discussion about reflects a general consensus that this would infringe on the
genethics must involve genetic counselling practice. autonomy of the future adult to make their own decision
about genetic testing.10
C O NS E N T A N D AU TO N O MY
R E P RO D U C T I VE AU TO N O MY
A key ethical commitment in clinical genetics is a respect
A N D ITS A P P RO P R I AT E L I M ITS
for individual autonomy, which manifests itself in a widely
agreed recognition of the importance of providing genetic The use of preconception, pre-implantation, and prenatal
counselling and ensuring informed consent prior to genetic testing to facilitate reproductive autonomy is a criti-
undertaking testing. This often involves gathering consent cal part of modern clinical genetics, and for many couples
for testing from potentially affected relatives, particularly who choose this option, it has substantially reduced the
where the individual referred for testing is not himself burden of serious inherited diseases.11 Individual autonomy,
affected. The obtaining of valid consent (or refusal) is, non-directive counselling, and patient empowerment12 are
however, not always a straightforward matter. Individuals central to supporting decisions that may include opting for
may struggle to fully comprehend the future implica- assisted reproduction, destruction of unwanted embryos,
tions of a test result, and obtaining informed consent or termination of affected pregnancies. Respecting the
from family members can sometimes be extremely chal- individual’s reproductive autonomy also means supporting
lenging; for example, due to difficulty in knowing how to and providing appropriate care to women and couples who
communicate genetics to relatives, possible differences in choose not to opt for these routes—for example, women
opinion about testing, or simply problems in even making who, on discovering that their pregnancy is “at risk” or
contact due to family breakdown. Moral dilemmas faced “affected,” opt to continue the pregnancy to term.
www.Ebook777.com
Ongoing political debate surrounding embryo research most of which have unknown clinical significance and are
and the ethics of abortion, amplified by the unpleasant unrelated to the reasons for which these tests were ordered.
historical specter of eugenics, means that developments in In twentieth-century genetics, the majority of variants
this area continue to be somewhat contentious. The scope seen clinically were rare and assumed to be pathogenic;
of individual reproductive choice remains unclear, and the however, twenty-first–century genomics has shown
majority of genetic tests available remain within the con- that non-pathogenic genetic variation is far more com-
fines of highly penetrant clinical diseases. Controversies mon. Knowledge of normal population genetic varia-
have arisen relating to what constitutes a “disease” and to tion is therefore crucial to interpreting genomic data.
what extent the autonomous choices of parents—to choose Whole-genome sequencing has the potential to reveal,
to have a deaf child using preimplantation genetic diagnosis, not only rare highly predictive variants with heritable
for example13—should be respected. Sex selection for social consequences, but also novel and common variants with
and cultural reasons or family balancing is generally viewed unknown or no clinical or phenotypical consequence.
as unacceptable in most countries and is only permitted to Some variants will be risk factors for common complex
prevent X-linked diseases. diseases; others may play a role in drug metabolism and
toxicity; a number will relate to behavioral phenotypes;
but many will have no discernable effect. In an era of
I N C I D E N TA L FI N D I N G S
multi-gene panel testing and clinical whole-genome
The issue of obtaining informed consent for genetic testing sequencing, most variants are likely to be assumed to be
is further complicated by the potential for uncovering inci- benign until proven otherwise.
dental findings (IFs)—unexpected results that do not relate Whilst genetic counselling and the ethical prac-
to the original clinical inquiry but that may nonetheless tices developed in the rare-disease-genetics field offer a
have equivalent or greater clinical or personal significance. solid foundation upon which to build, their relevance
This is a familiar problem within the medical imaging com- is weakened when we consider whole-genome testing.
munity, where scans may often reveal unexpected findings of The paradigm of genetics as deterministic and familial is
unknown significance, many of which turn out—after fur- unconvincing in the context of common variation, minor
ther investigation—to be benign.14 Genetic examples range genetic risk factors, and somatic mutations. In reality, most
from discovering an adult-onset cancer-predisposition gene human traits and diseases are complex and multifactorial,
in a child being investigated for developmental disorders,15 most variants have variable penetrance due to environmen-
or uncovering misattributed paternity (or maternity)16 in tal interactions or other genetic modifiers, and the major-
the course of a routine test. Although the use of targeted ity of germline genetic variation has little or no predictive
molecular testing for specific variants largely mitigates this power for individuals or their families. All genomes con-
problem for many conditions, use of genome-wide tech- tain some loss of function variants and recessive alleles,18
nologies such as karyotyping and DNA microarrays have and a whole-genome analysis could yield reams of infor-
made IFs a more frequent clinical occurrence. To date, there mation pertaining to a multitude of traits, providing risk
is very little consensus on how to handle these findings, and figures that are either small, weakly predictive, or uninter-
practice tends to vary between services and clinicians, often pretable. One might expect that such benign information
based on perceived clinical utility. will have minimal emotional or psychosocial value for the
individual or their family. This contrasts enormously with
a single test for a highly predictive, serious, life-threatening
G E N O M ET H I C S condition. Therefore, although the ethical, legal, and social
issues around the rare-disease component of genomics are
The move from genetics to genomics will bring about a well established, the framework required for genomethics
profound change in the practice of clinical genetics, pri- necessarily has a broader scope, due to the unprecedented
marily due to the dramatic fall in costs and the impending scale and range of genomic data as well as the seemingly less
data tsunami. Since every individual has around 3 to 4 mil- evocative nature of it.
lion genomic variants (versus the reference sequence),17
data management and interpretation will be an enormous
R E S P O N S I B L E DATA S T EWA R D S H I P
challenge. High-resolution DNA microarrays have already
given laboratories and clinicians a glimpse of the problem: a The first and most obvious principle in genomethics,
plethora of genetic variants present in every individual, stemming from respect for autonomy and the importance
INTRODUCTION T H E L E G A L F R A M EWO R K F O R
MEDICAL RESEARCH
A key event in the regulation of medical research was the
Nuremberg Trials that were held after World War II, which
T H E C O N S T I T U T I O NA L S T RU C T U R E
found Nazi physicians and researchers guilty of atroci-
ties carried out in the name of medical research. Since In order to understand the legal framework for medical
then, research ethics committees and oversight mecha- research, it is important to understand the role of law and
nisms for medical research on human beings have become its origins in each jurisdiction. The legal frameworks in the
the accepted norm in most countries. The regulation of United Kingdom and the United States have been developed
genomic research, like other subsets of medical research, is through different processes. In the United Kingdom, there
subject to the same regulatory controls that apply to medi- is a national Parliament that makes law for the whole coun-
cal research in general. Therefore, the focus of this chapter try. As a member of the European Community, the United
will be on the regulation of medical research on human Kingdom is also subject to European law. Regulations are
beings (excluding the approval of medicines and medical directly applicable in all European Union member states,
devices) in order to discuss the ways in which these more but EU directives need to be implemented by the member
general requirements apply to genomic research. This state in order to have effect. The Westminster Parliament
chapter will also make reference to some of the specialist is obliged to implement directives that are passed by the
bodies that have been established as expert advisory bod- European Parliament into U.K. law. Failure to do so will
ies for genomics. result in legal proceedings against the United Kingdom
I will compare the regulatory systems in the United in the Court of Justice of the European Union. However,
Kingdom and the United States in order to provide a gen- each of the member states has its own “margin of apprecia-
eral overview, while at the same time illustrating some of tion” in the way that it implements and interprets European
the significant differences that exist between them. These Directives, and this has led to discrepancies in the legislation
countries have quite different ways of regulating medi- of each member state. The United Kingdom can also sign
cal research on human beings. For example, the United Conventions of the European Council, but this is entirely
Kingdom is increasingly influenced by European law and is voluntary. The most significant legal instrument regarding
subject to increasing harmonization across Europe, whereas medical research in terms of the European Community
the United States has its own unique regulatory regime. In is the Convention on Human Rights and Biomedicine
order to elucidate these differences, this chapter will com- (1997). However, the United Kingdom has yet to sign this,
pare the legal framework between the two jurisdictions, the due to the provisions that relate to stem cell research.
institutions that have been established to oversee medical In contrast, the U.S. Congress (consisting of the House
research on human beings, and the powers that they have. By of Representatives and the Senate) is the highest law-making
highlighting the differences between two well-established body in the United States. There is no obligation to sign
legal systems in the Western world, it is hoped that this or ratify the law of another authority, such as the United
review may assist readers in other countries where the issues Nations or an international treaty, unless this has been a
of the regulation of human genomics research are also decision of the American Congress. The United States is a
actively debated and considered. federal system, which means that the Congress and the 50
259
state legislatures have different Constitutional powers. For 5. Research and demonstration projects which are
medical research there is federal legislation, but the states conducted by or subject to the approval of department
also have considerable powers to legislate. The lack of a or agency heads;
national health service has also resulted in differences and
6. Taste and food quality evaluation and consumer
anomalies in the provision of healthcare in different parts
acceptance studies.2
of the United States.
There have been significant advances in scientific
T H E RO L E O F S TAT U T E research since the Common Rule was developed, and it
is no longer seen to be as effective and comprehensive as
The United States has specially drafted federal legisla- it could be. Therefore, in July 2011, the U.S. Department
tion that covers the regulation of medical research, which of Health and Human Services announced a proposal to
applies across the United States. In 1991, the Department improve the rules protecting human research subjects.3 The
of Health and Human Services issued the “Common suggested amendments would make considerable changes
Rule.”1 The philosophical underpinnings of the Common to the Common Rule. They propose to amend seven aspects
Rule come from the Belmont Report—Ethical Principles of the current framework, as follows:4
and Guidelines for the Protection of Human Subjects. The
Common Rule has a very narrow ambit, as it only applies to
1. Refining the existing risk-based regulatory framework;
medical research that is conducted by a federal department
or agency or is federally funded. Research that comes under 2. Establishing a single IRB review of record for domestic
this scope must be reviewed and approved by an institu- sites in multiple locations;
tional review board (IRB) that operates in accordance with
3. Improving consent forms and the consent process;
the requirements of this policy.
However, under the Common Rule, the following 4. Establishing mandatory data security and information
research does not require IRB approval: protection standards for all studies involving
identifiable or potentially identifiable data;
1. Research conducted in established or commonly 5. Establishing an improved system for managing data on
accepted educational settings, involving normal unanticipated problems and adverse events;
educational practices;
6. Extending all federal regulatory protections to include
2. Research involving the use of educational tests all research at a U.S. institution that is wholly or even
(cognitive, diagnostic, aptitude, achievement), survey partly federally funded for research on human subjects;
procedures, interview procedures, or observation of and
public behavior, unless an individual can be identified
or the information provided by the individual will 7. Harmonization of regulations and related agency
incriminate or be detrimental to him or her; guidance.
G E N O M I C S I N C L I N I C A L P R AC T I C E
19.
GENETIC AND GENOMIC APPROACHES
TO CLINICAL MEDICINE
Dhavendra Kumar
A
ncient civilizationsin India, China, Greece,and proper practice of medicine than to give our minds
Egypt recognized that heredity played an impor- to the discovery of the usual law of Nature, by care-
tant part in human health. Writings, paintings, ful investigation of cases of rarer forms of disease.
and sculptures belonging to people from these regions, For it has been found, in almost all things, that
and dating back several thousand years, provide evidence when they contain of useful or applicable is hardly
that they knew of a close association between heredity and perceived unless we are deprived of them, or they
health. Hippocrates in his medical texts noted that “hered- become deranged in some way.
ity affects health.”It is likely that this concept and its prac-
tical applications were recognized by other societies but Garrod, in his seminal paper on the patient with
lacked documentary evidence. But there are some concrete “alkaptonuria,” provided the necessary evidence to sup-
examples of this awareness, such as the reference in ancient port the “one gene—one enzyme” theory. This laid the
Jewish texts to a young boy given exemption from circumci- foundation for molecular medicine, and showed the way
sion because he had a male relative who died from bleeding forward for genetic medicine. Surprisingly, all this hap-
following ritual circumcision. This reflects the practice of pened well before the elucidation of the chromosomal and
heredity medicine. However, scientific support for hered- nucleic acid basis of inheritance. However, it was widely
ity and its biological relevance was not clearly understood agreed that the individual genetic complement was locked
until the mid–nineteenth century, when the “evolution of in the nucleus.
species” theory proposed by Charles Darwin, and Gregor The science of human genetics probably began with the
Mendel’s seminal observations on cross-breeding various discovery of the full chromosomal complement of man by
forms of green pea plants, laid the foundations of the sci- Tjio and Evans in 1956 (see Chapters 1and 2).[1] This led
ence of genetics. Darwin’s theory elucidated the biological to the rapid expansion of the whole field of human genet-
importance of genetic variation, and Mendel put in place ics, and its divergence into various specialist fields such as
the fundamental principles of heredity based on individual medical and clinical genetics. Developments in medical
hereditary factors that were eventually to be identified as genetics and subsequent applications in clinical practice
genes. Unfortunately, the medical applications of these provided a strong base for wider applications of genetics in
major discoveries remained unrecognized for almost half a medicine. Genetic medicine gained wide recognition over a
century. period of 40 years and led to the historical decision of 1996
Archibald Garrod, one of the pioneers of “genetics in to sequence the whole human genome.[2] This mammoth
medicine” first began applying this knowledge to human task, called “The Human Genome Project,” was imme-
health at the start of the twentieth century. In his Harveian diately acknowledged as similar to, or even larger than,
Oration of 1924, he quoted from a letter written by William the Manhattan Project to build the atomic bomb, or the
Harvey, the Founder of Modern Medicine: Apollo Mission for the trip to the Moon and back.[3] Some
enthusiasts regard this as the “genomic period,” followed
Nature is nowhere accustomed more openly to dis- by the “post-genomic” phase. This is probably incorrect.
play her secret mysteries than in cases where she The genomic era has only just begun. The first glimpse of
shows traces of her workings apart from the beaten the human genome was made possible with the publication
path; nor is there any better way to advance the of the full sequence of the human mitochondrial genome.[4]
269
The technology to sequence the human genome was pro- and analytical methods were based on hypotheses. As
vided by the publication of the genome of the bacterium we enter into the genomic phase (present), the majority
Haemophilus influenzaeand other microbes.[5] This and of experiments are still hypothesis-based, but analysis is
many subsequent achievements were acknowledged as “a now more systematic. It is envisaged that in the future
point of entry into genomics.” (post-genomic era), experiments will be more systematic,
Perhaps it is reasonable to identify the period prior leading to automatic analytical methods.
to the completion of the human genome project as the
pre-genome era. However, Dr. Francis Collins, direc-
tor of the Human Genome Project and a keen enthusi- A N EW TAXO N O M Y F O R
ast for genomic medicine, dismissed the use of the term HUM AN DISE ASE
“post-genomic era” while discussing the scope of pro-
teomics following the completion of sequencing the Realizing the full beneficial potential of the human genome
human genome.[6] He queries whether this means that sequence will ultimately depend on its applications in
from the beginning of the universe until 2001 we were clinical medicine.[8] Many aspects of modern-day clinical
in the “pre-genome era,” and then suddenly, “Bang!” we practice will change with technological advances and the
moved into the post-genome era, leading one to won- understanding of disease mechanisms, developments in
der: what happened to the genome era? He suggested diagnosis, and new drugs and therapeutic procedures. These
that it was presumptuous to say that the Human Genome advances have largely influenced our approach towards
Project is already behind us. He pointed out that pro- human disease and have contributed to developing a new
teomics is a subset of genomics, and genomics is more taxonomy for human disease (see also Chapter 20).[9]
than sequencing genomes, which will be continuing for Genetic etiology in human disease—for example, con-
decades to come. The most appropriate term would be the genital heart disease—is recognized and includes diseases
post–Human Genome Project era, which can be referred to resulting from chromosomal abnormalities, mutations in
as the post-HGP era. Whatever the argument, this is truly nuclear and mitochondrial DNA gene sequences, and inter-
the “genomic era.” Perhaps “post-Mendelian” seems more action of environmental factors with several low-penetrance
appropriate as we move from an era in which genetics has genes carrying a small additive effect.[10] Advances in genomic
been rooted in monogenic diseases with high penetrance, technology have now made it possible to unravel pathogenic
to a greater awareness (but limited understanding) of mechanisms in certain genetic disorders that result from
polygenic diseases (and traits), often with relatively low unusual genetic factors that do not comply with the tradi-
penetrance. tional concept of genetic etiology. The pathogenic mecha-
The “genomic era” holds phenomenal promise for nisms include a number of complex alterations distributed
identifying the mechanistic basis of anatomical develop- across various genomic regions. These are now being termed
ment, metabolic processes, and disease. This is supported “genomic disorders.”[11] Distinct categories of genomic disor-
by bioinformatics research, which will have a dramatic ders include epigenomic diseases, disorders of genome archi-
impact on improving our understanding of such diverse tecture, trinucleotide repeat disorders, and diseases associated
areas as the regulation of gene expression, protein struc- with complex genomic polymorphisms.[12] It is likely that in
ture determination, comparative evolution, and drug dis- the future, more groups of genomic diseases will come to
covery. The availability of virtually complete data sets also light. It is essential that the underlying pathogenic mecha-
makes negative data informative: when we can map entire nisms in these diseases be understood in order to facilitate
pathways, for example, it becomes interesting to ask, not diagnosis and development of targeted specific therapy.
only what is present, but also what is absent.[7] This meta-
phor can also refer to the increasing emphasis on func-
TOX I C O G E N O M I C S
tional genomics. With an increasing number of organisms
for which we have (more or less) complete genomes, we Toxicogenomics is a scientific field that aims to study the com-
are just beginning to glimpse the potential power of fully plex interaction between the whole genome, chemicals in
mapped sequences. the environment, and disease. When the cells are exposed to
Perhaps we could delineate the various phases of genom- a stress, drug, or toxin, they respond by altering the pattern
ics more easily by examining the evolution of the scientific of expression of genes within their chromosomes. Genes are
methods that led to the development of different analytical transcribed into messenger RNA (mRNA), the chemical mes-
methods. In the past (pre-genomic era), both experiments sage by which information encoded in genes is translated into
proteins that serve a variety of cellular functions in response • Responses to complex chemical mixtures can
to the exposure.[13] The production of protein encoded by a be more easily elucidated and defined by their
given gene may be increased, decreased, or remain unchanged, gene-expression-profiling signatures;
depending upon the type of toxic exposure and the cellular
• Responses to chronic low doses of toxicants
requirements. In a general way, this could be directly attrib-
or environmental pollutants can be defined by
uted to post-transcriptional or post-translational effect, a
gene-expression-profiling;
fundamental step in the peptide chain assembly (see also
Chapter 3).[14] Another important effect of a toxic exposure • Specific gene polymorphisms can be defined that are
could be abnormal chromatin structure. Abnormal chromatin characteristic of an increased susceptibility to the pathol-
structure spread over successive generations could lead to epi- ogy of environmental health diseases.
genetic or epigenomic changes manifesting as developmental
or systemic disorders (see Chapter 4).[15] Toxicologists and environmental health scientists have
A technology that is central to the field of toxicogenom- studied the effects of the environment on human health for
ics is known as micro-arrays(http://www.ohsu.edu/croet/ several decades. Many adverse environmental effects have
research/centers/toxicogenomics/whatis.html), which been identified, and important progress has been made in
enable scientists to simultaneously monitor interactions preventing exposure to harmful agents such as γ-radiation,
among thousands of genes within the genome. This tech- ultraviolet light, lead, pesticides, and dioxins. Toxicological
nology (see also Chapter 10) will help define the complex research has attempted to develop an efficient, cost-effective,
regulatory circuitry within a cell, tissue, or organ and give comprehensive strategy for predicting and preventing toxic
scientists a global perspective on how an organism responds responses in humans. However, progress towards this goal
to a stress, drug, or toxin. The data generated will provide has been proportionate to the existing technologies and
information about cellular networks of responding genes, level of scientific knowledge. The field of toxicology could
define important target molecules associated with the toxic- not have risen to this challenge if it had only had access to
ity mechanism, and provide biomarkers for epidemiological the less efficient technologies of the past several decades.
studies. Ultimately, this information may allow us to iden- One challenge is to use the human genome sequence
tify ways to reduce or prevent disease by pinpointing bio- as a first step toward understanding the genetic and bio-
chemical and molecular functions that have been perturbed logical basis of complex biological traits and diseases such
by environmental chemicals. DNA micro-array technology as cancer, diabetes, Alzheimer’s disease, and Parkinson’s dis-
will undoubtedly become a major tool in environmental ease. Another challenge is to utilize the increased volume
medicine, because it will improve our diagnostic and prog- of toxicological data to construct genetic and biochemical
nostic capabilities for specific diseases, as well as our ability pathways that will explain the mechanism of toxic responses.
to examine drug interactions, sensitivities, and effective- Advances in combinatorial chemistry and molecular biology
ness. This technology will also aid research on alternative have dramatically accelerated the rate of drug discovery and
model-testing procedures and support the development of availability, and the rate at which populations are exposed to
new toxicity screening processes. new drugs. Such advances intensify the burden of exposure
It is envisioned that DNA micro-array technologies will in the population, making it critical that we rapidly increase
permit the design of experiments in the occupational and our understanding of the consequences of such exposure.
environmental sciences[16] that will clarify whether: The National Center for Toxicogenomics (NCT) of the
National Institutes of Health (NIH) in the United States
•
is leading the development of a unified strategy for toxi-
Specific toxicants have unique gene-expression profiling
cogenomics studies and a public knowledge-base. This will
signatures;
have an informatics infrastructure that will allow all part-
• Different cells in different tissues have profoundly differ- ners in this unprecedented enterprise to share equally in
ent response signatures for a given toxicant; its benefits and products. By providing a focus for techno-
•
logical coordination and basic research, a centralized public
Different species show similar, overlapping, or distinct
knowledge-base, and a center for coordination for all the
patterns of gene responses to a toxicant;
partners in the pharmaceutical and chemical industries, the
• A specific toxicant signature is altered depending upon NCT will facilitate this diverse national effort. The NCT
the stage in the developmental process or defined health will not only achieve economies of time, cost, and effort,
condition; but will help ensure the successful development of a broad
www.Ebook777.com
scientific consensus on the application of toxicogenomics to (see also Chapter 12).[20] The European Nutrigenomics
the improvement of human health. Organisation (NuGO) has recently taken over the ambi-
In brief, toxicogenomics combines the conventional tious challenge to translate the nutrigenomics data into
tools of toxicology (such as enzyme assays, clinical chem- an accurate prediction of the beneficial, or adverse, health
istry, pathology, and histopathology) with the new effects of dietary components. This organization and asso-
approaches of transcriptomics, proteomics, metabolomics, ciated agencies have set out to address important issues,
and bioinformatics.[17] This marriage of toxicology and including nutrigenomics technology standardization and
genomics has created not only opportunities, but also new innovation, bioinformatics environment harmonization,
informatics challenges. This field is likely to be of major andintegrated information-system development.
importance in genomic medicine. The integration of genomics and nutritional sciences
has led to the field of nutritional genomics. This provides
a very important base from which we can study the com-
M ETA B O N O M I C S A N D
plexity of genome responses to nutritional exposure while
M ETA B O L O M I C S
offering opportunities to enhance our understanding of the
Genomics measures the entire genetic makeup of an organ- effectiveness of dietary interventions, at both individual
ism, while proteomics measures all the proteins expressed and population levels. Nutrients influence multiple physi-
under given conditions. Metabonomics, as the name implies, ological responses that affect genome stability, imprinting,
is defined as measurement of the complete metabolic expression, and viability. Nutritional genomics challenges us
response of an organism to an environmental stimulus or to understand the complex interactions between the human
genetic modification. Some people prefer the term metabo- genome and dietary components in normal physiology and
lomics, which refers to a holistic metabolic profile, to meta- pathophysiology.[21] An understanding of these interactions
bonomics, which focuses at single-cell level.[18] Essentially, will enable us to assess the benefits and risks of various dietary
there is no biological difference between these two concepts. recommendations, minimizing the risk of unintended con-
The -omics can provide information for basic biological sequences. Furthermore, nutritional genomics will enable
research and for pharmaceutical and clinical applications. the design of effective dietary regimens for the prevention
One of the challenges is integrating the information from and management of complex chronic diseases. New perspec-
the various omics: in the process, yet another term is coined, tives in the nutritional sciences in the light of advances in
systeomics, which refers to the integration of genomics, pro- genomics are reviewed elsewhere (see Chapter 12).
teomics, metabolomics, and metabonomics.[19]
Metabolomics may be one of the most recently included
P H A R M AC O G E N O M I C S
members of the omics family; however, it is probably the old-
est. In fact, it dates back to old-fashioned biochemistry, with The study of the role of genetic inheritance in individual
its emphasis on metabolism, the sum of the processes that variation in drug response and toxicity is referred to as
operate to acquire and use energy in an organism, to biosyn- pharmacogenetics. Convergence during the past decade of
thesize cellular components, and to catabolize waste. Many advances in pharmacogenetics and human genomics has
toxicological and disease diagnostics are based on metabolic led to the emergence of the field of pharmacogenomics.[22]
profiling. This methodology has been in existence for around Pharmacogenomics is the study of the relationship between
50 years, well before the advent of genomics or proteomics. the specific drug, DNA-sequence variation, and drug
It is probably true that metabolomicsis “more closely related response (see Chapter 7).
to things in the clinical world” than either genomics or pro- Pharmacogenomics holds great promise for the future of
teomics, owing to the fact that metabolic signatures reflect medicine and is one of the major principles for the practice
both genetic information and environmental influences.[3] of personalized medicine.[23] This relies on genomic biomark-
ers for disease susceptibility, including both Mendelian and
complex diseases. Applications of human genomics will
NU T R I G E N O M I C S
result in improved understanding of the pathophysiology
In the past decade, nutrition research has undergone an of disease, identification of new therapeutic targets, and
important shift in focus from epidemiology and physiol- improved molecular classification of disease. The promise
ogy to molecular biology and genetics. Nutrigenomics is the of individualized therapeutic interventions largely depends
application of transcriptomics, proteomics, metabolomics/ on the identification of drug toxicity genomic biomarkers
metabonomics, and bioinformatics in nutrition research that will enable differentiation of individuals likely to show
Figure 19.1
Two complementary approaches in functional genomics: the gene-driven and the phenotype-driven, artificially separated along the
diagonal axis. Different levels of information are interconnected with main tools of analysis and routes of investigations. Adapted, with permission, from Trends
in Molecular Medicine.[28]
Large-cell
2004 2009
KRAS KRAS
Unknown
Unknown EGFR EGFR
EML4ALK
HER2
BRAF
MET
AKT1
MAP2K1
PI3KCA
Mutations associated with drug sensitivity
EGR Gly719X, exon 19 deletion, Leu858Arg, Leu861Gln
Mutations associated with primary drug resistance
EGR exon 20 insertions
Mutations associated with acquired drug resistance
EGR Thr790met, Asp761Tyr, Leu747Ser, Thr854Ala
Figure 19.2 Stratified approaches in planning and executing treatment for common cancers: the case for non–small cell lung cancer (NSCLC).[36]
• Diagnostic applications to accommodate new disease twin studies, support the importance of host–genotype
categories interactions in wide-ranging infectious-disease clinical
phenotypes, population differences in susceptibility or
• Challenges facing the pharmaceutical industry
resistance, and favorable or adverse reactions to antimi-
• Role of the regulatory and statutory agencies crobial therapy. Population studies have also contributed
evidence supporting the selective advantage of human
EXAMPLES OF GENOMIC AND evolution and population genetic structure. For instance,
M O L E C U L A R A P P R OAC H E S a high frequency of heterozygotes for sickle-cell anemia
and thalassemias in some populations confers a selective
advantage for malaria in the face of deleterious effects in
I N FEC T I O US D I S E A S E S
homozygotes.[46] Interplay between the Duffy locus muta-
Genetic factors have been recognized to influence infec- tions and sickle-cell anemia in certain populations is rec-
tious disease susceptibility, resistance, and response to ognized to confer malaria resistance. The molecular basis
antimicrobial therapy. Numerous studies, including of this phenomenon lies in the erythrocyte chemokine
13 14 15 16 17 18 19 20 21 22 23 24
Thus, with easy access to a well-annotated human To reach these key long-term goals, NIH-USA, the
genome and the availability of cheap, accurate, National Institute of Health Research (NIHR-UK),and
whole-genome-sequencing technology, an individual could many other organizations are actively pursuing and pro-
acquire either a specific or a complete genetic health profile, moting research in the above areas. These organizations
including risk and resistance factors. The information could are strategically investing in research to further our under-
then be used to improve and guide important medical deci- standing of the fundamental causes of diseases at their
sions, to assess the risk of possible future exposures, and to earliest genetic, genomic, and molecular stages. The cen-
select preventive treatments for improved health.[100] tral theme of the personalized medicine is based on the
In brief, the practice of personalized or specifically simple basic concept that individuals respond differently
individualized medicine will become the central focus of to environmental factors, including therapeutic inter-
the future practice of clinical medicine. However, this will ventions, according to their genetic/genomic endow-
demand lot of commitment, perseveration, and investment ment and their own behavior and lifestyle. In the future,
at the personal, family, community, and public or state lev- applied and translational genomic and molecular research
els. Inevitably and understandably, this approach will raise will allow us to predict how, when, and in whom a dis-
several ethical and social concerns: the fears of inequity, dis- ease will develop.[102] We can envision a time when we
crimination (primarily due to enormous costs and afford- will be able to precisely target or stratify treatment on a
ability), and potential misuse or abuse (malpractice). The personalized(individualized) basis to those who need it,
practice of personalized medicine shall not be allowed to avoiding treatment of those who do not. Ultimately, this
develop without relevant professional and statutory safe- individualized approach will allow us to preempt disease
guards put in place. This approach should be one of the other before it occurs, utilizing the participation of individuals,
major ingredients of clinical practice pathway, what is often communities, and healthcare providers in a proactive fash-
referred to The 4 P’s of Medicine: medicine that will be more ion, as early as possible, and throughout the natural cycle
Predictive, Personalized, Pre-emptive, and Participatory.[101] of a disease process.[103]
T
he philosophy of disease is complex and reflects the for their underlying mechanisms. Terms like degenerative or
way an abnormal body state is perceived and under- autoimmune disorders are not uncommonly used to describe
stood against the background of sociocultural, psy- an organ-system dysfunction. However,whether there is a
chological, and biomedical interpretation. During the late “cause and effect” relationship remains unclear. Nevertheless,
phase of the last millennium, new scientific discoveries in associated pathological changes often provide a firm and pre-
genetics offered entirely new approaches[1]. Nevertheless, cise basis to categorize a disease or disorder. This has helped
environmental and sociocultural factors remain important in delineating distinct categories of diseases such as immuno-
in influencing the outcome of disease. The psychological logical and metabolic diseases.
and biomedical factors provide the innate framework for Developments in genetics and molecular biology have
generating and developing symptoms and signs of the dis- provided a vast amount of data and information to support
ease, depending on the pathology. On this basis, a “disease” the view that most human diseases have a significant genetic
could be best defined as an “overall perception of an abnor- component. Characterization of the genetic determinants
mal body state and appreciation of the ensuing psychologi- of disease provides remarkable opportunities for clinical
cal and physical impact.” medicine through an improved understanding of patho-
Historically, clinical practice has always faced its inabil- genesis, diagnosis, and therapeutic options. An understand-
ity to differentiate events that mediate the disease process ing of the genetic basis of human disease has opened the
from resulting clinical, biochemical, and pathological way for a new taxonomy of human disease that will be free
changes. Rapid advances in biomedical sciences, particu- from limitations and bias in developing diagnostic criteria
larly genomics, have opened entirely new horizons[2][3]. related to events that are often secondary and peripheral to
However, despite having made tremendous advances, cli- its cause[3]. For instance, genetic information has allowed us
nicians continue to rely on phenotypical manifestations of to identify distinct forms of diabetes mellitus: defining an
the disease process to define most diseases. Inevitably, this auto-immune form associated with highly diverse and com-
approach often obscures the underlying mechanisms; thus plex human leukocyte antigens (HLA) and other inherited
a clinician may fail to identify significant heterogeneity. factors that affect both expression and modification of
Concerns have been raised that most human disease pro- gene products in mediating the adult form of the disease[5].
vides only “insecure and temporary conceptions”[4]. Apart Similarly, many genetically determined molecules and path-
from infectious diseases, there are alarmingly few diseases ways have been characterized that are crucial in the patho-
that have a truly mechanism-based nomenclature. genesis of bronchial asthma[6]. It is now widely believed that
The classification of human disease depends on several a clearer understanding of the mechanisms and pathways of
factors, ranging from perception and analysis of symptoms, a disease will assist us in delineating distinct disease subtypes
sociocultural interpretation of varied manifestations of the and may resolve many questions relating to variable disease
disease, and biological considerations,to therapeutic inter- symptoms, progression, and response to therapy. This might
ventions. Conventionally, “disease” refers to a particular help in revising the current diagnostic criteria. Eventually,
organ or system dysfunction resulting from one or more caus- genetics may contribute a new taxonomy of human disease
ative factors such as physical trauma, infection, exposures to a in clinical practice.
toxic substance, or malnutrition. A large number of diseases Although genetics is acknowledged to be an important
remain unaccounted for due to the lack of a clear explanation aspect of understanding the pathogenesis of disease, the
294
Malformations Childhood diseases Adult diseases Infections Trauma
100% 0%
Proportion of genetic factors
genetic classification of human disease has not yet received Table 20.1 THE CLASSIFICATION OF GENETIC
full recognition. There is ample evidence in support of the DISORDERS
argument that genetic factors are probably associated with
Chromosomal: Numerical—aneuploidy
all human diseases except for trauma (Figure 20.1). However,
Structural—deletion, duplication, inversion,
underlying genetic and genomic factors such as genetically isochromosome
determined connective-tissue disorders, host-response to
Ring chromosome, reciprocal or Robertsonian
infection, and tissue damage or inflammation could influ- translocation
ence the outcome of trauma. Various categories of genetic Mendelian: Autosomal recessive
disorders are considered to be rare, with a tendency to be Autosomal dominant
included under the broad title of “organ-system diseases.”
X-linked recessive
Often these are listed as simply etiological factors rather than
X-linked dominant
as a distinct disease category. This concept and approach is
Epigenetic: Imprinting/parent of origin effect; indirect
now rapidly being outdated, however, and replaced with influence on gene function
new classes of diseases. This progress is seriously hampered
Oligogenic: Distinct phenotype due to 2 or more genes
by the lack of formal education at all levels and integration
Polygenic: Environmental interaction with several hun-
of appropriate technologies into the modern medical diag- dreds of low-risk alleles, genetic polymor-
nostic and therapeutic infrastructure. phisms, and genomic copy number variations
Traditionally, genetic diseases are classified as chromo- Mitochondrial: Deletion; point mutations; polymorphic vari-
somal (numerical or structural), Mendelian or single-gene ants in mtDNA
disorders, multifactorial/polygenic complex diseases or Genomic variation: Copy number variation; single-nucleotide
polymorphisms
congenital anomalies, and diseases associated with specific
mitochondrial gene mutations (Table 20.1). Apart from
chromosomal disorders, essentially all genetic disorders
result from some form of alteration or mutation occurring C H R OMOSOMA L DISO R DE R S
in a specific gene (single-gene diseases) or involving multiple
loci spread across the human genome (polygenic disorders). The entire human genome is spread around 23 pairs of
The major impact of chromosomal disorders occurs before chromosomes, including one pair specifically assigned
birth and inflicts a serious health burden throughout child- to male (XY) or female (XX) gender, designated the
hood and during the early years of life (Figure 20.2). On “sex-chromosome pair.” The chromosomal constitution of
the other hand, single-gene diseases can pose a real medical man is complex and comprises variable amounts of euchro-
and health burden from the perinatal period to adult age, matin and heterochromatin that exhibit with a character-
with a peak around mid-childhood. In contrast, the poly- istic “banding-pattern,” and is essential for the physical
genic/multifactorial diseases tend to present late, except for and distinctive appearance of a particular chromosome.
developmental anomalies that will require active multidis- Typically, a chromosome pair includes two homologues,
ciplinary care during a child’s early life. A brief description each comprising a short arm (p) and a long arm (q) sepa-
of the major types of genetic diseases is included here. Any rated by the central heterochromatin-G-C–rich region
leading medical genetics textbook will contain a detailed designated the “centromere.” A detailed account of the
description of all these group of genetic disorders. chromosome structure and fundamental changes that occur
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 2 9 5
often quite striking, characterized by growth retardation,
developmental delay, and a variety of somatic abnormali-
ties. Many chromosomal syndromes are now recognizable.
The diagnosis and genetic management of these disor-
Chromosomal ders fall within the scope of the sub-specialty “clinical
cytogenetics.”
Multifactorial
The management of chromosomal disorders requires a
Single gene coordinated and dedicated team approach involving a wide
(Mendelian)
range of clinicians and health professionals. A typical exam-
ple is Down syndrome, resulting from either three copies
of chromosome 21 (trisomy) (Figure 20.3) or an addition
to the long arm of chromosome 21, usually resulting from
Birth Puberty Adult
an unbalanced meiotic rearrangement of a parental chro-
mosomal translocation between chromosome 21 and one
Figure 20.2 Distribution of different genetic disorders in various age of the other acrocentric (centromere located at the end)
groups. Adapted with permission from Principles of Medical Genetics by Thomas D. Gelehrter, Francis
S. Collins, and David Ginsburg .
[1]
chromosomes (Robertsonian translocation). Down syn-
drome occurs in about one in 800 live births and increases
in frequency with advancing maternal age. It is character-
during meiosis and mitosis can be found in any leading text- ized by growth and developmental delay, moderate to severe
book on basic genetics. mental retardation, and the characteristic facial appearance
Chromosomal disorders are essentially disorders of the recognized with upward-slanting eyes. A major cause of
genome resulting from either loss or addition of a whole death in these individuals is associated congenital heart
chromosome (aneuploidy) or parts of chromosomes defects that can complicate the clinical management in a
(structural). A chromosome abnormality results in major significant proportion of Down syndrome cases. Prenatal
disturbance in the genomic arrangement, since each chro- diagnosis and prenatal assessment of the maternal risk for
mosome or part thereof consists of thousands of genes Down syndrome employing a variety of imaging and bio-
and several non-coding polymorphic DNA sequences. chemical parameters is now established clinical and public
The physical manifestations of chromosome disorders are health practice in most countries.
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 XX
2 9 6 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
Clinically significant chromosome abnormalities occur to be defined at the molecular level. This has revolution-
in nearly 1% of live-born births and account for about 1% ized the diagnosis and management of these disorders. The
of pediatric hospital admissions and 2.5% of childhood single-gene disorders are inherited in a simple Mendelian
mortality[7]. The loss or gain of whole chromosomes is manner, and hence justifiably called “Mendelian disorders.”
often incompatible with survival, and such abnormalities The genetic transmission of altered genes or traits follows
are a major cause of spontaneous abortions or miscarriages. principles set out by the Austrian monk Gregor Mendel in
Almost half of the spontaneous miscarriages are associated 1865, based on his seminal work on garden pea plants[8].
with a major chromosomal abnormality. It is estimated that Mendel inferred that “those characteristics that are trans-
about a quarter of all conceptions may suffer from major mitted entire, or almost unchanged by hybridization, and
chromosome problems, because approximately 50% of all therefore constitute the characters of the hybrid, are termed
conceptions may not be recognized as established pregnan- dominant, and those that become latent in the process,
cies, and 15% of these end in a miscarriage. Essentially, the recessive.”
major impact of chromosomal disorders occurs before birth The nomenclature of these disorders reflects their
or during early life (Figure 20.2). gender-specific transmission and is supported by localiza-
The delineation of rare and uncommon chromosomal tion of an altered gene on either an autosome (1–22) or the
disorders has been crucial in the gene-mapping of several X chromosome. Mendelian disorders are described as auto-
Mendelian (single-gene) disorders such as the X-linked somal dominant, autosomal recessive, and X-linked reces-
Duchenne muscular dystrophy and type 1 neurofibromato- sive (Figure 20.4) or X-linked dominant (Figure 20.5). The
sis. The chromosomal regions involved in deletion, dupli- latter pattern differs from the X-linked recessive by having
cation, inversion, and break points involved in a complex an excess of affected females in a family because the hetero-
chromosomal rearrangement provide an important clue zygous mutation on the X chromosome can be transmit-
and assist the keen researcher in focusing on genes located ted to the daughter from an affected mother as well as the
within the chromosomal segment. affected father. Sporadic X-linked dominant diseases are
predominantly encountered in a female rather than a male
due to being lethal in the latter. A detailed family history
MENDE L IAN ( SIN G L E - G ENE ) and careful interpretation of the pedigree are essential pre-
DISO R DE R S requisites in the diagnosis of a Mendelian disease. Accurate
risk estimates, for use in genetic counseling, are impossible
About 4,000 human diseases are caused by mutations in without a reliable and comprehensive pedigree. The major
single genes, and these constitute a major health burden. features of the individual inheritance pattern are described
Single-gene disorders account for approximately 5–10% in leading genetic textbooks[1]. All human disorders and
of pediatric hospital admissions and childhood mortality. traits that follow the Mendelian principles are listed in a
The major impact of these disorders occurs in the newborn major resource—“McKusick’s Catalogue of Mendelian
period and early childhood. However, these also constitute Inheritance of Man.” An online version (Online Mendelian
a significant proportion of adulthood diseases, notably Inheritance in Man; www.OMIM.org) is available and reg-
late-onset neurodegenerative diseases and various forms of ularly updated.
familial cancer. Although the majority of single-gene dis-
eases are rare, some are relatively common and pose a major
health problem. For example, familial hypercholesterolemia, P O LYG ENIC O R
a major predisposing factor in premature coronary artery MU LTIFACTO R IA L DISO R DE R S
disease, occurs in one in 500 people. Other good examples
would be familial breast and colorectal cancers, which affect This group of disorders includes the most common and least
approximately one in 300. Some single-gene disorders are understood human genetic diseases. These diseases result
specific for certain populations, like Tay-Sachs disease from the interaction of certain environmental factors with
among Ashkenazi Jews, or cystic fibrosis in Caucasians, multiple genes, some of which may have a major effect, but
thalassemias among people from Southeast Asia and the the majority of which carry only a relatively minor effect.
Mediterranean countries, and sickle-cell disease in people The minor additive effect of these multiple loci lowers the
of Western African origin. Techniques in molecular biology “threshold” of an organ or body system’s ability to with-
have enabled the characterization of a number of mutated stand environmental pressures, resulting in either a devel-
genes. Sickle-cell disease was the first single-gene disorder opmental anomaly or an abnormal disease state. Examples
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 2 9 7
(A) Autosomal dominant
I
1 2
II
1 2 3
III
1 2 3 4 5 6 7 8
I
1 2 3
II
1 2 3 4 5 6 7 8
II
1 2 3 4 5 6
III
1 2 3 4 5 6 7 8
Figure 20.4 Typical pedigree appearances in Mendelian inheritance. Key to symbols: blank square = unaffected male; open circle = unaffected female; black-filled = affected (homozygous);
half black-filled = carrier (heterozygous)
II 3 3
III 6 2 3 4 6
Figure 20.5
A pedigree with an X-linked dominant disorder—note absence of “male–male” transmission; all daughters of an affected male would be
heterozygous and thus could be symptomatic. Adapted with permission from Principles of Medical Genetics by Thomas D. Gelehrter, Francis S. Collins, and David Ginsburg . [1]
2 9 8 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
include common congenital anomalies such as cleft lip, cleft recurrence risks for a given population group are estimated
palate, neural tube defects, and most congenital heart dis- to equal the square root of the birth incidence. For instance,
eases. The common chronic medical diseases fall within this birth incidence of ventricular septal defect is approximately
category of genetic disorders, including diabetes mellitus, three per 1000, the recurrence risk to a first-degree relative,
coronary heart disease, hypertension, arthritis, and schizo- such as the next child, would be the square root of 0.003,
phrenia. Understanding the genetic basis of common dis- or 3%. These figures are useful in giving a family genetic
eases remains the major challenge facing modern genetics counseling following the birth of a child with a congenital
and genomics. anomaly.
The clinical impact of multifactorial diseases is signifi- This group of diseases poses the challenge of working
cant both in the neonatal period as well as in adult life. It out the mechanisms that determine the additive or inter-
is estimated that about 25–50% pediatric hospital admis- active effects of many genes creating predisposition to
sions are related to these groups of disorders and are associ- diseases, which in turn manifest only in the presence of
ated with 25–35% of childhood mortality. There is an even certain environmental factors. It is hoped that a combina-
greater medical and health burden from these disorders tion of molecular genetic approaches, gene mapping, and
during adult life due to the sufferers’ chronic natural history functional genomics will enable a clearer definition of these
of resulting medical diseases. For instance, diabetes melli- genetic diseases. Several sections in this book will address
tus and obesity account for about 40% of the adult medical this issue at length and focus on specific disease groups and
problems in the developed and developing world. systems.
Identification of any such disorder or condition is
important in assessing risks to close relatives. A compari-
son of general population and multiple cases in a family MITOC H OND R IA L G ENETIC
would indicate a shift of the bell-shaped Gaussian curve to DISO R DE R S
the right, reflecting a lowered threshold with an increased
incidence (Figure 20.6). The precise additional risk would Apart from nuclear DNA (nDNA), a small proportion of
dependon the degree of relationship with the index case in DNA is also found in mitochondria in the cytoplasm of cells
the family. In addition, the gender of the index case is also (mtDNA). Each cell contains 2–100 mitochondria, each
important in assessing the liability. The genetic liability is of which contains 5–10 circular chromosomes. The 16.5kb
estimated to be greater if the index case is of the gender with mtDNA molecule is free from any non-coding intronic
lowest incidence. For example, in the case of pyloric steno- regions and encodes two ribosomal RNA (rRNA) genes, 22
sis, greater risk would be applicable if the index case were transfer RNAs (tRNA), and 13 polypeptides that are parts of
a female, which carries the lowest birth prevalence.Finally, multi-subunit enzymes involved in oxidative phosphorylation
Liability of
general Threshold
population
Number of individuals
Affected
individuals
Liability of
first degree
relatives
Liability
Figure 20.6
The “Gaussian” bell-shaped curve to illustrate “genetic threshold,”indicated by liability in the general population (shown in black). A shift
.
to the right (in gray) indicates increased liability in first-degree relatives with an increased risk of recurrence. With permission from Oxford University Press, U.K.[9]
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 2 9 9
(see also Chapter 9; Figure 20.7). In comparison to the simply or directly related to mtDNA genotype, but reflects
nuclear DNA, the mtDNA is 20 times more prone to recur- several factors, including the overall capacity for oxidative
rent mutations, resulting in generation of mutagenic oxygen phosphorylation determined by mtDNA and nuclear DNA
radicals in the mitochondria. The inheritance of mtDNA genes, the accumulation of somatic mtDNA mutations and
is exclusively maternal, due to its cytoplasmic location. The degree of heteroplasmy, tissue-specific requirements of oxi-
mature sperm head contains very little mtDNA, since it is dative phosphorylation, and age.
almost completely lost during the fertilization process, appar- Several mitochondrial diseases have now been char-
ently with the loss of the tail that carried the bulk mtDNA acterized (Table 20.2). One of the best-characterized is
in the cytoplasm. Due to the wholly maternal cytoplasmic Leber’s hereditary optic neuropathy (LHON), which
location, only females can transmit mitochondrial diseases to exclusively affects males. There is loss of central vision sec-
their offspring of either gender (see Figure 20.7). ondary to optic nerve degeneration. The vision loss usu-
Since mtDNA replicates separately from the nuclear ally occurs in the patient’s 20s and can progress rapidly in
DNA, and mitochondria segregate in daughter cells inde- some men. Eleven different missense mtDNA mutations
pendently of the nuclear chromosomes (replicative seg- in three different mitochondrial genes encoding respira-
regation), the proportion of mitochondria carrying the tory chain enzyme subunits have been described. The phe-
mtDNA mutation can differ among somatic cells. This notype in other mitochondrial diseases tends to include a
mitochondrial heterogeneity is also called heteroplasmy and combination of heart, muscle, and central nervous system
plays an important part in the variable and tissue-specific manifestations, with considerable intra-/inter-familial vari-
phenotype of mitochondrial disease. Since different tis- ability for the same mtDNA mutation. In addition, mito-
sues have varying degrees of dependence on oxidative phos- chondrial dysfunction can be part of the phenotype in
phorylation, with heart, muscle, and central nervous system some Mendelian diseases where the mutant gene-product
being the most dependent, the common manifestations of presumably has a pathogenic influence on the mitochondri-
mitochondrial disease include cardiomyopathy, myopathy, ally mediated metabolic pathway. Examples of this are the
and encephalopathy (see Figure 20.1). Furthermore, oxida- autosomal recessive respiratory enzyme disorders. Genetic
tive phosphorylation declines with age, probably related to counseling and decision for prenatal diagnosis can be diffi-
the accumulation of successive mtDNA mutations. Thus cult in mitochondrial disorders due to difficulty in predict-
the clinical phenotype in a mitochondrial disease is not ing the phenotype in the affected pregnancy.
OH
12S CY
T-b
N
S
D
16
-6
LHON (3460)
Complex-1 genes
(NADH dehydrogenase)
LHON Complex-III genes
(11,778) (ubiquinol:cytochrome-c oxidoreductase)
N D-
-4
Complex-V genes
OL
ND
2
NARP
MERRF or Leigh (ATP synthase)
(8344) (8993)
Complex-IV genes
-4L
ND
-3
(cytochrome-c oxidase)
CO
I ND
III
CO
CO I I A8 A6
tion
5-kb dele
Figure 20.7 The human mitochondrial DNA molecule with examples of point mutations with their associated clinical phenotypes. Adapted from Neurogenetics
by Stefan-M. Pulst, Oxford University Press, New York, 2000, with permission[10].
3 0 0 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
Table 20.2 GENETIC CLASSIFICATION OF MITOCHONDRIAL DISORDERS
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 0 1
Table 20.3 CLASSIFICATION OF GENOMIC 1
I
DISORDERS
MITOCHONDRIAL
principles of inheritance as outlined in the previous sec- MYOPATHY DEAFNESS
3 0 2 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
is evidence that DNA methylation depends on methyla- since the regulatory gene sequences for the pathogenic gene
tion of H3-K9 and can also be a trigger for its methylation. would be missing from the other parent. This characteris-
Recently, evidence has accumulated on the role of RNA tic abnormality is commonly referred to as “uni-parental
in post-transcriptional silencing. In addition, RNA in the disomy” or UPD. This could either be isodisomy (simi-
form of antisense transcripts (Xist or RNAi) can also lead to lar parental homologues) or heterodisomy (parental and
mitotically heritable transcriptional silencing by the forma- grandparental homologues) (Figure 20.9). The origin of
tion of heterochromatin. For example, transcription of anti- UPD is believed to result from the loss of the additional
sense RNA led to gene silencing and to the methylation of chromosomal homologue, failing which the conceptus
the structurally normal α-globin gene in patients with alpha would be trisomic. This mechanism is also called “trisomic
thalassemia. This could be one of the many human diseases rescue.”
resulting from epigenetic silencing due to antisense RNA For a maternally imprinted disorder, paternal UPD
transcripts[16]. would be confirmatory and maternal UPD diagnostic for
Mutations in genes that affect genomic epigenetic pro- the paternally imprinted condition. For example, mater-
files can give rise to human diseases that can be inherited or nal UPD is diagnostic for Prader-Willi syndrome, and
somatically acquired (Table 20.4). These epigenetic muta- paternal UPD for Angelman syndrome, both conditions
tions can be due either to hypermethylation (silencing) of being associated with a microdeletion of the 15q11 region.
a regulating gene or to loss of methylation (LOM) (activa- The parental origin of the 15q microdeletion follows the
tion) of another gene that has a positively modifying effect expected epigenetic pattern and is in keeping with the clini-
on the phenotype. The parental imprinting effect can be cal diagnosis. Recurrence risk estimates vary, depending on
inferred by demonstrating the parental origin of the mutant the specific epigenetic pattern. This information is crucial
allele. Similarly, either a loss or a gain of a chromosomal to obtain in order to offer accurate genetic counseling in any
segment can result in the same situation. Confirmation genomic imprinting disorder.
of a specific chromosomal deletion or duplication is usu- Many epigenetic diseases are associated with chromo-
ally possible by using the fluorescent insitu hybiridization somal alterations and manifest with physical and learning
(FISH) method. The paternal imprinting in this situation difficulties. For example, mutations in X-linked mental retar-
is commonly demonstrated by genotyping a set of polymor- dation with the alpha thalassemia phenotype (ATRX) result
phic markers located within the chromosomal segment. in consistent changes in the methylation pattern of ribosomal
Inheritance of the whole chromosomal homologue from DNA, Y-specific repeats, and subtelomeric repeats. Another
one parent effectively confirms imprinting phenomenon, X-linked recessive mental retardation syndrome, associated
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 0 3
Eye
Heart Optic neuropathy
Conduction disorder Ophthalmoplegia
Wolft-Parkinsons-White Retinopathy
Skeletal muscle
Weakness Cardiomyopathy
Fatigue
Myopathy
Neuropathy Liver
Hepatopathy
Brain
Seizures Kidney
Myoclonus Fanconi's syndrome
Ataxia Glomerulopathy
Stroke
Dementia
Migraine
ATP
Nuclear DNA
Pancreas
subunits OX PHOS
mt DNA Diabetes mellitus
O2 H O
2
Defects in intergenomic
communication
Multiple mtDNA deletions and
mtDNA depletion
Blood
Pearsson's syndrome
Inner ear
Colon
Sensorineural
Pseudo-obstruction
hearing loss
OX PHOS = OXIDATIVE PHOSPHORYLATION
Figure 20.9
The origin of uniparental disomy 15 in Prader-Willi syndrome through trisomic rescue during early embryogenesis—note different
homologues (maternal heterodisomy).
with a visible “fragile site” on the terminal part of the long arm disorder[18]. On a wider genomic level, mutations in the
of the X chromosome (fragile-X syndrome), results from de DNMT3b gene, causing the ICF (immunodeficiency,
novo silencing of the pathogenic gene FMR1. The syndrome centromeric region instability, and facial anomalies) syn-
is characteristically associated with an abnormal expansion of drome, result in deregulation of DNA methylation pat-
CGG triplet repeats in the FMR1 5′ untranslated terminal terns. A notable example is that of Beckwith-Wiedemann
region. Methylation of the expansion leads to silencing of the syndrome (BWS), an overgrowth syndrome predisposing
FMR1 gene and under certain cultural conditions creates the to Wilms’ tumor and other childhood tumors, which is
visible “fragile site” on the X chromosome. associated with duplications and rearrangements of a small
Epigenetic silencing is probably also significant in other chromosomal region on the short arm of the chromosome
neurodevelopmental disorders. For example, in Rett syn- (11p15.5). This region contains a cluster of genes, which
drome, a common cause of intellectual disability in young is susceptible to a number of epigenetic alterations, mani-
girls, mutations of the MeCP2 gene are seen in about 80% festing with the BWS phenotype and tumorigenesis, par-
of cases. The MeCP protein binds to methylcytosine resi- ticularly Wilms’ tumor and other childhood embryonal
dues and causes de-repression of genes normally suppressed tumors (Figure 20.10). Loss of methylation in imprinting
by DNA methylation. Despite the lack of firm evidence, it control regions (such as KvDMR1) can cause deregulation
is thought likely that MeCP2 might have a key role in the of imprinting, and either biallelic expression (IGF2 and
control of neuronal gene activity resulting in the pathology H19) or silencing (such as CDKN1C) of imprinted genes,
of Rett syndrome[17]. Interaction with another pathogenic which is seen in most sporadic BWS cases[19].
gene (CTKL5 or STK9) in Rett syndrome is likely to be The epigenetic phenomenon is probably significant for
important in the pathogenesis of this neurodevelopmental the phenotypical manifestations in some other hereditary
3 0 4 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
species have enabled investigators to view genetic informa-
RNA tion in the context of the entire genome. As a result, it is now
possible to recognize mechanisms of some genetic diseases
at the genomic level. Amongst the several biological pro-
cesses, duplication of genes, gene segments, and repetitive
Gene gene clusters have helped in the evolution of mammalian
genomes[22]. This aspect of genome architecture provides
Histone modification DNA methylation recombination hot spots between non-syntenic regions of
chromosomes that are distributed across the whole genome.
These genomic regions become susceptible to further DNA
Figure 20.10
The cluster of genes on 11p15.5 associated with the rearrangements that may be associated with an abnormal
phenotype of Beckwith-Wiedemann syndrome. The methylated region phenotype. Such disorders are collectively grouped under
KvDMR1 is indicated by the gray box within the gene KCNQ1OT1
and marked CH3 on the maternal homologue. The methylated region
the broad category of “genome architecture disorders”[11].
between the IGF2 and H19 genes is indicated by the hatched box and The term genome architecture disorder refers to a disease
marked CH3 on the paternal homologue. With permission from Oxford University that is caused by an alteration of the genome that results
Press[23].
in complete loss, gain, or disruption of the structural integ-
rity of a dosage sensitive gene(s) (Figure 20.12). Notable
examples include a number of chromosome deletion/
tumors. For example, transmission of autosomal dominant
duplication syndromes (Table 20.5). In these conditions,
familial chemodectomas (non-chromaffin paragangliomas
there is a critical rearranged genomic segment flanked by
or glomus tumors) is exclusively via the paternal line (Figure
large (usually >10 kb), highly homologous low copy repeat
20.11)[20]. The maternally derived gene is inactivated during
(LCR) structures that can act as recombination substrates.
oogenesis and can be reactivated only during spermatogen-
Meiotic recombination between non-allelic LCR copies,
esis. This genetically heterogeneous cancer family syndrome
also known as non-allelic homologous recombination, can
is associated with germline mutations in succinate dehydro-
result in deletion or duplication of the intervening segment.
genase subunits B (SDHB) and D (SDHD)[21].
Similarly, other chromosomal rearrangements, includ-
Thus epigenetic changes are probably significant in a
ing reciprocal, Robertsonian, and jumping translocations;
number of other complex phenotypes, particularly those
inversions; isochromosomes; and small marker chromo-
associated with cancer and a number of degenerative dis-
somes, may also involve susceptibility to rearrangement
eases (see “Complex Genomic Diseases”).
related to genome structure or architecture. In several cases,
LCRs, A-T–rich palindromes, and pericentromeric repeats
D I S O R D E R S O F G E N O M E A RC H IT E C T U R E are located at such rearrangement breakpoints. This suscep-
tibility to genomic rearrangements is implicated not only in
Recent completion of the Human Genome Project and
disease etiology, but also in primate genome evolution[25].
sequencing of the total genomes of yeast and other bacterial
An increasing number of Mendelian diseases
(Table 20.6) are recognized to result from recurrent
inter- and intra-chromosomal rearrangements involving
Disomy 15 Haploid unstable genomic regions facilitated by low-copy repeats
oocyte sperm
(LCRs)[26]. These genomic regions are predisposed to
non-allelic homologous recombination (NAHR) between
Trisomy 15 conceptus
paralogous genomic segments. LCRs usually span approxi-
mately 10–400 kb of genomic DNA, share 97% or greater
Loss of one homologue
sequence identity, and provide the substrates for NAHR,
thus predisposing to rearrangements. LCRs have been
Maternal shown to facilitate meiotic DNA rearrangements asso-
or heterodisomy 15
(Prader-Willi syndrome)
ciated with several multiple malformation syndromes
Biparental Biparental Uniparental and some disease traits (Table 20.6). Seminal examples
include microdeletion syndromes (Williams-Beuren
Figure 20.11
Pedigree showing paternal transmission of paraganglioma
in a family: note no maternal transmission among “at-risk” family syndrome[7q11del], DiGoerge syndrome[22q11del]);
members[20]. autosomal dominant Charcot-Marie-Tooth disease type
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 0 5
Centromeric domain Telomeric domain
Maternal
Expressed
Silenced
kvDMR
H19DMR
Figure 20.12
Molecular mechanisms for genomic disorders—dashed lines indicate either deleted or duplicated region; the rearranged genomic interval
is shown in brackets; gene is depicted by filled horizontal rectangle; regulatory gene is shown as horizontal hash-marked rectangle; asterisks denote
point mutations[24].
1A (PMP22 gene duplication); hereditary neuropathy of involvement of either the axonal or myelinated segments
pressure palsy (HNPP: PMP22 gene deletion) mapped to of the peripheral nerve. Genetically autosomal dominant,
17p11.2; and Smith-Magenis, a contiguous gene syndrome autosomal recessive, and X-linked dominant types are rec-
(CGS) with del (17)(p11.2p11.2). Dominantly inherited ognized. The disorder is not uncommon, affecting approxi-
male infertility related to AZF gene deletion follows a mately one in 2,500 of the adult population. This could be
similar mechanism. In addition, this LCR-based complex an underestimate, since medically the condition is benign,
genome architecture appears to play a major role in primate often not requiring any medical or surgical intervention.
karyotype evolution, the pathogenesis of complex traits, However, some affected individuals experience increasingly
and human carcinogenesis. progressive neuromuscular weakness of distal muscles of
A notable example includes genetically heterogeneous lower legs, feet, distal forearms, and hands, with onset in
Charcot-Marie-Tooth disease (CMTD). The disorder is the early teens, and causing severe locomotor restrictions.
also known as “hereditary motor and sensory neuropathy” An affected person usually presents late with relative
(HMSN) by virtue of being a peripheral neuropathy due to hypertrophy of the upper calf muscles, described as an
3 0 6 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
Table 20.6 MENDELIAN GENOMIC DISORDERS [11]
“inverted Champagne bottle”appearance (Figure 20.13), counseling is limited because both types are genetically het-
associated with pes cavus due to wasting of the small muscles erogeneous. For instance, molecular characterization and
of the feet. Similarly, wasting of the small muscles of hand gene mapping have confirmed the existence of at least four
leads to “clawhands.” Neurophysiological studies remain an types of type 1 CMTD: autosomal dominant types 1a, 1b,
essential method of differentiating the two major types of and 1c, and the X-linked type (XCMT). Similarly, there are
CMTD. A reduced motor-nerve-conduction velocity of distinct genetic types within the type 2 CMTD group.
less than 35 m/sec helps in differentiating type 1 CMTD Approximately two-thirds of cases of CMT1 have a
from type 2 CMTD, in which the motor-nerve-conduction detectable 1.5 Mb duplication within a proximal chro-
velocity is usually normal but the sensory-nerveconduction mosomal segment of the short arm of chromosome 17
is often slow. Whilst this distinction is undoubtedly helpful (17p12)[27]. This duplicated chromosomal segment contains
in determining clinical management, application for genetic a gene for peripheral myelin protein called PMP22. This
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 0 7
4
2 4 2 3
1 4 2 2 1 3 1
Figure 20.13 Lower legs and feet in Charcot-Marie-Tooth disease—note characteristic lower-leg appearance and pes cavus.
duplication results in the disruption of the gene, leading to duplicated 17p12 region. It soon became apparent that a 500
abnormal myelination of the peripheral nerves, an essential kb allele co-segregated with 17p duplication in all affected
molecular pathological step resulting in the CMT1 phe- individuals. This suggested a stable mutation and followed
notype designated as CMT1A. The CMT1A duplication a precise recombination mechanism. However, in de novo
was visualized by multiple molecular methods, including duplication, the presence of repeated flanking marker
fluorescence in-situ hybridization (FISH), pulsed-field gel alleles indicated the mechanism of unequalcrossing-over
electrophoresis (PFGE), and dosage differences of het- leading to duplication. Indeed, this was confirmed when
erozygous alleles by restriction-fragment-length polymor- a highly homologous >20 kb–size repeat sequence was
phisms (RFLPs) (Figure 20.14). This finding led to further confirmed flanking the 17 p duplication. It was appropri-
molecular studies on the origin of the 1.5 Mb duplicated ately named “CMT1A-REP.” As predicted by the unequal
17p12 segment[28]. crossing-over model, CMT1A-REP was found to be pres-
Studies by several investigators have revealed a sig- ent in three copies on the CMT1A duplication-bearing
nificant variation in the size of marker alleles flanking the chromosome. Interestingly, the presence of only one copy
del/dup
(B) Gene interruption
The 1.5 Mb duplicated chromosomal region of 17p12 including the PMP22 gene—note 500 Kb junction fragment allele flanking the
Figure 20.14
CMT1A gene detected by PFGE and Southern analysis. Note additional 17p segment (red color) by metaphase (top two pictures) and interphase
(lower two pictures) FISH[11].
3 0 8 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
Figure 20.15 The unequal meiotic recombination (crossing-over) resulting in duplication (CMT1A) and deletion (HNPP)[11].
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 0 9
Table 20.7 DISORDERS WITH TRINUCLEOTIDE (TRIPLET) REPEAT EXPANSION [11]
of the last exon of the DM protein kinase (myotonin) manifestations as in fragile-X syndrome. An expanded allele
gene (DM). in the pre-mutation range in a male would not be associ-
Each class of trinucleotide repeats exists in normal indi- ated with any clinical manifestations (normal transmitting
viduals. A pathogenic expansion is the one that is seen in male NTM), but this could further expand, resulting in all
clinically symptomatic individuals. Carriers for an X-linked his daughters’ being carriers. However, recent studies have
disease also have an expanded allele (pre-mutation), which provided data on the existence of late-onset gait ataxia in
does not usually result in an abnormal phenotype. However, NTMs[32]. On the other hand, a normal-size CGG repeat
it is likely that some carrier females might exhibit some in a normal male could undergo further expansion during
CMT1A
junction fragment CMT1A/HNPP
Analysis
PMP22
1.5 Mb
17p12
17
Figure 20.16 Location of four classes of triplet repeats in human diseases. Exons are shown in light pink with intervening introns as a pink solid line.
The translation site AUG and termination signal TAA are indicated by red vertical bars. Adapted with permission from Principles of Medical Genetics by Thomas D. Gelehrter,
Francis S. Collins, and David Ginsburg[1].
3 1 0 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
(sperm and egg), and can be transmitted to subsequent gen-
Deletion erations. In contrast, a genetic abnormality present only in
PMP22 HNPP specific somatic cells could not be transmitted. The genetic
abnormality in a somatic cell can occur any time from the
post-conception stage to late adult life. The paradigm of
PMP22 somatic cell genetic disorder is cancer, where the develop-
ment of malignancy is often the consequence of mutations
Duplication
CMT1A
in genes that control cellular growth. There are several such
genes, and these are designated oncogenes. It is now accepted
Figure 20.17
Schematic diagram of the polyglutamine tract resulting from that all human cancer results from mutations in the nuclear
abnormal expansion of CAG trinucleotide repeats[33]. DNA of a specific somatic cell, making it the most com-
mon genetic disease. The various genetic mechanisms that
meiosis, leading to a carrier daughter. This usually comes to can result in cancer are discussed in the chapter on cancer
light when a symptomatic grandson is confirmed to have genomics (see Chapter 36).
pathogenic FRAXA expansion. Prior to availability of the The clinical course and outcome of treatment in a num-
molecular testing in FRAXA, this kind of unusual pedi- ber of acute and chronic medical conditions depend upon
gree pattern in fragile-X syndrome was called the “Sherman various factors. For instance, there is overwhelming evi-
paradox” (Figure 20.18). Detailed molecular studies in the dence that highly polymorphic cytokine, interferon, and
family are often necessary to offer accurate genetic coun- interleukin families of complex proteins influence the host’s
selling to “at-risk” carrier females. Carrier females are at an response to acute infection and physical injury or inflam-
additional risk for developing premature ovarian failure, mation. Several genes encode these inflammatory pathway
usually diagnosed when investigated for secondary infertil- proteins. Similarly, association of human leucocyte antigens
ity (see Chapter 46). in the pathogenesis of a number of acute and chronic medi-
Genetic counselling in other neurodegenerative disorders cal disorders is well known (see Chapter 38). In addition,
with triplet repeats is often complicated. In particular, the interaction of mutations within these genes and with several
clinical prediction in “borderline” expanded triplet repeats other genomic polymorphisms, such as single-nucleotide
(intermediate allele) in HD is extremely difficult due to lack polymorphisms (SNPs) is probably important in several
of reliable data. However, recent studies have produced some acute medical conditions, including trauma. This will
data that are likely to be helpful in genetic counselling. have a major impact in critical care and acute medicine
(see Chapter 48). The role of SNPs in modulating com-
plex medical disorders, such as diabetes mellitus, coronary
COMPLEX GENOMIC DISEASES
heart disease, hypertension, and various forms of cancer, is
All inherited disorders have a genetic abnormality present unclear. However, the complexity of interaction of SNPs
in the DNA of all cells in the body, including germ cells with other genetic traits and loci is probably important in
AUG TAA
CGG GAA CAG CTG
gln
SPINOCEREBELLAR
ATAXIA TYPE I
HUNTINGTON DISEASE
DENTATORUBRAL-
PALLIDOLUYSIAN ATROPHY
(HAW RIVER SYNDROME)
MACHADO-JOSEPH DISEASE
Figure 20.18 The Sherman paradox: a hypothetical pedigree showing affected members (red) and carrier females (pink); individual III.1 is a normal
transmitting male; the % risk for mental retardation is given for respective size of the triplet (CGG) repeats. Adapted with permission from Principles of Medical
Genetics by Thomas D. Gelehrter, Francis S. Collins, and David Ginsburg[1,2].
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 1 1
1
I
60–70
III 1 2 3 4
T
60–70 5% 9%
70–90
1 2 3 4
IV
0% 40% 16%
70–90 >90
1 2 3 4
V
40% 16% 50% 28%
Figure 20.20
Schematic diagram of the polyglutamine tract resulting from
abnormal expansion of CAG trinucleotide repeats. Adapted from Perutz et al.,
1994, with permission.
3 1 2 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
Table 20.8 TAXONOMY OF HUMAN DISEASE: CORRELATION OF CLINICAL PHENOTYPES, MOLECULAR
PATHWAYS, AND GENOTYPES
I: Phenotypes:
Clinical—symptoms and physical signs
I: Correlation and interpretation of the above with one or more of the following parameters with the aim of arriving at a diagnosis or most
likely underlying mechanism of the disease:
Biochemical; e.g., urea/electrolytes, blood gases
Metabolic; e.g., sugar/lipid/endocrine profiles
Radiological—X-rays, ultrasound, CT/MRI scans, magnetic resonance spectroscopy, radioisotope, etc.
Pathological—histopathology, histochemistry, immune-histology, fluorescence microscopy, and electron microscopy
Hematological—hemoglobin, hematocrit, coagulation profile, etc.
Immunological—immunoglobulins (IgG, IgM, IgA, etc.); antibody profiles, e.g., lupus; specific immunological investigation
Microbial/ Pathogens—battery of tests for bacterial, viral, protozoal, parasitic, and fungal infections, including specific pathogen profiles
Toxicology—poisons, alcohol, therapeutic and recreational drugs
Environmental (Ecological)—Temperature extremes, high altitude, supersonic flying, and space travel
II: Correlation of the above phenotypes with one or more of the biological/ molecular pathways:
Growth factors/ growth factor receptors (e.g., EGF/EGFRs, FGF/FGFRs, TGF/TGFRs,and VGF/VGFRs)
Dynamic cell/tissue factors (e.g., protein kinase families such as P13, RAS/MAPK; tumor suppressor systems, etc.)
Respiratory chain/ oxidative pathways (e.g., mitochondrial and cyclooxygenase systems)
Cell/ tissue response systems (e.g., cytokines, interleukins, tissue necrosis factors, complement factors, etc.)
Apoptosis/ senescence systems (e.g., apoptotic pathways)
Scavenger/ housekeeping systems (e.g., lysosomal enzymes, alpha 1 antitrypsin, DNA repair genes, etc.)
Metabolic regulatory systems (e.g., insulin/glycemic regulation, lipids/ hepato-biliary systems, Krebs cycle, etc.)
Energy regulation (e.g., temperature regulation, energy conservation, nutritional state, etc.)
Hormonal regulation (e.g., endocrine pathways, autocrine and paracrine pathways)
Vascular pathways (e.g., angiogenetic systems, clotting/ coagulation pathways, and platelet factors)
III: Correlation and/or interpretation with ALL of the above with one (or more) individual’s genetic/ genomic pathology:
Chromosomal aberration—aneuploidy (e.g., trisomy 21, 18, 13; triploidy); structural changes (micro-deletion/duplication, inversion, ring
chromosome, etc.)
OR
Specific gene mutations in a Mendelian disorder (e.g., betal thalassemia, cystic fibrosis, Duchenne muscular dystrophy)
OR
Mutations in 2 or more genes (oligo- or multigenic) belonging to a gene/ molecular family (e.g., sarcomere genes in hypertrophic
cardiomyopathy)
OR
Interaction of several hundreds and thousands of low risk alleles/genes with one or more environmental factors, including the lifestyle—
polygenic/ multifactorial
OR
Mitochondrial gene mutations and/or polymorphisms—several multisystem disorders that follow matrilineal inheritance pattern.
DISEASE S P ECT RUM , B IO L O G ICA L could be logically linked with one or more phenotypes. This
PAT H WAYS , AND G ENOTY P ES could be demanding and challenging, as it might involve
in-depth analysis and understanding of the complex bio-
In modern medicine, diagnosis of any disease or morbid logical (e.g., metabolic or molecular) pathways implicated
state relies on establishing the phenotype along the lines in the disease process (Table 20.9). Once this was achieved,
of agreed parameters (Table 20.8). The next logical step is then a correlation could be looked for with specific protein
to find evidence for likely pathophysiological changes that or enzyme systems recognized to be essential component(s)
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 1 3
Table 20.9 GENETIC AND GENOMIC PATHOLOGY IN HUMAN DISEASE
The following genetic/ genomic pathology might be associated with one or more clinical phenotypes. Interpretation and precise diagnosis
would depend on the natural history, family history, and sensitivity/ specificity of genetic/genomic analyses:
• Epigenetic/ epigenomics changes—mutations/ deletions/ duplication/ inversion of genes or genomic segments adjacent to the promoter
region of certain genes; genetic imprinting abnormality involving specific genes demonstrating “parent of origin effect,” including com-
plete, partial, or mosaic uniparental disomy.
• Genome-wide abnormalities—pathogenic or disease-modification effect of structural variation across the genome; for example,
single-nucleotide polymorphisms, copy number variations, deletions/ duplications, nucleotide repeats (e.g., trinucleotide repeats).
• Gene function/ expression—specific “gain of function” or “loss of function” gene mutations (e.g., increased risk of cancer/tumor due to
mutation in a tumor-suppressor gene); mutations in transcription factors associated with a range of developmental anomalies; abnormali-
ties in RNA interference system associated with exaggerated or blunted therapeutic response; post-translational modification/ changes in
the gene product associated with one or more phenotypes consistent with a disease diagnosis.
of the core and successive biological pathways. Finding explain complex pathogenesis in some disorders. The spec-
structural or functional abnormalities of any given protein trum of these disorders is wide and includes both acute and
or enzyme system would require undertaking investigations chronic medical and surgical diseases. Perhaps it is reasonable
targeted at specific gene(s) or genomic regions harboring to identify these disorders on the basis of their underlying
particular gene(s). Establishing the precise genotype would molecular pathology, including genomic imprinting, genomic
then be the final piece in the complex jigsaw puzzle that is rearrangements, and gene–environment interactions involv-
collectively labeled as a disease or syndrome. The individual ing multiple genes and genomic polymorphisms. This chapter
genotype could be in any form (Table 20.8,III) including has reviewed the genetic and genomic approaches in the clas-
gross chromosomal changes, specific genes or gene clus- sification of human disease. A stepwise approach is presented
ters, and extremely small segments of the genome. Thus the based on correlations of the clinical phenotype, supporting
whole landscape of the disease or diagnosis involves a closely investigative phenotypes and specific evidence from targeted
linked network of three domains in the order of genotype– genetic and genomic analyses. This approach would enable a
pathway–phenotype. In other words, a diagnosis of any dis- modern clinician to finally arrive at the final determining fac-
ease or morbid state (including the mortal state) would be tor in the causation of human disease. The new taxonomy of
a cumulative process that should take into account all three human disease is likely to have a major impact on the practice
of these domains: disease= genotype + molecular pathway of clinical medicine in the future.
+ genotype.
R EFE R ENCES
SUMMA RY
1. Gelehrter, T.D., F.S. Collins, and D. Ginsburg, Principles of medical
Developments in genetics and subsequently the sequencing genetics. 1998: Lippincott Williams & Wilkins, Philadelphia.
2. Kumar, D., Clinical medicine in the genome era: an introduc-
of the human genome have provided us with an opportunity tion. Genomics and Clinical Medicine, Oxford University Press,
to review the taxonomy of human disease. Conventionally, New York. 2008: p. 145.
3. Bell, J.I., The double helix in clinical practice. Nature, 2003.
the causation of human disease includes malformations, 421(6921): p. 414–416.
trauma, infection, immune dysfunction, metabolic abnor- 4. Pearce, J.M.S., Sir Thomas Lewis MD, FRS(1881-1945). Journal of
mality, malignancy, and degenerative conditions associated neurology, 2006. 253(9): p. 1246–1247.
5. Ioannidis, J.P.A., and F.K. Kavvoura, Concordance of functional in
with aging. Genetic factors have long been recognized in vitro data and epidemiological associations in complex disease genetics.
all of these disease groups. The traditional genetic catego- Genetics in Medicine, 2006. 8(9): p. 583–593.
ries of diseases include chromosomal disorders, single-gene 6. Daniels, S.E., et al., A genome-wide search for quantitative trait loci
underlying asthma. Nature, 1996. 383(6597): p. 247–250.
or Mendelian diseases, and several forms of multifactorial/ 7. Hall, J.G., et al., The frequency and financial burden of genetic disease
polygenic conditions. In addition, somatic genetic changes in a pediatric hospital. American journal of medical genetics, 1978.
and mutations of the mitochondrial genome probably 1(4): p. 417–436.
8. Henig, R.M., The monk in the garden: the lost and found genius
account for a small, albeit important, number of diseases. of Gregor Mendel, the father of genetics. 2000: Houghton Mifflin
These groups of disorders are well recognized and have an Harcourt, London, UK.
established place in the classification of human disease. 9. Weatherall, D.J., The new genetics and clinical practice. Vol. 12.
1991: Oxford University Press, Oxford, UK.
Recent developments in genome research have provided 10. Pulst, S.M., Neurogenetics: XA-GB. 2000: Oxford University Press,
a wealth of data indicating different genomic mechanisms to New York.
3 1 4 • G e n o m i c s i n Cl i n i c a l P r ac t i c e
11. Kumar, D., Disorders of the genome architecture: a review. Genomic alterations including imprinting defects of KCNQ1OT1. Human
medicine, 2008. 2(3–4): p. 69–76. Molecular Genetics, 2001. 10(26): p. 2989–3000.
12. Egger, G., et al., Epigenetics in human disease and prospects for epigen- 24. Lupski, J.R., and P. Stankiewicz, Genomic disorders: molecular mech-
etic therapy. Nature, 2004. 429(6990): p. 457–463. anisms for rearrangements and conveyed phenotypes. PLoS genetics,
13. Jaenisch, R., and A. Bird, Epigenetic regulation of gene expres- 2005. 1(6): p. e49.
sion: how the genome integrates intrinsic and environmental signals. 25. Gu, W., F. Zhang, and J.R. Lupski, Mechanisms for human genomic
Nature genetics, 2003. 33: p. 245–254. rearrangements. Pathogenetics, 2008. 1(1): p. 4.
14. Jones, P.A., and S.B. Baylin, The fundamental role of epigenetic events 26. Zhang, F., C. Carvalho, and J.R. Lupski, Complex human chro-
in cancer. Nature Reviews Genetics, 2002. 3(6): p. 415–428. mosomal and genomic rearrangements.Trends in Genetics, 2009.
15. Lachner, M., R.J. O’Sullivan, and T. Jenuwein, An epigenetic road 25(7): p. 298–307.
map for histone lysine methylation. Journal of cell science, 2003. 27. Garcia, C.A., et al., Clinical variability in two pairs of identical twins
116(11): p. 2117–2124. with the Charcot-Marie-Tooth disease type 1A duplication. Neurology,
16. Lapidot, M., and Y. Pilpel, Genome-wide natural antisense transcrip- 1995. 45(11): p. 2090–2093.
tion: coupling its regulation to its different regulatory mechanisms. 28. Lupski, J.R., Genomic disorders: recombination-based disease result-
EMBO reports, 2006. 7(12): p. 1216–1222. ing from genome architecture. American journal of human genetics,
17. Martinowich, K., H. Manji, and B. Lu, New insights into BDNF 2003. 72(2): p. 246.
function in depression and anxiety. Nature neuroscience, 2007. 29. Murakami, T., et al., Charcot-Marie-Tooth disease and related inher-
10(9): p. 1089–1093. ited neuropathies. Medicine, 1996. 75(5): p. 233–250.
18. Carmona-Mora, P., et al., Mouse models of genomic syndromes as tools 30. Potocki, L., et al., DNA rearrangements on both homologues of chro-
for understanding the basis of complex traits: an example with the mosome 17 in a mildly delayed individual with a family history of
Smith-Magenis and the Potocki-Lupski syndromes. Current genom- autosomal dominant carpal tunnel syndrome. American Journal of
ics, 2009. 10(4): p. 259. Human Genetics, 1999. 64(2): p. 471–478.
19. Murrell, A., et al., An association between variants in the IGF2 31. Margolis, R.L., et al., Trinucleotide repeat expansion and neuropsychi-
gene and Beckwith-Wiedemann syndrome: interaction between atric disease. Archives of general psychiatry, 1999. 56(11): p. 1019.
genotype and epigenotype. Human Molecular Genetics, 2004. 32. Greco, C.M., et al., Neuropathology of fragile X–associated tremor/
13(2): p. 247–255. ataxia syndrome (FXTAS). Brain, 2006. 129(1): p. 243–255.
20. Petropoulos, A.E., et al., Genetic analysis in the diagnosis of familial 33. Chen, Y.W., Local protein unfolding and pathogenesis of
paragangliomas. The Laryngoscope, 2000. 110(7): p. 1225–1229. polyglutamine-expansion diseases. Proteins: Structure, Function, and
21. Dannenberg, H., et al., Frequent germ-line succinate dehydroge- Bioinformatics, 2003. 51(1): p. 68–73.
nase subunit D gene mutations in patients with apparently sporadic 34. Jones, P.A., and P.W. Laird, Cancer-epigenetics comes of age. Nature
parasympathetic paraganglioma. Clinical cancer research, 2002. genetics, 1999. 21(2): p. 163–167.
8(7): p. 2061–2066. 35. Jones, L.K., and V. Saha, Chromatin modification, leukaemia and
22. Ovcharenko, I., et al., Evolution and functional classification of verte- implications for therapy. British journal of haematology, 2002.
brate gene deserts. Genome research, 2005. 15(1): p. 137–145. 118(3): p. 714–727.
23. Weksberg, R., et al., Tumor development in the Beckwith-Wiedemann 36. Reisman, D., S. Glaros, and E.A. Thompson, The SWI/SNF complex
syndrome is associated with a variety of constitutional molecular 11p15 and cancer. Oncogene, 2009. 28(14): p. 1653–1668.
G e n et i c a n d G e n o m i c Taxo n o m y o f H u m a n D i s e a s e • 3 1 5
21.
GENOMICS OF COMPLEX CARDIOVASCULAR DISEASE
Foram N. Ashar and Dan E. Arking
316
Table 21.1 COMMON GENETIC VARIANTS THAT INFLUENCE SUSCEPTIBILITY TO CARDIOVASCULAR EVENTS
efforts to identify common genetic variants associated with APOE (termed E2, E3, and E4, and with frequencies of 8%,
CAD and MI have produced a list of over 30 loci showing 77%, and 15%, respectively, in Caucasian populations) on
robust association that has been replicated across multiple type III familial hyperlipoproteinemia, in which more than
studies and cohorts (Kathiresan et al., 2009; Samani et al., 95% of affected individuals were homozygous for the rare
2007; Schunkert et al., 2011). allele. The common genotype, E3/E3, is used as the reference
We highlight the progress made in the search for genes group in most studies, and individuals who carry the E2 allele
associated with MI by both the candidate gene approach have ~14 mg/dl lower LDL levels, and E4 carriers have ~7
and the genome-wide association approach, by discussing mg/dl higher LDL levels (Motulsky and Brunzell, 2002).
four specific examples: apolipoprotein E (APOE), coagula- Numerous studies have examined the association
tion proteins, 9p21, and SORT1, in detail. between the E2/E3/E4 variants and coronary heart disease
(CHD), including a large-scale meta-analysis incorporating
121 studies with 37,850 cases and 82,727 controls (Bennet
A P O L I P O P ROT E I N E (A P O E)
et al., 2007). The meta-analysis, which was stratified by
Apolipoprotein E, a low-density lipoprotein (LDL) receptor the number of participants in an individual study at a cut-
ligand, is an important player in the metabolism of cholesterol off of at least 1,000 healthy controls and 500 cases, dem-
and triglycerides, where it mediates the clearance of chylomi- onstrated a moderate increased risk for E4 carriers (odds
cron and very low-density lipoprotein (VLDL) from plasma. ratio [OR] 1.06, 95% confidence interval [CI], 0.99–1.13),
Utermann and colleagues (Utermann et al., 1977) first and a significant decreased risk for E2 carriers (OR 0.80,
described the effects of three common allelic variants of 95% CI 0.70–0.90) in the group with larger study size. In
CHR GENES OF INTEREST WITHIN OR SNP ODDS RATIO (95% CI) ASSOCIATION OF SNP OR PROXY WITH
NEAR ASSOCIATED INTERVAL PER RISK ALLELE OTHER CARDIOVASCULAR PHENOTYPES
1p13 CELSR2, PSRC1, SORT1 rs599839 1.11 (1.08–1.15) eQTL for SORT1, CELSR2, and PSRC1 tran-
script levels in liver
1p32 PCSK9 rs11206510 1.08 (1.05–1.11)
1p32 PPAP2B rs17114036 1.17 (1.13–1.22)
1q41 MIA3 rs17465637 1.14 (1.09–1.20)
2p21 ABCG5, ABCG8 rs4299376 1.07 (1.04–1.11) Serum phytosterols
2q33 WDR12, NBEAL1 rs6725887 1.14 (1.09–1.19) eQTL for NBEAL1 transcript level in aortic
media
3q22 MRAS rs9818870 1.12 (1.07–1.16) eQTL for MRAS transcript level in aortic media
6p21 ANKS1A rs17609940 1.07 (1.05–1.10)
6p24 PHACTR1 rs12526453 1.10 (1.06–1.13) Coronary artery calcification
6p24 c6orf105 rs6903956 1.65 (1.44–1.90)
6q23 TCF21 rs12190287 1.08 (1.06–1.10) eQTL for TCF21 transcript level in liver, fat
6q25 LPA rs3798220 1.51 (1.33–1.70) Lipoprotein(a)
6q26 LPA rs10455872 1.68 (1.43–1.98) Lipoprotein(a)
7q22 BCAP29, DUS4L rs10953541 1.08 (1.05–1.11)
7q32 ZC3HC1 rs11556924 1.09 (1.07–1.12)
8q24 TRIB1 rs17321515 1.06 (1.03–1.10) Triglycerides, HDL cholesterol
9p21 CDKN2A, CDKN2B, ANRIL rs4977574 1.29 (1.23–1.36) Coronary artery calcification, intracranial aneu-
rysm, abdominal aortic aneurysm, among others
9q34 ABO rs579459 1.10 (1.07–1.13) Venous thromboembolism, ACE enzyme activ-
ity, plasma E-selectin level, plasma vWF level,
among others
10p11 KIAA1462 rs2505083 1.07 (1.04–1.09)
10q11 CXCL12 rs1746048 1.09 (1.07–1.13) Plasma CXCL12 level
10q23 LIPA rs1412444 1.09 (1.07–1.12) eQTL for LIPA transcript level in monocytes
10q24 CYP17A1, CNNM2, NT5C2 rs12413409 1.12 (1.08–1.16) Intracranial aneurysm
11q22 PDGFD rs974819 1.07 (1.04–1.09) eQTL for PDGFD transcript level in aortic
media
11q23 ZNF259, APOA5, APOA1 rs964184 1.13 (1.10–1.16) Triglycerides, HDL cholesterol
12q24 SH2B3 rs3184504 1.07 (1.04–1.10) LDL cholesterol, platelet count, plasma eosino-
phil count, among others
13q34 COL4A1, COL4A2 rs4773144 1.07 (1.05–1.09)
14q32 HHIPL1 rs2895811 1.07 (1.05–1.10)
15q25 ADAMTS7 rs3825807 1.08 (1.06–1.10)
17p11 RASD1, PEMT, RAI1 rs12936587 1.07 (1.05–1.09) eQTL for RASD1 and PEMT transcript levels
in monocytes
17p13 SMG6 rs216172 1.07 (1.05–1.09)
17q21 UBE2Z rs46522 1.06 (1.04–1.08) eQTL for UBE2Z transcript level in blood
19p13 LDLR rs6511720 1.18 (1.11–1.25)
19q13 APOE rs2075650 1.14 (1.09–1.19)
21q22 SLC5A3, MRPS6, KCNE2 rs9982601 1.18 (1.12–1.24) eQTL for MRPS6 transcript level in blood
*Adapted from Kathiresan and Srivastava, 2012.
1 2 3
SCD
Figure 21.1
Potential genetic contributors to sudden cardiac death (SCD). Potential and documented elements of susceptibility are suggested in
three broad pathways: 1) those that lead to progressive atherosclerosis and frank coronary disease and the likelihood of an occlusive infarction and
ischemic arrhythmias; 2) those involved in electrogenesis and intromyocardial conduction pathways; and 3) those that may influence the initiation
of aberrant triggering events and the perpetuation of an arrhythmia. Adapted from Spooner et al., 2001, and Arking et al., 2004.
traits (Eijgelsheim et al., 2010; Sotoodehnia et al., 2010) has that additional genes play a role in SCD, and they are not
allowed a more comprehensive assessment of the role of likely to be identified through these approaches. Several
electrocardiogram (ECG)-associated SNPs in SCD risk. In GWAS with SCD as the phenotype of interest have been
one such study, using data generated from an SCD GWAS published, two of which have reported genome-wide sig-
composed of 1,283 SCD cases and >20,000 controls, Arking nificant findings. The AGNES (Arrhythmia Genetics in
and colleagues (Arking et al., 2011) examined 49 SNPs the Netherlands) cohort, which is composed of individuals
associated (p < 5 × 10–8) with QRS, QT, and RR inter- with a first myocardial infarction and ventricular fibrillation
vals. In a test looking at direction of the genetic effect, the (VF) who survived to hospital admission (n = 515) com-
ECG-trait-prolonging allele was significantly more often asso- pared with individuals with myocardial infarction alone
ciated with increased risk of SCD (31 of 49, p = 0.03), with (n = 457), reported a SNP, rs2824292 in the 21q21 locus,
this effect almost entirely due to the QRS/QT-associated with an OR of 1.78 (95% CI 1.47–2.13, p = 3.36 × 10–10),
SNPs (28 of 40, p = 0.006). Three loci, including PLN, and an OR of 1.49 (95% CI 1.14–1.95, p = 0.004) in a
KCNQ1, and NOS1AP, showed nominal association with replication sample of 146 out-of-hospital SCD cases and
SCD, while a fourth locus, TKT/CACNA1D/PRKCD, was 391 controls (Bezzina et al., 2010). This SNP is a common
significant even after Bonferroni correction to account for the variant (allele frequency of 47%) located in an intergenic
number of loci tested. The TKT/CACNA1D/PRKCD asso- region. The nearest gene, CXADR (~100 kb away), encodes
ciation is particularly intriguing due to the observation that a viral receptor implicated in viral myocarditis (Bowles
the QRS-prolonging allele was protective for risk of SCD, et al., 1986), but it is not directly implicated by this study.
which is counter to the effect observed with the measured
trait (longer QRS duration is associated with increased risk
BA Z2B
of SCD). This result raises the possibility that the effect of the
SNP variant on risk of SCD may not be mediated through A second GWAS, comprising a total of 4,400 SCD cases and
its effect on QRS interval. A similar result was also seen in >30,000 controls, all of European ancestry, reported a sig-
NOS1AP, where one of the alleles that were associated with nificant signal at the 2q24.2 locus, with the strongest SNP,
SCD had no effect on QT interval (Kao et al., 2009). rs4665058 (p = 1.8 × 10–10), mapped to an intron of the
BAZ2B gene. This locus contains three genes expressed in
the heart but not previously known to play a role in cardiac
21q21
biology (BAZ2B, WDSUB1, and TANC1). The risk allele
While focusing on candidate genes and endophenotypes had a study-size-weighted frequency of 1.4% and increased
has yielded several compelling candidates, there is no doubt the risk for SCD by 1.92-fold per allele (95% CI 1.57–2.34).
337
Genetic factors Genetic factors
1 PPAR g gene (Pro12Ala) Deeb et al., 1998; Lyssenko et al., 2005 Pro12Ala polymorphism protective in Caucasians
2 PPAR g gene (Pro12Ala) Radha et al., 2006 Pro12Ala polymorphism not protective in South Asians
3 PGC-1 α gene (Gly482Ser) Ek et al., 2001 Associated with relative risk of type 2 diabetes in a
European population
4 PGC-1 α gene (Thr394Thr) Vimaleswaran et al., 2005, 2006 Associated with type 2 diabetes and with body fat
5 PGC-1 α gene (Gly482Ser) Liang et al., 2006 Associated with the development of insulin resistance and
type 2 diabetes
6 PGC-1 α gene (Thr394Thr, Bhat et al., 2007 Associated with type 2 diabetes
Gly482Ser)
7 PC-1 gene (K121Q) Bacci et al., 2005 Associated with insulin resistance/atherogenic phenotypes
8 PC-1 gene (K121Q) Abate et al., 2005 Associated with type 2 diabetes
9 TCF7L2 gene(rs7903146) Grant et al., 2006; Scott et al.; T allele of rs7903146 associated with an increased risk of
Cauchi et al. type 2 diabetes
10 TCF7L2 gene(rs7903146) Chandak et al., 2007 Associated with type 2 diabetes
11 TCF7L2 gene(rs12255372; Bodhini et al., 2007 Associated with type 2 diabetes in Asian Indians
rs7903146)
12 TCF7L2 gene(rs7903146) Sanghera et al., 2008 Associated with type 2 diabetes in Asian Indians
13 Adiponectin gene (+45T/G; Hara et al., 2001; Menzaghi et al., 2002; Significantly associated with type 2 diabetes and adiponec-
+276G/T) Vasseur et al., 2002 tin level in Japanese population and with insulin resistance
in some Caucasian populations
14 Adiponectin gene Stumvoll et al., 2002 SNP 45 is associated with obesity in a German population
15 Adiponectin gene Vimaleswaran et al., 2008 +10211T→G Associated with type 2 diabetes
16 FTO gene Frayling et al., 2007 Associated with body mass index
17 FTO gene Dina Associated with obesity-related traits
18 FTO gene Yajnik et al., 2009 Associated with type 2 diabetes in south Asian Indians
19 FTO gene Ramya et al., 2010 Associated with type 2 diabetes and obesity in South Asian
Indians
increased insulin sensitivity and decreased risk of T2D in The SNP E23K of KCNJ11 has now been convincingly
a Finnish and a second-generation Japanese cohort. Since associated with T2D. Although initial smaller studies failed
this initial work, the preponderance of evidence has sup- to replicate the association of the E23K polymorphism
ported PPARG’s association with T2D, with an odds ratio with T2D, large-scale studies and meta-analyses have con-
(OR) of ~1.2. The risk of T2D conferred by this SNP sistently associated the lysine variant with T2D, with an
has been studied prospectively in the Finnish Diabetes OR of 1.15 (Moore et al., 2008).
Prevention Study and the larger Botnia Prevention Study. Adiponectin, encoded by the ADIPOQ gene, is one
In the Finnish study (Uusitupa et al., 2001), 500 subjects of the adipocyte-expressed proteins that enhances insulin
with impaired glucose tolerance, the relative risk of devel- sensitivity and functions in regulating the homeostatic con-
oping diabetes was doubled in alanine carriers, contradict- trol of glucose, lipid, and energy metabolism (Diez et al.,
ing the prior evidence that the alanine allele was protective. 2003). Genome-wide scans have mapped a susceptibility
In the larger Botnia study, comprising >2000 subjects, pro- locus for type 2 diabetes and obesity/metabolic syndrome
line homozygotes were 1.7 times more likely to develop to chromosome 3q27, where the ADIPOQ gene is located
diabetes than alanine carriers. In contrast in our group we (Kissebah et al., 2000; Vionnet et al., 2000; Comuzzie
found that the Pro12Ala polymorphism of the PPAR–G et al., 2001; Lindsay et al., 2003) SNPs of the ADIPOQ
gene, which is protective against diabetes in Caucasians, gene have been genotyped in large datasets from various
does not offer protection in two cohorts of South Asians ethnic groups, and several SNPs associated with hypoadi-
studied at Chennai, India, and Dallas in the United States ponectinemia, obesity, and type 2 diabetes have been iden-
(Radha et al., 2006). tified (Menzaghi et al., 2002; Vasseur et al., 2003; Gibson
1 CDKAL1 gene Chidambaram et al., 201 Associated with type 2 diabetes in south Indians
CDKN2A/B gene
HHEX gene
BAZ1B gene
2 GRB14 gene Kooner et al., 2011 Associated with type 2 diabetes in south Asian
ST6GAL1 gene ancestry including south Indians
VPS26A gene
AP3S2 gene
HMG20A gene
HNF4A gene
3 TCF7L2 gene Tabassum et al., 2012 Associated with type 2 diabetes in Indians
TMEM163 gene
TMEM163 gene
MAP3K1 gene
TGFBR3 gene
FLJ35379 gene
The GWAS that discovered the first loci associated linked through obesity (Ramya et al., 2011). FTO is still
with BMI was part of the Welcome Trust Case Control considered the locus with the largest effect on BMI, which,
Consortium studies examining the genetics of type 2 dia- together with its high minor allele frequency, implies that
betes (Frayling et al., 2007). Genetic variation in 1,924 it is a low-hanging fruit that could be easily identified by
individuals with type 2 diabetes was compared with moderately powerful studies.
that in 2938 controls. A SNP in the FTO (fat mass and The second wave of discoveries was possible because of
obesity-associated) gene was found to show strong associa- the increase in their sample size. In order to increase the
tion with type 2 diabetes, which was replicated in the sec- statistical power in the analysis of gene variants associated
ond stage using a larger sample size. Once the analyses were with BMI, a multinational collaboration known as the
adjusted for BMI, the association with T2D was abolished. Genetic Investigation of ANthropometric Traits (GIANT)
The effect of FTO on T2D was on a pathway through BMI. Consortium was established. This study confirmed the
The association with BMI was subsequently established strong association with variation in FTO and identified one
by replicating in a sample comprising 19,424 adults from new locus near melanocortin 4 receptor (MC4R), a gene in
seven studies and 10,172 children from two different stud- which mutations are a common cause of extreme childhood
ies (Grant et al., 2008). Thus came the conclusion that the obesity (Cauchi et al., 2009; Heid et al., 2010). At the same
FTO locus affects diabetes through its effect on adiposity. time, the GWAS in subjects of Asian Indian origin identi-
Subsequently, numerous replication studies also confirmed fied the same locus to be associated with waist circumfer-
the association of the FTO locus with BMI and related obe- ence and related traits. This finding was further replicated in
sity traits. Earlier studies in Asian Indians on rs9939609 South Asians, East Asians, various white European popula-
T→A and rs7193144 C→T variants of the intron 1 of the tions, and African Americans (Chambers et al., 2008). The
FTO gene showed an association with type 2 diabetes that sample size was doubled in a third wave of discoveries by
was independent of BMI (Chauhan et al., 2011). Our recent the GIANT Consortium, using 32,387 individuals of white
association study showed that the rs8050136 C→A vari- European descent, and identifying six new loci robustly
ant is associated with both Generalised Obesity is the one associated with BMI. Ten loci were near TMEM18, near
in which BMI is taken as the measure of obesity. Central KCTD15, near GNPDA2, in SH2B1, in MTCH2, near
Obesity is the one in which waist circumference and waist NEGR1, near FAIM2, near SEC16B, near ETV5, and in
hip ratio are the measures of obesity. The rs8050136 C/A BDNF (Willer et al., 2009; Thorleifsson et al., 2009).
polymorphism was associated with generalized obesity. The In the fourth wave of discoveries, the GIANT
odds ratio for obesity for the CA genotype was significant Consortium expanded its GWAS stage to comprise 123,865
even after adjustment for age, sex, and diabetes [OR: 2.0 individuals from 46 populations of white European descent.
(95% CI: 1.57 –2.76), p <0.0001]. This study demonstrated By the end of the fourth wave, GWAS had identified 32 loci
that there is no independent association of rs8050136 C→A unequivocally associated with BMI (Speliotes et al., 2010;
with T2DM, as its ((association with T2DM appears to be Day et al., 2011; Loos, 2012). Although the number of
352
Physical risk factors Environmental risk factors
Secondary
osteoporosis
Hyperthyroidism Hyperparathyroidism
Cushing’s syndrome Osteomalacia
Glucocorticoid therapy Bone metastasis of cancer
Rheumatoid arthritis Multiple myeloma
Diabetes mellitus
Figure 23.1
Pathogenesis of osteoporosis. Primary osteoporosis is a bone disorder of relatively unknown origin that occurs with aging and
accelerates after menopause. There is no direct or singular cause for primary osteoporosis. There are two types of primary osteoporosis. Type
I (postmenopausal) osteoporosis develops when estrogen levels decline in postmenopausal women. Type II (senile) osteoporosis occurs in elderly
people when bone formation cannot keep up with the natural process of bone loss. Secondary osteoporosis is caused by the use of corticosteroids
or by various medical conditions, including hormonal imbalances, certain cancers, diabetes mellitus, and rheumatoid arthritis. In addition to
environmental and physical risk factors, genetic factors and interactions between multiple genes and environmental factors are important in the
development of primary osteoporosis.
Functional analysis
Figure 23.2
Strategies for identifying disease susceptibility genes. There are two basic strategies for identifying genes that influence common diseases or
other complex traits, the genome-wide approach and the candidate-gene approach, both of which rely on linkage analyses and association studies. In
the genome-wide association study (GWAS), single-nucleotide polymorphisms (SNPs) or copy number variations (CNVs) distributed throughout
the entire genome are used to identify genomic regions that harbor genes that influence the trait of interest with a detectable effect size. The
candidate-gene approach involves the direct examination of whether an individual gene or genes might contribute to the trait of interest.
3 5 4 • G e no m ic s in C l inic a l Pr actic e
candidate-gene approach is not, however, able to identify linkage disequilibrium with nearby genes. In the following
disease-associated polymorphisms in unknown genes. In sections, two candidate genes (VDR and COL1A1) that
the genome-wide scan, single-nucleotide polymorphisms are of particular interest in the genetics of osteoporosis are
(SNPs) or copy number variations (CNVs) distributed reviewed.
throughout the entire genome are used to identify genomic
regions that harbor genes that influence the trait of interest
Vitamin D Receptor (VDR)
with a detectable effect size. This is a hypothesis-generating
approach, allowing the detection of previously unknown Vitamin D is a potent regulator of bone and calcium homeo-
potential trait loci. stasis as well as of cellular differentiation and proliferation
in many tissues. Its active form, 1,25-dihydroxyvitamin D3
(calcitriol), interacts with the highly specific vitamin D
L I N K AG E A NA LYS I S O F B M D O R
receptor (VDR) and thereby affects the expression of target
O S T EO P O RO S I S
genes. A Bsm I restriction fragment length polymorphism
The published results of whole-genome and partial-genome of VDR was associated with BMD in Australian women[67].
linkage analyses for BMD or osteoporosis are summa- Of the many studies performed subsequently, some[68–70]
rized in Table 23.1. Initial linkage analyses found that a have supported this association, whereas others[71–73] have
locus linked to low BMD mapped to chromosomal region not. These apparently contradictory results are possibly
1p36.3-36.2[59–61], whereas a locus linked to high BMD attributable to differences in several factors among the
mapped to 11q12-13[62]. Genes responsible for other studies, including size, age, and ethnic background of the
BMD-related phenotypes, such as autosomal recessive study populations, as well as environmental aspects such as
osteoporosis-pseudoglioma syndrome[63] and autosomal diet, especially vitamin D and calcium intake. In addition,
dominant osteopetrosis[64], are also located at 11q12-13. the effects of VDR genotype on BMD appear to be rela-
Osteoporosis-pseudoglioma syndrome and high bone tively small. A meta-analysis of 16 studies concluded that
mass syndrome are caused by loss-of-function and activat- BMD was 2.5% lower for the spine and 2.4% lower for the
ing mutations, respectively, of LRP5 at this locus[34–36], as femoral neck in individuals with the BB genotype (B allele,
described in detail in a later section below. absence of the Bsm I restriction site) than in those with the
Additional studies have indicated that BMD is linked bb genotype (b allele, presence of the restriction site)[74].
to multiple chromosomal loci and that such loci differ Another meta-analysis of 75 studies, including those exam-
for BMD at different sites (Table 23.1). Osteoporosis ining four polymorphisms (Bsm I, Apa I, Taq I, and Fok I),
has also been linked to chromosomal region 20p12.3[65]. confirmed the association of VDR polymorphisms with
Family-based genome-wide scans have revealed that the BMD[75]. Furthermore, a more recent meta-analysis of the
loci that affect BMD are largely gender, age, and site relation between VDR Bsm I genotype and either BMD
specific[66]. or osteoporosis in women revealed that individuals with
the BB genotype had a lower BMD than did those with
the Bb or bb genotypes at baseline, which led to a greater
C A N D I DAT E - G E N E A S S O C I AT I O N
proportional loss in BMD in those with the BB genotype
S T U D I E S O F B M D, O S T EO P O RO S I S ,
over time[76].
O R FR AC T U R E S
Determination of the nucleotide sequence of human
Polymorphisms of a variety of candidate genes have been VDR[77] revealed two potential translation initiation sites, the
associated with BMD or with susceptibility to osteoporosis most 5′ of which is affected by a T→C SNP (ATG→ACG).
or fractures (Table 23.2). It is also possible that polymor- Individuals with the T allele of this SNP thus have two start
phisms in these genes are in linkage disequilibrium with sites and are able to initiate translation from the first ATG
other polymorphisms in nearby genes that are the actual codon, whereas those with the C allele initiate translation at
determinants of BMD or susceptibility to osteoporosis or the second ATG. The predicted protein produced by initia-
fracture. Despite the identification of a variety of candidate tion at the first ATG is three amino acids longer than is that
genes related to osteoporosis or BMD, the replicability of generated by initiation at the second start site. The T→C SNP
such findings is relatively low, mainly because of the lim- of VDR has been associated with BMD in postmenopausal
ited population size of the studies, the ethnic diversity of Mexican-American women[78], premenopausal American
gene polymorphisms, gene–gene interactions, complicating white women[79] and Japanese women[80], with the TT geno-
environmental factors, gene–environment interactions, and type implicated as a risk factor for reduced BMD. However,
3 5 6 • G e no m ic s in C l inic a l Pr actic e
Table 23.2 Candidate Gene Association Studies of BMD, Osteoporosis, or Fractures
no association of this SNP with BMD was detected in pre- Collagen, Type I, Alpha 1 (COL1A1)
menopausal American black[79] or premenopausal French[81]
Collagen, type I is the most abundant protein of bone
women. The T→C SNP of VDR was found to affect not
matrix. Mutations in the coding regions of the genes for
only the molecular mass of the encoded protein (T allele, 50
the two collagen, type I chains (COL1A1 and COL1A2)
kDa; C allele, 49.5 kDa), but also the transcriptional activa-
result in a severe autosomal dominant pediatric condition
tion of the gene by vitamin D (T allele < C allele)[80]. This
known as osteogenesis imperfecta[87]. A G→T SNP at the
latter observation was not independently confirmed, how-
first base of a consensus binding site for the transcription
ever[82]. The functional impact of this SNP thus remains to
factor Sp1 in the first intron of COL1A1 was associated not
be determined.
only with BMD in white women[52] but also with osteopo-
A large meta-analysis for 26,242 subjects failed to
rotic fracture in postmenopausal women[53, 88]. Other stud-
demonstrate any relation of Bsm I, Apa I, Taq I, or Fok
ies, however, showed only a weak association of this SNP
I polymorphism to BMD or fracture[83]. In addition, the
with BMD or osteoporotic fracture in premenopausal
candidate-gene meta-analysis in 19,195 subjects showed
French women[89], or a lack of association in postmeno-
no significant association of VDR genotypes with BMD or
pausal Swedish women[90], in American women[91], or in
fracture[84].
postmenopausal Danish women[92]. The T allele of the Sp1
A –3731A→G SNP that affects the binding site of the
binding site polymorphism affects collagen gene regulation
caudal-related homeodomain protein Cdx-2 in the VDR
in such a manner that it increases the production of the
promoter was associated both with transcriptional activ-
α1(I) collagen chain relative to that of the α2(I) chain and
ity of the promoter and with BMD for the lumbar spine
leads to reduced bone strength by a mechanism that is partly
in Japanese women, with the G allele corresponding to
independent of bone mass[93]. A –1997G→T SNP in the
reduced transcriptional activity and low BMD[85]. The A
COL1A1 promoter was also associated with BMD for the
allele of this polymorphism exhibited a protective effect
lumbar spine[94]. The –1997G→T SNP and the G→T SNP
on fracture, consistent with the association of the G allele
of the Sp1 binding site were in linkage disequilibrium[94].
with reduced BMD[86]. An association of the –3731A→G
A meta-analysis of the effect of the Spl binding site poly-
polymorphism with BMD and fracture was also detected
morphism of COLIA1 on the prevalence of fractures in
in a large meta-analysis[83]. These various observations
3641 subjects revealed that the risk was 1.7-fold greater in ss
thus suggest that VDR may be a susceptibility gene for
(TT) homozygotes versus SS (GG) homozygotes, 1.4-fold
osteoporosis or fractures, although a VDR locus has not
greater in ss homozygotes versus Ss (GT) heterozygotes, and
attained a significant association with these conditions by
1.3-fold greater in Ss heterozygotes versus SS homozygotes.
GWASs.
3 5 8 • G e no m ic s in C l inic a l Pr actic e
The effects of this polymorphism were slightly increased genes. GWASs represent a substantial advance in the search
when the analysis was limited to vertebral fractures for genetic variants that confer susceptibility to multifacto-
(odds ratios, 2.1, 1.5, and 1.3, respectively)[95]. Another rial polygenic diseases. GWASs, however, have the disad-
meta-analysis of an association of this polymorphism vantages that currently available marker sets are designed to
with BMD or osteoporotic fracture in 7849 participants identify common alleles and are not well suited to study-
revealed that BMD for the lumbar spine was significantly ing the effects of rare polymorphisms (< 1% to 5% popula-
lower in individuals with the Ss genotype than in those tion frequency) within a gene of interest[25]. The published
with the SS genotype. BMD for the femoral neck was also results of GWASs for BMD, osteoporosis, and fractures are
lower in individuals with the Ss genotype or those with the summarized in Table 23.3.
ss genotype than in those with the SS genotype. Analysis of In the following sections, six genes (TNFRSF11B,
the prevalence of fractures showed an increased odds ratio TNFSF11, TNFRSF11A, LRP5, ESR1, and SOST) that
for any fracture in subjects with the Ss genotype (1.3) and were identified by GWASs and are of particular interest in
an even greater increase in those with the ss genotype (1.8). the genetics of osteoporosis are reviewed.
Increased risk was largely attributable to vertebral fracture,
for which the odds ratio was 1.4 for individuals with the
Tumor Necrosis Factor Receptor Superfamily,
Ss genotype and 2.5 for those with the ss genotype[96]. In
Member 11b (TNFRSF11B)
the study of 20,786 subjects, the Sp1 polymorphism of
COL1A1 was associated with BMD for the spine and hip Tumor necrosis factor (TNF) receptor superfamily, mem-
and vertebral fractures in women[97]. These observations ber 11b (TNFRSF11B, or osteoprotegerin) is a secreted
thus suggest that genetic variants that affect collagen, type glycoprotein that was independently identified by three
I metabolism are important determinants of the develop- groups of research workers[99–101]. In vitro studies suggest
ment of osteoporosis and fractures, although COL1A1 has that TNFRSF11B inhibits osteoclastogenesis by inter-
not attained a significant association with these conditions rupting intercellular signaling between osteoblastic stromal
by GWASs. cells and osteoclast progenitors[99]. TNFRSF11B-deficient
mice exhibit a condition similar to juvenile Paget’s disease
that is characterized by a marked decrease in trabecular and
G E N O M E -WI D E A S S O C I AT I O N
cortical bone density, pronounced thinning of the parietal
S T U D I E S O F B M D, O S T EO P O RO S I S ,
bone of the skull, and a high incidence of fractures[102],
O R FR AC T U R E S
whereas hepatic expression of TNFRSF11B in transgenic
Candidate-gene association studies have substantial limita- mice results in osteopetrosis and a coincident decrease in
tions for detecting the genetic basis of osteoporosis because the proportion of osteoclasts in the later stages of differ-
this approach relies on the selection of the genes for associa- entiation[99]. The systemic administration of recombinant
tion studies based on either a biological hypothesis or the TNFRSF11B also results in a marked increase in BMD in
location of a particular gene in implicated linkage regions. normal rats as well as in prevention of bone loss in ovari-
In addition, most candidate-gene association studies have ectomized rats[99,103]. Furthermore, a single subcutaneous
generated inconsistent or inconclusive results[98]. The injection of TNFRSF11B reduced bone resorption in post-
recent development of high-density genotyping arrays has menopausal women[104]. Similar treatment with a recombi-
improved the resolution of unbiased genome-wide scans for nant TNFRSF11B construct suppressed bone resorption in
common variants associated with multifactorial diseases. patients with multiple myeloma or breast cancer with bone
Currently, the GWAS makes use of high-throughput geno- metastases[105].
typing technologies that include up to 4.3 million markers Osteoclastogenesis is regulated by three TNF– or TNF
for SNPs and CNVs to examine their relation to clinical receptor–related proteins: tumor necrosis factor recep-
conditions or measurable traits. As of January 20, 2012, the tor superfamily, member 11a (TNFRSF11A, or recep-
Catalog of Published Genome-Wide Association Studies tor activator of nuclear factor–κB [RANK])[106,107], tumor
(National Human Genome Research Institute, National necrosis factor superfamily, member 11 (TNFSF11, or
Institutes of Health [NIH]; http://www.genome.gov/ RANK ligand [RANKL])[108,109], and TNFRSF11B[99–101].
gwastudies/) included 1152 publications and 5637 SNPs TNFSF11 expressed on the surface of bone marrow stromal
associated with various diseases or traits, many in genes cells induces the differentiation of osteoclasts, enhances the
not previously suspected of having a role in the condition activity of mature osteoclasts, and inhibits osteoclast apop-
studied, and some in genomic regions containing no known tosis by binding to its functional receptor, TNFRSF11A,
3 6 0 • G e no m ic s in C l inic a l Pr actic e
Table 23.3 CONTINUED
CHROMOSOMAL DBSNP NUCLEOTIDE GENE (NEARBY GENE) PHENOTYPE REFERENCE
LOCUS SUBSTITUTION
11q13.4 rs3736228 C→T (Ala1330Val) LRP5 BMD, osteoporo- 121
sis, fracture
rs599083 T→G BMD 124
12q13 rs10876432 A→G SP7 BMD 130
rs2016266 G→A 124
13q14 rs9594738 C→T, TNFSF11 BMD 122
rs9594759 C→T
rs1021188 T→C 127
rs9533090 C→T AKAP11 BMD 124
14q32 rs2010281 A→G MARK3 BMD, fracture 130
14q32.12 rs1298989 T→C, CATSPERB BMD 163
rs1285635 A→G
16p13 rs13336428 A→G CLCN7 BMD 157
16p11 rs8057551 A→G, IL21R BMD 162
rs8061992 C→A,
rs7199138 C→G
16q23 rs16945612 T→C, ADAMTS18 BMD 156
rs11859065 G→A,
rs11864477 T→C,
rs11860781 A→T
16q24 rs10048146 A→G FOXL1 BMD 124
17q12 rs9303521 G→T CRHR1 BMD 124
17q21 rs1513670 G→A, SOST BMD, fracture 130
rs7220711 A→G,
rs1107748 C→T
rs228769 C→G HDAC5 BMD 124
18q11.2 rs7227401 G→T OSBPL1A BMD 123
18q21 rs3018362 G→A TNFRSF11A BMD, fracture 130, 122
rs884205 G→T BMD 124
20p12.1-p11.23 rs2273061 A→G JAG1 BMD 164
expressed on osteoclasts or their progenitors[108–111]. The Several SNPs have been detected in TNFRSF11B, some
interaction between TNFSF11 and TNFRSF11A is antag- of which have been shown to be associated with BMD or
onized by TNFRSF11B, which acts as a decoy receptor for fractures. Both 209G→A and 245T→G SNPs of the
TNFSF11. The biological effects of TNFRSF11B include TNFRSF11B promoter were found to be associated with
inhibition of the later stages of osteoclastogenesis[99,103,112], BMD for the lumbar spine in postmenopausal Slovenian
suppression of the activation of mature osteoclasts[108,110], women, with the 209GA/245TG genotype representing
and induction of osteoclast apoptosis[113]. The balance a risk factor for reduced BMD[116]. Both 163A→G and
between TNFRSF11B and TNFSF11 may thus represent 245T→G SNPs were associated with vertebral fractures
an important determinant of bone resorption[112,114]. The in Danish women and men, with the G allele of each poly-
importance of TNFRSF11B in the regulation of bone morphism representing a risk factor for fracture[117]. The
remodeling in humans has been indicated by the occurrence 245T→G SNP was associated with BMD at various sites
of juvenile Paget’s disease, characterized by rapid remodel- in postmenopausal Japanese women, with the GG genotype
ing of woven bone, osteopenia, fractures, and progressive representing a risk factor for reduced BMD[118]. A 950T→C
skeletal deformity, in Navajo individuals homozygous for a SNP was also associated with BMD in postmenopausal
deletion of ~100 kb in TNFRSF11B[115]. Japanese women[119]. In contrast, SNPs of TNFRSF11B
3 6 2 • G e no m ic s in C l inic a l Pr actic e
remodeling. Some forms of osteopetrosis, especially those Ala1330Val) of LRP5 were significantly associated with
attributable to osteoclast dysfunction, are associated with BMD and fracture[139]. The LRP5 locus was also identified
increased fracture risk. However, this is not seen in the high as a significant determinant of BMD in the GWAS[121] and
bone mass syndrome. Four mutations of LRP5 (331G→T in the GWAS meta-analysis[124].
[Asp111Tyr], 511G→C [Gly171Arg], 724G→A The most likely functional candidates among LRP5
[Ala242Thr], and 758C→T [Thr253Ile]) have been identi- variants are a Val667Met (rs4988321) and an Ala1330Val
fied in five families affected with osteopetrosis[134]. (rs3736228). Promoter reporter assays showed that differ-
LRP5 polymorphisms have also been associated with ent haplotypes for the Val667Met and Ala1330Val poly-
endosteal hyperostosis and its variant, van Buchem dis- morphisms differently activated reporter-gene transcription,
ease. Like osteopetrosis, these conditions are marked by suggesting that they may be functional[140]. The various stud-
excessive accrual of bone, although, in this instance, such ies showed that rare mutations in LRP5 have major effects
accrual is limited to the inner (endosteal) surface, leading to on BMD, whereas common polymorphisms have smaller
obliteration of the medullary space. Two coding polymor- effects on BMD and fracture risk in the general popula-
phisms of LRP5 (640G→A [Ala214Thr] and 724G→A tion[25]. LRP5 may thus be an important regulator of bone
[Ala242Thr]) have been identified in individuals with these mass and determinant of predisposition to osteoporosis.
conditions[134].
Several studies have implicated common polymor-
Estrogen Receptor 1 (ESR1)
phisms of LRP5 in BMD variation in the general popula-
tion. A 2047G→A (Val667Met) SNP in exon 9 was thus The importance of estrogen receptor 1 in the regulation of
significantly associated with bone mineral content of the bone mass was indicated by the occurrence of osteoporo-
lumbar spine, with bone area, and with stature in a cohort sis in a man with a nonsense mutation in ESR1[33] as well
of 889 healthy white men and women, with the association as by the observation that BMD in mice lacking a func-
being most pronounced in adult men[56]. Haplotype analy- tional ESR1 allele is 20 to 25% less than that in wild-type
sis of five SNPs at the LRP5 locus suggested that additional mice[141]. Two SNPs have been identified in the first intron
genetic variation at this locus might also contribute to of ESR1: a T→C polymorphism that is recognized by the
determination of bone mass and size. SNPs of LRP5 were restriction endonuclease Pvu II (T and C alleles correspond
associated with BMD for the lumbar spine, total hip, or to the presence [p allele] and absence [P allele] of the restric-
femoral neck in a cohort of 909 British Caucasian adults[135]. tion site, respectively), and an A→G polymorphism that
Family studies identified one SNP (171346C→A in intron is recognized by Xba I (A and G alleles correspond to the
21) with a significant relation to BMD, with the association presence [x allele] and absence [X allele] of the restriction
being stronger in men than in women. The association of site, respectively). These SNPs, alone or in combination,
LRP5 polymorphisms with BMD was also observed as the have been associated with BMD in postmenopausal[54,55] or
over-representation of variants of three SNPs in osteopo- premenopausal[142,143] women or with the response to hor-
rotic probands compared with unrelated postmenopausal mone replacement therapy[144]. However, other studies did
women with increased BMD. In addition, an association not confirm these observations[145–150]. Although no asso-
between haplotypes of SNPs and BMD was detected by ciation was observed between BMD in women and either
comparison of the osteoporotic subjects with the indi- of these two ESR1 polymorphisms alone, an association
viduals with a high BMD. Association of SNPs in LRP5 was detected between BMD for the lumbar spine or femo-
with BMD has also been observed in Japanese[136] and ral neck and the haplotype of the SNPs[55]. In addition, a
Australian[137] women. Although a significant association microsatellite (TA repeat) polymorphism in the promoter
was observed between LRP5 SNPs and BMD for the hip region of ESR1 was associated with BMD and with the
or spine in a study of healthy premenopausal white women, prevalence of fractures in studies in which such an associa-
only a small proportion of the total variation in BMD was tion with the T→C or A→G SNPs in the first intron was
accounted for by these SNPs[138]. The genotyped SNPs thus not apparent[55,145,149]. No association was detected between
accounted for ~0.8% of the variation in BMD for the femo- estrogen responsiveness of BMD and ESR1 genotype in
ral neck and 1.1% of that for the lumbar spine, suggesting postmenopausal Korean women who had undergone hor-
that natural variation in and around LRP5 is not a major mone replacement therapy[148]. In contrast, women with
contributing factor to the observed variability in peak the TT genotype (Pvu II SNP) have been suggested to be
BMD at either of these sites in white women. In a study of relatively estrogen insensitive; those with the C allele thus
37,534 subjects, nonsynonymous variants (Val667Met and appeared to benefit more from the protective effect of
3 6 4 • G e no m ic s in C l inic a l Pr actic e
to predict the future risk for osteoporosis in each individual 19. Slemenda CW et al.: The genetics of proximal femur geometry,
distribution of bone mass and bone mineral density. Osteoporos Int
on the basis of measurement of BMD and genetic analyses. 6:178–182, 1996.
Furthermore, it may be possible to prevent an individual from 20. Keen RW et al.: Family history of appendicular fracture and risk of
having osteoporotic fractures by medical intervention based osteoporosis: a population-based study. Osteoporos Int 10:161–166,
1999.
on his or her genotypes for specific SNPs. In the future, we may 21. Deng HW et al.: Genetic determination of Colles’ fracture and dif-
have the ability to use specific agents particularized for certain ferential bone mass in women with and without Colles’ fracture.
genetic susceptibility factors, thereby increasing their efficacy J Bone Miner Res 15:1243–1252, 2000.
22. MacGregor AJ et al.: Genetic factors and osteoporotic fractures in
and limiting any side effects of treatment. Identification of sus- elderly people. Br Med J 320:1669–1670, 2000.
ceptibility genes for osteoporosis and clarification of the func- 23. Kannus P et al.: Genetic factors and osteoporotic fractures in elderly
tional relevance of genetic variants to this condition will thus people: prospective 25 year follow up of a nationwide cohort of
elderly Finnish twins. Br Med J 319:1334–1337, 1999.
contribute to the personalized prevention, early diagnosis, and 24. Ralston SH: Genetic control of susceptibility to osteoporosis. J Clin
treatment of osteoporosis. Endocrinol Metab 87:2460–2466, 2002.
25. Ralston SH et al.: Genetics of osteoporosis. Endocr Rev 31:629–
662, 2010.
26. Rauch F et al.: Osteogenesis imperfecta. Lancet 363:1377–1385,
2004.
R EFE R E N C ES 27. Glorieux FH et al.: Osteogenesis imperfecta type VI: a form of
brittle bone disease with a mineralization defect. J Bone Miner Res
1. Kanis JA et al.: The diagnosis of osteoporosis. J Bone Miner Res 17:30–38, 2002.
9:1137–1141, 1994. 28. Marini JC et al.: Consortium for osteogenesis imperfecta mutations
2. Ralston SH: Genetics of osteoporosis. Rev Endocr Metab Disord in the helical domain of type I collagen: regions rich in lethal muta-
2:13–21, 2001. tions align with collagen binding sites for integrins and proteogly-
3. Melton LJ III: How many women have osteoporosis now? J Bone cans. Hum Mutat 28:209–221, 2007.
Miner Res 10:175–177, 1995. 29. Barnes AM et al.: Deficiency of cartilage-associated protein in reces-
4. Cummings SR et al.: Bone density at various sites for prediction of sive lethal osteogenesis imperfecta. N Engl J Med 355:2757–2764,
hip fractures. The Study of Osteoporotic Fractures Research Group. 2006.
Lancet 341:72–75, 1993. 30. Cabral WA et al.: Prolyl 3-hydroxylase 1 deficiency causes a reces-
5. Dargent-Molina P et al.: Fall-related factors and risk of hip frac- sive metabolic bone disorder resembling lethal/severe osteogenesis
ture: The EPIDOS prospective study. Lancet 348:145–149, 1996. imperfecta. Nat Genet 39:359–365, 2007.
6. Faulkner KG et al.: Simple measurement of femoral geometry pre- 31. Barnes AM et al.: Lack of cyclophilin B in osteogenesis imperfecta
dicts hip fracture: the study of osteoporotic fractures. J Bone Miner with normal collagen folding. N Engl J Med 362:521–528, 2010.
Res 8:1211–1217, 1993. 32. Morishima A et al.: Aromatase deficiency in male and female sib-
7. Holbrook TL et al.: Dietary calcium and risk of hip fracture: 14 year lings caused by a novel mutation and the physiological role of estro-
prospective population study. Lancet ii:1046–1049, 1988. gens. J Clin Endocrinol Metab 80:3689–3698, 1995.
8. Hannan MT et al.: Risk factors for longitudinal bone loss in elderly 33. Smith EP et al.: Estrogen resistance caused by a mutation in the
men and women: The Framingham Osteoporosis Study. J Bone estrogen-receptor gene in a man. N Engl J Med 331:1056–1061,
Miner Res 15:710–720, 2000. 1994.
9. Krall EA et al.: Smoking increases bone loss and decreases intestinal 34. Gong Y et al.: LDL receptor-related protein 5 (LRP5) affects bone
calcium absorption. J Bone Miner Res 14:215–220, 1999. accrual and eye development. Cell 107:513–523, 2001.
10. Cooper C et al.: Physical activity, muscle strength, and calcium 35. Boyden LM et al.: High bone density due to a mutation in
intake in fracture of the proximal femur in Britain. Br Med J LDL-receptor-related protein 5. N Engl J Med 346:1513–1521, 2002.
297:1443–1446, 1988. 36. Little RD et al.: A mutation in the LDL receptor-related protein 5
11. Arden NK et al.: The heritability of bone mineral density, ultra- gene results in the autosomal dominant high-bone-mass trait. Am J
sound of the calcaneus and hip axis length: a study of postmeno- Hum Genet 70:11–19, 2002.
pausal twins. J Bone Miner Res 11:530–534, 1996. 37. Janssens K et al.: Mutations in the latency-associated peptide of
12. Flicker L et al.: Bone density in elderly women: a twin study. J Bone TGF-β1 cause Camurati-Engelmann disease. Nat Genet 26:273–
Miner Res 10:1607–1613, 1995. 275, 2000.
13. Howard GM et al.: Genetic and environmental contributions to the 38. Kinoshita A et al.: Domain-specific mutations in TGFB1 result in
association between quantitative ultrasound and bone mineral density Camurati-Engelmann disease. Nat Genet 26:19–20, 2000.
measurements: a twin study. J Bone Miner Res 13:1318–1327, 1998. 39. Janssens K et al.: Transforming growth factor-β1 mutations in
14. Pocock NA et al.: Genetic determinants of bone mass in adults: a Camurati-Engelmann disease lead to increased signaling by alter-
twin study. J Clin Invest 80:706–710, 1987. ing either activation or secretion of the mutant protein. J Biol Chem
15. Gueguen R et al.: Segregation analysis and variance components 278:7718–7724, 2003.
analysis of bone mineral density in healthy families. J Bone Miner 40. McGowan NW et al.: A mutation affecting the latency-associated
Res 10:2017–2022, 1995. peptide of TGFβ1 in Camurati-Engelmann disease enhances osteo-
16. Cummings SR et al.: Risk factors for hip fracture in white women. clast formation in vitro. J Clin Endocrinol Metab 88:3321–3326,
Study of Osteoporotic Fractures Research Group. N Engl J Med 2003.
332:767–773, 1995. 41. Balemans W et al.: Increased bone density in sclerosteosis is due to
17. Stewart A et al.: Prediction of fractures in perimenopausal women: a the deficiency of a novel secreted protein (SOST). Hum Mol Genet
comparison of dual energy X-ray absorptiometry and broadband 10:537–543, 2001.
ultrasound attenuation. Ann Rheum Dis 55:140–142, 1996. 42. Balemans W et al.: Identification of a 52 kb deletion downstream of
18. Garnero P et al.: Genetic influence on bone turnover in postmeno- the SOST gene in patients with van Buchem disease. J Med Genet
pausal twins. J Clin Endocrinol Metab 81:140–146, 1996. 39:91–97, 2002.
3 6 6 • G e no m ic s in C l inic a l Pr actic e
88. Langdahl BL et al.: An Sp1 binding site polymorphism in the 109. Yasuda H et al.: Osteoclast differentiation factor is a ligand for
COLIA1 gene predicts osteoporotic fractures in both men and osteoprotegerin/osteoclastogenesis-inhibitory factor and is identi-
women. J Bone Miner Res 13:1384–1389, 1998. cal to TRANCE/RANKL. Proc Natl Acad Sci U S A 95:3597–
89. Garnero P et al.: Collagen Iα 1 Sp1 polymorphism, bone mass, 3602, 1998.
and bone turnover in healthy French premenopausal women: The 110. Burgess TL et al.: The ligand for osteoprotegerin (OPGL) directly
OFELY Study. J Bone Miner Res 13:813–817, 1998. activates mature osteoclasts. J Cell Biol 145:527–538, 1999.
90. Liden M et al.: Polymorphism at the Sp1 binding site in the col- 111. Fuller K et al.: TRANCE is necessary and sufficient for
lagen type Iα 1 gene does not predict bone mineral density in osteoblast-mediated activation of bone resorption in osteoclasts. J
postmenopausal women in Sweden. Calcif Tissue Int 63:293–295, Exp Med 188:997–1001, 1998.
1998. 112. Takai H et al.: Transforming growth factor-β stimulates the pro-
91. Hustmyer FG et al.: Polymorphism at an Sp1 binding site of duction of osteoprotegerin/osteoclastogenesis inhibitory factor by
COL1A1 and bone mineral density in premenopausal female bone marrow stromal cells. J Biol Chem 273:27091–27096, 1998.
twins and elderly fracture patients. Osteoporos Int 9:346–350, 113. Akatsu T et al.: Osteoclastogenesis inhibitory factor suppresses
1999. osteoclast survival by interfering in the interaction of stromal cells
92. Heegaard A et al.: Lack of influence of collagen type Iα 1 Sp1 bind- with osteoclast. Biochem Biophys Res Commun 250:229–234, 1998.
ing site polymorphism on the rate of bone loss in a cohort of post- 114. Hofbauer LC et al.: Clinical implications of the osteoprotegerin/
menopausal Danish women followed for 18 years. Calcif Tissue Int RANKL/RANK system for bone and vascular diseases. JAMA
66:409–413, 2000. 292:490–495, 2004.
93. Mann V et al.: A COL1A1 Sp1 binding site polymorphism predis- 115. Whyte MP et al.: Osteoprotegerin deficiency and juvenile Paget’s
poses to osteoporotic fracture by affecting bone density and qual- disease. N Engl J Med 347:175–184, 2002.
ity. J Clin Invest 107:899–907, 2001. 116. Arko B et al.: Sequence variations in the osteoprotegerin gene
94. Garcia-Giralt N et al.: Two new single-nucleotide polymorphisms promoter in patients with postmenopausal osteoporosis. J Clin
in the COL1A1 upstream regulatory region and their relationship Endocrinol Metab 87:4080–4084, 2002.
to bone mineral density. J Bone Miner Res 17:384–393, 2002. 117. Langdahl BL et al.: Polymorphisms in the osteoprotegerin gene are
95. Efstathiadou Z et al.: Association of collagen Iα 1 Sp1 polymor- associated with osteoporotic fractures. J Bone Miner Res 17:1245–
phism with the risk of prevalent fractures: a meta-analysis. J Bone 1255, 2002.
Miner Res 16:1586–1592, 2001. 118. Yamada Y et al.: Association of polymorphisms of the osteoprote-
96. Mann V et al.: Meta-analysis of COL1A1 Sp1 polymorphism in gerin gene with bone mineral density in Japanese women but not
relation to bone mineral density and osteoporotic fracture. Bone men. Mol Genet Metab 80:344–349, 2003.
32:711–717, 2003. 119. Ohmori H et al.: Linkage and association analysis of the osteopro-
97. Ralston SH et al.: Large-scale evidence for the effect of the tegerin gene locus with human osteoporosis. J Hum Genet 47:400–
COLIA1 Sp1 polymorphism on osteoporosis outcomes: the 406, 2002.
GENOMOS Study. PLoS Med 3:e90, 2006. 120. Brändström H et al.: Single nucleotide polymorphisms in the human
98. Xu XH et al.: Molecular genetic studies of gene identification for gene for osteoprotegerin are not related to bone mineral density or
osteoporosis: the 2009 update. Endocr Rev 31:447–505, 2010. fracture in elderly women. Calcif Tissue Int 74:18–24, 2004.
99. Simonet WS et al.: Osteoprotegerin: a novel secreted protein 121. Richards JB et al.: Bone mineral density, osteoporosis, and
involved in the regulation of bone density. Cell 89:309–319, 1997. osteoporotic fractures: a genome-wide association study. Lancet
100. Tan KB et al.: Characterization of a novel TNF-like ligand and 371:1505–1512, 2008.
recently described TNF ligand and TNF receptor superfamily 122. Styrkarsdottir U et al.: Multiple genetic loci for bone mineral den-
genes and their constitutive and inducible expression in hemato- sity and fractures. N Engl J Med 358:2355–2365, 2008.
poietic and non-hematopoietic cells. Gene 204:35–46, 1997. 123. Hsu YH et al.: An integration of genome-wide association study
101. Tsuda E et al.: Isolation of a novel cytokine from human fibroblasts and gene expression profiling to prioritize the discovery of novel
that specifically inhibits osteoclastogenesis. Biochem Biophys Res susceptibility loci for osteoporosis-related traits. PLoS Genet
Commun 234:137–142, 1997. 6:e1000977, 2010.
102. Bucay N et al.: Osteoprotegerin-deficient mice develop early onset 124. Rivadeneira F et al.: Twenty bone-mineral-density loci identified
osteoporosis and arterial calcification. Genes Dev 12:1260–1268, by large-scale meta-analysis of genome-wide association studies.
1998. Nat Genet 41:1199–1206, 2009.
103. Yasuda H et al.: Identity of osteoclastogenesis inhibitory fac- 125. Kim JG et al.: Association between osteoprotegerin (OPG), recep-
tor (OCIF) and osteoprotegerin (OPG): a mechanism by which tor activator of nuclear factor-κB (RANK), and RANK ligand
OPG/OCIF inhibits osteoclastogenesis in vitro. Endocrinology (RANKL) gene polymorphisms and circulating OPG, soluble
139:1329–1337, 1998. RANKL levels, and bone mineral density in Korean postmeno-
104. Bekker PJ et al.: The effect of a single dose of osteoprotegerin in pausal women. Menopause 14:913–918, 2007.
postmenopausal women. J Bone Miner Res 16:348–360, 2001. 126. Hsu YH et al.: Variations in genes involved in the RANKL/
105. Body JJ et al.: A phase I study of AMGN-0007, a recombinant RANK/OPG bone remodeling pathway are associated with bone
osteoprotegerin construct, in patients with multiple myeloma or mineral density at different skeletal sites in men. Hum Genet
breast carcinoma related bone metastases. Cancer 97:887–892, 118:568–577, 2006.
2003. 127. Paternoster L et al.: Genome-wide association meta-analysis of
106. Anderson DM et al.: A homologue of the TNF receptor and its cortical bone mineral density unravels allelic heterogeneity at the
ligand enhance T-cell growth and dendritic-cell function. Nature RANKL locus and potential pleiotropic effects on bone. PLoS
390:175–179, 1997. Genet 6:e1001217, 2010.
107. Hsu H et al.: Tumor necrosis factor receptor family member 128. Choi JY et al.: Genetic polymorphisms of OPG, RANK, and ESR1
RANK mediates osteoclast differentiation and activation induced and bone mineral density in Korean postmenopausal women.
by osteoprotegerin ligand. Proc Natl Acad Sci U S A 96:3540– Calcif Tissue Int 77:152–159, 2005.
3545, 1999. 129. Koh JM et al.: Identification of novel RANK polymorphisms and
108. Lacey DL et al.: Osteoprotegerin ligand is a cytokine that regulates their putative association with low BMD among postmenopausal
osteoclast differentiation and activation. Cell 93:165–176, 1998. women. Osteoporos Int 18:323–331, 2007.
3 6 8 • G e no m ic s in C l inic a l Pr actic e
24.
GENETICS AND GENOMICS OF CHRONIC KIDNEY
DISEASE
Albert C. M. Ong and Alexander P. Maxwell
369
(A)
GN (biopsy proven)
PKD 10% 17%
Diabetes Pyelonephritis
15% 12%
(B) Aetiology
unknown 4%
0% Other
13% GN 15%
0%
Hypertension
25%
Diabetes 38%
PKD 5%
Figure 24.1
Primary diagnosis in prevalent patients on renal replacement therapy (RRT) in England and Wales (E&W) in 2010 and in the United
States in 2009. Diabetes was the second most common cause in the U.K. (A) and the most common cause of ESRD in prevalent RRT patients in the
U.S. (B). In the U.K., the major change in the over-65 age group was an increase in renovascular disease (from 1.1% to 7.8%). The male:female ratio
was greater than unity for all groups except for PKD. Source: UK Renal Registry data, 2010, US Renal Data System 2011.
mutations reported to date (ADPKD mutation database, 2003).As yet, no clear phenotype–genotype correlations
http://pkdb.mayo.edu). Unlike PKD2, PKD1 is highly poly- have been reported for PKD2 (Magistroni et al., 2003).
morphic, with on average 10–13 changes described per indi-
vidual (Garcia-Gonzalez et al., 2007; Rossetti et al., 2007).
Gene Syndromes
The majority of mutations are private (i.e., unique to a single
family) and indicate that a significant level of new mutation Patients with a contiguous gene deletion of both PKD1
is occurring (Rossetti et al., 2001). For PKD1, mutations 5′ and the neighboring tuberous sclerosis gene (TSC2) allele
to the median are associated with more severe disease (aver- have more severe early-onset PKD than those with TSC2
age age at onset of ESRD: 5′, 53 yrs; 3′, 56 yrs) and a sig- mutations alone (Brook-Carter et al., 1994). Most probably
nificantly greater risk of developing intracranial aneurysms this indicates a synergistic role between polycystin-1 and
(Rossetti et al., 2002a; Rossetti et al., 2003).This association tuberin (the TSC2 protein) in cyst development: tuberin
is not related to mutation type and may be due to the pro- may play a role in trafficking polycystin-1 to the lateral cell
posed cleavage of polycystin-1 via a G-protein coupled recep- membrane in kidney epithelial cells (Kleymenova et al.,
tor proteolytic site (GPS) into two different proteins (Figure 2001; Ong et al., 1999).
24.2) with mutations to each half having potentially different A family with bilineal inheritance of germline PKD1
phenotypical consequences (Qian et al., 2002; Rossetti et al., and PKD2 mutations has been described and recently
Key
Transmembrane region
GPS
NH2 COOH
COOH
Polycystin-1 Polycystin-2
Figure 24.2
Predicted structure and topology of the ADPKD proteins, polycystin-1 and polycystin-2. Several functional protein, carbohydrate, and
lipid-binding functional domains of both proteins are depicted (Ong and Harris, 2005). Both proteins interact via a coiled coil domain in their
C-termini. A G-protein coupled receptor proteolytic site (GPS) may allow polycystin-1 to be cloven into two halves that remain tethered by
non-covalent bonds.
NKCC1 TRAM-34
Sorafenib KCa3.1
EGFR SR
CFTR inhibitors Cl– + Metformin
RAS B-Raf MAPK – + – AMPK
Bosutinib + Somatostatin CFTR
Src cAMP Cl– cAMP SR
Tolvaptan +
cAMP + – + Somatostatin
+ V2R
Fluid secretion –
Cilia – Tolvaptan
Proliferation +
2+ Cilia V2R
Ca ATP ATP
– Roscovitine Ca2+
+
PC1 PC2 Genz-123346 HDAC inhibitors
PC1 PC2 PC2
+ – AMPK Ca2+
mTOR Calcimimetics + ER
GlcCer – + + +
TSC1/TSC2 CaSR
Sirolimus Metformin Triptolide
Everolimus
PC1 PC2 PC1 PC2
Figure 24.5
Mechanism-based therapeutics in ADPKD. The major signaling pathways involved in the pathogenesis of ADPKD and the targets of
candidate drugs (Chang and Ong, 2011). (A) The “secretory” pathways; (B) the “proliferative” pathways. Abbreviations: AMPK, AMP-activated protein kinase; CaSR,
calcium-sensing receptor; CFTR, cystic fibrosis transmembrane regulator; ER, endoplasmic reticulum; GlcCer, glucosylceramide; HDAC, histone deacetylase; KCa3.1, endothelial Ca2+-activated K+-channel; MAPK,
mitogen-activated protein kinase; mTOR, the mammalian target of rapamycin; NKCC1, Na-K-Cl cotransporter; PC, polycystin; SR, somatostatin receptor; TSC, tuberous sclerosis;V2R, vasopressin V2 receptor.
are associated with the loss of the normal barrier function clinical proteinuria (dipstick urinalysis positive), and pro-
of the glomerular basement membrane; consequently, gressive renal failure (Figure 24.6c).
albumin can pass from the glomerular capillaries through Diabetic nephropathy is arguably the most important
“leaky” glomerular basement membranes and disrupted microvascular complication of diabetes in terms of health-
slit diaphragms into the urinary space (Gross et al., 2005a) care costs, morbidity, and premature mortality (Alicic
(Figures 24.6a and 24.6b ). Initially, this abnormal urinary and Tuttle, 2010; Groop et al., 2009; Hex et al., 2012).
albumin excretion is low in quantity and not detected on Individuals with diabetes have an increased mortality rate
dipstick urinalysis. Microalbuminuria, a term that describes due to cardiovascular disease, such as coronary heart disease
this small amount of urinary albumin excretion, is important and stroke (Morgan et al., 2000; Williams and Airey, 2002;
since it is often the first detectable clinical sign of diabetic Ninomiya et al., 2009; Drury et al., 2011). The develop-
kidney disease (Shields and Maxwell, 2010). Persistent glo- ment of diabetic nephropathy further amplifies this risk of
merular injury is associated with renal tubular dysfunction, macrovascular disease which is increased two- to four-fold
(B)
(A) Ang II
Afferent Efferent
arteriole arteriole Angiotensin II induced
P vasoconstriction of the
efferent arteriole
Ang II
P
Tubulointerstitial
Glomerulus fibrosis
Renal tubules
Glomerulus
Glomerulus
Glomerular
sclerosis
Neutrophil
Glomerulus Interstitial
inflammatory Lymphocyte
infiltrate
Macrophage
Figure 24.6c
Progressive glomerular sclerosis and tubulointerstitial fibrosis are associated with loss of renal function. An inflammatory interstitial
infiltrate is seen in association with tubular atrophy.
in early renal disease (microalbuminuria), nine times with system with angiotensin-converting enzyme inhibitors
established nephropathy (proteinuria) and up to 50-fold and/or angiotensin II receptor blockers (Brenner et al.,
in persons with ESRD (Deckert et al., 1996; Dinneen 2001; Lewis et al., 1993; Lewis et al., 2001). These agents
and Gerstein, 1997; Tuomilehto et al., 1998; Fuller et al., have beneficial effects beyond blood pressure lowering,
2001). Diabetic nephropathy is associated with prolonged including lowering glomerular capillary pressure and reduc-
duration of diabetes, poor glycemic control, and raised ing proteinuria (Chiurchiu et al., 2005). Smoking cessa-
blood pressure, and is more common in diabetic individu- tion, aspirin therapy, and lipid-lowering drugs are also part
als of Asian or African ancestry (The Diabetes Control of the multiple-risk-factor management to reduce cardio-
and Complications Trial (DCCT) Research Group, 1993; vascular events in persons with nephropathy (Fioretto
UK Prospective Diabetes Study [UKPDS] Group 1998; and Solini, 2005; Gaede et al., 2003; Gaede et al., 2008).
Nelson et al., 1997). Dietary protein restriction (to 1g/kg body weight) can also
Primary prevention of diabetic nephropathy requires slow progression in proteinuric diabetic patients, although
excellent glycemic control that is often difficult to achieve the clinical utility of this strategy is limited (Pedrini et al.,
in practice (The Diabetes Control and Complications Trial 1996; Robertson et al., 2007).
(DCCT) Research Group,1993; UK Prospective Diabetes Unfortunately, despite these interventions, kidney dis-
Study (UKPDS) Group 1998; Barnett, 2004). Intensive ease still develops in up to 30% of individuals with type 1
control of diabetes reduces the incidence of diabetic diabetes and up to 50% of persons with type 2 diabetes, and
nephropathy in type 1 diabetes (de Boer et al., 2011) and progresses in approximately 30% of diabetics (Krolewski
type 2 diabetes (Stratton et al., 2000). Nevertheless, efforts et al., 1996; Pambianco et al., 2006; Parving et al., 2006).
to intensively manage glycemic control in patients with Additional novel strategies are therefore urgently required
both type 2 diabetes and cardiovascular disease have been to fight the epidemic of diabetic renal disease (Wolf and
reported recently to be associated with an increased overall Ritz, 2003). Of interest, diabetic nephropathy is not an
patient mortality despite improvements in renal outcomes inevitable complication of a prolonged duration of dia-
(Dluhy and McMahon, 2008). Progression of renal disease betes (Bain et al., 2003). Also, whilst the cumulative inci-
can be modified by inhibition of the renin-angiotensin dence of nephropathy is greatest in those persons with the
393
TF expressing cell
TF TF
TF
TF
FVII TF
TF
FIX FX
FXIa TF-FVIIa-FXa
TFPI
AT
FIXa
FVIIIa
FXIa-AT FIXa-AT
PT KR KR SP
FXa
FVa FVai
FVIIIai
Serpin AT
Thrombin
FXa-AT
FVIII FV
AT
cc Thrombin-AT
α PS
cc APC
β FBG FXIIIa FXIII
cc
γ FBG
Fibrinogen Fibrin
TG
TG
PC
TM
bin
Endothelial cell
rom
Th
CR
Thrombin
EP
Platelet
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Serpin
TAFIa
Serpin
PAI-1
Plasmin- α2AP
α2AP
tPA-PAI-1
PLG
Figure 25.1
The hemostasis network. The process of blood coagulation is initiated by the exposure of cells expressing TF to flowing blood.
Thrombin generation is propagated by a series of positive-feedback loops, leading to fibrin deposition. This process is controlled by a series of
negative-feedback steps: The initiation complex is inhibited by the formation of the quaternary complex TF-FVIIa-FXa-TFPI, and the active
proteases FIXa, FXa, and thrombin are inactivated by the serpin AT. In addition, thrombin initiates a negative-feedback pathway by activating PC,
leading to the inhibition of FVa and FVIIIa. The fibrin clot is degraded by plasmin.All abbreviations are given in Table 25.1: Protein names are
colored thus: functionally active proteins, red; inactivated or inhibited proteins, blue; precursors and fibrinogen-derived components, black. Lines
and arrows are colored thus: dashed red lines, positive procoagulant feedback loops; dashed blue lines, negative feedback and inhibitory loops; solid
black lines, interactions or processes.
The importance of providing a negatively charged in Man [OMIM]: 262890), which is characterized by a fail-
phospholipid surface fortheassembly of the procoagulant ure toexposephosphatidylserine on the outer leaflet of the
response is seen in the extremely rare bleedingdisorder plasma membrane and isassociated with a moderate bleed-
known as Scott syndrome (Online Mendelian Inheritance ing tendency.
COMMON NAME ABBREV- SUB-UNIT GENE GENE GENE NO. OF MRNA AMINO MR OF OMIM MAIN ACTION LMD
IATION SYMBOL LOCATION SIZE EXONS SIZE ACIDS MONOMER
(KBP) (BP) (MATURE) (KDA)
Tissue Factor TF F3 1p13 12.6 6 1852 263 44 134390 Cofactor for FVII/FVIIa
Prothrombin FII F2 11p11.1 20.3 14 2028 579 72 176930 Clots FBG, activates PC,
FXI, TAFI
Factor V FV F5 1q23 72.4 25 7009 2196 330 227400 Cofactor for FXa
Factor VII FVII F7 13q34 15.1 8 2880 416 50 227500 Activates FIX & FX
Factor VIII FVIII F8 Xq28 187.1 26 8957 2332 330 306700 Cofactor for FIXa http://hadb.org.uk/
Factor IX FIX F9 Xq27 32.8 8 2831 415 56 306900 Activates FX http://factorix.org/
Factor X FX F10 13q34 26.8 8 1884 445 59 227600 Activates prothrombin
Factor XI FXI F11 4q35 25.9 15 2979 607 80* 264900 Activates FIX http://www.factorxi.com/
Factor XIII** FXIII A F13A1 6p25 177.8 15 3834 731 75** 134570 Crosslinks fibrin http://www.f13-database.de/
(A chain) (xhgmobrswxgori45zk5jre45)/
content.aspx?menu=1,6
Factor XIII** FXIII B F13B 1q31 28 12 2190 641 80** 134580 Stabilizes FXIII A chain http://www.f13-database.de/
(B chain) (xhgmobrswxgori45zk5jre45)/
content.aspx?menu=1,6
Fibrinogen FGN Α FGA 4q32 7.8 6 3828 866 68*** 134820 Mechanical stabilization http://www.geht.org/
(α chain)*** of clot databaseang/fibrinogen/
Fibrinogen FGN Β FGB 4q32 9.8 8 3693 491 52*** 134830 Mechanical stabilization http://www.geht.org/
(β chain)*** of clot databaseang/fibrinogen/
Fibrinogen FGN Γ FGG 4q32 23.6 10 1753 453 49*** 134850 Mechanical stabilization http://www.geht.org/
(γ chain)*** of clot databaseang/fibrinogen/
von Willebrand factor VWF VWF 12p13 176 52 9028 2050 255 193400 Cell adhesion & FVIII http://www.vwf.group.shef.
carrier ac.uk/
Protein C PC PROC 2q14.2 10.8 9 1759 419 62 176860 Inactivation of FVa and
FVIIIa
Protein S PS PROS1 3q11.2 101.9 15 3477 676 69 176880 Inactivation of FVa and http://www.isth.
FVIIIa org/?ProteinSDeficiency
Antithrombin AT SERPINC1 1q23 21 9 1684 464 58 107300 Inhibits thrombin, FIX,
FX, FXI
Plasminogen PLG PLG 6q27 51.1 14 2905 791 92 173350 Dissolution of clot in
wound repair
Tissue plasminogen TPA PLAT 8p11.1 32.7 14 2933 562 69 173370 Plasma activator of
activator plasminogen
Plasminogen activator PAI-1 SERPINE1 7q22 12.3 9 3320 379 52 173360 Inhibition of TPA and
inhibitor 1 UPA
Alpha2-antiplasmin α2-AP SERPINF2 17p13 13.3 9 2210 452 67 262850 Inhibition of plasmin
Thrombin-activatable TAFI CPB2 13q14 52.4 11 1984 401 60 603101 Inhibition of fibrinolysis
fib. inhibitor
Deletions
Insertions Promoter rather than X-linked inheritance. VWD is an autosomal
1% 6%
10% Splice site disease with variableexpressivity and reduced penetrance. It
9% has been classified into six distinct types:types 1 and 3 result
from quantitative deficiency, whereas the four type 2 vari-
Nonsense
4% ants (2A, 2B, 2M, and 2N) are characterized by qualitative
abnormalities of VWF. Types 1,2A, 2B, and 2M VWD are
inherited as autosomal dominant, whereas types 2N and 3
areinherited in autosomal recessive manner. Four hundred
and three unique variants have been described that cause
VWD (https://grenada.lumc.nl/LOVD2/VWF/home.
php?select_db=VWF; Oct 2012).
FVII deficiency is an autosomal recessive disorder
with a highly variablephenotype. In contrast to the other
bleeding disorders, the residual clotting-factoractivity does
not accurately predict the clinical phenotype. Complete
absence ofFVII activity is usually incompatible with life,
Missense and individuals die shortly afterbirth, due to severe hemor-
70%
rhage. The majority of individuals with mutations intheir
Figure 25.2
Spectrum of mutation types causing FVII deficiency. Data F7gene(s) are asymptomatic and only come to light during
obtained from the FVII mutation database (McVey et al., 2001).
preoperative coagulation-factor screening. A severe bleed-
ing phenotype is only observed inindividuals with residual
represented inthe FVII mutation database (McVey et al., FVII activity below 2%; however, a considerableproportion
2001) is typical of thosefound in other coagulation factor of individuals with a mild to moderate bleedingphenotype
gene disorders (Figure 25.2). Recurrent mutations inunre- have similarFVII activity levels. The severity of the bleed-
lated individuals arise at functionally important residues ing in individuals with no residualFVII activity is consistent
and in CpGdinucleotides, which are known hotspots for with the key role of the TF-FVIIa complex in initiating-
mutation due to deamination ofmethylated cytosine resi- blood coagulation in vivo. Surprisingly, no congenital
dues. Arginine residues occupy critically important func- abnormalities in the geneencoding TF have been described
tional sites in the coagulation factors, and four of the six to date. Mutations in the TF gene might havebeenpredicted
codons encoding arginine contain the dinucleotide CpG. to lead to either a bleeding (loss of function) or prothrom-
A mutation hotspot, specific to hemophilia A, is therecur- botic (gainof function) phenotype. Targeted disruption
rent rearrangement responsible for 50% of all severe cases of the mouse TF gene (f3) results inembryonic lethality
in whichhomologous recombination of a sequence within of f3–/– embryos at embryonic days 9.5–10.5 (Bugge et al.,
intron 22 of the F8gene and twoextragenic and telomeric 1996;Carmeliet et al.,1996;Toomey et al., 1996). Hence,
copies of the sequence results in an inversion that disrupt- loss in early pregnancy mostprobably accounts for the lack
sthe F8gene. of TF-null individuals in clinical practice.
Deficiencies of FVIII (hemophilia A) and FIX (hemo-
philia B) share an indistinguishable clinical phenotype char-
THE BLEEDING DISORDER S
acterized by bleeding into muscles, joints,and other organs,
The most common inherited bleeding disorder is von with consequent damage. They are both X-linked recessive-
Willebrand’s disease (VWD). Von Willebrand factor disorders affecting one in 5000 and one in 30,000 males,
(VWF), which is the protein deficient or defective inpa- respectively. Clinically,hemophilia is noted to be severe,
tients with VWD, has two distinct roles in hemostasis. First, moderate, or mild, and this correlates with theresidual level
it is responsiblefor enabling platelets to adhere to collagen of activity of the affected factor, which in turn relates to the
exposed at wound sites under conditionsof shear. Second, precisegene defect. Thus, severely affected patients bleed
VWF is the plasma carrier for FVIII, thus deficiency of spontaneously and have lessthan 2% of the normal level
VWFleads to deficiency of FVIII. Mutations in the FVIII of factor in their blood; moderately affected patientsbleed
binding site within VWF may result in structurally and func- after minor trauma and have 2–5% of the normal level; and
tionally normal VWF, except that it does not bind FVIII and mildly affectedindividuals have more than 5% of the nor-
presents as mild hemophilia A, but with autosomal recessive mal level and bleed only after severe trauma.
int22h-2 int22h-3
int22h-3 int22h-2
Xq28 Xq28
3' 3'
FVIII FVIII
Gene int22h-3 Gene int22h-2
int22h-2 int22h-3
5' 5'
Xq28 Xq28
FVIII FVIII
Gene Gene
int22h-3 Ex23-26 int22h-2
Ex23-26
int22h-2 int22h-3
5' 5'
FVIII FVIII
Gene Gene
Ex1-22 Ex1-22
int22h-1 int22h-1
Figure 25.3
Inversions in Xq28 causing hemophilia A. Proposed mechanism causing polymorphic inversion and recombination between int22h
repeats leading to inversion of the F8gene. The F8 gene is indicated by the large arrow. Inversions disrupting the F8gene result from recombination
between int22h-1 region (in intron 22 of the F8gene) and either int22h-2 or int22h-3, which lie 400 kb distal to F8. The three copies of the int22h
sequence are indicated by the red, blue, and green colored boxes, and their orientation relative to the intron 22 copy is indicated by the arrowhead.
The distal copies (int22h-2 and int22h-3) are in opposite orientation to each other and are flanked by a large, imperfect palindrome, indicated by
blue boxes. Proposed recombination between the arms of the palindrome would generate alleles where either the in22h-2 or the int22h-3 is most
telomeric and in the opposite orientation to int22h-1.
G20210A). The population frequency of FV Leiden is 5%in of thrombosis in either the arterial or the venous vascular
ethnic Europeans. FVa is inactivated by activated PC through system is complex and multifactorial, involving both envi-
proteolyticcleavage at R506–S507 and R1765–L1766. The ronmental and genetic factors. The GWAS approach is
cleavage at R506 is rate limiting,and the mutation FV R506Q particularly suited to identifying the genetic component of
is resistant to cleavage by activated PC. The mutationin the complex diseases—however,in order to provide adequate
prothrombin gene occurs in the 3′ untranslated region of the statistical power to identify common variants,these studies
gene and isassociated with raised prothrombin levels. Its fre- require the genotyping of unprecedented numbers of indi-
quency in Europeans is approximately 2%. viduals. Atherothrombosis represents an advanced stage
Elevated levels of FVIII are a strong independent risk of atherosclerosis, in which plaque rupture or erosion has
factor for the development of venousthromboembolism; triggered coagulation leading to the formation of an arte-
however, the cause is unknown (reviewed in Seligsohn and rial thrombosis. It is perhaps not surprising,therefore, that
Lubetsky, 2001;Feinbloom and Bauer, 2005).In contrast, a the susceptibility loci identified for coronary artery disease
failure to appropriately regulate fibrinolysis and thus clot- and myocardial infarction are related to the initiation and
disintegration may also be associated with a thrombotic progression of the disease rather than to the thrombotic
tendency; however, theseconditions (deficiency of tPA and end-point (Lotta, 2010).Despite the success of GWAS in
plasminogen; increased PAI-1 activity) are veryrare. identifying over 34 different genetic loci at which common
variants have been associated with coronary artery disease,
it is likely that these loci only explain a modest fraction of
C O M PA R AT I VE S E Q U E N C E A N A LYS I S the total heritability, and it is also likely that some common
coagulation-factor variants may contribute to inherited risk
The impact of the completed sequence of the human of arterial disease (Peden and Farrall, 2011).
genome on our understandingof hemostasis has been much In contrast, GWAS studies on venous thrombosis have
less than for other inherited disorders, since most ofthe rel- to date been extremely limited and involved the analy-
evant genes were cloned before the genomic era. However, sis of much smaller cohorts(Morange and Trgouet, 2011;
the identification andcharacterization of the genes respon- Germain et al, 2011). Prior to GWAS studies, as discussed
sible for combined FV and FVIII deficiencyand of Type II above, a number of variants associated with venous throm-
multiple coagulation factor deficiency were aided by this bosis had been identified in the genes encoding SERPINC1,
genomicinformation. The completion of the sequence of PROC, PROS1, F2, FGG, F5, and ABO.The first three
other vertebrate genomes has,however, provided a fascinat- genes harbor multiple private mutations rather than com-
ing insight into the evolution of the coagulationnetwork. mon variants. In contrast, the latter four genes harbor more
In addition, thecomparative sequence information has been common variants,which result in variations in qualitative
invaluable ininterpreting naturally occurring mutations or quantitative differences in coagulation factors. In the
in blood coagulation deficiencies(Gomez et al., 2004). GWAS study, only the common variants in the F5 gene and
Comparison of the amino acid sequence alignments has ABO genes reached statistical significance, although a num-
alsoprovided important insights into structure–function ber of other loci were identified (HIVEP1, C4BPA, TC2N,
relationships of the coagulationfactors. This information GP6 and F11). The novel gene loci, however, confer a mod-
may aid the design of novel recombinant coagulationfac- est increase in the risk of venous thrombosis (odds ratios
tors for replacement therapy, in particular FVIII molecules of 1.10–1.35). GWAS studies have also been carried out
with modified domains (Ward et al., 2011).Comparative on quantitative traits believed to be associated with venous
sequence analysis of non-coding sequences may also fur- thrombosis: activated partial thromboplastin time, protein
therour understanding of regulatory gene sequences. S and VWF plasma levels have contributed to the identifica-
tion of novel susceptibility loci for venous thrombosis.
G E N O M E -W I D E A SS O C I AT I O N
STUDIES P H A R M AC O G E N O M I C S A N D
PER SONALIZED MEDICINE?
Genome-wide association studies (GWAS) have made an
enormous impact on the identification of genetic loci asso- Expectations are high that increasing knowledge of the
ciated with human diseases and their quantitative risk fac- genetic basis of disease will eventually lead to person-
tors (http://www.genome.gov/gwastudies/). The etiology alized medicine; that is, preventative and therapeutic
404
1 KILOBASE
ζ ψζ ψα2 ψα1 α2 α1 θ1 ε Gγ Aγ ψβ δ β
CHROMOSOME 16 CHROMOSOME 11
Figure 26.1 Genetic control of hemoglobin production (from Weatherall and Clegg, 2001, ref. 6).
functions that lie outside the genes themselves. At the 5´ different genes in the β globin gene cluster during different
non-coding (flanking) regions of the globin genes there stages of maturation.
are blocks of nucleotide homology. The first, the ATA The mechanisms that control the switches from embry-
box, is about 30 bases upstream of the initiation codon. onic to fetal, and fetal to adult, hemoglobin production are
The second, the CCAAT box, is about 70 base pairs still not understood. A number of loci have been identified
upstream from the 5´ end of the genes. About 80–100 that are involved in the regulation of persistent fetal hemo-
bases further upstream, there is the sequence GGGGTG globin production in adult life. These are the subject of sev-
or CACCC, which may be inverted or duplicated. These eral recent reviews4,5 and are considered further in a later
three highly conserved DNA sequences, or promoter ele- section. A more extensive review of the genetic control of
ments, are involved in the initiation of the transcription hemoglobin production in general has also been published
of the individual genes. Finally, in the 3´ non-coding recently.3
region of all the globin genes, there is the sequence
AATAAA, which is the signal for cleavage and polyA
addition to RNA transcripts. T H E C L A S S I F I C AT I O N O F T H E
The globin gene clusters also contain several sequences HEMOGLOBIN DISORDER S
that constitute regulatory elements, which interact to pro-
mote erythroid-specific gene expression and coordination The hemoglobin disorders are broadly divided into the thal-
of the changes in globin gene activity during develop- assemias, diseases characterized by a reduced rate of synthe-
ment. These include the globin genes themselves and their sis of one or another of the globin chains, and the structural
promoter elements, enhancers, and “master” regulatory hemoglobin variants, which alter the morphology, stability,
sequences, called, in the case of the β globin gene cluster, the or function of the hemoglobin molecule.6 These conditions
“locus control region” (LCR), and, in the case of the α genes, are summarized in Table 26.1.
“HS40” (a nuclease-hypersensitive site 40 kb upstream from The thalassemias are divided into the α and β thalas-
the α globin genes). Each of these sequences has a modular semias, in which there is defective α or β chain synthesis,
structure made up of an array of short motifs that represent respectively. The β thalassemias are divided into the β+ and β0
the binding sites for transcriptional activators or repressors. thalassemias in which there is either a reduced or a complete
The regulatory regions contain sequence motifs for vari- absence of β chain synthesis, respectively. Other important
ous ubiquitous and erythroid-restricted transcription fac- forms of β thalassemia include the compound heterozygous
tors. There are binding sites for these factors in each of the states for β thalassemia and structural hemoglobin variants
globin gene promoters and the hypersensitive site regions such as hemoglobins S, C, or E. By far the commonest of
in the various regulatory elements. While the ubiquitous these is hemoglobin E (HbE) β thalassemia. Hemoglobin
factors are found in a variety of cell types, there are many E, which occurs at a high frequency throughout many parts
transcription factors that appear to be restricted to red-cell of Asia, results from a mutation that activates a new splice
precursors, including GATA-1, EKLF, and NF-E2, and oth- site; hence, it is produced at a reduced rate and behaves like
ers. Binding of the specific factors activates the LCR, and it a mild form of β thalassemia. HbE β thalassemia is by far
seems likely that this complex becomes approximated to the the commonest form of the disease in many Asian countries
412
morphological grounds such as degree of lineage, commit- Balanced translocations result in the expression of
ment, and differentiation.11 In this classification, the AML fusion proteins, which frequently contain a truncated tran-
FAB subtypes (M0–M7) were determined by cell morphol- scription factor joined with another, unrelated protein.
ogy with particular subtypes such as M3 defining the acute The balanced translocations result in a deregulation of
promyelocytic leukemia (APL). In 2002, the World Health transcription factors and can be seen as a main pathogenic
Organization (WHO) introduced a new classification of event.18 The simultaneous occurrence of more than one of
hematopoietic and lymphoid neoplasm by combining mor- these recurrent, balanced translocations is rare. The detec-
phology, genetic, immunophenotype, with biologic and tion of subtype-specific abnormalities such as the t(15;17)
clinical information.12 or t(8;21) translocations has resulted in the characteriza-
The 2008 WHO released an updated Classification of tion of the involved genes, the identification of abnormal
Tumors of Hematopoietic and Lymphoid Tissues,13 which fusion proteins, and provided new diagnostic targets, but
included new criteria as well as clarification and refinement has also identified distinct AML subtypes.
of the defining criteria for the all the hematopoietic malig- Based only on cytogenetic abnormalities, patients were
nancies, including AML. The 2008 revised WHO classifi- classified into three distinct prognostic subgroups or risk
cation placed the most common and well-defined recurring groups: those with favorable or good risk, intermediate or
cytogenetic abnormalities into a separate group: t(8;21) standard risk, and unfavorable, adverse, or poor risk disease.
(q22;q22); inv(16)(p13q22) or t(16;16)(p13;q22); The favorable risk group includes approximately 15% of
t(15;17)(q22;q12); and 11q23 (MLL) abnormalities. Some patients, with an age below 60 years, and is defined by the
of these chromosomal abnormalities correlate with specific presence of either t(15;17), t(8;21) and inv(16), or t(16;16)
FAB subtypes, for instance, the translocation t(8;21) is cytogenetic translocations. Patients presenting with these
present in approximately 40% of the FAB-M2 leukemia.14 mutations usually have an increased rate of complete remis-
Translocations involving the retinoic receptor α (RARα) on sion and are at relatively low risk of relapse.19 The poor prog-
chromosome 17 (t(15;17)) is present in 98% of cases15 of nosis group of AML patients have cytogenetic abnormalities
APL (APML) FAB-M3 and inv(16) or t(16;16) are involved such as deletion of chromosomes 5 or 7 (or the short arms of
in 80% of the FAB-M4-Eo cases.16,17 the chromosomes), abnormalities in the long-arm of chro-
In addition, in the 2008 WHO classification, new cytoge- mosome 3, and AML with complex karyotype (three or
netic entities were added: t(6;9)(p23;q34); inv(3)(q21q26.2) more cytogenetic aberrations).20,21 Cytogenetic mutations
or t(3;3)(q21;q26.2); t(1;22)(p13;q13). Furthermore, for the involving the 11q23 region are also considered a marker
first time, two provisional molecular mutation entities are for poor prognosis. Overall, patients with adverse risk-type
added: AML with mutated NPM1; and AML with mutated mutations generally have a probability of five-year survival
CEBPA. It was recommended that FLT3 mutations also be of approximately 20%. Patients presenting with AML with
examined in cases of cytogenetically normal AML. other types of cytogenetic abnormalities or without any
apparent cytogenetic abnormalities (i.e., an apparently
normal karyotype) are considered to have an intermediate
C Y TO G E N ET I C S : C L A S S I F I C AT I O N S risk.22,21,23
A N D R E C U R R E N T T R A N S L O C AT I O N S In 2010, a refinement of the cytogenetic classification
for AML was published24 (Figure 27.1). In this refinement,
Cytogenetics including fluorescence in situ hybridization almost 6000 patients, 18–59 years old at diagnosis, were
(FISH) provided an insight into the heterogeneity of the classified into 54 cytogenetic subgroups. The favorable cyto-
disease by identifying two major groups of AML. In one genetic group maintained the t(15;17), inv(16) and t(8;21)
group, chromosomal aberrations occurred in approxi- balanced reciprocal translocations, whilst in the adverse group
mately 50% of de novo AML. Within this group, at least the list of abnormalities included t(3;5)(q25;q34), inv(3)
two further subtypes could be further distinguished: one (q21q26)/t(3;3)(q21;q26), add(5q)/del(5q), -5, 7, add(7q)/
that had balanced aberrations mainly consisting of t(8;21), del(7q), t(6;11)(q27;q23), t(10;11)(p11_13;q23), other
t(15;17), and inv(16); and a second subgroup with unbal- t(11q23) [excluding t(9;11)(p21_22;q23) and t(11;19)
anced aberrations such as 5q-, 7q-, -5, -7, or complex cyto- (q23;p13)], t(9;22)(q34;q11), _17, and abn(17p). Note that
genetic abnormalities. The second major group consisted this list now moved several of the 11q23 (MLL) translocations
of approximately 40–50% of AML cases in whom the from the previous intermediate group into the adverse group.
karyotype appears to be normal (cytogenetically normal, In addition, any patients without the favorable or adverse
or CN). cytogenetic abnormalities, but with four or more unrelated
$: t(15;17) excluded
Figure 27.1
Comparison of different risk stratification classifications, showing the changes from solely cytogenetic based classification to the molecular
risk stratification model.
abnormalities, were designated as a “complex” group. The 10% of all AML cases.27 A striking feature of APL is that is
refinement of the cytogenetic risk categories resulted in the very uncommon in children younger than 10 years of age;
reassignment of 299 patients (15% of the patients with abnor- then its incidence increases steadily during the teen years to
mal, non-favorable cytogenetic abnormalities) from interme- reach a plateau during early adulthood, and remains con-
diate to adverse (275 patients) or adverse to intermediate (24 stant and then decreases after age 60 years.28 There is a sug-
patients).24 gestion that APL may arise as a complication of previous
The European Leukemia Network (ELN) also pro- exposure to chemotherapy (particularly drugs targeting
posed a standardized reporting system for genetic abnor- topoisomerase II) or radiotherapy, particularly in patients
malities that would allow for a better comparison of studies with a history of breast cancer.29,30
by including data from cytogenetic analysis and NPM1, APL is characterized by translocations involving the
CEBPA, and FLT3 gene mutation analyses.25 RARα gene located at 17q12. In the majority of the cases,
A fully molecular-derived risk stratification model the partner chromosome in the translocation is chromo-
has also been developed26 (Figure 27.1) in which conven- some 15, resulting in the fusion of large parts of RARα to
tional cytogenetic analysis was substituted with molec- most of PML coding sequence, generating the PML-RARα
ular analysis of the fusion transcripts (PML-RARA, fusion protein.31,32 Rarer cases of APL show translocations
RUNX1-RUNX1T1, CBFB-MYH11) and integrated with involving t(11;17)(q23;q12), t(5;17)(q35q12), t(17;17)
mutation analysis for CEBPA, NPM1, RUNX1, ASXL1, (q11;q12), or t(11;17)(q13;q12). These result in fusion
FLT3-ITD and MLL-PTD. This molecularly based stratifi- proteins that also contain RARα but with different partner
cation led to five categories’ being formulated. genes: PLZF (promyelocytic leukemia zinc finger), NPM
(nucleophosmin), Stat5b or NuMA.33–39
RARα (retinoic acid receptor alpha) is a nuclear hor-
T(15;17)—PM L-RA Rα
mone receptor that acts as a hormone-dependent tran-
Acute promyelocytic leukemia (APL: FAB subtype M3) is scriptional switch and in the absence of the ligand retinoic
a relatively rare malignancy and accounts for approximately acid transcription is repressed through the recruitment of
Sin
NCoR
3
HAT
H
RXR
D
AC
RARA
DNA DNA
Physiological levels of RA
(10–9 M)
Nucleosomes with
deacetylated histone tails
Nucleosomes with
acetylated histone tails
PML-RARA
DNA DNA
Pharmacological levels of RA
(10–6 M)
Figure 27.2
Schematic of the different molecular responses between RARA and PML-RARA in response to physiological and pharmacological levels
of ATRA.
the “CBF (core binding factor) complex,” and these trans- and recruits a range of co-repressors to facilitate transcrip-
locations account for 12% of AML cases, particularly in tional repression.
patients less than 60 years of age at diagnosis.24 The CBF The C-terminus of the transcriptional activator AML1
consists of two subunits: RUNX1 (also referred to as is replaced by the transcriptional repressor ETO, resulting in
AML1, CBPα2, and PEBP2αB) and is located on chro- the fusion protein AML1-ETO (CBFA2-CBFA2T1).61 The
mosome 21q22 and CBFβ, which is encoded by a gene on ETO protein binds to co-repressors (e.g., NcoR, SMRT, and
16q22. The two subunits form a heterodimer transcrip- mSin3) and histone deacetylases (HDACs).62,63 The recruit-
tion activator, although only the AML1 subunit contains ment of HDACs to the complex changes the histone acety-
DNA-binding ability. lation status and as a result alters the DNA conformation,
The most frequent translocation is the t(8;21) transloca- making it less accessible for the basal transcription machin-
tion, which has been reported to be present in 7% of younger ery and results in a repression of AML1 target genes. Target
(<65 yrs. old) adult AML patients with this disease24 and genes include granulocyte–macrophage colony-stimulating
results in the formation of the fusion protein AML1-ETO factor (GM-CSF),64 for interleukin-3 (IL-3) and NP3.65
or RUNX1-RUNXT1. The Runt-related transcription fac- AML1-ETO influences various other transcription fac-
tor 1 (RUNX1) (also known as AML1) or core-binding fac- tors, including MEF, c/EBPα, AP-1, and Pu.1.66–68 c/EBPα
tor subunit alpha-2 (CBFA2) regulates the differentiation is a major regulator of early granulocytic differentiation, and
of hematopoietic stem cells into mature blood cells. The AML-ETO physically interacts with c/EBPα and downregu-
RUNX1T1, also known as AML1T1, CBFA2T1, ETO, or lates it at the transcriptional level,67 which has an important
MTG8 is a member of the myeloid translocation gene fam- influence on granulocytic differentiation. An additional
ily, which interacts with DNA-bound transcription factors function of AML1-ETO in leukemia transformation is
Mutation landscape map for six genes (NPM1, FLT3 ITD, DNMT3A, CEPBA, MLL-PTD, and RUNX1) across ~350 AML samples.
Figure 27.4
(Data combined from89,97)
Tissues.13 However, the number of mutated genes in AML an NPM1 mutation very rarely, if ever, have a mutation in
has dramatically expanded in recent years as consequence of TP53. Other combinations frequently occur; for example,
the accessibility of next-generation sequencing. NPM1 mutations with FLT3-ITD or NPM1 mutation with
The first two cancer patients to have their genome a DNMT3A gene mutation. A study of 200 AML patients
fully sequenced were AML patients, and from the analysis, analyzed by next-generation sequencing showed that eight
IDH1/2 and DNMT3A were identified as novel frequently patients had mutations in all three of the genes frequently
mutated genes.87,88 Mutated genes have subsequently been mutations: NPM1, FLT3, and CEPBA. Furthermore, one
identified in tumor-suppressor genes, genes associated patient did not have any mutations in 33 genes or groups of
with DNA methylation or chromatin modifications, genes genes with similar functions (e.g., spliceosomes or Ser-Thr
involved in signaling and transcription factors.89–96 Some of kinases).97 Two ways of illustrating the interactions are
the genes are mutually exclusive; for example, patients with shown in Figures 27.4 and 27.5.
Figure 27.5 Circos plot visualizing the data from Figure 27.4 showing how the different mutations co-occur with each other. (Combined data taken
from89,97)
Cell membrance
Kinase domain
Figure 27.6 Molecular structure of the FLT3 protein showing the location of the possible sites of mutation: ITD and activating point mutations.
N PM1
The NPM1 gene encodes for a nuclear-cytoplasmic protein t(5;17) t(3;5) Insertions
APL MDS/AML AML
that is involved in translocations and mutations in vari-
ous hematological malignancies. The discovery of NPM1
mutations arose from the observation of ectopic expres-
sion of nucleophosmin in the cytoplasm of leukemia cells.
This immune-histochemical finding led to the discovery
of several gene mutations mainly located in exon-12 that
usually result in changes at the C-terminus of the native
Figure 27.7
Molecular structure of NPM1 highlighting the location of the
NPM1 protein; i.e., the disruption of the nucleolar localiza- t(3;5) and t(5;17) translocation and the different common types of 4
tion signal and the generation of a new additional nuclear base pair insertions seen in AML.
DNA
Transcription activation binding Dimerisation
domains domain region
TAD1 TAD2 DBD bZIP
Wild type
ATG 1 ATG 1
42 kD 30 kD
Mutant
IN FRAME
C Terminal missense
mutation
BCOR Mutations
M LL MU TAT I O NS
Mutations in the BCOR (BCL6 co-repressor) gene were
In addition to the MLL gene being involved in 11q24 chro- discovered by the whole-exome sequencing of a single
mosomal translocations (see above and Figure 27.3), partial CN-AML patient.89 As a result of follow-up studies, BCOR
tandem duplications (PTD) have been described in AML. mutations were identified in ~4% of all CN-AML. BCOR
PTDs result in the duplication of a portion of the gene mutations may act by interfering with epigenetic mechanism
and have been associated with trisomy 11145,146 and around and are frequently associated with DNMT3A mutations.
MO L EC U L A R EVO LU T I O N I N A M L
RAS
The advent of next-generation sequencing (NGS) has been
Activating RAS mutations have been described in many able to give valuable insights into the clonal mutational
reports. These are usually at codon 12 and 13 of N-RAS,152 spectrum and evolution in AML. Several of the mutations
also codon 61 of N-RAS, but rarely in K-RAS, and very discussed above were initially identified by next-generation
rarely in H-RAS. Overall, the incidence of N-RAS muta- sequencing and then subsequently validated and the inci-
tions is between 10% and 27%.153 The presence of RAS dence confirmed in larger cohorts of AML patients.
mutations has been reported to correlate with low blast However, the NGS analysis has also identified that each
counts and an improved survival of patients, but the con- patient has a unique spectrum of random, preexisting back-
trast has also been reported in patients with RAS mutations ground mutations in the hematopoietic cell that acquired
having a worse outcome compared to outcomes in patients the potentially initiating mutation.158–161 These studies have
without tyrosine kinase or RAS mutations.153 also suggested that relapse clone in AML originates not
RAS mutations MDSs patients correlated with an only from the larger dominant clone present at diagnosis,
increased risk of progression to AML,153,154 but the RAS but from the rarer clones, at diagnosis, that evolve into more
mutation status in patients at first diagnosis and at relapse resistant and aggressive clones. Alternative therapies that
revealed a marked instability of these mutations, with loss can completely eradicate all the leukemic clones need to be
of mutations in some cases, sometimes replaced by muta- developed to improve outcomes by reducing relapse.
tions in different codons.155 This would suggest that RAS
mutations are not an initiating event but may be important
G E N E -E X P R E S S I O N C L A S S I FI C AT I O N
for disease progression.
IN AML
Mutations within the FLT3, KIT, and RAS genes have
been shown to have a strong involvement in the develop- The first approach to disease classification, based on micro-
ment of AML. But it is likely that mutations and the sub- arrays, was described in 1999162 and used an unsupervised,
sequent activation of other, as yet unidentified, signaling class discovery approach, identifying previously unrecog-
pathways also occur. nized subtypes, to uncover the distinction between AML
The over-expression of non-mutated receptor tyrosine and acute lymphoblastic leukemia (ALL). The authors
kinases may result in ligand-independent activation, such speculated that162 a larger sample would enable finer
as might occur in cases of FLT3 over-expression.156 High sub-classifications to be identified that would correspond to
expression of wild-type FLT3 in AML blasts can induce existing sub-classifications for AML or define new group-
spontaneous FLT3 auto-phosphorylation,156 which may ings. This concept sparked numerous publications163,164
contribute to leukemia transformation. trying to identify gene-expression profiles or signatures
431
Table 28.1 COMPARISON OF CROHN’S DISEASE WITH ULCERATIVE COLITIS
mice affected with colitis, the increased risk of cancer can increasing incidence and prevalence of IBD in industrialized
be due to the toxicity of the altered microbial composition countries (see Figure 28.1). The clinical manifestation of IBDs
(Arthur et al. 2012) as well as modulation by interleukin is likely to be due to a provocation of the mucosal immune sys-
22 (IL-22), a cytokine of the IL-10 superfamily (Huber tem by commensal intestinal flora in susceptible individuals—
et al. 2012). in most individuals, these would be non-pathogenic (Xavier
and Podolsky 2007). Accumulating evidence from genetic
epidemiology implies that the host genome predisposes to
E P I D E M I O L O GY
autoimmune and/or inflammatory diseases (Jostins et al.
Over the last century, CD and UC have manifested an increas- 2012). However, genetics alone is not adequate in explaining
ing incidence in various global populations, with a current esti- the disease, since animal models that are genetically altered
mated prevalence of 0.15% in northwest Europe and North for predisposition to colitis do not develop the phenotype
America (Binder 2004). The environmental influences associ- when kept in a germ-free environment (Blumberg et al. 1999;
ated with modern urban developments have coincided with Strober 1985).
Figure 28.1
The regions known for high incidence and prevalence of inflammatory bowel diseases are highlighted on the global atlas. The allele
frequencies (shown in the lighter font showing the cases and the darker font showing the controls) indicate association of the heterozygote
NOD2 mutations Arg702Trp (SNP8), Gly908Arg (SNP12), and the frameshift 3020 insC (SNP13) among populations of European descent
and European admixture populations affected with Crohn’s disease (see section “Epidemiology of NOD2 in CD”; Hampe et al., 2002; Cavanaugh
et al., 2003; Fidder et al., 2003; Palmieri et al., 2003). In contrast, a core TNFSF15 haplotype is associated with CD in both Caucasians and
Japanese populations (see section “TNFSF15”). Furthermore, the HNF4A locus is also associated with UC in both Caucasian and Japanese
populations.
Meta-Analyses of IBDs
1 2 3 4 5 6 7 8 9 10 11
M
H
C
IL23R
TNFSF15
CARD9
IL10
ATG16L1
12 13 14 15 16 17 18 19 20 21 22
NOD2 HNF4A
Figure 28.2
The meta-analysis of GWAS has confirmed 110 (dark shaded line) IBD loci, and 30 CD-specific (medium shaded line) and 23
UC-specific (light shaded line). Linkage to IBD at the MHC locus (6p) was also confirmed by a meta-analysis (adapted from van Heel et al.,
2004).
Arg703Cys Met863Val
Arg311Trp Ser431Leu Val793Met Asn852Ser
28 124 127 220 273 577 744 1020 1040
Nuclear
Leucine rich
CARD1 CARD2 binding
repeat region
domain
N terminal C terminal
Leu469Phe Gly908Arg
Arg702Trp
Arg334Glu
Arg334Trp
Leu1007finsC
Mutations in vitro show an in Mutations in vitro show a
increase in NF-kβ activation decrease in NF-kβ
Figure 28.3
The three NOD2 domains include the following: caspase recruitment, nuclear binding, and leucine-rich domain. The three relatively
common CD-associated variants display a loss of function, but the rare mutations that are associated with Blau syndrome reveal a gain in function
from in vitro studies. Other relatively rare CD-associated variants are also shown.
448
Box 29.1 GLOSSARY OF GENETIC TER MS (SEE ALSO CHAPTER 50)
3C Chromosome conformation capture. This is a technique that can study the spatial 3D conformation
of chromatin within a cell.
Epigenetics The study of heritable changes to gene function that do not involve a change in the DNA sequence.
eQTL Expression quantitative-trait loci are genomic loci that that explain variation in gene-expression levels.
Exome sequencing A sequencing strategy to obtain the DNA sequence corresponding to all exons, excluding introns and
non-coding genomic sequence.
GWA or GWAS Genome wide association study. A study in which specific genetic markers (usually SNPs) across the entire
genome are scanned in both cases and controls in order to find genetic variations associated with disease.
Genome-wide Statistical term indicating that a genetic association is probably real, rather than being a false positive. To
significance account for multiple testing, 5 x 10-8 is used to claim statistical significance for genetic association. This is
based on testing 1 million SNPs with Bonferroni correction.
HapMap A catalogue of common genetic variants to help identify similarities and differences between human beings.
HLA Human leukocyte antigen. Human version of the major histocompatibility complex, a region on chromo-
some 6 containing many genes of immunological function.
Genotype imputation The prediction of genotypes at SNPs that have not been assayed directly in an association study.
Linkage disequilibrium Non-random association of alleles at genetic loci such that combinations of alleles occur more or less fre-
quently in a population than would be expected from the random formation of haplotypes.
MAF Minor allele frequency. Frequency at which the least common allele occurs in a given population.
Non-synonymous SNP A nucleotide mutation that alters the amino acid sequence of a protein.
SNP Single nucleotide polymorphism. A single base change in the DNA sequence. The most common type of
genetic variation among people.
manifestations by months or years. These antibodies include simultaneously to better predict the diagnosis and progno-
the classic immunoglobulin M (IgM) rheumatoid factor sis of patients with RA.
(RF) found in approximately 85% of the patients, and the Bone loss is often a complication of RA; it can be
more specific anti-cyclic citrullinated peptide antibodies both focal and more generalized.3 Juxta-articular osteope-
(anti-CCP), the latter being detectable early in the course nia adjacent to inflamed joints and the presence of focal
of disease and also shown to be strong predictors of disease
severity and radiological damage.2 Both tests are often used
Figure 29.1
Rheumatoid arthritis of the hands showing volar subluxation
at the wrist and metacarpophalangeal joints with associated soft-tissue Figure 29.2
Classic juxta-articular erosions in rheumatoid arthritis
swelling. occurring at the point of synovial insertion (arrows).
SCORE
Target population (Who should be tested?): patients who—
Have at least one joint with definite clinical synovitis (swelling)*
With the synovitis not better explained by another disease†
Classification criteria for RA (score-based algorithm: add score of categories A–D; a score of 6/10 is needed for classification
of a patient as having definite RA)‡
A. Joint involvement§
1 large joint¶ 0
2–10 large joints 1
1–3 small joints (with or without involvement of large joints) 2
4–10 small joints (with or without involvement of large joints) 3
>10 joints (at least 1 small joint)** 5
B. Serology (at least 1 test result is needed for classification)††
Negative RF and negative anti-CCP 0
Low-positive RF or low-positive anti-CCP 2
High-positive RF or high-positive anti-CCP 3
C. Acute-phase reactants (at least one test result is needed for classification)‡‡
Normal CRP and normal ESR 0
Abnormal CRP or abnormal ESR 1
D. Duration of symptoms§§
<6 weeks 0
≥6 weeks 1
* The criteria are aimed at classification of newly presenting patients. In addition, patients with erosive disease typical of rheumatoid arthritis (RA) with a history com-
patible with prior fulfilment of the 2010 criteria should be classified as having RA. Patients with longstanding disease, including those whose disease is inactive (with
or without treatment) and who, based on retrospectively available data, have previously fulfilled the 2010 criteria, should be classified as having RA.
† Differential diagnoses vary among patients with different presentations, but may include conditions such as systemic lupus erythematosus, psoriatic arthritis, and
gout. If it is unclear which relevant differential diagnoses to consider, an expert rheumatologist should be consulted.
‡ Although patients with a score of 6/10 are not classifiable as having RA, their status can be reassessed and the criteria might be fulfilled cumulatively over time.
§ “Joint involvement” refers to any swollen or tender joint on examination, which may be confirmed by imaging evidence of synovitis. Distal interphalangeal joints, first
carpometacarpal joints, and first metatarsophalangeal joints are excluded from assessment. Categories of joint distribution are classified according to the location and
number of involved joints, with placement into the highest category possible based on the pattern of joint involvement.
¶ “Large joints” refers to shoulders, elbows, hips, knees, and ankles.
# “Small joints” refers to the metacarpophalangeal joints, proximal interphalangeal joints, second through fifth metatarsophalangeal joints, thumb interphalangeal
joints, and wrists.
** In this category, at least one of the involved joints must be a small joint; the other joints can include any combination of large and additional small joints, as well as
other joints not specifically listed elsewhere (e.g., temporomandibular, acromioclavicular, sternoclavicular, etc.).
†† “Negative” refers to international unit (IU) values that are less than or equal to the upper limit of normal (ULN) for the laboratory and assay; “low-positive” refers
to IU values that are higher than the ULN but 3 times the ULN for the laboratory and assay; “high-positive” refers to IU values that are 3 times the ULN for the labo-
ratory and assay. When rheumatoid factor (RF) information is only available as positive or negative, a positive result should be scored as low-positive for RF.
‡‡ “Normal/abnormal” is determined by local laboratory standards. CRP—C-reactive protein; ESR—erythrocyte sedimentation rate.
§§ “Duration of symptoms” refers to patient self-report of the duration of signs or symptoms of synovitis (e.g., pain, swelling, tenderness) of joints that are clinically
involved at the time of assessment, regardless of treatment status.
SOURCE: Aletaha et al. “2010 Rheumatoid Arthritis Classification Criteria.” Arthritis & Rheumatism. 2010; 62(9):2574. DOI 10.1002/art.27584. Copyright 2010,
American College of Rheumatology. Reprinted with permission by John Wiley & Sons.
of inflammatory arthritis is challenging. Radiographic for early treatment intervention and higher chances of
changes are often absent at presentation, and conventional remission. In the new criteria, the diagnostic and prog-
radiographic methods do not provide much information nostic importance of anti-CCP as a serological marker in
about the soft tissues and synovium.20 Ultrasonography and addition to RF has been realized. Anti-CCP antibodies
magnetic resonance imaging (MRI) have shown promising can present early in the disease, often before clinical onset,
results for detecting synovitis and early erosive change, but and the titer of anti-CCP has been shown to increase over
their utility in the management of RA is currently limited time, predating the diagnosis of RA.24,25 The anti-CCP
by cost and availability.21–23 The new classification criteria test has high specificity (95% for RA), and the presence of
home in on RA characteristics that emerge early in the dis- anti-CCP antibodies has also been shown to predict the
ease course, with the aim of identifying patients early in development of erosive disease with a higher risk of joint
their disease in order to widen the window of opportunity damage.26 A strong association of anti-CCP with the major
462
be the next frontiers in genetic and epigenetic investigations. also been previously associated with include SNPs within
I will conjecture about the implications of specific loci for PDE4D20 and DENND1B.21
therapeutic intervention and interpret the overall findings The study of populations with diverse ancestry and
in the context of investigations of the airway microbiome. environmental exposures adds depth to genetic studies.
A GWAS of pediatric asthma in the Japanese population
showed pediatric asthma to be associated primarily to the
major histocompatibility complex (MHC), with asso-
G E N ET I C S ciation to HLA-DP alleles DPA1*0201 and DPB1*0901,
which are in strong linkage-disequilibrium with each other
G E N O M E -WI D E A S S O C I AT I O N S T U D I E S were strongly associated with pediatric asthma.22 This study
found no effects from the ORMDL3 locus, although asso-
Large-Scale Studies
ciation to this locus had been confirmed in other Japanese
Adequately powered genome-wide association studies subjects (Hirota PMID 18155279). A study of asthma in
(GWAS), in which hundreds of thousands of genetic mark- Japanese adults showed the strongest associations to the
ers are genotyped in thousands of cases and controls, have MHC, between the complement genes and the HLA-DRA
been the method of choice for identifying complex dis- locus, and confirmed association to the TSLP locus.23 These
ease genes.12 Several of these have now been completed for results suggest a differing balance of etiologies to asthma in
asthma, with gratifying and consistent results. Japan and Europe. The HLA-DP association may indicate
A first-generation GWAS of approximately 1000 the presence of an antigen driving the process, and it may
children with asthma and 1000 controls showed in 2007 be relevant that HLA-DP associations to aspirin-induced24
that markers on chromosome 17q21 were strongly asso- and isocyanate asthma25 have been observed previously,
ciated with childhood-onset asthma.13,14 The GWAS both involving small molecules.
was extended by measuring global gene expression in A well-powered meta-analysis of GWAS in ethnically
Epstein-Barr virus–transformed lymphoblastoid cell diverse North American populations from the EVE consor-
lines (LCL) from asthmatic children and their siblings.15 tium (meta-analysis of genome-wide association studies of
This showed that the asthma-associated SNPs were also asthma in ethnically diverse North American populations)
strongly associated (P = 10–23) with transcript abundance confirmed the previously reported loci near ORMDL3,
of ORMDL3.13 Subsequently, we and others have shown IL1RL1, TSLP, and IL33, and showed that these loci were
that the locus also regulates expression of the neighboring associated with asthma risk in Hispanic and African-American
gene, GSDMB.16,17 groups. The authors also identified association to PYHIN1,
A multidisciplinary study to identify the genetic and with the association being specific to individuals of African
environmental causes of asthma in the European commu- descent.26 This result has yet to be confirmed.
nity (GABRIEL collaboration) was a large international More recently, an Australian study that combined new
initiative that completed a second-generation GWAS of data with published results showed new associations to the
10,000 asthmatics and 16,000 controls, including those interleukin-6 receptor (IL6R) gene and markers on chro-
with adult-onset asthma as well as childhood disease.18 mosome 11q13.5 near the leucine-rich repeat containing 32
Associations were found on chromosome 2 within IL1RL1/ gene (LRRC32, also known as GARP). The LRRC32 locus
IL18R1; on chromosome 6 within HLA-DQ; on chromo- was significantly associated with atopic status among asth-
some 9 flanking IL33; on chromosome 15 within SMAD3, matics.27 It had first been found to be a susceptibility locus
and on chromosome 22 within IL2RB. Association within for atopic dermatitis28 and was shown in the ALSPAC pop-
the chromosome 17q21 ORMDL3/GSDMB locus was ulation to be a novel susceptibility factor for atopic asthma
confined to childhood-onset disease, and the HLA-DQ and hay fever.29
locus was the only significant hit in the adult-onset group. The asthma genetics community is currently consoli-
Markers at two additional genes, SLC22A5 and RORA, dating all the data from these international GWAS, so the
were identified just below genome-wide significance in the number of asthmatics included in a single meta-analysis
whole group, and markers within TSLP were seen at a simi- will rise from 10,000 to 25,000. The experience with other
lar level of significance in the patients with the most severe diseases30 indicates that this exercise may double the num-
disease. Associations within IL1RL1 and IL33 had been ber of verified asthma-susceptibility loci. It is likely, how-
identified previously, through the intermediate phenotype ever, that the strongest and most consistent genetic effects
of elevated eosinophil counts,19 and TSLP variants had on asthma have already been identified.
4 6 4 • G e no m ic s in C l inic a l P r actic e
in the gastrointestinal and bronchial epithelium (http:// Thymic stromal lymphopoietin (TSLP)
www.proteinatlas.org/). Members of the gene family may
Despite the initial identification of TSLP in the culture
have a role in the regulation of apoptosis.46
supernatant of a thymic stromal cell line, this cytokine is
expressed mainly by epithelial cells at barrier surfaces (skin,
IL33, IL18R1, and IL1RL1 gut and lung).56,57 The cell populations with the highest
known co-expression of the TSLP receptor (TSLPR) and
IL33, IL18, and IL1 belong to the IL1 family of cytokines
its associated subunit IL-7Rα are myeloid dendritic cells
that alter host responses to inflammatory and infectious
(DCs).56 Treatment of human DCs with TSLP induces
challenges. They exert their functions through a family of
improved survival, upregulation of major histocompatibil-
receptors that belong to the Toll-like receptor-IL-1 recep-
ity complex class II and co-stimulatory molecules, and the
tor (TLR-IL-1R) superfamily. An important feature of IL1
production of a variety of chemokines.56 It promotes TH2
receptor signaling is the activation of transcription factor
cytokine–associated inflammation by directly promot-
NF-κB and mitogen-activated protein (MAP) kinases p38,
ing the effector functions of CD4+ TH2 cells, basophils,
JNK, and ERK1/2.47
and other granulocyte populations while simultaneously
IL33 was originally identified as a nuclear factor in vas-
limiting the expression of DC-derived proinflammatory
cular endothelial cells,48 and was subsequently detected in
cytokines and promoting regulatory T-cell responses in
airway epithelial cells.49,50 The activities of IL33 as a nuclear
peripheral tissues.57
factor remain unclear,51 but it may repress gene expression
of pro-inflammatory factors. In contrast to inducible cyto-
kines, IL33 is constitutively expressed and is thought to SMAD3
function as an endogenous danger signal or alarmin to alert
the immune system after endothelial or epithelial cell dam- SMAD3 encodes SMAD (“mothers against decapentaple-
age during trauma or infection.52 IL33 is formed as a prodo- gic homolog”) family member 3 and came to attention
main containing polypeptide, which after activation with for its role in modifying tumour growth57,58,59 through
caspase-1, is realized as mature IL33. A mouse gene knock- the transforming growth factor-beta (TGFB) pathway.60
out has shown Il33 works as a crucial amplifier of innate SMAD3 is concentrated in the nuclei of bronchial epithe-
immunity.53 IL-33 expression is induced by a range of envi- lial cells and macrophages (http://www.proteinatlas.org/)
ronmental and endogenous triggers, suggesting an essential and functions as a transcriptional modulator activated by
role during infection, inflammation, and tissue damage.54 TGFB. TGFB family members exert fundamental effects
IL33 activates a heterodimeric receptor complex con- on the maintenance of immune function in the lung,61
taining IL1RL1 (ST2) and IL-1 receptor accessory pro- and TGFB signaling pathways are activated after allergen
tein (IL1RAP), leading to activation of NF-κB and MAP challenge in mild asthma.62 A mouse knockout of Smad3
kinases and drives production of Th2-associated cytokines showed accelerated wound healing and an impaired local
such as IL4, IL5, and IL13.49 inflammatory response,63 even though mice lacking Smad3
The IL18R1 gene, along with four other members of may exhibit increased baseline levels of proinflammatory
the interleukin 1 receptor family (IL1R2, IL1R1, ILRL2 cytokines in their lungs.64 Smad3 signaling is required for
[IL-1Rrp2], and IL1RL1 [T1/ST2]), form a cluster on myogenic differentiation of myoblasts,65 perhaps pointing
chromosome 2q. IL18R1 and IL1RL1 flank each other to a role in airway smooth muscle hypertrophy.
with the same orientation of translation. They are within
the same island of linkage disequilibrium, and it has not yet
IL2RB
been possible to assign the genetic effects at this locus to
one gene or the other. It is possible that both genes may be IL12RB encodes the beta receptor of IL2. IL2 is secreted by
co-regulated. antigen-activated T cells. It controls the survival and prolif-
As stated above, IL1RLI encodes the receptor of IL33. eration of regulatory T cells66 and plays a crucial role in the
IL18 is closely related to IL3349 and synergizes with IL12 to maintenance of natural immunological self-tolerance.67 The
induce the production of interferon gamma and to promote IL2 receptor is composed of α (CD25), β (CD122), and γ
TH1 responses.55 These loci therefore identify a pathway for chains.66 The β chain (IL2RB) is a signal transduction ele-
the communication of epithelial damage to the adaptive ment that is also present in the IL15 receptor. It belongs to
immune system and a potential switch point for choosing the type I cytokine receptor family and is devoid of intrinsic
between TH1 or TH2 responses. kinase activity.68 The receptor modulates T cell–mediated
4 6 6 • G e no m ic s in C l inic a l P r actic e
epithelium, a worthwhile understanding of the causes of airways.95–97 Murine studies have shown that sterility of the
asthma needs to reconcile consistent epidemiological indi- airways and intestinal tract results in markedly enhanced
cations of the importance of the microbiome to the disease. inflammatory responses to a variety of stimuli.96,98
These include the protection afforded by a rich microbial The study of the role of the microbiome of asthma is in
environment in early life,10,86 observations that the bron- its infancy. Even at this early stage, however, there are many
chial tree contains characteristic flora that are disturbed by questions that suggest a structured set of experiments that
the presence of pathogens such as Haemophilus infuenzae in will elucidate how the microbiome operates in health and
asthma,11,87 birth cohort studies showing that the presence disease.
of the same pathogens in throat swabs predicts the later
development of asthma,88 and recognition that these bac-
teria have consistently been associated with exacerbations C L I N I C AL I M P L I C AT I O N S
of asthma.89
Ninety percent of the cells in the human body are
G E N ET I C R I S K
microorganisms including bacteria, parasites, and archaea.90
These microorganisms are commensal on body surfaces The primary motivation for genetic investigations of
exposed to the external environment, including the gut, asthma has been to develop a systematic understanding
respiratory tract, and skin. Although most bacteria are not of otherwise impenetrable disease processes. The success
cultivable with standard methods,91 the membership of of GWAS for complex diseases and the rapidly declining
complex microbial communities can be quantified and clas- costs of sequencing individual human genomes have led to
sified by DNA sequencing of the conserved bacterial 16S a high level of public interest in using DNA to provide an
rRNA gene.92,93 Bacteria are classified by these sequences estimated individual risk of disease, with the expectation
into operational taxonomic units (OTUs). OTUs approxi- that high-risk individuals may benefit from preventive and
mate closely, but not completely, to taxonomy derived from therapeutic interventions.99 It is appropriate that the value
classical techniques, but sequences of other regions may be of individual genetic screening has been questioned on the
necessary for precise discrimination at the species level. grounds of interpretation of statistical risk estimates,99 the
A small study from our group compared the airway quality control of genotyping, and the need to establish
microbiota at three levels in adult patients with asthma, the validity of predictive tests from appropriately designed
the related condition of COPD, and controls.11 Bronchial trials.100
lavage from asthmatic children and controls was also inves- The outstanding finding from asthma GWAS has been
tigated.11 The bronchial tree was not sterile, and contained the GSDMB-ORMDL3 locus on chromosome 17q21.13,18
a mean of 2,000 bacterial genomes per square centimeter In the initial GWAS,13 the odds ratio (OR) for the most
surface sampled. Pathogenic proteobacteria, particularly strongly associated SNP in cases of childhood asthma
Haemophilus spp., were much more frequent in bronchi of recruited through hospital clinics and compared to con-
adult asthmatics or patients with COPD than in controls. trols was 1.84 per additional risk allele. The per-allele OR
Highly significant increases in proteobacteria were also seen at this locus for childhood-onset asthma was 1.32 in the
in asthmatic children. GABRIEL consortium GWAS, which contained a high
A recent study used a microarray technique to test for proportion of asthmatics from epidemiological surveys,18
relationships between the composition of the airway bacte- and was 2.02 in severe therapy-resistant asthmatics attend-
rial microbiota and clinical features of asthma, establishing ing tertiary referral clinics.101 These substantial differences
that bacterial burden and diversity were higher among asth- in ORs reflect the different ascertainment criteria and types
matic patients compared to controls, although not identify- of asthma for each study, and show that genetic risk is high-
ing the organisms responsible for this effect.87 est in children with early-onset disease that is resistant to
Of potential importance is the finding that organisms therapy, and that environmental factors are important mod-
commonly found in healthy airways (such as Bacteroidetes, ifiers of this risk.
particularly Prevotella spp.) were significantly reduced in The size of these effects can be put into perspective
asthmatic airways.11 As described above, genetic studies by considering well-known epidemiological effects on
indicate an overlap in mechanisms underlying asthma and asthma. The OR for the protective effect of farming on the
IBD. It is increasingly recognized that a normal bacterial prevalence of childhood asthma in the PARSIFAL study
flora is essential in maintaining a healthy bowel mucosa,94 (Prevention of allergy risk factors for sensitization in chil-
and similar mechanisms are likely to be important in the dren related to farming and anthroposophic lifestyle) was
4 6 8 • G e no m ic s in C l inic a l P r actic e
receptor in inflammation have been difficult to define with C O N C LU S I O N S A N D F U T U R E
precision. IL33 knockouts have been difficult to develop, P E R S P E C T I VES
the nuclear binding sites for IL33 are not known, there has
not yet been a clear demonstration of free IL33 in inflamed GWAS of asthma have provided a rich harvest of knowledge
airways, and it is not yet a definite target for therapy. Based about the mechanisms of asthma. Next steps in genetics will
on the results of murine knockouts, IL18 and its receptor include a comprehensive meta-analysis of all GWAS data
IL18R may be more tractable therapeutic targets, but the internationally, currently being undertaken by the applied
field requires the development of an appropriate model sys- genetics core (TAGC). This exercise may double the num-
tem for IL18/IL18R interactions. ber of loci proven to be influencing asthma.
TSLP has a well-defined biology, and it represents a The contribution of rare mutations to asthma is not
good target for biologic therapies. It may be anticipated known, and full genomic sequencing of multiple diseased
that an anti-TSLP antibody would have its most profound individuals is the gold standard to be applied to this prob-
effects on dendritic cells and granulocytes. lem. Whilst the community waits for sequencing costs to
The IL2RB and SMAD3 proteins may provide a means fall to acceptable levels before undertaking full sequenc-
of downregulating inflammation and promoting healing. ing, whole-exome sequencing with enrichment of exonic
SMAD3 is, however, concentrated in the nucleus, and noth- sequences by hybridization before ultra-deep sequencing115
ing is yet known of its nuclear binding sites and partners. has proven to be extremely effective in the identification of
IL2RB is expressed at the cell surface, and murine knock- rare mutations in Mendelian diseases,116,117 and merits appli-
down studies indicate that biologics directed against it may cation to complex disorders such as asthma. Another interim
have therapeutic benefits. solution may come from the Illumina exome chip, which is
derived from assembled information on 12,000 sequenced
genomes and exomes and catalogues, for each variant that
M A N I P U L AT I N G T H E M I C RO B I O M E
potentially affects protein structure, the total number of
The finding of a disordered microbiome in asthma inevi- times it was seen, and the total number of datasets that
tably leads to direct manipulation of the airway bacterial included the variant. Non-synonymous variants seen at least
community. Clinical investigators have supported these three times across two datasets are to be included on the chip.
findings with direct interventions with long courses of anti- It is unlikely that many important genetic effects will
biotics.109 A conclusion from these studies is that chronic remain undiscovered after completion of global meta-analysis
bacterial infections are relevant in a subgroup of preschool and identification of any important rare mutations. The
children with persistent wheezing, and such children ben- source of the missing asthma heritability is most likely to be
efit significantly from antibiotic therapy.109 Persistent bacte- polygenic aggregation on multiple small effects. The large
rial bronchitis (PBB) is a syndrome of childhood wheezing number of loci identified through candidate gene studies118
with a chronic productive cough that is often diagnosed as that do not show up in GWAS may point to this eventuality.
asthma.110 It may be relevant that Haemophilus influenzae Molecular genetic studies identify unexpected pathways
and Streptococcus pneumoniae are the most commonly iso- and potential targets on the basis of sequence variation in
lated organisms, and that long courses of antibiotics are very DNA. The essential components of many disease-related
effective in treating the disease.110 pathways may, however, not be subject to genetic varia-
Many patients with asthma receive antibiotics at dif- tion. Epigenetic variation may be another means to identify
ferent times without noticeable changes in the course of novel disease pathways and therapeutic targets.
their disease. Rational antibiotic therapy would probably Whatever the outcome of future genetic studies, the
necessitate direct sampling of the airway mucosa with biggest advances in understanding and treating asthma
brushings or lavage, together with assessment of the key will come from full functional analyses of the genes already
elements of the local inflammatory response. Antibiotics identified, from reagents that manipulate their actions, and
have profound effects on microbial communities, and from disentangling their relationship to the microbial part-
restoration of a normal bacterial flora may be a neces- ners on the other side of the respiratory mucosa.
sary part of the therapeutic approach. In this context, it
is remarkable that replacement of a normal bacterial com-
R EFE R E N C ES
munity in the bowel has been very effective in patients
with IBD,111,112 and curative in patients with Clostridium 1. ISAAC. Worldwide variation in prevalence of symptoms of
difficile necrotizing enteroclotitis.113,114 asthma, allergic rhinoconjunctivitis, and atopic eczema: ISAAC.
4 7 0 • G e no m ic s in C l inic a l P r actic e
45. Tamura M, Tanaka S, Fujii T, et al. Members of a novel gene family, the homeostasis of T-cell subsets. J Allergy Clin Immunol
GSDM, are expressed exclusively in the epithelium of the skin and 2009;123:758–762.
gastrointestinal tract in a highly tissue-specific manner. Genomics 67. Setoguchi R, Hori S, Takahashi T, Sakaguchi S. Homeostatic main-
2007;89:618–629. tenance of natural Foxp3(+) CD25(+) CD4(+) regulatory T cells
46. Saeki N, Usui T, Aoyagi K, et al. Distinctive expression and function by interleukin (IL)-2 and induction of autoimmune disease by IL-2
of four GSDM family genes (GSDMA-D) in normal and malig- neutralization. JEM 2005;201:723–735.
nant upper gastrointestinal epithelium. Gene Chromosome Canc 68. Gaffen SL. Signaling domains of the interleukin 2 receptor.
2009;48:261–271. Cytokine 2001;14:63–77.
47. Dunne A, O’Neill LA. The interleukin-1 receptor/Toll-like recep- 69. Montes de Oca P, Malarde V, Proust R, Dautry-Varsat A,
tor superfamily: signal transduction during inflammation and host Gesbert F. Ectodomain shedding of interleukin-2 receptor beta
defense. Science’s STKE [electronic resource] 2003;171:2003, pre3. and generation of an intracellular functional fragment. J Biol
48. Baekkevold ES, Roussigne M, Yamanaka T, et al. Molecular charac- Chem;285:22050–22058.
terization of NF-HEV, a nuclear factor preferentially expressed in 70. Doganci A, Karwot R, Maxeiner JH, et al. IL-2 receptor beta-chain
human high endothelial venules. Am J Pathol 2003;163:69–79. signaling controls immunosuppressive CD4+ T cells in the draining
49. Schmitz J, Owyang A, Oldham E, et al. IL-33, an interleukin-1-like lymph nodes and lung during allergic airway inflammation in vivo. J
cytokine that signals via the IL-1 receptor-related protein ST2 Immunol 2008;181:1917–1926.
and induces T helper type 2-associated cytokines. Immunity 71. Lee GR, Fields PE, Griffin TJ, Flavell RA. Regulation of the Th2 cyto-
2005;23:479–490. kine locus by a locus control region. Immunity 2003;19:145–153.
50. Carriere V, Roussel L, Ortega N, et al. IL-33, the IL-1-like cytokine 72. Marsh DG, Meyers DA, Bias WB. The epidemiology and genetics of
ligand for ST2 receptor, is a chromatin-associated nuclear factor in atopic allergy. N Engl J Med 1981;305:1551–1559.
vivo. Proc Natl Acad Sci U S A 2007;104:282–287. 73. Marsh DG, Blumenthal MN, Ishikawa T, et al. HLA and spe-
51. Kurowska-Stolarska M, Hueber A, Stolarski B, McInnes IB.
cific immune respoinsiveness to allergens. In: Tsuji K, Aizawa
Interleukin-33: a novel mediator with a role in distinct disease M, Sazazuki T, eds. Proceedings of the Eleventh International
pathologies. J Intern Med 2011;269:29–35. Histocompatability Workshop and Conference. Oxford,
52. Moussion C, Ortega N, Girard JP. The IL-1-like cytokine IL-33 is UK: Oxford University Press; 1992:765–767.
constitutively expressed in the nucleus of endothelial cells and epi- 74. Moffatt MF, Young A, Schou C, Faux J, et al. Involvement
thelial cells in vivo: a novel “alarmin”? PLoS One 2008;3:e3331. of TCR-a/d and HLA-DR in specific allergy. Allergy Suppl
53. Oboki K, Ohno T, Kajiwara N, et al. IL-33 is a crucial ampli- 1995;50:164.
fier of innate rather than acquired immunity. Proc Natl Acad Sci 75. Moffatt MF, Cookson WO. Tumour necrosis factor haplotypes and
U S A;107:18581–18586. asthma. Hum Mol Genet 1997;6:551–554.
54. Lloyd CM. IL-33 family members and asthma—bridging
76. Link E, Parish S, Armitage J, et al. SLCO1B1 variants and
innate and adaptive immune responses. Curr Opin Immunol statin-induced myopathy—a genomewide study. N Engl J Med
2010;22:800–806. 2008;359:789–799.
55. Fukao T, Matsuda S, Koyasu S. Synergistic effects of IL-4 and IL-18 77. Tantisira KG, Lasky-Su J, Harada M, et al. Genomewide associa-
on IL-12-dependent IFN-gamma production by dendritic cells. J tion between GLCCI1 and response to glucocorticoid therapy in
Immunol 2000;164:64–71. asthma. N Engl J Med 2011;365:1173–1183.
56. Reche PA, Soumelis V, Gorman DM, et al. Human thymic stromal 78. Bird A. DNA methylation patterns and epigenetic memory. Genes
lymphopoietin preferentially stimulates myeloid cells. J Immunol Dev 2002;16:6–21.
2001;167:336–343. 79. Bird AP. CpG islands as gene markers in vertebrate nucleus. Trends
57. Ziegler SF, Artis D. Sensing the outside world: TSLP regulates bar- Genet 1987;3:342–347.
rier immunity. Nat Immunol 2010;11:289–293. 80. Eckhardt F, Lewin J, Cortese R, et al. DNA methylation profiling of
58. Yang YA, Zhang GM, Feigenbaum L, Zhang YE. Smad3 reduces sus- human chromosomes 6, 20 and 22. Nat Genet 2006;38:1378–1385.
ceptibility to hepatocarcinoma by sensitizing hepatocytes to apopto- 81. Feinberg AP. Phenotypic plasticity and the epigenetics of human
sis through downregulation of Bcl-2. Cancer Cell 2006;9:445–457. disease. Nature 2007;447:433–440.
59. Tian F, DaCosta Byfield S, Parks WT, et al. Reduction in Smad2/3 82. Liu J, Morgan M, Hutchison K, Calhoun VD. A study of the
signaling enhances tumorigenesis but suppresses metastasis of breast influence of sex on genome wide methylation. PLoS One 2010;5:
cancer cell lines. Cancer Res 2003;63:8284–8292. e10028.
60. Daly AC, Vizan P, Hill CS. Smad3 protein levels are modulated by 83. Christensen BC, Houseman EA, Marsit CJ, et al. Aging and envi-
Ras activity and during the cell cycle to dictate transforming growth ronmental exposures alter tissue-specific DNA methylation depen-
factor-beta responses. J Biol Chem 2010;285:6489–6497. dent upon CpG island context. PLoS Genet 2009;5:e1000602.
61. Lloyd CM, Hawrylowicz CM. Regulatory T cells in asthma.
84. Boks MP, Derks EM, Weisenberger DJ, et al. The relationship of
Immunity 2009;31:438–449. DNA methylation with age, gender and genotype in twins and
62. Kariyawasam HH, Pegorier S, Barkans J, et al. Activin and trans- healthy controls. PLoS One 2009;4:e6767.
forming growth factor-beta signaling pathways are activated 85. Heijmans BT, Kremer D, Tobi EW, Boomsma DI, Slagboom PE.
after allergen challenge in mild asthma. J Allergy Clin Immunol Heritable rather than age-related environmental and stochastic fac-
2009;124:454–462. tors dominate variation in DNA methylation of the human IGF2/
63. Ashcroft GS, Yang X, Glick AB, et al. Mice lacking Smad3 show H19 locus. Hum Mol Genet 2007;16:547–554.
accelerated wound healing and an impaired local inflammatory 86. Eder W, Ege MJ, von Mutius E. The asthma epidemic. N Engl J Med
response. Nat Cell Biol 1999;1:260–266. 2006;355:2226–2235.
64. Anthoni M, Wang G, Leino MS, Lauerma AI, Alenius HT, Wolff 87. Huang YJ, Nelson CE, Brodie EL, et al. Airway microbiota and
HJ. Smad3—signalling and Th2 cytokines in normal mouse airways bronchial hyperresponsiveness in patients with suboptimally con-
and in a mouse model of asthma. Int J Biol Sci 2007;3:477–485. trolled asthma. J Allergy Clin Immunol 2011;127:372–381 e3.
65. Ge X, McFarlane C, Vajjala A, et al. Smad3 signaling is required for 88. Bisgaard H, Hermansen MN, Buchvald F, et al. Childhood asthma
satellite cell function and myogenic differentiation of myoblasts. after bacterial colonization of the airway in neonates. N Engl J Med
Cell Res.2011;21(11):1591–1604. doi:10.1038/cr.2011.72. 2007;357:1487–1495.
66. Letourneau S, Krieg C, Pantaleo G, Boyman O. IL-2—
89. Kraft M. The role of bacterial infections in asthma. Clin Chest Med
and CD25-dependent immunoregulatory mechanisms in 2000;21:301–313.
4 7 2 • G e no m ic s in C l inic a l P r actic e
31.
GENETICS AND GENOMICS OF NEURO-PSYCHIATRIC
DISEASES, I: SEIZURE DISORDER S
William Owen Pickrell
473
Table 31.1 SOME DIFFERENT TYPES OF EPILEPTIC SEIZURE
GENERALIZED SEIZURES
Tonic-clonic Loss of consciousness with an initial period of stiffness (tonic) followed by rhythmic generalized jerking
(clonic). Gradual recovery of consciousness, with post-ictal confusion lasting minutes to hours.
Absence (typical) Sudden, brief (generally <10 sec) periods of loss of awareness with behavioral arrest (staring episodes) with
rapid recovery, occasional eye movements, automatisms.
Absence (atypical) Longer-than-typical absences and frequently associated with myoclonic or atonic attacks. Start and finish more
gradually; focal features more prominent, and more retained awareness than in typical absences.
Clonic Rare; loss of consciousness with repetitive symmetrical and rhythmical jerking. Rapid onset and cessation.
Tonic Sustained contraction of musculature, generally sudden-onset <1 min in duration. Abrupt onset and rapid
recovery.
Atonic Sudden loss of muscle tone, frequently causing a fall to the ground. Rapid recovery.
Myoclonic Brief muscle contractions with sudden onset and cessation. Single muscle to generalized jerking. Consciousness
generally not impaired.
FOCAL SEIZURES a
Frontal lobe Frequently nocturnal; brief, rapid onset and cessation. Prominent motor features, sometimes with posturing
and head version. Frequent bizarre automatisms/behaviors and vocalization.
Temporal lobe Generally longer in duration than frontal lobe seizures. Variety of sensory disturbances, including psychic
(déjà-vu, jamais-vu, fear), gustatory and olfactory hallucinations. Sensation of epigastric disturbance. Oro-facial
automatisms (e.g., chewing, sucking) or fidgety hand movements. Frequently, altered awareness (complex partial
seizures) and post-ictal confusion. Auditory features may point to lateral temporal lobe involvement.
Occipital lobe Visual disturbance/hallucinations, typically elementary color images in the contralateral eye field.
OTHER SEIZURES
Unknown Sometimes there is insufficient evidence available to classify a seizure as generalized, focal, or both.
a
These are the main types of focal seizure; there are other types. See also Berg et al. 2010.13
cause an increased tendency to experience epileptic seizures. environmental mechanisms underpins an individual’s
There are several different ways of classifying the epilepsies. seizure threshold. Figure 31.1 illustrates the spectrum of
A useful distinction is that between the symptomatic epi- some of the causes of epilepsy and the relevant influence
lepsies, where the epilepsy occurs with a concurrent under- of genetic and environmental factors.
lying structural or metabolic problem, such as tuberous
sclerosis; and the idiopathic epilepsies, where the epilepsy
occurs in otherwise normal individuals with no detectable MONOGENIC INHERITED
metabolic or structural problem, such as childhood absence EPILEPSIES
epilepsy (these were previously known as genetic epilep-
sies).13 There is some overlap, however, between genetic and Monogenic inherited epilepsies are rare, accounting for
symptomatic epilepsy, and in some cases the dividing line a small fraction of all epilepsies, and are predominately
is not so clear. For example, epilepsy due to glucose trans- caused by mutations in single genes encoding ion-channels.
porter (GLUT-1) deficiency is a symptomatic epilepsy, as it Despite their rarity, they are of interest as their study has
is caused by a metabolic problem, but it also can be thought led to a better understanding of some of the molecular
of as a genetic epilepsy, as it is caused by mutations in the mechanisms underpinning epilepsy. Despite being pre-
SLC2A1 gene. dominately monogenic, these epilepsies show locus hetero-
The concept of a seizure threshold can be used to geneity (i.e., the same phenotype caused by mutations in
explain the effect of genetics on the tendency to develop different genes); variable expression (mutations in the same
epilepsy. People with a tendency to develop seizures gene causing different phenotypes); and their causative
spontaneously can be thought of as having a low seizure mutations are rarely found in the more common sporadic
threshold, whereas people who develop seizures only genetic epilepsies. Table 31.2 lists the main types of mono-
after extremely provoking circumstances; for example, genetic genetic epilepsy, and the next few sections describe
severe head trauma patients can be thought of as hav- some of the more important of these epilepsies and the
ing a high seizure threshold. The interplay of genetic and mechanisms involved.
Trauma
Tumour
Infection
CVD
Metabolic
Neuro-
degenerative
Monogenetic
symptomatic
epilepsies e.g. NCLs
Increasing Increasing
Genetic Environmental
Influences Influences
Genetic focal &
generalised epilepsies
e.g. JME
Monogenetic
genetic epilepsies
e.g. BNFS
Genetic Epilepsies
Figure 31.1
Some of the causes of epilepsy, with the relative influence of genetic and environmental factors and the “division” between symptomatic
and genetic epilepsies. In some cases the distinction between symptomatic and genetic epilepsy is blurred; for example, disorders of cortical
malformation clearly cause a structural change in the brain (symptomatic), but they may be caused by a mutation in one gene and thus be genetically
predetermined (genetic). The figure is not to scale, and the size of the captions does not reflect the number of patients with epilepsy due to each
cause. BNFS = benign neonatal familial seizures; CVD = cerebrovascular disease; JME = juvenile myoclonic epilepsy; NCL = neuronal ceroid lipofuscinoses.
seizures.16 Retigabine, a new anti-epileptic drug used as to an inhibitory system after the neonatal period might
an additional therapy for refractory partial-onset sei- provide the additional “inhibition” that is required to
zures, acts by enhancing the activity of KV7.2–5 channels, prevent seizures in patients with KCNQ2/3 mutations.7
increasing M-current and therefore reducing neuronal Recently, KCNQ2 mutations have been discovered in
excitability.17 10% of patients with a neonatal/early infantile encepha-
A perplexing question is, why do the seizures in BFNS lopathy,15,18 suggesting a role for this gene in other types of
stop after a few months of life? Heterozygous mutations in more severe epilepsy.
KCNQ2 and KCNQ3 typically only cause a small reduc-
tion of 25–30% in the M-current,32 and it may be that,
S O D I UM C H A N N E L E P I L E P S I E S
after the neonatal period, the additional ion-channels
expressed in the brain mean that this small reduction in Voltage-gated sodium channels (VGSC) are well charac-
current is less problematic. Alternatively, the fact that terized to date and play an important part in initiating and
the GABAergic system switches from being an excitatory propagating action potentials in electrically excitable tissue
Severe myoclonic epilepsy in infancy (SMEI) or Dravet’s Gamma-aminobutyric acid type A receptors (GABAA)
syndrome is a severe epilepsy syndrome usually starting are ligand-gated ion-channels and are the major inhibitory
at around 6 months of age, usually with a febrile seizure ion-channels within the human brain. Binding of GABA to
and progressing to multiple drug-refractory seizures and these receptors causes an influx of chloride, which results in
developmental delay. Around 80% of all Dravet’s syn- inhibition of neuronal activity. GABAA receptors have a het-
drome cases are caused by mutations in SCN1A, with eropentameric structure formed from at least seven sub-unit
around 95% of these mutations occurring de novo.17,36 The classes, but most GABAA receptors found within the brain
fact that SCN1A mutations cause the relatively mild phe- contain two α subunits, two β subunits, and a γ or δ sub-
notype of GEFS+ as well as the devastatingly severe SMEI unit.25 Mutations in the GABRA1, GABRG2, and GABRD
is partly explained by the fact that the specific, mostly genes, which encode the α1, γ2 and δ subunit, respectively,
de novo, nonsense mutations associated with SMEI give have been implicated in a wide variety of epilepsy pheno-
rise to more severe truncated or non-expressed proteins. types, including febrile seizures, absence seizures, GEFS+,
This is compared to the mostly missense mutations that and familial juvenile myoclonic epilepsy ( JME).
are associated with GEFS+. However, there are probably The majority of mutations seem to cause a loss of
additional factors that explain variable expression such as function of the GABAA receptor and a reduction in
modifier genes, genetic mosaicism, or other undiscovered GABAA-mediated current. This would then cause a reduc-
influences. tion in inhibition, which seems a plausible mechanism for
Both loss- and gain-of-function mutations in SCN1A seizure generation.
have been described in GEFS+, although it generally seems
that loss-of-function mutations are more common. The fact
LGI1-A S S O C I AT E D E P I L E P S I E S
that NaV1.1 is predominantly expressed in inhibitory inter-
neurons may explain how loss-of-function mutations can The leucine-rich glioma-inactivated 1 (LGI1) gene
lead to increased seizures.35 derives its name from the leucine-rich regions on the
Lissencephaly and Subcortical Band Heterotopia The mutation results in defective transport of glucose across
the blood–brain barrier, depriving the brain of essential
Lissencephaly literally means “smooth brain” and is a neuro-
glucose. The disorder was discovered in 1991 and is char-
nal migration disorder in which there are absent or reduced
acterized by hypoglycorrhachia (low CSF glucose) with
gyra on the surface of the brain. Subcortical band heteroto-
normoglycemia, drug-refractory epilepsy with a variety of
pia (SBH) is a less severe form wherein bands of gray mat-
seizure types, acquired microcephaly, and episodic move-
ter are present within the white matter of the brain.35 Both
ment disorders.40 Although GLUT-1 DS is rare, it is an
cause learning difficulties and epilepsy. Most lissencephaly is
important diagnosis not to miss, as treatment with some
caused by mutations in PAFAH1B1 (LIS1) or DCX.36 DCX
conventional anti-epileptic drugs and other medications
is located on the X-chromosome, and mutations in females
can inhibit GLUT-1 and worsen symptoms, whereas treat-
cause SBH, whereas mutations in males cause lissenceph-
ment with a ketogenic diet can reduce seizure frequency
aly.37 Deletion of PAFAH1B1 as well as neighboring genes
dramatically.41
causes a more severe phenotype with characteristic dysmor-
phic features (Miller-Dieker syndrome). Both PAFAH1B1
and DCX regulate microtubule function, important in the Pyridoxine-Dependent Epilepsy
regulation of migrating neurones.38
Pyridoxine-dependent epilepsy is an autosomal recessive
disease, usually presenting in neonates with seizures, which
Periventricular Nodular Heterotopia respond to treatment with pyridoxine (OMIM 266100).
The disease is caused by homozygous or compound hetero-
Periventricular nodular heterotopia is a neuronal migration
zygous mutations in the ALDH7A1 gene, which encodes
disorder in which grey matter is positioned abnormally within
antiquitin, an aldehyde dehydrogenase. This causes an
the brain, typically along the walls of the lateral ventricles.
elevation in α-aminoadipic semialdehyde (AASA), result-
Mutations in FLN1 on the X chromosome account for most
ing in an intracellular reduction in pyridoxal-5′-phosphate
“classical cases.” Most females with the condition develop
(PLP)—the active vitamin B6 co-factor essential for neuro-
epilepsy, while males tend to die before birth. FLN1 encodes
transmission.42,43 Although rare, this is another example of
a protein with a role in neuronal maturation. There is also an
an important disease to diagnose, as lifelong treatment with
autosomal recessive form of the disease associated with micro-
pyridoxine causes seizure freedom.
cephaly and mutations in the ARFGEF gene (see Table 31.4).39
several disorders (Chapter 10). A recently published study exome-sequenced.53 No statistically significant variants
used WGS to identify a mutation in the sodium-channel were identified. From the list of 87,255 variants identified
gene SCN8A as a cause of a severe epileptic encephalopa- at the exome-sequencing stage, 3897 were identified and
thy.51 These NGS studies have used small sample sizes genotyped in a larger cohort of 878 patients with IGE and
(nuclear family units of three or four, or very small cohorts). 1830 controls. Again none of these candidate variants met
significance criteria (p <0.05) after Bonferonni correction
for the large number of variants tested. The study had 80%
Next-Generation Sequencing Panels
power to detect single-nucleotide variants (SNVs) with a
Lemke et al.52 used NGS technology to test a “panel” of frequency of 0.5% and a relative risk of 5.4. Although the
genes with a known association with epilepsy in 33 patients study did not produce any statistically significant findings,
with a wide range of sporadic and familial epilepsy pheno- the list of nearly 2000 variants observed exclusively in cases
types (see Chapters 6 and 10). They found, on average, 326 with IGE is likely to contain several real risk factors and will
variants per patient from within the panel of 265 genes. be a useful reference for future studies. It is interesting to
Filtering, using criteria such as including only variants note that, as in the GWAS studies described above, none of
within coding regions and not present on publically avail- the variations detected were in ion channel genes or known
able databases, reduced the average number of variants to ion-channel modifiers. Several studies are now underway
15 per patient. Software was then used in order to predict to use NGS technology with different methodologies in
the pathogenicity of the variants. This led to mutations large-scale cohorts, such as Epi4k.54
being found in 16 of the patients, including all eight of the
patients with a well-defined phenotype (some of whom had
EPIGENOMICS
already had candidate genes sequenced using traditional
methods with no success). Mechanisms other than changes in nucleotide sequence;
NGS gene panel methods have the advantage of being that is, epigenomic changes, may play a part in the etiology
able to offer increased coverage of genes of interest com- and hereditability of epilepsy (see also Chapter 4).55 Some
pared to WGS or WES methods. They also produce less examples of epigenetic mechanisms associated with epilepsy
‘noise,’ albeit at the cost of missing variation in genes not follow.
present on the panel. The cost of gene panels continues DNA methylation has a role in regulating gene expres-
to decrease and is already cheaper than sequencing several sion, and several genes associated with DNA methylation
genes using traditional methods. Gene panels are beginning have been associated with epilepsy; for example, MECP2
to be used in clinical practice and are likely to be increas- and the X-linked Rett syndrome (autism, learning difficul-
ingly used in the future. ties, and epilepsy—OMIM 312750).56
Histone proteins have a crucial role in gene regu-
lation as well as in organizing and packaging DNA.
Large-Scale Next-Generation Sequencing Studies
Kainate-induced seizures in animals are important models
Heinlein and colleagues have recently published the for research. Kainate administration causes changes in his-
results of a study in which 118 unrelated patients with tone proteins; for example, hyperacetylation of histones on
genetic generalized epilepsy (GGE), and 242 controls were the brain-derived neurotrophic factor (BDNF) promoter.
INTRODUCTION E P I D E M I O L O GY O F MU LT I P L E
SCLEROSIS
Multiple sclerosis (MS) is both an inflammatory and
a degenerative disease of the central nervous system The frequency of MS demonstrates considerable geographic
(CNS) and the commonest non-traumatic cause of neu- variation, and a latitudinal gradient has been observed for
rological disability in young adults in Western Caucasian some decades1; a less marked longitudinal gradient is also
populations. Its etiology is complex and incompletely seen from East to West. In European populations, the preva-
understood, but related to a combination of genetic and lence increases with distance from the equator, up to approx-
environmental factors.1 Although the disease’s frequency imately 55° north, with the highest worldwide frequency of
has increased over time, MS remains relatively rare, with MS being reported in Orkney, Scotland, where prevalence
a variable and geographically dependent prevalence was recently estimated at 402 per 100,000 (95% confidence
affecting approximately one per 685 of the population in interval [CI] 319 to 500).3 However, the frequency of MS
Northern Europe.2 Diagnosis can be problematic, clinical appears to decrease at latitudes further north.4 This pattern
manifestations of disease diverse, and the long-term prog- is replicated in the Southern Hemisphere, although the
noses extremely variable. Improvements in disease man- effect is somewhat attenuated when corrected for age and
agement have led to gradually increasing life expectancy, sex.5 The cause of these disease gradients remains obscure,
which now varies little from that of the general popula- and some observers have questioned their origins,6 sug-
tion, so patients commonly live with the consequences of gesting they may occur not as a true environmental effect
the disease rather than dying from it. However, despite sig- but rather as an artifact of historic population movements
nificant recent therapeutic advances in MS, there remains over the centuries. Recently, a detailed meta-analysis has
no cure; therefore the impact of the disease on the indi- examined these issues in more detail and concluded that
vidual is often substantial and prolonged. the latitudinal variance in disease prevalence occurs only
A fundamental epidemiological characteristic of MS in populations of European descent,4 therefore providing
is familial recurrence, which was recognized more than some convincing evidence for an interaction between genes
100 years ago and provided an early indication of the and environment that contributes to disease susceptibility.
importance of genetics to its frequency, distribution, and More formal analysis from population migration
outcome. Subsequently, an understanding of the genetic studies in people affected by MS has helped shed further
contribution to MS has proved valuable in counselling light on this variable geographical distribution of MS and
patients, understanding pathological mechanisms of dis- indicates that the environmental contribution to the risk
ease, identifying novel therapeutic targets, and determining of developing disease occurs at a relatively early stage in
disease outcome. Genetic studies have also proved infor- life. These studies suggest that migrants moving before
mative in unravelling environmental contributions to the their mid-teenage years appear to adopt the risk profile of
disease as well as the nature of potential gene–environment their host country, while migrants moving after this age
interactions. These combined factors may now offer the retain the risk of their country of origin.7–9 This phenom-
exciting future prospect of a more individualized approach enon is operative for migrants moving from areas of high
to patient management and treatment. prevalence to areas of low prevalence,7 and vice versa.8,9
487
Although there remains some debate about the underlying to identify protective factors relevant in this population
mechanism of this phenomenon and also about the meth- have so far been unsuccessful.18
odology of studies (which may have identified migrants Finally, a small number of large families with more
who were not genetically representative of the indigenous cases of MS than would be expected by chance alone have
population10), it is clear that environmental influences been identified, including a Canadian pedigree containing
can significantly affect the risk of developing MS and 15 affected individuals over four generations, which sug-
even override an underlying genetic predisposition to dis- gested a pattern of autosomal dominant inheritance (albeit
ease.11 An interesting exception to this is the unexpectedly with incomplete penetrance) leading to MS in this kin-
low rate of MS in second- and third-generation Japanese dred20; and a Swedish kindred with 22 cases (albeit related
Americans, born in the United States but with Japanese more distantly).21 However, detailed genetic analysis has
parents, suggesting that environmental factors may not so far failed to identify any such genes,22 and it now seems
be able to override genetic background in all cases, and unlikely that an autosomal dominant form of MS exists.
that a protective factor or factors may be active in specific
populations.12
Studies of some genetically and geographically isolated C L I N I C A L C H A R AC T E R I S T I C S
populations which have identified a disease frequency that O F MU LT I P L E S C L E R O S I S
does not conform to the more generally accepted pattern
of distribution have also provided valuable insights into As in other autoimmune diseases, MS is more common in
factors important in the development of MS. For example, females, with a ratio of approximately 1 male to 2.5 females,
the population of Sardinia, Italy, which although situated which appears to be increasing over time,23 and a peak age
in Southern Europe where rates of MS are generally low, of onset in the third and fourth decades.24 However, the age
has a relatively high prevalence of MS, estimated at 151.9 range of presentation is broad, and patients have been known
per 100,00013 and comparable to prevalence rates more to present in almost every decade of life (Figure 32.1).
commonly observed in Northern Europe. Sardinia has a In the majority of patients, the initial disease course fol-
genetically homogenous population that has been relatively lows a relapsing pattern, where episodes of neurological dys-
unaffected by recurrent waves of immigration, principally function are interspersed with periods of clinical stability25
because the interior of the island is remote and difficult to at a rate of approximately one significant relapse per year in
access.13 Consequently, Sardinians have an HLA profile the early stages. However, in around 10–20% of patients,
that is distinct from other European populations’14 but rela- there is progressive neurological deterioration from the
tively free of population stratification, making them a useful onset of the disease,26–28 commonly termed primary pro-
population in which to study complex genetic traits.14 Due gressive disease.25 Of patients whose initial disease course
to the relatively high prevalence of MS within a small popu- is relapsing, at least 65% will go on to develop progressive
lation and the fact that the Sardinian population has been neurological deterioration at a later date, which is known
well characterized over a number of years, it has proved as the secondary progressive phase of disease1,25 (Figure 32.2).
a particularly valuable population in which to examine MS relapses manifest as a subacute onset of neurologi-
genetic traits in MS.15–17 cal dysfunction in the absence of intercurrent illness (such
Conversely, the Hutterite community of west- as infection), evolving over days or weeks, and may affect
ern and central Canada and the United States has a any part of the nervous system. Common symptoms fre-
lower-than-expected frequency of disease and therefore quently relate to involvement of clinically eloquent areas
also has potential for the study of the effect of rare genes of the CNS and include visual loss, sensory disturbance, or
in MS. The Hutterites are an Anabaptist community that motor weakness, either alone or in combination. Less com-
emigrated from Europe and now comprises around 28,000 monly, brainstem or cerebellar dysfunction may be seen1,29,30
individuals in Canada alone.18 Marriage rarely occurs out- (Table 32.1). Younger patients are more likely to experience
side of these relative socially isolated Hutterite commu- bouts of optic neuritis than are older patients, while the fre-
nities; as a result, there is a high rate of consanguinity, so quency of certain other symptoms, such as lower limb motor
all Hutterites are related to each other as second cousins. disturbance, sphincter disturbance, and facial weakness,
However, MS remains rare in this population, despite its increases with age.31 Although older patients have fewer
location in an area generally considered to be a high risk for relapses32 and are less likely to experience a severe relapse,33
the development of MS,19 with a prevalence of only 21 per they are more likely to be left with residual disability as a
100,000 (although absolute numbers are small). Attempts consequence of the relapse, while younger patients make a
300
250 All
Number of patients
Female
200 Male
150
100
50
0
1–5 6–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45 46–50 51–55 56–60 61–65 66–70 71–75
Age at onset (years)
more complete recovery.31 In most patients, recovery from Two other commonly reported symptoms that are char-
relapse occurs within two months, although improvements acteristic of MS and also provide interesting clinical insights
may continue for up to a year following relapse.34 Overall into the pathology of disease, and in particular the effects
relapse rates within the MS population occur at approxi- of demyelination, are Lhermitte’s phenomenon (distal tran-
mately 0.3 relapses per patient per year32 and decrease grad- sient sensory symptoms on neck flexion, thought to occur
ually with age and disease duration, so that even in a group as the result of ephaptic transmission), and Uhtoff ’s phe-
selected for highly active disease, it is unusual to find more nomenon (transient neurological deterioration experienced
than two significant relapses per year.35 in sites of prior demyelination symptoms when core body
KEY
Initial relapsing course Current relapsing course
Disability
Time Time
Disability
Time Time
Figure 32.2 Changes in level of disability over time for each disease course group.
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 489
Table 32.1 FREQUENCY OF NEUROLOGICAL SYMPTOMS AT ONSET OF MS
temperature is raised, such as following a hot bath or shower, of MS, and the syndromic nature of the diagnostic process,
or after exercise1). Despite the widespread nature of pathol- inevitably lead to some significant difficulties in managing
ogy in the MS-affected brain, frank dementia is seen only patients in the very earliest stages of disease. In particular,
rarely, and usually in the latter stages of disease. However, the need to establish evidence for clinical or radiologi-
certain subtler cognitive deficits are frequently observed in cal change in order to demonstrate dissemination in space
earlier stages, and these are particularly related to speed of and time commonly leads to a delay from the onset of first
processing and executive dysfunction, although verbal and symptoms to a firm diagnosis. Prior to the widespread use
visuospatial memory may also be impaired. In contrast to of MRI in diagnosis, the mean time from disease-onset to
Alzheimer’s disease or vascular dementia, language function diagnosis was approximately eight years, but more recently
is often well preserved. However, overall cognitive dysfunc- this has been reduced to four years27 with more widespread
tion does not correlate well with levels of physical disability availability of imaging technology. However, even with
and may be confounded by other common symptomatic these advances, interval review of historical cohorts from
manifestations of MS, including fatigue and depression.37 specialist centers consistently reveals a small proportion of
patients in whom a previously established diagnosis of MS
has been reversed.27,42 Whilst absolute numbers of patients
D I AG N O S I S O F MU LT I P L E S C L E RO S I S
with initially erroneous diagnoses are small, it does result
From the earliest description of the disease, establishing a in a degree of fallibility in diagnosing MS, which may also
diagnosis of MS has necessitated a demonstration of dis- in turn inject some inconsistency into the interpretation of
semination of disease activity in both space and time38 genetic studies of MS.
and remains a cornerstone of initial disease assessment,
although this is now usually augmented by the results of
P RO G N O S I S O F MU LT I P L E S C L E RO S I S
paraclinical tests. On occasion the diagnosis remains chal-
lenging, despite the technological advances now available to Although relapses are a characteristic feature of early dis-
clinicians, principally because of the wide range of diseases ease, the evidence from natural history studies suggests that
that can mimic clinical features of MS (Table 32.2), but also relapse characteristics and frequency do not contribute sig-
because of the particular challenges of establishing diagnos- nificantly to longer-term outcome.43 Prognosis is therefore
tic certainty in patients with primary progressive disease. best assessed using fixed disability, the most commonly used
Diagnostic criteria have gradually been refined over time measure of which is the Expanded Disability Status Scale
as the understanding of information provided by magnetic (EDSS),44 a 20-point scale ranging from “no disability” to
resonance imaging (MRI) in MS has improved,38–41 along “death from MS” (Table 32.4). Commonly used disability
with the contribution of informative paraclinical tests. milestones include EDSS 4 (inability to walk 500 meters
At present, a diagnosis can still be made on clinical unaided), EDSS 6 (requiring unilateral support to mobi-
grounds alone if a patient has had two attacks, at least one lize), and EDSS 8 (wheelchair-bound). Alternative mea-
of which can be corroborated with objective clinical signs, sures of disability exist, although these are either derived
and in the absence of evidence for an alternative diagnosis. from the EDSS—such as the MS Severity Score (MSSS),
However, in practice, most clinicians still depend on sup- which incorporates EDSS and disease duration,45 or its pre-
port from the additional investigations described below, decessor the Progression Index46—or are not widely used,
and in the absence of clinical evidence, MRI is the test that making them less helpful for direct comparisons; for exam-
is perhaps most commonly employed to supplement clini- ple, the MS Functional Composite (MSFC).47
cal information38 (Figure 32.3; Table 32.3). However, the Most patients with MS will eventually develop progres-
lack of a single specific and sensitive test for the diagnosis sion of disability even when the initial disease course has
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 491
(A) (B) (C)
(G) (H)
Figure 32.3
MRI of brain in multiple sclerosis: (a) Axial T2 image showing periventricular and juxtacortical hyperintense lesions. (b) Axial T2
image showing juxtacortical hyperintense lesions. (c) Axial T1 image post-contrast showing gadolinium-enhancing lesions.(d) Axial T1 image
post-contrast showing gadolinium-enhancing lesions. (e) Axial T1 image showing periventricular and juxtacortical black holes. (f ) Axial T1 image
showing juxtacortical black holes (g) Sagittal T2 image of spinal cord showing single lesion at C4/C5. (h) Coronal T2 image showing brainstem
hyperintense lesions (as well as periventricular and juxtacortical lesions).
arthritis, psoriasis, and inflammatory bowel disease, as well axons, enabling rapid propagation of neural signaling).
as MS.53 Under the influence of IL-23, Th0 cells differenti- Subsequently, proinflammatory cytokines amplify the
ate into Th17 cells, which are then responsible for producing response, recruit activated macrophages and microglia,
a number of pro-inflammatory cytokines, including IL-6, and compromise the integrity of the blood–brain barrier,
IL-17, IL-22, TNF-α, and IFN-γ,54 which are key factors allowing ingress of further activated T cells.55 Propagation
in the development of pathology. Interestingly, IL-23–defi- of these effects leads to continuing tissue damage, and
cient mice have been shown to be resistant to developing oligodendrocytes are significantly depleted in active MS
mouse models of MS and inflammatory arthropathies.54 lesions and in chronic stable lesions,56 as well as in clini-
Th17 cells are activated in the periphery and cross the cally silent lesions.57 However, more recently it has become
blood–brain barrier into the CNS, where a local inflam- apparent that B cells also have an important role to play,
matory reaction results in damage to axons and oligo- which may have considerable relevance for the develop-
dendrocytes (which form the myelin sheath surrounding ment of novel treatments.1
EDSS SCORE
0 Normal neurological examination
1 No disability, signs present in one functional neurological system
1.5 No disability, signs present in more than one functional system
2.0 Minimal disability in 1 functional system
2.5 Minimal disability in 2 functional systems
3.0 Moderate disability in one functional system, or mild disability in 3 or 4 functional systems
3.5 Fully ambulatory but with moderate disability in one functional system and mild disability in one or two functional systems, or
moderate disability in two functional systems, or mild disability in five functional systems
4.0 Fully ambulatory without aid, up and about 12 hours a day, significant disability in one functional system, or combinations of
lesser grades exceeding previous steps. Able to walk 500 meters without aid or rest
4.5 Able to walk 300 meters without aid or rest
5.0 Able to walk 200 meters without aid or rest
5.5 Able to walk 100 meters without aid or rest
6.0 Intermittent or unilateral assistance required to walk 100 meters with or without rest
6.5 Constant bilateral assistance required to walk about 20 meters without resting
7.0 Unable to walk more than 5 meters even with aid, essentially restricted to wheelchair, but transfers alone
7.5 Unable to take more than a few steps, essentially restricted to wheelchair, may need help to transfer
8.0 Restricted to chair but up and about most of the day, has effective use of arms
8.5 Restricted to bed or chair, some effective use of arms
9.0 Helpless bed patient, can communicate and eat
9.5 Helpless bed patient, unable to communicate or eat
10 Death due to MS
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 493
Proportion of patients not yet reaching disability milestone 1.0 a neurodegenerative process that is independent of neuroin-
EDSS 4 flammation.62 However, recent histopathological evidence
EDSS 6
EDSS 8
has suggested that neurodegeneration and neuroinflam-
0.8
mation may coexist, with neuroinflammation driving the
process so that active lesions in relapsing, acute, and pro-
0.6 gressive MS show ongoing inflammation and axonal loss,
with no observable difference between clinical subgroups
of MS.61 As will be discussed below, the recent advances in
0.4 genetics in MS further implicate the immune system in MS
pathogenesis and do not provide any additional evidence
that neurodegeneration is a separate, or indeed the primary
0.2
underlying, process (Figure 32.5).
Finally, radiological correlates of both neuroinflam-
0.0 mation and neurodegeneration have been demonstrated
with gadolinium-enhancement of T2 lesions, considered
0 10 20 30 40 50 60
to reflect the level of inflammation and breakdown of the
Years since onset of disease
blood–brain barrier; T1 black holes representing areas of
Figure 32.4
Kaplan-Maier survival curves showing time to Expanded neurodegeneration and atrophy; and in vivo magnetic reso-
Disability Status Scales 4, 6, and 8 in a UK cohort of patients nance spectroscopy that is able to measure axonal damage
(unpublished data).
using levels of N-acetyl aspartate.63
WM
PP
(C) (D)
WM
BV
Figure 32.5
Histopathological changes in multiple sclerosis: (a) Histopathological specimen stained for myelin with Luxol Fast Blue, showing
loss of myelin within a slowly expanding plaque. 4x magnification. (WM = white matter, GM = gray matter, P = plaque) (b) High-power view
of peri-plaque area, showing demyelination in the plaque compared to surrounding white matter. 40x magnification. (PP = peri-plaque) (c)
Histopathological specimen stained for HLA, showing activated macrophages within the peri-plaque rim (arrows) of a slowly expanding MS
plaque. 4x magnification. (BV = blood vessel) (d) High-power view within the plaque showing few HLA-positive macrophages (arrows) in
relatively acellular, demyelinated plaque. 40x magnification.
reduce relapse rates by approximately 30%,66,72 although have an increased risk of disease compared to the general
a commonly used alternative is glatiramer acetate, which population, proportional to their degree of relatedness to
has a similar efficacy.73 More recently, several new treat- the proband.78–80 In general, male relatives have a lower risk
ments have been developed, including natalizumab,68 fin- than female relatives with the same degree of kinship78,79
golimod,69 alemtuzumab,35 laquinimod,70 BG-12,71 and (Figure 32.6). Similar patterns of inheritance are observed
teriflunomide74 (Table 32.5). Nevertheless, none of these in both high- and low-prevalence populations.80 The sibling
treatments has conclusively been shown to have a convinc- recurrence risk measures the risk to siblings relative to that
ing impact on long-term progression of disability,75–77 and for the general population and provides a measure of genetic
indeed there are currently no effective treatments for the risk in that population. It lies between 20 and 30 for MS in
neurodegenerative component of disease. both Europe17 and Australia,80 suggesting that genetic risk
remains similar, and observed variation in prevalence may
be more likely to relate to environmental factors affecting
G E N ET I C S A N D MU LT I P L E S C L E R O S I S the rate of MS within different populations with the same
genetic backgrounds.
Classical twin studies have also been useful in fur-
FA M I LY S T U D I E S
ther quantifying the genetic contribution to MS etiology.
For nearly a century, a genetic role in the etiology of MS Despite some early controversy,81 a substantially increased
has been suggested by epidemiological studies and observed risk to monozygotic twins of around 25% has been con-
patterns of familial recurrence. Relatives of MS probands firmed as compared to dizygotic twins, who have the same
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 495
Table 32.5 CURRENT TREATMENTS FOR MULTIPLE SCLEROSIS AND THEIR MECHANISMS OF ACTION
risk as non-twin siblings.82–84 The increased but imperfect It is apparent from these studies that the genetic con-
concordance between monozygotic twins adds weight to tribution to MS disease susceptibility is complex and likely
the hypothesis that there is both a genetic and an environ- to involve multiple genes and epistatic effects. More formal
mental contribution to MS risk. modelling using the familial risks as outlined above has sug-
Studies of half-siblings have proven to be another useful gested that at least six genes (but likely to be more, with no
method for understanding the genetic contribution to MS. upper limit to the model) contribute to disease susceptibil-
There have been two large studies of half-siblings in Canada, ity, and that an autosomal dominant pattern of inheritance
which showed an increased risk to half-siblings of patients fits the observed patterns much more closely than recessive
with MS as compared to the general population, but less models.92 Finally, it is possible and indeed likely that genetic
than that for full siblings85,86 (Figure 32.6). Additionally, heterogeneity exists, such that the risk alleles in one patient
there is also evidence for a parent-of-origin effect, with may be different from those in another.92
half-siblings who have maternal genes in common having a
greater risk of developing MS than paternal half-siblings,86
raising the possibility of parental imprinting and epistatic
T H E M A J O R H I S TO C O M PAT I B I L IT Y
effects, although evidence at the genetic level for such
COMPLEX
mechanisms has not been available until very recently.87,88
Furthermore, epidemiological studies of MS probands who The association of the HLA region with susceptibility to
were either adopted themselves, or who have adopted rela- MS was first identified in 1972,93 and has since been con-
tives, show no increase in risk of MS in the non-biological firmed using a variety of genetic analysis methods and
relatives of MS patients despite cohabitation,89 although in a wide range of populations. The HLA DRB1*1501–
risk for biological relatives of the same patients is increased HLA-DQA1*0102–HLA-DQB1*0602 extended haplo-
in a pattern similar to that of non-adopted probands, sug- type has been shown to be most strongly associated with
gesting that familial recurrence in MS is purely a genetic MS, increasing risk of MS three-fold, or by 17-fold in homo-
phenomenon.89 Furthermore, there is no increased risk zygous patients.94 Disentangling the effect of each individ-
to spouses90 or step-siblings91 of MS probands, making it ual allele has proven to be difficult, as HLA-DRB1*1501,
unlikely that a shared environmental factor or transmissible HLA-DQA1*0102, and HLA-DQB1*0602 are in strong
agent is responsible for disease ignition. linkage disequilibrium in European populations where MS
MZ: monozygotic
DZ: dizygotic
0.6% 3.2%
Figure 32.6 Pedigree diagram summarizing lifetime recurrence risk to siblings, half-siblings, and offspring of MS patients, according to published
data.78,79,81–86
is most prevalent, and it was not clear for many years whether nuclear families with at least one affected member, and studies
the effect was due to one allele or a combined effect.94 In involving this familial construct were initially the only ones
African-American populations, the HLA-DRB1*1501 able to detect an effect for HLA-DRB1*1501,104–107 although
allele is more frequently inherited alone, which allowed they remained largely underpowered. However, transmis-
confirmation that it is this allele that is responsible for the sion disequilibrium analysis (TDT) of MS families was also
impact on disease susceptibility; African-American patients able to confirm that HLA-DRB1*1501 is over-transmitted
carrying this allele were at similar increased risk as European to MS patients as compared to their unaffected siblings,16,96,98
patients carrying the HLA DRB1*1501–DQA1*0102– even in populations with a low prevalent frequency of
DQB1*0602 extended haplotype.95 The HLA-DRB1*1501 HLA-DRB1*1501.16 These early linkage studies also sug-
allele has now been shown to be associated with risk of gested that association with MS might be present within other
MS in a wide range of populations, including Northern genetic regions,104–107 albeit with a smaller effect size.106
European and European-descended populations,96–100 With the advent of the publication of the Human
Sardinian,16 Turkish,101 Mexican,102 Ashkenazi Jewish,103 Genome Project,108,109 effective international collaborations,
and African-American populations,95 although association and availability of low-cost genome-wide typing, explora-
is more difficult to demonstrate in populations where the tion of the complexities of the association of the MHC with
frequency of HLA-DRB1*1501 is low.16,95 MS susceptibility has become a more realistic proposition,
A recent meta-regression confirmed a significant and in recent years substantial advances have been made.
positive association of disease prevalence with latitude The association of HLA-DRB1*1501 has been unequivo-
in populations of European descent, and when corrected cally confirmed in genome-wide association studies
for HLA-DRB1 carriage, the association is enhanced.4 (GWAS) in MS in large-scale studies of unrelated MS cases
However, although HLA-DRB1 is clearly an important and controls.100 The genome-wide approach has also been
variable in the European population, it does not account able to identify other MHC susceptibility loci in addition
for all observed variations in prevalence, suggesting a signif- to the HLA-DRB1*1501 locus; in particular HLA-A*0201
icant and possibly essential role for environmental factors in has been shown to exert a protective effect,97,100 and if pres-
disease evolution.4 ent in conjunction with HLA-A*0201, it reduces the risk
The modest effect on MS susceptibility conferred by sin- from an odds ratio (OR) of 3.6 to 1.9.97 Independent effects
gle genes, including HLA-DRB1*1501, makes the study of of other MHC loci have also been observed, including a
the genetic contribution to MS difficult. Prior to more wide- protective effect from HLA-C*05,110 and increased risk
spread availability of the technology for genome-wide analy- associated with HLA-A*0301,97 HLA-DRB1*0301, and
sis, the most powerful methods for analysis were those using HLA-DQB1*0201.100
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 497
Large-scale genome-wide studies have also allowed linkage screens.104 Subsequently, IL-7R on chromosome
more detailed dissection of the MS risk associated with 5p13 was confirmed as a susceptibility locus for MS in both
the HLA-DRB1 locus, and it has become apparent that candidate-gene analyses116,117 and the first genome-wide
trans-epistatic effects between HLA-DRB1 alleles modu- study in MS,110 thus unequivocally establishing IL-7R as an
late the overall risk. The presence of HLA-DRB1*1501 in MS susceptibility locus. This first genome-wide study also
combination with HLA-DRB1* 17 confers a risk for MS identified two single-nucleotide polymorphisms (SNPs)
in excess of that seen from HLA-DRB1*1501 in conjunc- within the IL-2RA gene on chromosome 10p15 as a second
tion with alternative HLA-DRB1 alleles.111 HLA-DRB1*08 non-MHC MS susceptibility locus.110 Both these genes have
is independently associated with a small increased risk of a much smaller effect size than HLA-DRB1*1501: IL-2RA
MS, but if present together with HLA-DRB1*1501, the has an odds ratio of 1.25 (95% CI 1.16–1.36), while IL-7R
risk of MS doubles.111,112 In contrast, HLA-DRB1*14 exerts has an odds ratio of 1.18 (95% CI 1.11–1.26).110
a protective effect,111,112 such that if inherited in conjunc- More recently, an international collaborative GWAS
tion with HLA-DRB1*1501, it almost entirely negates the analyzed 9,772 MS cases and 17,376 controls. Nearly all pre-
risk associated with HLA-DRB1*1501.112 Epistatic effects vious associations were replicated, and further SNPs associ-
have also been demonstrated between HLA-DRB1*1501, ated with MS were identified, so that almost 50 SNPs outside
HLA-DQA1*0102, which increases the risk of MS if the MHC complex have now been shown to be associated
inherited together, and between HLA-DRB1*1501 and with MS,100 including SNPs within genes for cytokines
HLA-DQB1*0602, thus also implicating HLA-DQ in MS (CXCR5, IL2RA, IL7R, IL7, IL12RB1, IL22RA2, IL12A,
susceptibility.113 IL12B, IRF8, TNFRSF1A, TNFRSF14, TNFSF14), sig-
In concordance with the observation of a parent-of- naling molecules (CD37, CD40, CD58, CD80, CD86,
origin effect in epidemiological studies,86 over-transmission CLECL1), signal transduction (CBLB, GPR65, MALT1,
of HLA-DRB1*1501 from mothers to affected offspring RGS1, STAT3, TAGAP, TYK2), as well as genes relevant
has also been observed, and this effect is most marked for vitamin D (CYP27B1, CYP24A1), and response to MS
for daughters (OR 3.31, 95% CI 2.59–4.24), although therapies (VCAM1 and IL2RA).100 Furthermore, analysis
maternal transmission is also increased to sons (OR 2.13, of known genetic roles using a gene ontology approach has
95% CI 5 1.44–3.14), in contrast to paternal transmis- demonstrated that genes involved in immune system pro-
sion of HLA-DRB1*1501 (OR for daughter 2.14, 95% cesses, particularly lymphocyte function and specifically
CI 1.64–2.78; OR for son 2.16, 95% CI 1.37–3.39).87 T-cell activation and proliferation are over-represented in
Additionally, HLA-DRB1*08 is over-transmitted MS, confirming our previous understanding of MS as a pri-
from HLA-DRB1*1501-negative mothers to mary immunological disorder100 (Table 32.6).
HLA-DRB1*1501-positive daughters (OR 4.80, 95% CI Replication studies have since confirmed the asso-
1.43–16.06), and maternal transmission of HLA-DRB1*14 ciation of MS susceptibility with CD6,118–120, CD58,121
to HLA-DRB1*1501-negative is also distorted (transmis- CLEC16A,118 IL12A,122 IL12B,123 IRF8,119,120 KIF21B,124
sion to daughters: OR 0.40, 95% CI 0.11–1.43; sons: OR MPHOSPH9/CDK2AP1,122 RGS1,122 STAT3,121
0.14, 95% CI 0.02–0.84).87 TMEM39A,123,124 TNFRSF1A,119,120 and TYK2.125 The
importance of replication studies to confirm genetic asso-
ciation with MS was emphasized by a small GWAS that
N O N-M H C G E N E S
initially suggested association with KIF1B,126 which a
Early candidate-gene approaches exploring the association larger GWAS approach was unable to replicate.127 A further
of genes outside the MHC region with risk for MS were study combining several independent datasets and utilizing
unrewarding, and genome-linkage screens were underpow- non-equivalence statistical methods subsequently demon-
ered to detect the effect of other genes.104–106 Conflicting strated that this association was a false positive.128
results were obtained from studies of genes that seemed In addition, association of MS with a number of
likely candidates on a theoretical or functional basis to other genetic loci has been suggested, including chro-
be relevant to MS, such as the T-cell surface molecule mosome 5p13.1,129 chromosome 3p24.1, the MLANA
CTLA-4,114 or APOE-ε4,115 and later studies using more gene on chromosome 9p24.1, and chromosome 2p21.121
powerful methods of analysis have failed to confirm a role However, these have not been replicated in other cohorts
for these genes in MS susceptibility.100 and therefore remain unsubstantiated. Functional stud-
However, a putative association with MS for a region on ies have now started to unravel the possible mechanisms
chromosome 5p13 was suggested in one of the early genome underlying the association of IL-7R and IL-2RA with
(continued)
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 499
Table 32.6 CONTINUED
MS susceptibility. IL-7R is found in two forms: if exon Early studies to further characterize potential pheno-
6 is included in the transcript, this encodes a membrane- typical associations were limited by power,135 but associa-
bound form of IL-7R, while if exon 6 is not included, a tion between HLA-DRB1*1501 and a younger age at onset
soluble form of IL-7R is produced. When the risk allele of disease has now been confirmed by several independent
is present, it appears to modify exon splicing such that studies136–139; although it remains possible that this asso-
exon 6 is skipped twice as often, thus increasing the pro- ciation is simply a consequence of the strong association
duction of soluble IL-7R as compared to membrane- of HLA-DRB1*1501 with disease susceptibility, mak-
bound IL-7R. This is likely to affect IL-7 signaling and ing HLA-DRB1*1501-positive individuals more likely
in turn T-cell proliferation and maintenance.117 Similarly, to manifest disease earlier.138 Less robustly, an association
IL-2RA is also found in soluble and membrane bound between HLA-DRB1*1501 and female gender has also
forms. The two IL-2RA SNPs associated with MS have been observed,136,138,139 although other studies have failed to
been shown to be associated with variations in soluble confirm this.137 No association of HLA-DRB1*1501 with
IL-2RA levels in healthy controls and MS patients.130 disease course has so far been identified.24,136–138,140
Soluble IL-2RA can inhibit T-cell signaling, but can also Association of HLA with a more rapid progression of
enhance T-cell proliferation and expansion.130 Patterns of disability was recognized shortly after its association with
cell surface expression of IL-2RA have also been shown disease susceptibility was confirmed,141 and HLA-DQ poly-
to vary depending on IL-2RA genotype.131 Thus, genetic morphisms have been shown to influence the progression
variation may lead to alterations in T-cell development, of demyelination in viral mouse models of MS.142 However,
functioning, and signaling that may in turn lead to the deeper analysis of the contribution of HLA-DRB1*1501
development of MS. to accumulation of disability has been limited by power135
and by the difficulties inherent in the collection of dis-
ability data, including the slow rate of progression and the
G E N OT Y P E –P H E N OT Y P E
high rate of variability. Consequently, different measures
C O R R E L AT I O NS
of disability have been used in previous studies, making
A further question that is distinct from those relating to dis- direct comparisons difficult. However, no association of
ease susceptibility is whether there are genetic influences on HLA-DRB1*1501 with progression index135,138 or MSSS139
phenotypes. Some epidemiological studies of familial dis- has been found. Association between HLA-DRB1*01,
ease have provided some evidence for concordance within HLA-DRB1*04, and more rapid time to EDSS 6 has been
affected families, suggesting a genetic contribution to dis- observed using survival analysis, although no effect was
ease expression in addition to disease susceptibility.132–134 In seen for time to EDSS 3.140 Finally, HLA-DRB1*01 was
particular, concordance within families has been noted for over-represented in benign cases (EDSS ≤3 after ≥20 years
disease course132,134 and age at onset,134 but not for neuro- of disease) as compared to malignant cases (EDSS >6
logical symptoms at onset of disease.132 In addition there is within five years of disease onset)143 in a larger study uti-
also some limited evidence for familial concordance in the lizing an extremes-of-outcome approach (a more powerful
gender of affected patients.132 statistical method144).
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 501
lymphoblastoid cell lines are than non-immune cell types,88 EBV is currently thought to act by stimulating overactivity
such that vitamin D preferentially enhances transcription of the immune system, promoting a proinflammatory state,
of MS-associated genes within immunologically active cells. and allowing survival of a T-cell population with broad-
Smoking is more common in patients with MS than the ened specificity and higher likelihood of self-reactivity—
general population,158,159 and it has also been linked with the “hygiene hypothesis,” where a more hygienic early life
earlier conversion to secondary progression.158,160,161 The leads to late exposure to childhood viruses and a less bal-
incidence and prevalence of MS in women has been shown anced immune response as a consequence.159 Interaction
to have significantly increased over the last century, while between EBV and genetic profile has been shown to affect
rates in men have remained relatively constant.23 One possi- risk of MS.167,168 HLA-DRB1*1501-positive patients with
ble explanation for this may be the changing smoking habits high titers of antibody against the EBV nuclear antigen
of women during the twentieth century.159 Tobacco taken in (EBNA) have a nine-fold increased risk for MS com-
other forms, such as the chewing tobacco that is popular in pared to HLA-DRB1*1501-negative patients with low
Sweden, is not associated with an increase in the frequency anti-EBNA antibody titres.167 Additionally, the interaction
of MS, suggesting that risk is specific to the inhalational between high anti-EBNA titers and the presence of genetic
mode of administration of tobacco and not tobacco per risk factors (presence of HLA-DRB1*1501 or absence of
se.162 It has been postulated that cigarette smoke has direct HLA-A*02) has been shown to confer extra risk in addi-
cytotoxic effects in the lungs, leading to local immune tion to the risk conferred by each individual factor.168
reactions and sensitization-to-self antigens, which in turn It is becoming apparent that the interaction between
increases the likelihood of developing autoimmune dis- environmental factors and underlying genetic susceptibility
eases, including MS.159 Interaction of smoking with genetic may be as important as, or more important than the role
risk factors has been observed, such that smokers with two of individual genes in predisposing individuals to MS. It is
genetic risk factors (presence of HLA-DRB1*1501 and likely that multiple factors will be relevant in any individual
absence of the protective factor HLA-A*02) have an odds patient, and disentangling the relative contribution of each
ratio of 13.5 for developing MS, compared to nonsmokers factor will require well-designed, large epidemiological
without either genetic risk factor, suggesting that priming studies that incorporate genetic information into analyses.
the immune system in the lungs by cigarette smoking may There has been no investigation as of this writing of the con-
tip the balance towards development of MS in genetically tribution of gene–environment interactions to the clinical
susceptible people.163 However, a study comparing rates of variability of disease, and this may also be a fruitful line of
smoking and MS in MS patients and their healthy siblings investigation.
found no difference.164 This study did not take into account
HLA profiles of participants, which might be expected to
be more similar than in the general population, therefore C O N C LU S I O N S
further replication of these findings in other populations is
required in order to understand the interactions between Evidence for a polygenic etiology for MS is strong. HLA-
smoking and genetic profile in MS. DRB1*1501 undoubtedly carries the largest effect, con-
The clear association of MS with underlying autoim- ferring increased risk of MS and affecting the clinical
mune dysfunction has led to interest in possible infec- manifestations of MS, including age at onset, gender ratios,
tious triggers of disease, although the only infectious and accumulation of disability. More recently, other genes
agent with consistent findings of possible association with have been shown to be associated with MS, and there is early
MS is the Epstein-Barr virus (EBV).159 EBV seropositiv- evidence that these may also play a role in altering the dis-
ity is significantly higher in MS patients than in the gen- ease course, although the contribution of individual genes is
eral population, but there appears to be a relationship likely to be modest. Additionally, epistasis and gene–envi-
between seroconversion in the teenage years and increased ronment interactions are important in MS disease suscep-
risk of MS,165,166 contrasting with populations with a low tibility. We are only at the very early stages of unravelling
frequency of MS where EBV is acquired much earlier the complexities of the role of genetics in MS, but impor-
in life.159 It seems unlikely that EBV is a direct cause of tant advances have been made in recent years, particularly
MS, as viral loads are variable and not always increased after the publication of the Human Genome Project. A
in patients with newly diagnosed MS,159 and prospective deeper understanding of the mechanisms underlying the
studies have found a prolonged lag period of several years onset of MS and the subsequent accumulation of disability
between acquisition of EBV and development of MS.166 may allow development of new therapeutic targets with the
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 503
38. Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi 58. Bjartmar C, Kidd G, Mörk S, Rudick R, and Trapp BD, 2000.
M, et al., 2011. Diagnostic criteria for multiple sclerosis: 2010 revi- Neurological disability correlates with spinal cord axonal loss and
sions to the McDonald criteria. Ann Neurol, 69(2); 292–302. reduced N-acetyl aspartate in chronic multiple sclerosis patients.
39. Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ann Neurol, 48(6); 893–901.
Ebers GC, et al., 1983. New diagnostic criteria for multiple 59. Bitsch A, Schuchardt J, Bunkowski S, Kuhlmann T, and Brück W,
sclerosis: guidelines for research protocols. Ann Neurol, 13(3); 2000. Acute axonal injury in multiple sclerosis. Correlation with
227–231. demyelination and inflammation. Brain, 123(6); 1174–1183.
40. McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, 60. Lassmann H, 2011. Review: the architecture of inflammatory
Lublin FD, et al., 2001. Recommended diagnostic criteria for demyelinating lesions: implications for studies on pathogenesis.
multiple sclerosis: guidelines from the International Panel on the Neuropathol Appl Neurobiol, 37(7); 698–710.
Diagnosis of Multiple Sclerosis. Ann Neurol, 50(1); 121–127. 61. Frischer JM, Bramow S, Dal-Bianco A, Lucchinetti CF, Rauschka
41. Polman CH, Reingold SC, Edan G, Filippi M, Hartung HP, Kappos H, Schmidbauer M, et al., 2009. The relation between inflammation
L, et al., 2005. Diagnostic criteria for multiple sclerosis: 2005 revi- and neurodegeneration in multiple sclerosis brains. Brain, 132(5);
sions to the “McDonald Criteria.” Ann Neurol, 58(6); 840–846. 1175–1189.
42. Hirst C, Swingler R, Compston DAS, Ben-Shlomo Y, and
62. Lucchinetti C, Brück W, Parisi J, Scheithauer B, Rodriguez M,
Robertson NP, 2008. Survival and cause of death in multiple scle- and Lassmann H, 2000. Heterogeneity of multiple sclerosis
rosis: a prospective population-based study. J Neurol Neurosurg lesions: implications for the pathogenesis of demyelination. Ann
Psychiatry, 79(9); 1016–1021. Neurol, 47(6); 707–717.
43. Scalfari A, Neuhaus A, Degenhardt A, Rice GP, Muraro PA,
63. Lucchinetti C and Brück W, 2004. The pathology of primary pro-
Daumer M, et al., 2010. The natural history of multiple sclerosis: a gressive multiple sclerosis. Mult Scler, 10(Suppl 1); S23–S30.
geographically based study 10: relapses and long-term disability. 64. Brusaferri F and Candelise L, 2000. Steroids for multiple sclerosis
Brain, 133(7); 1914–1929. and optic neuritis: a meta-analysis of randomized controlled clinical
44. Kurtzke JF, 1983. Rating neurologic impairment in multiple sclero- trials. J Neurol, 247(6); 435–442.
sis: an expanded disability status scale (EDSS). Neurology, 33(11); 65. Zajicek JP, Hobart JC, Slade A, Barnes D, Mattison PG, and on
1444–1452. behalf of the MUSEC Research Group, 2012. MUltiple Sclerosis
45.. Roxburgh RHSR, Seaman SR, Masterman T, Hensiek AE, Sawcer and Extract of Cannabis: results of the MUSEC trial. J Neurol
SJ, Vukusic S, et al., 2005. Multiple Sclerosis Severity Score: using Neurosurg Psychiatry, 83(11); 1125–1132.
disability and disease duration to rate disease severity. Neurology, 66. PRISMS Study Group, 1998. Randomised double-blind
64(7); 1144–1151. placebo-controlled study of interferon beta-1a in relapsing/
46. Poser S, Ritter G, Bauer HJ, Grosse-Wilde H, Kuwert EK, and Raun remitting multiple sclerosis. PRISMS (Prevention of Relapses
NE, 1981. HLA-antigens and the prognosis of multiple sclerosis. and Disability by Interferon beta-1a Subcutaneously in Multiple
J Neurol, 225(3); 219–221. Sclerosis) Study Group. Lancet, 352(9139); 1498–1504.
47. Fischer JS, Rudick RA, Cutter GR, and Reingold SC, 1999. The 67. Johnson KP, Brooks BR, Cohen JA, Ford CC, Goldstein J, Lisak
Multiple Sclerosis Functional Composite Measure (MSFC): an RP, et al., 1995. Copolymer 1 reduces relapse rate and improves
integrated approach to MS clinical outcome assessment. Mult Scler, disability in relapsing-remitting multiple sclerosis: results of a
5(4); 244–250. phase III multicenter, double-blind placebo-controlled trial. The
48. Debouverie M, 2009. Gender as a prognostic factor and its impact Copolymer 1 Multiple Sclerosis Study Group. Neurology, 45(7);
on the incidence of multiple sclerosis in Lorraine, France. J Neurol 1268–1276.
Sci, 286(1–2); 14–17. 68. Polman CH, O’Connor PW, Havrdova E, Hutchinson M, Kappos
49. Simone IL, Carrara D, Tortorella C, Liguori M, Lepore V, Pellegrini L, Miller DH, et al., 2006. A randomized, placebo-controlled
F, et al., 2002. Course and prognosis in early-onset MS: comparison trial of natalizumab for relapsing multiple sclerosis. N Engl J Med,
with adult-onset forms. Neurology, 59(12); 1922–1928. 354(9); 899–910.
50. Renoux C, Vukusic S, Mikaeloff Y, Edan G, Clanet M, Dubois B, 69. Kappos L, Radue EW, O’Connor P, Polman C, Hohlfeld R,
et al., 2007. Natural history of multiple sclerosis with childhood Calabresi P, et al., 2010. A placebo-controlled trial of oral fingoli-
onset. N Engl J Med, 356(25); 2603–2613. mod in relapsing multiple sclerosis. N Engl J Med, 362(5); 387–401.
51. Qiu W, Wu JS, Castley A, James I, Joseph J, Christiansen FT, et al., 70. Comi G, Jeffery D, Kappos L, Montalban X, Boyko A, Rocca MA,
2010. Clinical profile and HLA-DRB1 genotype of late onset et al., 2012. Placebo-controlled trial of oral laquinimod for multiple
multiple sclerosis in Western Australia. J Clin Neurosci, 17(8); sclerosis. N Engl J Med, 366(11); 1000–1009.
1009–1013. 71. Gold R, Kappos L, Arnold DL, Bar-Or A, Giovannoni G, Selmaj
52. Confavreux C and Vukusic S, 2006. Age at disability milestones in K, et al., 2012. Placebo-controlled phase 3 study of oral BG-12 for
multiple sclerosis. Brain, 129(3); 595–605. relapsing multiple sclerosis. N Engl J Med, 367(12); 1098–1107.
53. Korn T, 2008. Pathophysiology of multiple sclerosis. J Neurol, 72. Rice GP, Incorvaia B, Munari L, Ebers G, Polman C, D’Amico R,
255(Suppl 6); 2–6. et al., 2001. Interferon in relapsing-remitting multiple sclerosis.
54. Boniface K, Blom B, Liu YJ, and de Waal Malefyt R, 2008. From Cochrane Database Syst Rev, (4); CD002002.
interleukin-23 to T-helper 17 cells: human T-helper cell differentia- 73. Mikol DD, Barkhof F, Chang P, Coyle PK, Jeffery DR, Schwid SR,
tion revisited. Immunol Rev, 226; 132–146. et al., 2008. Comparison of subcutaneous interferon beta-1a with
55. Lindquist S, Hassinger S, Lindquist JA, and Sailer M, 2011. The bal- glatiramer acetate in patients with relapsing multiple sclerosis (the
ance of pro-inflammatory and trophic factors in multiple sclerosis REbif vs Glatiramer Acetate in Relapsing MS Disease [REGARD]
patients: effects of acute relapse and immunomodulatory treatment. study): a multicentre, randomised, parallel, open-label trial. Lancet
Mult Scler, 17(7); 851–866. Neurol, 7(10); 903–914.
56. Lucchinetti C, Brück W, Parisi J, Scheithauer B, Rodriguez M, and 74. O’Connor P, Wolinsky JS, Confavreux C, Comi G, Kappos L,
Lassmann H, 1999. A quantitative analysis of oligodendrocytes Olsson TP, et al., 2011. Randomized trial of oral teriflunomide for
in multiple sclerosis lesions. A study of 113 cases. Brain, 122(12); relapsing multiple sclerosis. N Engl J Med, 365(14); 1293–1303.
2279–2295. 75. Wolinsky JS, Narayana PA, O’Connor P, Coyle PK, Ford C, Johnson
57. Mews I, Bergmann M, Bunkowski S, Gullotta F, and Brück W, 1998. K, et al., 2007. Glatiramer acetate in primary progressive multiple
Oligodendrocyte and axon pathology in clinically silent multiple sclerosis: results of a multinational, multicenter, double-blind,
sclerosis lesions. Mult Scler, 4(2); 55–62. placebo-controlled trial. Ann Neurol, 61(1); 14–24.
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 505
susceptibility, clinical course and progression of multiple sclerosis. subjects and the effect of soluble IL-2RA on immune responses.
J Neuroimmunol, 140(1-2); 188–193. J Immunol, 182(3); 1541–1547.
115. Burwick RM, Ramsay PP, Haines JL, Hauser SL, Oksenberg JR, 131. Dendrou CA, Plagnol V, Fung E, Yang JHM, Downes K, Cooper
Pericak-Vance MA, et al., 2006. APOE epsilon variation in mul- JD, et al., 2009. Cell-specific protein phenotypes for the autoim-
tiple sclerosis susceptibility and disease severity: some answers. mune locus IL2RA using a genotype-selectable human biore-
Neurology, 66(9); 1373–1383. source. Nat Genet, 41(9); 1011–1015.
116. Lundmark F, Duvefelt K, Iacobaeus E, Kockum I, Wallstrom E, 132. Robertson NP, Clayton D, Fraser M, Deans J, and Compston DA,
Khademi M, et al., 2007. Variation in interleukin 7 receptor alpha 1996. Clinical concordance in sibling pairs with multiple sclerosis.
chain (IL7R) influences risk of multiple sclerosis. Nat Genet, Neurology, 47(2); 347–352.
39(9); 1108–1113. 133. Brassat D, Azais-Vuillemin C, Yaouanq J, Semana G, Reboul J,
117. Gregory SG, Schmidt S, Seth P, Oksenberg JR, Hart J, Prokop A, Cournu I, et al., 1999. Familial factors influence disability in MS
et al., 2007. Interleukin 7 receptor alpha chain (IL7R) shows allelic multiplex families. French Multiple Sclerosis Genetics Group.
and functional association with multiple sclerosis. Nat Genet, Neurology, 52(8); 1632–1636.
39(9); 1083–1091. 134. Hensiek AE, Seaman SR, Barcellos LF, Oturai A, Eraksoi M,
118. Hoppenbrouwers IA, Aulchenko YS, Janssens AC, Ramagopalan Cocco E, et al., 2007. Familial effects on the clinical course of mul-
SV, Broer L, Kayser M, et al., 2009. Replication of CD58 and tiple sclerosis. Neurology, 68(5); 376–383.
CLEC16A as genome-wide significant risk genes for multiple scle- 135. Weinshenker BG, Santrach P, Bissonet AS, McDonnell SK,
rosis. J Hum Genet, 54(11); 676–680. Schaid D, Moore SB, et al., 1998. Major histocompatibility
119. De Jager PL, Jia X, Wang J, de Bakker PIW, Ottoboni L, Aggarwal complex class II alleles and the course and outcome of MS: a
NT, et al., 2009. Meta-analysis of genome scans and replication population-based study. Neurology, 51(3); 742–747.
identify CD6, IRF8 and TNFRSF1A as new multiple sclerosis 136. Celius EG, Harbo HF, Egeland T, Vartdal F, Vandvik B, and
susceptibility loci. Nat Genet, 41(7); 776–782. Spurkiand A, 2000. Sex and age at diagnosis are correlated with
120.. International MS Genetics Consortium (IMSGC), 2011. The the HLA-DR2, DQ6 haplotype in multiple sclerosis. J Neurol Sci,
genetic association of variants in CD6, TNFRSF1A and IRF8 to 178(2); 132–135.
multiple sclerosis: a multicenter case-control study. PLoS One, 137. Masterman T, Ligers A, Olsson T, Andersson M, Olerup O, and
6(4); e18813. Hillert J, 2000. HLA-DR15 is associated with lower age at onset in
121. Patsopoulos NA, Bayer Pharma MS Genetics Working Group multiple sclerosis. Ann Neurol, 48(2); 211–219.
Steering Committee of Studies Evaluating IFNβ-1band a 138. Hensiek AE, Sawcer SJ, Feakes R, Deans J, Mander A, Akesson
CCR1-Antagonist, Australia and New Zealand Genetics E, et al., 2002. HLA-DR 15 is associated with female sex and
(ANZgene) Consortium, GeneMSA, et al., 2011. Genome-wide younger age at diagnosis in multiple sclerosis. J Neurol Neurosurg
meta-analysis identifies novel multiple sclerosis susceptibility loci. Psychiatry, 72(2); 184–187.
Ann Neurol, 70(6); 897–912. 139. Smestad C, Brynedal B, Jonasdottir G, Lorentzen AR, Masterman
122. International Multiple Sclerosis Genetics Consortium (IMSGC), T, Akesson E, et al., 2007. The impact of HLA-A and -DRB1 on
2010. IL12A, MPHOSPH9/CDK2AP1 and RGS1 are novel mul- age at onset, disease course and severity in Scandinavian multiple
tiple sclerosis susceptibility loci. Genes Immun, 11(5); 397–405. sclerosis patients. Eur J Neurol, 14(8); 835–840.
123. Varade J, Comabella M, Ortiz MA, Arroyo R, Fernandez O, 140. Romero-Pinel L, Pujal JM, Martinez-Yelamos S, Gubieras L, Matas
Pinto-Medel MJ, et al., 2012. Replication study of 10 genes show- E, Bau L, et al., 2011. HLA-DRB1: genetic susceptibility and dis-
ing evidence for association with multiple sclerosis: validation of ability progression in a Spanish multiple sclerosis population. Eur J
TMEM39A, IL12B and CLBL genes. Mult Scler, 18(7); 959–965. Neurol, 18(2); 337–342.
124.. International Multiple Sclerosis Genetics Consortium 141. Jersild C, Fog T, Hansen GS, Thomsen M, Svejgaard A, and
(IMSGC), 2010. Comprehensive follow-up of the first Dupont B, 1973. Histocompatibility determinants in multiple
genome-wide association study of multiple sclerosis identifies sclerosis, with special reference to clinical course. Lancet, 2(7840);
KIF21B and TMEM39A as susceptibility loci. Hum Mol Genet, 1221–1225.
19(5); 953–962. 142. Pavelko KD, Drescher KM, McGavern DB, David CS, and
125. Ban M, Goris A, Lorentzen AR, Baker A, Mihalova T, Ingram G, Rodriguez M, 2000. HLA-DQ polymorphism influences progres-
et al., 2009. Replication analysis identifies TYK2 as a multiple scle- sion of demyelination and neurologic deficits in a viral model of
rosis susceptibility factor. Eur J Hum Genet, 17(10); 1309–1313. multiple sclerosis. Mol Cell Neurosci, 15(6); 495–509.
126. Aulchenko YS, Hoppenbrouwers IA, Ramagopalan SV, Broer L, 143. DeLuca GC, Ramagopalan SV, Herrera BM, Dyment DA,
Jafari N, Hillert J, et al., 2008. Genetic variation in the KIF1B locus Lincoln MR, Montpetit A, et al., 2007. An extremes of outcome
influences susceptibility to multiple sclerosis. Nat Genet, 40(12); strategy provides evidence that multiple sclerosis severity is deter-
1402–1403. mined by alleles at the HLA-DRB1 locus. Proc Natl Acad Sci U S
127. International Multiple Sclerosis Genetics Consortium (IMSGC), A, 104(52); 20896–20901.
Booth DR, Heard RN, Stewart GJ, Cox M, Scott RJ, et al., 2010. 144. Risch N and Zhang H, 1995. Extreme discordant sib pairs for
Lack of support for association between the KIF1B rs10492972[C] mapping quantitative trait loci in humans. Science, 268(5217);
variant and multiple sclerosis. Nat Genet, 42(6); 469–470; author 1584–1589.
reply, 470–471. 145. Brynedal B, Wojcik J, Esposito F, Debailleul V, Yaouanq J,
128. Gourraud PA and (IMSGC) IMSGC, 2011. When is the absence Martinelli-Boneschi F, et al., 2010. MGAT5 alters the severity of
of evidence, evidence of absence? Use of equivalence-based anal- multiple sclerosis. J Neuroimmunol, 220(1-2); 120–124.
yses in genetic epidemiology and a conclusion for the KIF1B 146. Watson CT, Ramagopalan SV, Morrison KM, Ebers GC, and
rs10492972*C allelic association in multiple sclerosis. Genet Breden F, 2010. IGHV4-39 deletion polymorphism does not asso-
Epidemiol, 35(6); 568–571. ciate with risk or outcome of multiple sclerosis. J Neuroimmunol,
129. Matesanz F, Gonzalez-Perez A, Lucas M, Sanna S, Gayan J, Urcelay 225(1-2); 164–166.
E, et al., 2012. Genome-wide association study of multiple sclerosis 147. Sombekke MH, Arteta D, van de Wiel MA, Crusius JBA, Tejedor
confirms a novel locus at 5p13.1. PLoS One, 7(5); e36140. D, Killestein J, et al., 2010. Analysis of multiple candidate genes
130. Maier LM, Anderson DE, Severson CA, Baecher-Allan C, Healy in association with phenotypes of multiple sclerosis. Mult Scler,
B, Liu DV, et al., 2009. Soluble IL-2RA levels in multiple sclerosis 16(6); 652–659.
G enetics a nd G enomics of N euro -P sychi atric D ise a ses , I I : Multiple S clerosis • 507
33.
GENETICS AND GENOMICS OF NEURO-PSYCHIATRIC
DISEASES, III: THE COMMON DEMENTIAS
Amy Gerrish, Rebecca Sims, and Julie Williams
508
calculated to be 5.3 per 100,000 individuals at risk (12.9% the first pathway, APP is cleaved by α-secretase, which cuts
of EOAD)[9]. within the Aβ domain. This precludes the generation of the
Aβ peptide and therefore the pathway involving α-secretase
cleavage of the APP protein is termed “non-amyloidogenic.”
G E N ET I C S O F E A R LY- O N S ET
In the alternative amyloidogenic pathway, APP is cleaved by
A L Z H E I M E R’ S D I S E A S E
β-secretase to generate soluble extracellular β-N-terminal
fragment (β-APPs) and the C-terminal APP fragment,
Many of the initial studies of the molecular genetics of AD
C99. Subsequent cleavage of C99 by γ-secretase results in
focused on EOAD in rare families where the disease is usu-
the formation of the Aβ peptide.
ally transmitted in an autosomal dominant fashion. The
Autosomal dominant mutations in the APP and pre-
identification of rare mutations in three genes that cause
senilin genes appear to cause AD through alterations in
EOAD has had a major impact on our understanding of the
APP processing. In APP, pathogenic mutations are found
pathogenesis of the disease.
in exons 16 and 17, clustered around the β- and γ-secretase
The first success in the search for AD genes was the
sites. Presenilin 1 and 2 (PSEN1 and PSEN2) have been
discovery of a positive genetic linkage between a locus on
implicated in γ-secretase function[21; 22]. Mutations in
chromosome 21q and autosomal dominant EOAD[10]. The
PSEN1 and PSEN2 are predominantly located in the
amyloid precursor protein (APP) gene was cloned and local-
highly conserved transmembrane domains and are pre-
ized to the long arm of chromosome 21, and a point mutation
sumed to distort the precise conformation of the molecule
was discovered in a family showing linkage on chromosome
within the membrane.
21[11]. Several pathogenic mutations have since been identi-
The discovery of the aforementioned mutations and
fied in autosomal dominant EOAD families, and there are
their mechanism of action led to the formulation of the
currently 32 APP mutations listed in the Alzheimer Disease
amyloid cascade hypothesis, which postulates that elevated
and Frontotemporal Dementia Mutation Database (http://
levels of Aβ result in the oligomerisation, fibrillogenesis,
www.molgen.ua.ac.be/ADMutations/default.cfm). These
and aggregation of the peptide. This is thought to initi-
act to increase production of Aß, specifically Aß42, the most
ate a cascade of injurious events, including direct neuronal
amyloidogenic form of the peptide[12; 13].
damage plus free radical production and glial activation.
It was noted that the APP mutations did not account for a
Soluble Aβ oligomers are themselves a neurotoxic species
substantial proportion of autosomal dominant EOAD cases,
prior to deposition in SPs[23], as they can lead to the genera-
meaning that additional pathogenic loci were still to be identi-
tion of hydroxyl ions and oxidative injury of phospholipid
fied. In 1992, an AD linkage study identified a candidate locus
membranes, and this loss of membrane integrity leads to
on the long arm of chromosome 14, and, a few years later, the
cell death[24]. Aβ may also exert toxicity through influ-
presenilin 1 (PSEN1) gene was isolated by a positional clon-
ences on calcium channels, exaggerating calcium influx and
ing[14; 15]. Similarly, evidence for a locus on chromosome 1q was
initiating cell death. The aggregation of Aβ in plaques is
reported in several related kindreds of German-Russian origin
also thought to trigger the inflammatory response and the
with multiple cases of autosomal dominant EOAD[16]. The
activation of microglia that surround extracellular lesions.
presenilin 2 (PSEN2) gene was subsequently identified based
These reactive changes may further compromise brain
on its homology to PSEN1[17; 18]. The AD & FTD Mutation
function.
Database lists 185 PSEN1 pathogenic mutations and 13
PSEN2 pathogenic mutations. The effect of these mutations
again seems to be the enhanced production of Aß42[19; 20]. It
G E N ET I C S O F L AT E - O N S ET
should be noted that around one-third of dominantly inher-
A L Z H E I M E R’ S D I S E A S E
ited AD cases are not associated with known mutations in
either the APP or PSEN genes, which implies the existence of
While much progress has been made in determining the
further disease loci[9].
causes of EOAD, the inheritance pattern of the more com-
mon, late-onset form of the disease appears to be much
APP PROCESSING AND THE AMYLOID more complex. Considerable evidence shows that familial
C A S C A D E H Y P OT H E S I S factors play an important role in the etiology of LOAD,
with the heritability estimated to be between 58% and
APP is a ubiquitously expressed trans-membrane protein 79%[25]. Numerous studies have investigated the role of
that is processed via one of two proteolytic pathways. In the early-onset genes in the more common late-onset form
521
and usually lifelong impacts on an individual’s capacity to disease are heritable (estimated heritability of 60–80% of
function in society. Schizophrenia and bipolar disorder are the liabilities). This study also found a significantly elevated
ranked twelfth and fourteenth, respectively, among the 20 risk of schizophrenia in first-degree relatives of probands
leading causes of disability worldwide.7 The burden of ill- with bipolar disorder and vice versa, which indicates that
ness associated with schizophrenia and bipolar disorder is schizophrenia and bipolar disorder partly share a common
substantial and may include unemployment, marital dys- genetic cause.11
function, suicide attempts, and increased health services
usage. Treatments are not uniformly effective, and the bio-
logical basis for the diseases is obscure, although several GENOMIC STUDIES IN
neurotransmitter, neuroendocrine, and second-messenger N E U R O -P SYC H I AT R I C G E N ET I C
systems appear to be involved. Thus, there is intense interest CONDITIONS
in determining the etiology of schizophrenia and bipolar
disorder, with the goal of developing more effective treat-
T H E P R E – G E N O M E -WI D E
ment and prevention strategies.
A S S O C I AT I O N S T U D I E S (GWA S ) E R A
In the past two decades, researchers have invested a large
T H E G E N ET I C E P I D E M I O L O GY O F amount of effort in finding risk gene(s) for psychiatric dis-
SCHIZOPHRENIA AND BIPOLAR orders. Thousands of linkage studies and candidate-gene
DISORDER association studies were performed to find the genetic
underpinnings of schizophrenia and bipolar disorder.
The pathophysiological mechanism underlying schizophre- Genome-wide linkage studies have identified multiple
nia and bipolar disorder is not well understood yet. Family, linked chromosomal regions, and association studies have
twin, and adoption epidemiological studies have unequiv- implicated many candidate genes, but the results have been
ocally demonstrated a strong contribution of inherited inconsistent.
genetic variation to the risk for both diseases. In the case
of schizophrenia, the combined data from a large number
G E N O M E -WI D E L I N K AG E S T U D I E S
of European family and twin studies showed that the risk
to first-degree relatives of schizophrenia patients is approxi- Linkage analysis is a powerful method of mapping the loca-
mately 12-fold higher than in the general population,8 and tion of disease-causing loci by identifying genetic markers
the estimated heritability of schizophrenia is between 70% that are co-inherited with a phenotype of interest (e.g.,
and 90%, based on meta-analyses of twin studies.9 In the disease status) in families. Evidence of genetic linkage is
case of bipolar disorder, previous family studies have found widely regarded as the most powerful statistical evidence
that first-degree relatives of bipolar disorder patients are at for the existence of an underlying genetic basis for a dis-
least 5 to 10 times more likely to develop the illness than ease. To that end, many genome-wide linkage studies have
relatives of control subjects, and the estimated heritabil- been conducted since the 1990s. These studies differ not
ity of bipolar disorder is between 70% and 90%, based on only with respect to ascertainment criteria, but also in their
twin studies.10 Schizophrenia and bipolar disorder are both marker density and method of analysis. Genome-wide sig-
considered to be among the most heritable major mental nificant evidence at various chromosomal regions has been
illnesses. reported, although none of those regions has been consis-
A more recent, landmark, large-scale epidemiological tently replicated across independent studies. Nevertheless,
study of nearly 2 million nuclear Swedish families clearly researchers have reported linkage signals in many partially
demonstrated that first-degree relatives of patients with overlapping regions. The lack of consistent significant find-
either schizophrenia or bipolar disorder are at increased ings may reflect insufficient power, differing designs, the
risk of these diseases, while half-siblings had a significantly presence of genetic heterogeneity, or the absence of a sus-
increased risk, which is nevertheless substantially lower than ceptibility locus in the region. Since most individual studies
that of the full-siblings.11 Heritability for schizophrenia and were conducted in small numbers of families, a clearer pic-
bipolar disorder was estimated at 64% and 59%, respec- ture may emerge following combined analysis of individual
tively. These estimates are lower than previous estimates studies. In the case of schizophrenia, the largest genome
(~80%) based on twin studies. Nonetheless, both types scan meta-analysis (GSMA) was carried out on 32 inde-
of studies and results show that schizophrenia and bipolar pendent genome-wide linkage scan analyses that included
CHROMOSOME EXAMPLE CLOSEST GENE OMIM OVERALLP ODDS MAF ANCESTRY REFERENCE
LOCUS VARIANT VALUE RATIO
1p21.3 rs1625579 MIR137 614304 1.6E-11 1.12 0.20 European 34
1q24.2 rs10489202 BRP44 614737 9.5E-9 1.19 0.14 Chinese 28
2p15.1 rs2312147 VRK2 602169 1.9E-9 1.09 0.39 European 30
2q32.1 rs1344706 ZNF804A 612282 2.5E-11 1.10 0.41 European 31
2q32.3 rs17662626 PCGEM1 605443 4.3E-11 1.18 0.09 European 34
6p21-p22.1 rs13194053 HIST1H2AH 615013 9.5E-9 1.21 0.18 European 25, 27
6p21-p22.1 rs2021722 TRIM26 600830 2.2E-12 1.15 0.22 European 34
6p21-p22.1 rs1635 NKAPL n/a 6.9E-12 1.28 0.33 Chinese 32
6p21-p22.1 rs6932590 PRSS16 607169 1.4E-12 1.16 0.22 European 29
8p12 rs16887244 LSM1 607281 1.3E-10 1.20 0.32 Chinese 28
8p23.2 rs10503253 CSMD1 608397 1.5E-8 1.16 0.19 European 34
8q21.3 rs7004633 MMP16 602262 2.8E-8 1.10 0.18 European 34
10q24.32 rs7914558 CNNM2 607803 1.8E-9 1.10 0.41 European 34
10q24.33 rs11191580 NT5C2 600417 1.1E-8 1.15 0.09 European 34
11p11.2 rs11038167 TSPAN18 n/a 1.1E-11 1.29 0.40 Chinese 32
11p11.2 rs11819869 AMBRA1 611359 3.9E-9 1.25 0.15 European 26
11q24.2 rs548181 STT3A 601134 2.9E-8 1.20 0.12 European 34
11q24.2 rs12807809 NRGN 602350 2.4E-9 1.15 0.17 European 29
18q21.2 rs9960767 TCF4 602228 4.1E-9 1.23 0.06 European 29
18q21.2 rs12966547 CCDC68 n/a 2.6E-10 1.09 0.42 European 34
18q21.2 rs4309482 TCF4 602228 7.8E-9 1.09 0.42 European 30
Abbreviations: OMIM: Online Mendelian Inheritance in Man database; MAF: minor allele frequency
regarding the top associated genomic regions; however, Most of the chromosomal loci identified by schizophre-
meta-analysis of some of the GWAS revealed 16 association nia and bipolar disorder GWAS have been reported in pre-
signals surpassing genome-wide significance level (Table vious linkage studies. Most of the GWAS-implicated genes
34.2)36–39,41,42—among them, only four loci (10q21, 11q14, (Tables 34.1 and 34.2) could be considered as positional or
12p13, and 12q13) were implicated by different studies. functional candidates according to gene locations or puta-
The largest study to date, a meta-analysis of ~750,000 tive functions of their encoded proteins; however, none of
markers on a combined sample of ~18,000 subjects of those genes has been tested in a previous candidate-gene
European and Asian ancestry, identified association signals association study, except one gene—CACNA1C, which
in four chromosome regions, three of which were novel was previously reported in a large-scale candidate-gene
(1p31.1, 2q11.2, and 3p22.2).38 On the basis of overall evi- association study.48 Although the risk conferred by each
dence from large-scale GWAS or meta-analyses, the most individual locus is small (odds ratios that are conferred by
notable findings were with genetic variants located in or each additional risk allele are no greater than 1.10–1.30), it
near four genes (CACNA1C, ANK3, ODZ4, and DHH). is still notable that a small impact on risk does not detract
CACNA1C, located on chromosome 12p13, encodes an from the most important aspect of GWAS findings—novel
alpha subunit of the L-type, voltage-gated calcium channel, associations, provided that they are replicable, shed light
which mediates a variety of calcium-dependent processes in on the etiology of schizophrenia and bipolar disorder. For
neurons.41,46 ANK3, located on chromosome 10q21, encodes example, the results of schizophrenia GWAS implicated
ankyrin G, which is an adaptor protein found at axonal ini- MIR137-mediated dysregulation as a new etiological mech-
tial segment and has been shown to regulate the assembly of anism in schizophrenia,34 and the results of bipolar disorder
voltage-gated sodium channels.41 ODZ4, located on chromo- GWAS indicate that ion channelopathies may be involved
some 11, encodes a member of a family of cell-surface pro- in the pathogenesis of bipolar disorder.41
teins, the teneurins, and is related to the Drosophila pair-rule
gene ten-m (ODZ). These genes are probably involved in
C O P Y N UM B E R VA R I AT I O N S
cell-surface signaling and neuronal pathfinding.36 DHH,
located on chromosome 12q13, encodes desert hedgehog. As in many traits and diseases, the risk conferred by each
The hedgehog gene family encodes signaling molecules that common risk variant is small (odds ratios <1.5), and these
play an important role in regulating morphogenesis.42 GWAS “hits” (surpassing genome-wide significance level)
Table 34.3 DETAILS OF GENE REGIONS IDENTIFIED IN COMBINED GWAS FOR SCHIZOPHRENIA AND
BIPOLAR DISORDER
CHROMOSOME EXAMPLE CLOSEST GENE OMIM OVERALL ODDS MAF ANCESTRY REFERENCE
LOCUS VARIANT P VALUE RATIO
2q32.1 rs1344706 ZNF804A 612282 9.9E-9 1.12 0.41 European 23
3p21.1 rs2239547 ITIH4 600564 7.8E-9 1.12 0.28 European 34
10q21.2 rs10994359 ANK3 600465 2.5E-8 1.22 0.06 European 34
12p13.33 rs4765905 CACNA1C 114205 7.0E-9 1.11 0.33 European 34
12p13.33 rs1024582 CACNA1C 114205 1.9E-8 1.07 0.34 European 60
Abbreviations: OMIM: Online Mendelian Inheritance in Man; MAF: minor allele frequency
INTRODUCTION I N T E L L E C T UA L D I S A B I L I T Y
Compared to many areas of medicine, relatively little is The definition of intellectual disability (ID) requires there
known about the underlying cellular mechanisms that to be significantly sub-average general intellectual function-
lead to learning and behavioral disorders in humans. The ing (Criterion A), which is accompanied by limitations in
reasons for this are many: social, political, medical, and adaptive functioning in at least two of the following skill
practical. The advent of the Human Genome Project and areas: communication, self-care, home living, social/inter-
developments in molecular genetics over the last 20 years personal skills, use of community resources, self-direction,
now mean that from a practical point of view, this area functional academic skills, work, leisure, health, and safety
can be considered in detail, which was not possible (Criterion B). The onset must occur before the age of
previously. 18 years (Criterion C). General intellectual functioning is
Why should one even consider trying to understand the defined by the intelligence quotient, IQ. “Adaptive func-
genetic basis of disorders of learning or behavior? The over- tioning” refers to how effectively individuals cope with com-
whelming reason is that families who are caring for children mon life demands. These are less objective measures and rely
with a severe disability wish to understand how and why on information gathered from independent sources, such as
this has happened to their child. If a mechanism of disease teacher evaluations and educational, developmental, and
can be identified, there is an opportunity to understand medical history; nevertheless, these data are extremely use-
and to clarify the cause of the problems. Knowledge of the ful. In the United Kingdom, the ICD-10 Classification of
disease can also lead to a greater understanding of the dis- Mental and Behavioural Disorders (WHO, Geneva 1992) is
ease and provide prognostic information and, potentially, used, whilst in the United States, the Diagnostic and
therapeutic interventions. In many cases, having a diagnosis Statistical Manual of Mental Disorders’ diagnostic classifica-
means that the lengthy search for the cause of the disease can tion is used, which is broadly similar to the WHO classifi-
now cease, and further unnecessary investigations can be cation[1,2]. IQ across the population is normally distributed
avoided. Also, the identification of a specific diagnosis may and is set at 100, and an IQ of less than 70 is classified as
mean that other medical systems such as cardiac or nutri- indicating intellectual impairment or disability.
tional status now need closer surveillance than would have
been done if the diagnosis were not made. This can prevent
further morbidity and avoid early mortality. Increasingly,
C AUS E S O F I N T E L L EC T UA L
as the genetic basis of these diseases is understood, there is
D I S A B I L IT Y: E P I D E M I O L O GY
more potential to provide therapeutic trials that will ame-
liorate the underlying biological defect. The underlying causes of ID are extremely heterogeneous,
This chapter focuses on the identification of novel as the developing brain is susceptible to a wide variety of
genes and abnormalities of these genes that cause intel- environmental and genetic insults[3,4]. In 1995, a consensus
lectual disability and autism. However, close analysis of a conference held under the auspices of the American College
few conditions with striking behavioral phenotypes gives of Medical Genetics undertook a literature review of nine
us astonishing insights into the workings of the human surveys of the intellectually disabled that were carried out
brain. between 1977 and 1994[5]. This review enabled a general
530
distribution of the causes of mental retardation to be deter- defects. Chromosomal abnormalities are collectively the
mined. Table 35.1 summarizes their findings, together single most frequent identifiable cause of mental retarda-
with a more recent survey of 429 individuals with mental tion (Table 35.1). The most common chromosomal cause
retardation carried out as part of the Australian Child and of ID is trisomy of chromosome 21, which underlies Down
Adolescent (ACAD) study[6]. Overall, intellectual disability syndrome. Other abnormalities associated with intellectual
(IQ <70] is a common condition that affects approximately disability include autosomal and sex chromosome aneu-
1–1.5% of the population[7]. Moderate to profound ID (IQ ploidies; partial chromosomal deletions and duplications;
<50) is estimated to affect 0.3–0.5% of the population. The translocations; insertions; inversions; ring chromosomes;
causes of ID divide into environmental and genetic fac- and uniparental disomies. Single-gene disorders, both as
tors, although environmental and genetic causes frequently recognized syndromes and known monogenic disorders,
coexist in an individual. Environmental exposures can be collectively include a significant proportion of causes of ID
subdivided into prenatal, perinatal, and postnatal expo- [7–21%).
sure. The long-term consequences of extreme prematurity A striking statistic from Table 35.1 is that, in many cases
and the associated medical complications account for an (up to 40%), the underlying etiology remains unknown, but
increasingly significant proportion of children with ID, with improved technology and understanding this figure is
while the proportion of children suffering from postnatal gradually reducing.
infections is diminishing[4,8]. Perinatal injury due to birth The relative excess of males in the population with
problems remains common, as does the prenatal exposure severe ID (1.3–1.7:1) suggested that X-linked disease genes
to teratogens during pregnancy. These include fetal expo- are significant contributors to the overall genetic etiol-
sure to sodium valproate and other anticoagulants, alco- ogy[4,11]. Studies using modern cytogenetic techniques, to
hol, and high levels of blood glucose in diabetic mothers. exclude chromosome abnormalities, suggest that where
A small proportion of children also suffer from accidents two male siblings are severely retarded, in the absence of
or infections beyond the postnatal period that result in a chromosome abnormality, the likelihood of this being
intellectual disability. The genetic contribution to ID is due to an X-linked gene abnormality is as high as 80%[12].
high, as empirical recurrence risks for siblings of severely However, the contribution of X-linked disease genes to the
affected individuals are 5–8% if a single case is observed, male excess of cases with ID is probably overestimated, as
and 12–15% if two siblings are affected[9,10]. These figures surveys of singleton male cases shows a ten-fold reduction
include both chromosomal abnormalities and single-gene in expected monogenic cases[13].
Although a diagnosis (genetic or environmental) can-
Table 35.1 COMPARISON OF THE CAUSES OF not be reached in about 40% of patients with an IQ less
INTELLECTUAL DISABILITY REPORTED BY THE than 50, this figure is declining as genome-sequence analy-
CONSENSUS CONFERENCE (1997) AND THE ACAD sis is improving. It is likely that there will be a significant
STUDY (2000) (ADAPTED FROM PARTINGTON ET AL., increase in our ability to attribute causation in the near
2000) future for the severe cases. However, approximately 76%
CATEGORY/GROUP CONSENSUS ACAD
of cases with mild mental retardation (IQ 50–70) remain
CONFERENCE (%) STUDY (%) undefined, largely due to the multifactorial causation and
Chromosomal abnormalities 4–28 21 limits in our knowledge of the contribution of genes to the
Down syndrome 12.9–16.1 15 normal distribution of IQ in an assumed polygenic model
Recognizable syndromes 3–7 2 of disability[4].
Provisional unique syndromes 1–5 3
Known monogenic conditions 3–9 7 G E N ET I C C AUS E S
Structural CNS abnormalities 7–17 6 O F I N T E L L EC T UA L D I S A B I L I T Y
Complications of prematurity 2–10 8
Many chromosome abnormalities are detected by the
Environmental/teratogenic 5–13 8 visual inspection of the whole genome at the resolution
Metabolic/endocrine 1–5 0 of approximately 5–10 Mb. This will detect large gains or
Unknown 30–50 46 losses of chromosome material and rearrangements, which
Male/female ratio (all categories) 1.35–1.4 1.38 are almost always of clinical significance. Although the
Male/female ratio (unknown – 1.77 technique of karyotyping is ultimately limited by the reso-
diagnosis subgroup) lution of the microscope, the quality of the chromosome
541
Inherited Although highly penetrant genes are the ones most
5– likely to be clinically relevant, they only explain a small
10% Familial portion of the hereditability of cancer. As cancer is rela-
15–20%
tively common in the general population, it is possible
to have a chance clustering of the same or related can-
cers within a family (Figure 36.1). These familial cluster-
ings are most likely due to low-penetrance alleles with
risks that do not clearly exceed an action threshold.
70–80% Genome-wide association studies (GWAS) have further
Sporadic
identified large numbers of genetic variants that are,
individually, associated with very limited increases in
risk, and the clinical utility of identifying these variants
is completely unclear.6
Figure 36.1
Estimated frequency distribution of heritable, “familial,” For the reader’s reference, a table is included with the
and sporadic cancers. The majority of common cancers are believed more commonly encountered inherited cancer syndromes
to be sporadic; 5–10% are heritable and arise due to highly penetrant (Table 36.1). From this longer list, we have chosen four
germline mutations in Mendelian genes. An additional 10–15% are
referred to as “familial” and may be caused by low-penetrance genes,
syndromes to highlight how understanding the underlying
gene–environment interactions, or both. genetic cause of a disease allows for gene-enabled manage-
ment. We discuss in detail hereditary breast and ovarian
cancer syndrome (HBOC) and hereditary non-polyposis
limited resection in patients with colorectal cancer who
colorectal cancer syndrome (HNPCC), as targeted by
have Lynch syndrome. Germline changes associated with
Healthy People 2020, to illustrate the clinical utility of
susceptibility have also recently been predictive of differ-
germline genetic testing and risk-reduction strategies. We
ential response to treatment approaches. Recent success
highlight Cowden syndrome (CS), a condition associated
of inhibitors of poly(ADP-ribose) polymerase inhibitors
with a spectrum of protean malignant and benign clinical
in BRCA-mutated breast, ovarian, and pancreatic cancers
features, to illustrate the clinical diagnostic difficulty in
indicates a possible role for this in the future,4 as does the
identifying patients at risk. Genotype-specific recommen-
suggestion of sensitivity to cis-platinum in the neoadjuvant
dations are available for component neoplasias where there
treatment of BRCA1-mutated breast cancer.5 Perhaps the
are clear genotype–phenotype correlations, and we high-
most important application of genetic testing is in individ-
light multiple endocrine neoplasia 2 (MEN 2) to illustrate
ual risk assessment for developing cancer, allowing for suc-
how genotype–phenotype correlation fine-tunes clinical
cessful risk-reduction strategies for the at-risk patient whilst
management.
avoiding unwarranted surveillance in unaffected family
members.
H E R E D I TA RY B R E A S T C A N C E R
Sporadic Inherited
Increasing numbers of genes have been implicated in the
development of breast cancer. Although 5–10% of breast
cancers are believed to be caused by mutations in single
genes, including such mutations predispose carriers to a
very high lifetime risk of developing breast, ovarian and
other cancers. In addition to BRCA1 and BRCA 2, heredi-
tary cancer syndromes that include breast cancer risk are
caused by germline mutations in PTEN, TP53, STK11,
• Later ages of onset (60s–70s) • Early ages of onset (under 50s) CDH1, NF1, and SDHB/D, and germline promoter hyper-
• Little or no family history of • Multiple generations with methylation of KLLN (Figure 36.3). Importantly, for many
cancer cancer
• Single or unilateral cancers • Multiple or bilateral cancers of these genes, women who carry these germline mutations
• Clustering of certain cancers (e.g. are at a higher risk of developing breast cancer at a young
breast and ovarian)
age than the recommended age of mammogram screening
Figure 36.2 Clinical “red flags” suggestive of heritable cancers. for the general population.
5 4 2 • G e no m ic s in C l inic a l P r actic e
Table 36.1 HEREDITARY CANCER SYNDROMES ACCORDING TO ORGAN SYSTEMS
(continued)
Table 36.1 CONTINUED
(continued)
5 4 4 • G e no m ic s in C l inic a l P r actic e
Table 36.1 CONTINUED
5 4 6 • G e no m ic s in C l inic a l P r actic e
BRCA1, 20%
Unknown, 53–54%
Sporadic Familial
Hereditary
BRCA2, 20%
Relative contribution of predisposition genes to heritable breast cancer. The majority of cases of breast cancer are sporadic, with BRCA1
Figure 36.3
and BRCA2 germline mutations accounting for the majority of hereditary breast cancers. Other highly penetrant genes contributing to heritable
breast cancer syndromes together probably make up 5% of all hereditary and familial cases.
Hereditary Breast and Ovarian Cancer Syndrome than 36 years and in 4% of women diagnosed at ages from 36
to 45.28 Personal and family history characteristics suggestive
Hereditary breast–ovarian cancer (HBOC, MIM 113705) of HBOC are listed in Figure 36.4.
syndrome is the most common form of heritable breast can- The proportion of HBOC attributable to BRCA1
cer and is caused by germline mutations in BRCA1 on 17q11 or BRCA2 (BRCA1/2) mutations is poorly defined, and
and BRCA2 on 13q12-q13.7,8 The prevalence of BRCA1/2 estimates depend upon the population studied, the num-
mutations in the general population is estimated to be ber of breast and ovarian cancer cases in the family, and
between 1 in 500 and 1 in 1000,9,10 although this is much the mutation-detection technique utilized. In families
higher in certain founder populations, such as the Ashkenazi with two or more cases of breast cancer (<50 years) and at
Jewish and Icelandic populations (Table 36.2).11–27 In a least one case of ovarian cancer, more than 90% will carry
population-based study, BRCA1/2 mutations were found in a deleterious germline BRCA1/2 mutation. In families
6% of women diagnosed with breast cancer at an age of less with site-specific breast cancer, these figures are lower and
C l inic a l C a nc e r G e no m ic s • 5 4 7
Personal History
• Breast Cancer Diagnosed at less than 40 years of age
• Bilateral breast cancer, especially when first cancer is
diagnosed at less than 50 years of age
• History of breast and ovarian cancer
Family History
• Two or more family members diagnosed with breast cancer
at 50 years of age or younger
• Both breast and ovarian cancer in the family
• One or more family members diagnosed with breast cancer
at 50 years of age or younger and of Ashkenazi Jewish
ancestry
• One or more family members diagnosed with ovarian cancer
and of Ashkenazi Jewish ancestry
• Male breast cancer
Figure 36.4 Personal and family history characteristics suggestive of inherited breast and ovarian cancer syndrome.
range from 30% to 81%.29 In addition, a significant pro- p53,37 RB1,38 and MYC.39 The C-terminus contains a tran-
portion of young breast cancer cases unselected for family scription activation domain and interacts with RNA poly-
history will carry a BRCA1/2 mutation. This proportion merase II40 and BRCA2,41 as well as other proteins. BRCA1
is as high as 20% in women of Ashkenazi Jewish ances- has been shown to play a role in a multitude of cellular pro-
try diagnosed before the age of 40.14,30 The cancer risks in cesses, including, but not limited to, DNA transcription,
BRCA1-mutation-positive individuals are confined largely cell cycle regulation, DNA damage repair, and apoptosis.42
to the breast and ovaries31 (Figure 36.5), while pancreatic It acts as a tumor suppressor, which is supported by the fact
and prostate cancer, melanoma, and other cancer risks are that many BRCA1-associated tumors show loss of heterozy-
increased in BRCA-mutation-positive individuals.32 Male gosity (LOH) of the wild-type allele.43,44 The BRCA2 pro-
BRCA1/2-mutation-positive individuals have an increased tein has also been implicated in DNA repair and has been
risk of developing prostate cancer (16% lifetime risk by age shown to regulate the activity of RAD51, which is required
70), and male BRCA2-mutation-positive individuals are at for homologous recombination leading to double-stranded
risk for breast cancer (5–10% lifetime risk).12 DNA repair.45 Interestingly, biallelic inactivating mutations
The mutations in BRCA1/2 causing HBOC are inher- in BRCA2 were identified in patients with Fanconi anemia
ited in an autosomal dominant fashion and are highly pen- group D1 (FANC-D1)46, linking two seemingly unrelated
etrant. Mutations in BRCA1 typically confer a lifetime risk syndromes to a common DNA damage response pathway.
between 50% and 85% of developing female breast cancer Evidence for genotype–phenotype correlations exists.35,47,48
and a 15% to 40% lifetime risk of ovarian cancer.33 BRCA2 Mutations in and around exon 11 of BRCA2, which is often
germline mutations can lead to a lifetime breast cancer risk referred to as the “ovarian cancer cluster region” (OCCR),
between 6% and 7.5% for men, 50% and 85% for women, confer a higher ovarian cancer risk and lower breast cancer
and approximately 14% and 27% for ovarian cancer.32,34,35 risk than mutations in other areas of the gene.24,49
BRCA1 encodes a 1863-amino-acid polypeptide.7
The N-terminus contains a zinc-finger domain that inter-
Identifying Patients for Genetics Risk Assessment
acts both directly and indirectly with DNA.7 Exon 11 of
BRCA1 encodes over 60% of the protein, contains two Individuals with a greater than 5–10% probability of having
nuclear localization signals, and interacts with RAD51,36 a BRCA1/2 mutation are considered candidates for genetics
5 4 8 • G e no m ic s in C l inic a l P r actic e
Brain
General population <1%
HNPCC 1–3%
Melanoma
General population 2%
Thyroid PHTS 6%
General population 1%
PHTS 35%
Breast Stomach
General population 12% General population <1%
HBOC-BRCA1 50–85% HNPCC 11–19%
HBOC-BRCA2 50–85%
PHTS ~85%
Renal Cell
General population 1.6%
Hepatobiliary Tract PHTS 34%
General population <1%
HNPCC 2–7% Small Bowel
General population <1%
Pancreatic HNPCC 1–4%
General population <1%
HBOC-BRCA1 2.3% Ovarian
HBOC-BRCA2 2.3% General population 1.6%
HNPCC 4% HBOC-BRCA1 15–40%
Colon HBOC-BRCA2 14–25%
General population 5.5% HNPCC 3–13%
HNPCC 60–80%
Endometrial
PHTS 9%
General population 2.6%
Urinary Tract HNPCC 20–60%
General population <1% PHTS 28%
HNPCC 4–5%
Figure 36.5
Lifetime risks for cancer in patients with HBOC, HNPCC, and PHTS. Lifetime cancer risks for each hereditary cancer syndrome are
shown together with organ-specific lifetime risk for cancer in the general population.
referral for risk assessment and consideration of gene testing recommendations of a National Institutes of Health (NIH)
in the context of genetic counseling. Multiple models are consensus panel and other recommendations for screening
available to aid in determining good candidates for genetic women who test positive for a BRCA1 or BRCA2 mutation.
testing (Table 36.3).50 As mentioned above, most germline Ovarian cancer screening measures available (transvaginal
BRCA1/2 mutations are caused by point mutations or small ultrasound examination and serum CA-125 concentration)
deletions or insertions. However, large genomic rearrange- have limited sensitivity and specificity and have not been
ments appear to be the source of a significant minority of shown to reduce mortality.52 Thus, the standard of care is
BRCA1/2 mutations,51 and these should be considered if prophylactic bilateral salpingo-oophorectomy, with or
testing by full PCR-based Sanger sequencing is negative on without total hysterectomy.
an individual basis.
Risk-Reduction Strategies
Cancer Surveillance Recommendations
Several risk-reduction strategies have been proposed to
Enhanced surveillance is recommended to promote early help patients lower their risk of developing breast or ovar-
detection of breast cancer in women with a known or ian cancer. In the Breast Cancer Prevention Trial, the use of
suspected genetic mutation. Table 36.4 summarizes the tamoxifen for primary prevention of breast cancer reduced
C l inic a l C a nc e r G e no m ic s • 5 4 9
Table 36.3 PROBABILITY MODELS TO ESTIMATE THE LIKELIHOOD OF A BRCA MUTATION
the risk of invasive breast cancer by 49%.53 The women in 400 women who were mutation positive.58 Prophylactic
this trial were not known mutation carriers but rather were removal of the ovaries is presented as an option to women
determined to be at increased risk using the Gail model, a who are BRCA1/2-mutation-positive, supported by mul-
model devised to assess the five-year lifetime risk for breast tiple prospective studies confirming that such surger-
cancer by taking into account a woman’s personal and fam- ies not only decrease the incidence of subsequent breast
ily histories (Table 36.5 compares the two commonly used and ovarian cancer, but also find occult early-ovarian
breast cancer risk-assessment models). Tamoxifen has sub- neoplasms.59,60 Individuals at high a priori risk of being
sequently been shown to reduce breast cancer incidence in BRCA1/2-mutation-positive (Figure 36.4) should strongly
BRCA1/2 mutation carriers.43,54–56 Before starting women consider genetic testing for presence of BRCA1/2 mutation
on tamoxifen, its risks, including the increased risk of prior to surgery for their primary cancer diagnosis.
thrombosis, and its benefits should be discussed.
Prophylactic mastectomy and oophorectomy can pre-
C OWD E N S Y N D RO M E
vent cancers in mutation carriers, but surgical menopause
may have other deleterious effects, and mastectomy in Cowden syndrome (CS; [Mendelian Inheritance in Man]
particular is not acceptable to some women. A retrospec- MIM 158350) is an autosomal dominant disorder and part
tive study on a cohort of 639 women with a strong fam- of the PTEN hamartoma tumor syndrome (PHTS), which
ily history of breast cancer demonstrated that prophylactic also includes subsets of Bannayan-Riley-Ruvalcaba syn-
mastectomy reduced the risk of breast cancer by at least drome (BRRS), Proteus syndrome (PS), and Proteus-like
90%51 and increased life expectancy by three to five years.57 syndrome.61,62 PHTS is molecularly defined as having a
Similar results were noted in a prospective study of over germline PTEN mutation irrespective of clinical syndrome.
5 5 0 • G e no m ic s in C l inic a l P r actic e
Table 36.5 PROS AND CONS FOR GAIL AND CLAUS MODELS IN BREAST CANCER RISK ASSESSMENT
CS is a multiple hamartoma syndrome with a high risk of cancer) by age 20, and by age 30 nearly 100% of muta-
benign and malignant tumors of the thyroid, breast, endo- tion carriers are believed to have developed at least some
metrium, and other organs. BRRS is a congenital disorder of the mucocutanous signs. Affected individuals also
characterized by macrocephaly, intestinal polyposis, lipo- have an increased risk for several malignancies, including
mas, and pigmented macules of the glans penis. PS is a female breast cancer (85% lifetime risk), thyroid cancer
complex, highly variable disorder involving congenital mal- with an over-representation of follicular thyroid cancer
formations and overgrowth of multiple tissues. Proteus-like and endometrial cancer (Figure 36.6).66–68 Consensus diag-
syndrome is undefined but refers to individuals with signifi- nostic criteria for CS were first developed in 1996 by the
cant clinical features of PS who do not meet the diagnostic International Cowden Consortium (ICC), and they form
criteria for PS. the basis for the National Comprehensive Cancer Network
The estimated incidence of CS is approximately one in Guidelines (Table 36.7).65 Approximately 25% of individ-
200,000,63,64 but as many of the features of CS are common uals who meet the strict diagnostic criteria for CS have a
in the general population (e.g., fibrocystic breast disease, pathogenic PTEN mutation, including large deletions.66
uterine fibroids), the incidence may be even higher. The The CS susceptibility locus was mapped to 10q22-q23.63
most commonly reported manifestations of CS include Subsequently, germline mutations in PTEN, localized to
macrocephaly, mucocutanous lesions, fibrocystic breast dis- 10q23.3, were found in CS.69 PTEN spans nine exons,
ease, thyroid abnormalities, multiple uterine leiomyomata, with a transcript length of 3,417 bps, encoding for a 403
and gastrointestinal polyps (Table 36.6).65 amino acid protein (Figure 36.6). The C-terminal domain
CS is a highly penetrant genetic disorder. More than contains important subdomains that are common to other
90% of individuals with PTEN mutations are believed to signal-transducing molecules. First, PTEN contains a C2
manifest some feature of the syndrome (although rarely domain, which is associated with phospholipid-binding
C l inic a l C a nc e r G e no m ic s • 5 5 1
15 185 351 403
Domain Phosphatase domain C2 domain PDZ
PEST
Figure 36.6
Functional domains of PTEN. The domain structure of phosphatase and tensin homologue mutated on chromosome ten (PTEN). PTEN
is a 403-amino acid protein that comprises a phosphatase domain, a C2 domain, and a PDZ-binding domain. PDZ domains are significant regions
for protein–protein interactions that play a vital role in cellular signal transduction. The N-terminal domain contains the phosphatase domain (the
enzymatic activity of PTEN), and it is therefore not surprising that the majority of PTEN mutations occur within this domain.
regions, including phospholipid membranes. Additionally, PEST sequences, in addition to several phosphorylation
the C terminus features a PDZ-binding motif, which inter- sites located in the last 50 amino acids, which both appear
acts strongly with the phosphatase domain. PDZ domains to play important roles in PTEN protein stability. The
are significant regions for protein–protein interactions that N-terminal domain contains the phosphatase domain (the
play a vital role in cellular signal transduction.70 Removal enzymatic activity of PTEN) and it is, therefore, not sur-
of the PDZ domain reduces the ability of PTEN to inhibit prising that the majority of PTEN mutations occur within
one of its substrates, AKT. The C terminus also contains this domain.71,72
Recent research efforts to identify other susceptibility
genes in CS have resulted in six other germline susceptibil-
Table 36.7 INTERNATIONAL COWDEN SYNDROME ity genes. Approximately 10% of individuals with classic
CONSORTIUM OPERATIONAL CRITERIA FOR THE CS or CS-like (CSL) phenotypes carry germline variants
DIAGNOSIS OF COWDEN SYNDROME (VER. 2000) in the genes encoding three of the four subunits of suc-
cinate dehydrogenase (SDH) or mitochondrial complex
Pathognomonic criteria (mucocutaneous lesions):
Trichilemmomas, facial II.73 Single-exon KLLN, on 10q23, encodes KILLIN, and
Acral keratoses shares a bidirectional promoter with PTEN. Up to 30% of
Papillomatous papules individuals with CS/CSL phenotypes, without germline
Mucosal lesions
PTEN or SDHx mutations, were recently found to carry
Major criteria:
Breast carcinoma germline KLLN promoter hypermethylation.74 Another 9%
Thyroid carcinoma (non-medullary), especially follicular thyroid of unrelated CS individuals without germline PTEN muta-
carcinoma tions were found to have germline PIK3CA mutations, and
Macrocephaly (megalencephaly, e.g., ≥97th percentile)
Lhermitte–Duclos disease (LDD) 2% harbored germline AKT1 mutations.75 Functionally,
Endometrial carcinoma all mutation-positive protein lysates showed a significant
Minor criteria: increase in P-Thr308-AKT1 levels, further supporting their
Other thyroid lesions (e.g., adenoma or multinodular goiter) potential role as novel CS-susceptibility genes.75
Mental retardation (e.g., IQ ≤75)
Gastrointestinal hamartomas
Fibrocystic disease of the breast
Lipomas Identifying Patients for Genetics Risk Assessment
Fibromas
GU tumors (e.g., renal cell carcinoma or uterine fibroids) or In part because of phenotypical heterogeneity, and in part
malformation because of the high frequency of de novo PTEN germline
Operational diagnosis in an individual: mutations,76 a family history of associated cancers may not
1. Mucocutaneous lesions alone if (a) there are six or more facial be apparent. Additionally, many of the benign features of
papules, of which three or more must be trichilemmoma;
(b) cutaneous facial papules and oral mucosal papillomatosis; CS are common in the general population, making the
(c) oral mucosal papillomatosis and acral keratosis; (d) palmo- diagnosis of CS a challenge for most clinicians. To aid clini-
plantar keratosis, six or more cal diagnosis, a clinical predictor (Cleveland Clinic PTEN
2. Two major criteria, of which one must include macrocephaly or
LDD Risk Calculator) was developed based on clinical features
3. One major and three minor criteria derived from a prospective study of over 3000 patients
4. Four minor criteria suspected of having CS.66 The questionnaire-based clini-
Operational diagnosis in a family where one individual is diagnos- cal decision tool is available online (http://www.lerner.ccf.
tic for Cowden:
1. One or more pathognomonic criterion/ia org/gmi/ccscore/) to assist clinicians at the point of patient
2. Any one major criterion with or without minor criteria care. Based on a patient’s presentation, a score will be
3. Two minor criteria derived that corresponds to an estimated risk for germline
5 5 2 • G e no m ic s in C l inic a l P r actic e
PTEN mutation (Figure 36.7) to guide clinicians for refer- with a PTEN mutation is increased cancer surveillance to
ral to genetics professionals. detect any tumors at the earliest, most treatable stages.
Based on new information on lifetime risk estimates
of CS component cancers,68 expert opinion currently rec-
Surveillance and Management of Cowden Syndrome
ommends an annual comprehensive physical examination
The mucocutaneous manifestations of Cowden syndrome with special attention paid to skin and thyroid (all ages),
are rarely life-threatening. If the patient is asymptomatic, and breast (adults) as well as annual formal dermatologi-
observation alone is prudent. When symptomatic, topical cal examination, starting at 18 years. Screening for various
agents (e.g., 5-fluorouracil), curettage, cryosurgery, or laser component cancers may start at the ages listed in Table 36.6,
ablation may provide only temporary relief.77 Treatment or should start 5–10 years before the youngest diagnosis of
for the benign and malignant manifestations of PHTS is that particular cancer type in the family, whichever is earlier.
the same as for their sporadic counterparts. Some women
at increased risk for breast cancer consider prophylactic
H E R E D I TA RY C O L O R EC TA L C A N C E R
mastectomy, especially if breast tissue is dense or if repeated
breast biopsies have been necessary. The recommendation Most (70–80%) colorectal cancers (CRC) are sporadic.
of prophylactic mastectomy is a generalization for women Around 20% to 30% of cases have a familial component;
at increased risk for breast cancer from a variety of causes, that is, there are one or more affected first- or second-degree
not just from PHTS, and is best managed by breast sur- relatives (or both) and a potentially definable genetic
geons with a specialty interest in high-risk breast cancer basis.78,79 In the United States, around 10% to 15% of adults
patients. have a history of CRC in a first-degree relative.80 Molecular
The most serious consequences of PHTS relate to the genetics has made a significant impact by identifying germ-
increased risk of cancers, including breast, thyroid, endome- line and somatic mutations associated with the develop-
trial, and (to a lesser extent) renal cancers. In this regard, the ment of CRC.81 Single-gene germline mutations conferring
most important aspect of management of any individual a hereditary susceptibility (high risk) to CRC account for
Total Score
0 5 10 15 20 25 30 35 40
Risk for
PTEN mutations 0.005 0.01 0.02 0.03 0.05 0.1 0.15 0.30 0.6 0.8 0.95 0.99
Cleveland Clinic (CC) PTEN Risk Calculator: three illustrative cases. The CC score (Total Score) is first derived by a sum of the weights
Figure 36.7
of positive clinical features. To illustrate, three hypothetical cases are presented, each corresponding to a CC score of 5 points, 10 points, and 15
points, respectively. Case 1 may present with breast cancer at age 55 (1 point), with background of thyroid cancer at age 44 (4 points), with a
final score of 5 and corresponding prior probability <1% of having a PTEN mutation. Case 2 may present with breast cancer at age 38 (4 points)
and concurrent macrocephaly (6 points), with a final score of 10 and corresponding prior probability of 3%. Case 3 may present with a single
hamartomatous gastrointestinal polyp (10 points) found on endoscopy, Hashimoto’s thyroiditis (4 points), and lipomas (1 point), for a final score of
15 and corresponding prior probability of 10%. Adapted from Tan et al, American Journal of Human Genetics 2011.66
C l inic a l C a nc e r G e no m ic s • 5 5 3
round 6% to 7% of all CRCs, the most common of which for 63% of all disease-causing mutations in families with
is hereditary non-polyposis colorectal cancer (Table 36.1).79 HNPCC.95 Other founder mutations include those seen
in the Swiss, the Chinese, and Ashkenazi Jews.96–98 Notably,
genomic rearrangements accounted for 24% of all muta-
Hereditary Non-Polyposis Colorectal
tions in a series of 59 clinically selected North American
Cancer Syndrome
HNPCC families.99 Testing for large genomic rearrange-
Hereditary non-polyposis colorectal cancer (HNPCC, ments and germline mutations that affect promoter regions
MIM 114500) was described as a clinical entity, character- or deep intronic regulatory regions may be missed by the
ized by a strong family history of early-onset colon cancer use of conventional methods alone but may be warranted,
and associated cancers without numerous polyps, long especially for families for whom tumor immunohistochem-
before causative genes were identified.82 HNPCC is the istry testing suggests a germline mutation but PCR-based
most common form of hereditary colorectal cancer, affect- Sanger sequencing of germline DNA is negative.100–102
ing as many as one in 400 individuals, and it is caused by a
germline mutation in any one of five DNA mismatch-repair
Identifying Patients for Genetics Risk Assessment
(MMR) genes (MLH1, MSH2, MSH6, PMS1, and
PMS2). To honor the physician who described the first The International Collaborative Group on HNPCC estab-
families with HNPCC, Henry Lynch, individuals/families lished diagnostic criteria, known as “the Amsterdam cri-
with HNPCC who are found to have a germline MMR teria,” in 1991 (Table 36.8).103 These criteria were further
gene mutation are designated as having Lynch syndrome.83 modified in 1999 (Amsterdam II Criteria) in an attempt
Molecular screening of all patients with colorectal can- to incorporate extracolonic cancers.104 However, stud-
cer has shown that about 3% of these cancers are attribut- ies have shown that only 40–80% of families that meet
able to germline MMR gene alterations.84–87 Penetrance in the original criteria, and no more than 50% meeting the
HNPCC is typically high, and the phenotype is quite vari- modified criteria, have germline mutations in the associated
able within families. Males with HNPCC have virtually a mismatch-repair genes.105 For this reason, a third set of crite-
100% chance of developing colorectal cancer by age 70.88 ria, the Bethesda criteria, were developed. When compared
Women appear to have a higher lifetime risk of endometrial to the Amsterdam I and II criteria, the Bethesda criteria are
cancer (42–60%) than colorectal cancer (54%).88–91 The the most sensitive, especially in identifying HNPCC fami-
lifetime risk for stomach cancer is 13–19%, although this lies with pathogenic mutations.106 The modified Bethesda
may be much higher in families of Asian descent.92 Women criteria will identify people without a strong family history
with HNPCC have a 9–12% lifetime risk of developing of colorectal cancer who are considered at high-risk.107
ovarian cancer.88 Other cancers seen in HNPCC include
cancers of the small intestine, pancreas, brain, hepatobili-
Microsatellite Instability (MSI) Testing
ary tract, and urinary tract, for which the lifetime risk is in
the range of 2–4%88 (Figure 36.8). Sebaceous gland tumors, DNA replication errors characterize tumors with loss-of-
multiple keratoacanthomas, basal cell carcinomas, and pos- function mutations in MMR genes. These can be detected
sibly breast cancer may be more common in the Muir-Torre as microsatellite instability (MSI), which is the finding that,
variant (MIM 158320) of HNPCC. The Turcot variant in the same individual, the number of repeats in a given
(MIM 276300) should be considered when the family his- repeating sequence of DNA varies from cell to cell instead
tory includes glioblastoma multiforme. of being constant. More than 90% of colorectal cancers in
There are differences in phenotype, depending on the people with germline DNA MMR gene mutations have
MMR gene involved. Endometrial cancer may be more high MSI, whereas less than 15% of sporadic colorectal can-
common in families with germline MSH6 alterations as cers do (Figure 36.8).
compared to the other mismatch repair genes.79,93,94 In the Although the sensitivity of the Bethesda Guidelines is
Muir-Torre variant of HNPCC, the vast majority of germ- better than 72%, these guidelines are burdensome to recall
line mutations identified have been in MSH2, with a few and have been shown to be poorly implemented in clinical
families harboring MLH1 alterations. One-third of patients practice. In 2009, the Evaluation of Genomic Applications
with Turcot syndrome have mutations in MSH2. In certain in Practice and Prevention (EGAPP) Working Group rec-
populations, founder mutations in MLH1 and MSH2 have ommended that all newly diagnosed patients with colorec-
been identified in Lynch syndrome families. In the Finnish tal cancer be screened for HNPCC through MSI testing or
population, for example, two mutations in MLH1 account immunohistochemistry testing of tumor samples.108 While
5 5 4 • G e no m ic s in C l inic a l P r actic e
(A)
Normal Dysplastic
crypt
Inactivation of Mutation Adenomatous CIN 18q loss Inactivation Colon Cancer
APC (5q loss) of K-Ras lesion defect DCC, DPC4 of p53
(17p loss)
APC Mutation
β-catenin of B-Raf
Somatic MMR Axin2 (>K-Ras)
epigenetic inactivation
(MLH1)
Serrated Colon Cancer
adenomatous lesion
Sporadic MSI-H
pathways (CIMP) Inactivation of tumor suppresor genes
by promoter hyprmethylation
Figure 36.8 Genetic model of sporadic and MMR-related colorectal cancers. A: Sporadic colorectal cancers (CRC) are believed to arise from
adenomatous polyps, resulting from accumulation of mutations in oncogenes and tumor-suppressor genes. Mutations in adenomatous polyposis
coli (APC), β-catenin, or other components of this pathway mediate the transition of single pre-neoplastic cells to aberrant crypt foci and then
to adenoma and colorectal carcinoma. B: In about 20% of colorectal cancers, mismatch repair (MMR) function is inactivated, either by somatic
mutations or by epigenetic inactivation leading to microsatellite instability (MSI-H). Mutational inactivation of MMR genes is most commonly
observed as the second hit in patients who already carry germline mutations in MMR genes and fall under the hereditary nonpolyposis colorectal
cancer syndrome. Epigenetic inactivation of MMR genes most commonly affects hypermethylation of the MLH1 promoter. Adapted from Fearon and
Bommer, Devita, Hellman and Rosenberg’s Cancer: Principles and Practice of Oncology, 8 ed.179
th
the adoption of such universal institution-wide practice has malignancy of endocrine organs. While medullary thyroid
been problematic and varied across institution types,109 a carcinoma (MTC) represents only 10% of all thyroid can-
recent study has shown that successful implementation of cers, 10 to 25% of patients with MTC are due to mutation
universal screening can be achieved with a high uptake of in a high penetrance predisposition gene.72,114 Molecular
diagnostic genetic services. To achieve the greatest success, studies have led to increased appreciation of the heteroge-
the program at a minimum must have representation from neity of thyroid cancers, with hereditary predisposition,
the fields of colorectal surgery, gastroenterology, gynecol- somatic mutation, and epigenetic modulation all contribut-
ogy, pathology, and genetics.110 ing to tumor behavior.115 For example, the discovery of the
RET (REarranged during Transfection) proto-oncogene as
the causative gene for multiple endocrine neoplasia type 2
Risk Reduction for HNPCC-associated Cancers
(MEN 2) enabled gene-based molecular diagnosis, predic-
Colorectal cancer-risk reduction in HNPCC is mainly tive testing, and genotype-specific management for affected
through colonoscopy to identify and remove polyps. individuals.
Surveillance for various types of cancer in individuals with
HNPCC is based on expert recommendations and on case
MU LT I P L E E N D O C R I N E N EO P L A S I A
series that have demonstrated a reduction in CRC incidence
2 ( M E N 2)
with periodic colonoscopic surveillance111–113 (Table 36.9).
MEN 2 is an autosomal dominant cancer syndrome classified
into three clinical subtypes: MEN 2A (MIM 171400), famil-
H E R E D ITA RY T H Y RO I D C A N C E R
ial medullary thyroid carcinoma (FMTC; MIM 155240), and
While thyroid cancer only accounts for approximately 1% MEN 2B (MIM 162300) (Table 36.10). All three subtypes
of all newly diagnosed cancer cases, it is the most common involve high risk of developing MTC; MEN 2A and MEN
C l inic a l C a nc e r G e no m ic s • 5 5 5
Table 36.8 HEREDITARY NONPOLYPOSIS COLORECTAL CANCER (HNPCC) CRITERIA
Original Amsterdam criteria (Amsterdam criteria I)There should be at least three relatives with colon cancer and:
1. One should be a first-degree relative of the other two.
2. At least two successive generations should be affected.
3. At least one should be diagnosed before age 50.
4. Familial adenomatous polyposis should be excluded.
5. Tumors should be verified by pathological examination.
Revised Amsterdam criteria (Amsterdam criteria II)There should be at least three relatives with an HNPCC-associated cancer (cancer of the
colorectum, endometrium, small bowel, ureter, or renal pelvis) and:
1. One should be a first-degree relative to the other two.
2. At least two successive generations should be affected.
3. At least one should be diagnosed before age 50.
4. Familial adenomatous polyposis should be excluded.
5. Tumors should be verified by pathological examination.
Bethesda guidelines (ver. 2004)Tumors from individuals should be tested for microsatellite instability (MSI) in the following situations:
1. Colorectal cancer diagnosed in a patient who is less than 50 years of age
2. Presence of synchronous, metachronous colorectal, or other HNPCC-associated tumors, regardless of age
3. Colorectal cancer with the MSI-H histology diagnosed in a patient who is less than 60 years of age
4. Colorectal cancer diagnosed in one or more first-degree relatives with an HNPCC-related tumor, with one of the cancers being diagnosed
under age 50 years
5. Colorectal cancer diagnosed in two or more first- or second-degree relatives with HNPCC-related tumors, regardless of age
2B have an increased risk for pheochromocytoma (PC).116 a transmembrane domain, and an intracellular tyrosine
MEN 2A also has an increased risk for primary hyperpara- kinase domain134,135 (Figure 36.9). However, the RET recep-
thyroidism (HPT). Additional features in MEN 2B include tor is unique in that it requires a multicomponent complex
mucosal neuromas of the lips and tongue, distinctive facies to trigger activation. RET has to bind one of four GFRα
with enlarged lips, ganglioneuromatosis of the gastrointestinal family of GPI-linked co-receptors before binding one of
tract, and an asthenic “Marfanoid” body habitus.117 four members of the glial cell line–derived neurotrophic
All three subtypes are caused by germline RET muta- factor family of ligands (GDNF; neurturin, persephin, and
tions.118–121 RET was the first proto-oncogene to be impli- artemin) in a heterohexameric complex.136–138 RET signals
cated in an inherited cancer-susceptibility syndrome. The down a number of well-characterized signal transduction
frequency range of germline RET mutations reported for pathways, including the MAP kinase–RAS–RAF and
apparently sporadic MTC cases varies widely, from approxi- phosphoinositide 3-kinase pathways.139,140 MEN 2A- and
mately 5% based on direct testing to about 25% using fam- FMTC-related mutations affecting cysteine codons result
ily history–based assessments.122–127 in disulphide bond–mediated inter-receptor binding and
RET comprises 21 exons and encodes a transmembrane thus constitutive activation.141,142
receptor tyrosine kinase expressed in tissues and tumors
derived mainly from the neural crest.128–133 The structure
Identifying Patients for Genetics Risk Assessment
of the RET receptor is like that of most receptor tyro-
sine kinases, with a large extracellular domain, compris- The diagnostic criteria for MEN 2A are the presence
ing a cysteine-rich domain and a cadherin-like domain, of MTC, PC, and/or HPT, and a germline RET gene
5 5 6 • G e no m ic s in C l inic a l P r actic e
Table 36.10 CLINICAL FEATURES OF MEN 2 SUBTYPES
Codon
Mutation MEN 2A FMTC MEN 2B
609
611
Cysteine-rich 618
Doamain
620
630
634
Transmembrane
768
Tyrosine Kinase
Doamain 790
804
*(if in combination
with other variants)
883
Tyrosine Kinase
891
Doamain
918
Germline missense mutations in RET associated with MEN 2A, FMTC, and MEN 2B. Schematic diagram of the RET receptor and
Figure 36.9
distribution of mutated codons associated with different risk levels for aggressive MTC in MEN 2 syndromes. The most common MEN 2–
associated mutations are reported. For codon 804, there is observed phenotypical heterogeneity. There is an increased risk for aggressive MTC as
well as pheochromocytomas in patients harboring a codon 804 mutation with a second RET mutation or variant, and such individuals look MEN
2B–like. While this schematic illustrates the most common germline RET mutations causing MEN 2, note that somatic M918T mutations are very
common in sporadic MTC tissue.
C l inic a l C a nc e r G e no m ic s • 5 5 7
mutation. The mutation frequency is greater than 98%, thus advanced stage and increasing risk of metastases correlated
few families present a diagnostic dilemma when fewer than with mutation in codon position (609→620) the closer one
three clinical features are present. In cases in which only approached the juxtamembrane domain.154
one or two clinical features are present, the diagnosis can be
made either when a first-degree relative shows MEN 2A fea-
Surveillance in MEN 2
tures or when a RET mutation is identified. At a minimum,
the diagnosis can be made when two clinical features are MEN 2 showcases how genetic testing facilitates presymp-
present, even when an autosomal dominant pattern is not tomatic diagnosis and prevention, which are tailored to the
evident and a RET mutation has not been demonstrated.143 specific genotype.155 Codon-specific risks and age of onset
FMTC is, by definition, phenotypically limited to MTC.144 have been described for PC.156,157 It has been suggested that
The importance of identifying all MEN 2–associated annual screening for PC may be warranted, beginning at
MTC cases for optimal patient and family management age 10 years, for carriers of mutations in codons 918, 634,
has led to a recommendation to offer pre- and post-test and 630. Carriers with mutations at other codons could ini-
genetic counseling and genetic testing for RET mutations tiate screening at age 20 years.157
to all patients presenting with MTC, irrespective of other
features and family history.143
Management in MEN 2 Is Genotype-Dependent
Total thyroidectomy with central cervical lymph-node dissec-
Genetic Testing and Genotype–Phenotype
tion is the recommended surgical procedure for MEN2 muta-
Correlations in MEN 2
tion carriers.143,158 Earlier age at surgery results in less chance of
Germline RET mutations have been identified in approxi- recurrence, but there have been differing opinions about the
mately 95% of all MEN 2 cases, with 98% of MEN 2A pro- optimal timing of surgery. Genotype–phenotype correlations
bands found to have a mutation, 97% in MEN 2B, and 85% in MEN 2 are strong, both in terms of being mutation-specific
in FMTC.145–147 The characteristic mutational spectrum and as highly predictive of age of MTC-onset, which makes
found in MEN 2A includes missense mutations in one of genotyping an indispensable tool in planning the age of pro-
cysteine codons 609, 611, 618, 620 (exon 10), or 634 (exon phylactic surgery on an individual basis143,155,158 (Table 36.11).
11).145,148 Genotype–phenotype analyses reveal that codon An approach put forth by the National Comprehensive
634 mutations are associated with the presence of PC and Cancer Network143,158 is to stratify risk by genotype into the
hyperparathyroidism (HPT).145 In particular, the C634R categories, and to time surgery, as described in Table 36.11.72,143
mutation is probably associated with the development of Somatic RET mutations are also found in at least 50% of spo-
HPT.145,149,150 FMTC-associated mutations occur at the radic MTCs. Importantly, this link between MTC and RET
same cysteine codons as those in MEN 2A, although muta- has recently led to the USFDA’s approval of vandetanib in
tions at codons 609–620 are more proportionately frequent 2011, an orally bioavailable inhibitor of RET, VEGFR-2,
in FMTC than in MEN 2A. Germline M918T and A883F and EGFR, for use in advanced MTC, even cases arising in
mutations occur in 95% and ~2%, respectively, of MEN 2B the sporadic setting, following dramatic improvements in
patients.145–147 These two mutations are unique to MEN 2B progression-free survival over placebo.159
and have never been observed in MEN 2A or FMTC.
Genetic testing for RET mutations is therefore recom-
S O M AT I C C A N C E R G E N O M I C S : B EYO N D
mended to proceed in stepwise fashion, based on the high
I N H E R I T E D C A N C E R SY N D RO M E S
rate of exon 10 or 11 mutation in MEN 2A and character-
istically affected codons in MEN 2B.143,148,151 The various Early successes in targeted cancer therapy have helped
genotypes also have an impact on penetrance. For example, enlighten our understanding. The best known of these is
it has now become obvious that RET codons’ 918, 883, and extending imatinib’s use in chronic myeloid leukemia to
634 mutations have the highest penetrance, predisposing to gastrointestinal stromal tumors (GIST), the majority of
MEN 2B and MEN 2A with MTC, PC, and HPT involve- which sport a receptor tyrosine kinase (called c-Kit) that is
ment, respectively.152,153 In contrast, germline V804M activated by mutation,160,161 heralding an era of increasingly
mutations and perhaps cysteine codon 609/611 mutations extensive systematic sequencing of cancer genomes over the
have the lowest penetrance and the older ages of onset.145 past few years. This sequencing has identified several new
A recent study showed that codon-associated penetrance by mutated somatic cancer genes encoding for proteins that
age 50 ranged from 60% (codon 611) to 86% (620). More serve as targets for new therapeutics, particularly those that
5 5 8 • G e no m ic s in C l inic a l P r actic e
Table 36.11 GENOTYPE-GUIDED MANAGEMENT RECOMMENDATIONS FOR MEDULLARY THYROID CANCER
encode enzymes such as kinases and in which the muta- Drug Administration (FDA); the 21-gene recurrence
tions result in constitutive activation. Examples include score Oncotype DX; and the 76-gene outcome Rotterdam
BRAF, PIK3CA, AKT1, and IDH1,162–166 and searches signature.170–172 These profiles, which are obtained with
for inhibitors of the proteins encoded by many of these the use of tumor-derived mRNA assayed on either DNA
mutated genes are in progress. A notable example of a pro- microarrays or by quantitative reverse-transcriptase
tein activated through somatic mutation in cancer is BRAF, polymerase-chain-reaction (RT-PCR) assay, are being used
a serine–threonine kinase. Activating somatic mutations in in clinical practice to predict prognosis and hence to influ-
BRAF were identified by a systematic search of candidate ence therapeutic intervention (Figure 36.10). Reflecting
cancer genes in a panel of diverse cancers; BRAF was then this, the use of the Oncotype DX assay has been included
shown to be mutated in about two-thirds of malignant in the guidelines of the American Society of Clinical
melanomas.165 Because of the limited treatment options for Oncology for the evaluation of patients with node-negative,
metastatic malignant melanoma, the discovery of BRAF ER-positive disease to identify those who may obtain the
mutations triggered multiple drug-discovery programs in most benefit from adjuvant tamoxifen and those who may
the academic and pharmaceutical sectors. Vemurafenib, not require chemotherapy. This has led to a significant
a potent inhibitor of mutated BRAF,167 has marked anti- change in how medical oncologists manage breast cancer
tumor effects against melanoma cell lines with the BRAF that is at low clinical risk for recurrence. A recent multi-
V600E mutation, but not against cells with wild-type center study of physician practice showed that the treat-
BRAF.167,168 Vemurafenib produced improved rates of over- ment recommendation of medical oncologists has changed
all and progression-free survival in patients with previously for almost a third of patients in this subgroup on the basis
untreated melanoma with the BRAF V600E mutation169 of this assay.173
and is currently the first-line recommended therapy for
melanomas with the BRAF V600E mutation.
The use of gene-expression signatures has further C O N C LU S I O N S
empowered the identification of prognostic subclasses.
For example, gene-expression profiles have been used to The ultimate goal in oncology is to markedly decrease
define the risk of relapse in patients with early-stage breast death from cancer. The past 50 years of research have dem-
cancer. Three profiles have shown prognostic ability: the onstrated that the architecture of inherited genetic suscep-
70-gene profile MammaPrint, approved by the Food and tibility to cancers is defined by a collage of predisposition
C l inic a l C a nc e r G e no m ic s • 5 5 9
High Molecular Risk
Hormone Therapy
+
Chemotherapy
Figure 36.10
Use of a gene-expression profile to define the risk of relapse in patients with early-stage breast cancer and to aid choice of adjuvant
treatment. Gene-expression signatures can be incorporated with clinical decision-making. Patients with a gene-expression profile predicting a high
risk of relapse can be offered the more aggressive option of adjuvant hormonal and chemotherapy for early-stage hormone-responsive breast cancers;
in contrast, patients with low-risk gene-expression profiles may do as well with hormonal therapy alone.
alleles with different levels of risk and prevalence in the 4. Fong PC, Boss DS, Yap TA, et al. Inhibition of poly(ADP-ribose)
polymerase in tumors from BRCA mutation carriers. N Engl J Med.
population. Improved detection of patients harboring del- Jul 9, 2009;361(2):123–134.
eterious mutations has allowed tailored, gene-enabled man- 5. Byrski T, Gronwald J, Huzarski T, et al. Pathologic complete
agement both in terms of novel targeted therapy, as well as response rates in young women with BRCA1-positive breast
cancers after neoadjuvant chemotherapy. J Clin Oncol. Jan 20,
by a targeted-surveillance approach for affected individuals. 2010;28(3):375–379.
However, technology continues to move apace, dramati- 6. Offit K. Genomic profiles for disease risk: predictive or premature?
cally reducing costs and making genome sequencing fairly JAMA. Mar 19, 2008;299(11):1353–1355.
7. Miki Y, Swensen J, Shattuck-Eidens D, et al. A strong candidate for
routine. The optimal management of low to moderately the breast and ovarian cancer susceptibility gene BRCA1. Science.
penetrant genes conferring modestly increased relative risks Oct 7, 1994;266(5182):66–71.
remains incompletely defined. The next frontier would be to 8. Wooster R, Bignell G, Lancaster J, et al. Identification of the
breast cancer susceptibility gene BRCA2. Nature. Dec 21–28,
determine what the modifiers of specific risks for Mendelian 1995;378(6559):789–792.
high-penetrance alleles are and to better understand how 9. Claus EB, Risch N, Thompson WD. Genetic analysis of breast can-
low-penetrance susceptibility alleles “cross-talk” with cer in the cancer and steroid hormone study. Am J Hum Genet. Feb
1991;48(2):232–242.
one another, to facilitate greater precision in risk predic- 10. Ford D, Easton DF. The genetics of breast and ovarian cancer. Br J
tion. Additionally, recent efforts such as the Collaborative Cancer. Oct 1995;72(4):805–812.
Oncological Gene-Environment Study (COGS) shed light 11. Roa JC, Baeza R, Romeo E, Andrade L, Foradori A, Gonzalez S.
[Breast cancer: comparison between biochemical and immunohis-
on new genetic susceptibility loci for breast, ovarian, and tochemical determination of estrogen receptors]. Rev Med Chil.
prostate cancers.174–178 These newly identified susceptibility Mar 1996;124(3):307–312.
12. Struewing JP, Hartge P, Wacholder S, et al. The risk of can-
loci explain an increasing proportion of the familial risk of cer associated with specific mutations of BRCA1 and BRCA2
these cancers. Understanding the biological basis of how among Ashkenazi Jews. N Engl J Med. May 15, 1997;336(20):
these germline susceptibility genes interact with somatically 1401–1408.
13. Gao Q, Neuhausen S, Cummings S, Luce M, Olopade OI. Recurrent
acquired alterations will be key to ongoing efforts for novel germ-line BRCA1 mutations in extended African American
cancer-prevention and therapeutic strategies. families with early-onset breast cancer. Am J Hum Genet. May
1997;60(5):1233–1236.
14. Neuhausen S, Gilewski T, Norton L, et al. Recurrent BRCA2
6174delT mutations in Ashkenazi Jewish women affected by breast
R E F E R E N C ES cancer. Nat Genet. May 1996;13(1):126–128.
15. Tonin P, Weber B, Offit K, et al. Frequency of recurrent BRCA1 and
1. Broca P. Traits des tumeurs: Tome premier, Des tumeurs en general. BRCA2 mutations in Ashkenazi Jewish breast cancer families. Nat
1866; http://www.archive.org/stream/traitdestumeurs02brocgoog Med. Nov 1996;2(11):1179–1183.
- page/n12/mode/1up. Accessed Mar 2013. 16. Peelen T, van Vliet M, Petrij-Bosch A, et al. A high proportion
2. Warhin A. Heredity with reference to carcinoma as shown by of novel mutations in BRCA1 with strong founder effects among
the study of cases examined in the pathological laboratory of the Dutch and Belgian hereditary breast and ovarian cancer families.
University of Michigan. Arch Intern Med. 1913;12:546–555. Am J Hum Genet. May 1997;60(5):1041–1049.
3. U.S. Department of Health and Human Services. Healthy People 17. Petrij-Bosch A, Peelen T, van Vliet M, et al. BRCA1 genomic dele-
2020: http://www.healthypeople.gov/2020/default.aspx. Accessed tions are major founder mutations in Dutch breast cancer patients.
Mar 8, 2013. Nat Genet. Nov 1997;17(3):341–345.
5 6 0 • G e no m ic s in C l inic a l P r actic e
18. Simard J, Tonin P, Durocher F, et al. Common origins of BRCA1 40. Scully R, Chen J, Ochs RL, et al. Dynamic changes of BRCA1 sub-
mutations in Canadian breast and ovarian cancer families. Nat nuclear location and phosphorylation state are initiated by DNA
Genet. Dec 1994;8(4):392–398. damage. Cell. Aug 8, 1997;90(3):425–435.
19. Tonin PN, Mes-Masson AM, Narod SA, Ghadirian P, Provencher 41. Chen J, Silver DP, Walpita D, et al. Stable interaction between the
D. Founder BRCA1 and BRCA2 mutations in French Canadian products of the BRCA1 and BRCA2 tumor suppressor genes in
ovarian cancer cases unselected for family history. Clin Genet. May mitotic and meiotic cells. Mol Cell. Sep 1998;2(3):317–328.
1999;55(5):318–324. 42. Somasundaram K. Breast cancer gene 1 (BRCA1): role in cell cycle
20. Evans DG, Neuhausen SL, Bulman M, Young K, Gokhale D, regulation and DNA repair—perhaps through transcription. J Cell
Lalloo F. Haplotype and cancer risk analysis of two common Biochem. Apr 15, 2003;88(6):1084–1091.
mutations, BRCA1 4184del4 and BRCA2 2157delG, in high 43. Smith SA, Easton DF, Evans DG, Ponder BA. Allele losses in the
risk northwest England breast/ovarian families. J Med Genet. Feb region 17q12–21 in familial breast and ovarian cancer involve the
2004;41(2):e21. wild-type chromosome. Nat Genet. Oct 1992;2(2):128–131.
21. Thorlacius S, Olafsdottir G, Tryggvadottir L, et al. A single BRCA2 44. Cornelis RS, Neuhausen SL, Johansson O, et al. High allele loss rates
mutation in male and female breast cancer families from Iceland with at 17q12-q21 in breast and ovarian tumors from BRCAl-linked
varied cancer phenotypes. Nat Genet. May 1996;13(1):117–119. families. The Breast Cancer Linkage Consortium. Gene Chromosome
22. Andersen TI, Borresen AL, Moller P. A common BRCA1 mutation Canc. Jul 1995;13(3):203–210.
in Norwegian breast and ovarian cancer families? Am J Hum Genet. 45. Jasin M. Homologous repair of DNA damage and tumorigenesis: the
Aug 1996;59(2):486–487. BRCA connection. Oncogene. Dec 16, 2002;21(58):8981–8993.
23. Gorski B, Byrski T, Huzarski T, et al. Founder mutations in the 46. Howlett NG, Taniguchi T, Olson S, et al. Biallelic inac-
BRCA1 gene in Polish families with breast-ovarian cancer. Am J tivation of BRCA2 in Fanconi anemia. Science. Jul 26,
Hum Genet. Jun 2000;66(6):1963–1968. 2002;297(5581):606–609.
24. Gayther SA, Mangion J, Russell P, et al. Variation of risks of breast 47. Thompson D, Easton D. Variation in BRCA1 cancer risks by
and ovarian cancer associated with different germline mutations of mutation position. Cancer Epidemiol Biomarkers Prev. Apr
the BRCA2 gene. Nat Genet. Jan 1997;15(1):103–105. 2002;11(4):329–336.
25. Liede A, Cohen B, Black DM, et al. Evidence of a founder BRCA1 48. Gayther SA, Warren W, Mazoyer S, et al. Germline mutations of
mutation in Scotland. Br J Cancer. Feb 2000;82(3):705–711. the BRCA1 gene in breast and ovarian cancer families provide
26. Hakansson S, Johannsson O, Johansson U, et al. Moderate frequency evidence for a genotype–phenotype correlation. Nat Genet. Dec
of BRCA1 and BRCA2 germ-line mutations in Scandinavian famil- 1995;11(4):428–433.
ial breast cancer. Am J Hum Genet. May 1997;60(5):1068–1078. 49. Thompson D, Easton D. Variation in cancer risks, by mutation
27. Johannsson O, Ostermeyer EA, Hakansson S, et al. Founding position, in BRCA2 mutation carriers. Am J Hum Genet. Feb
BRCA1 mutations in hereditary breast and ovarian cancer in south- 2001;68(2):410–419.
ern Sweden. Am J Hum Genet. Mar 1996;58(3):441–450. 50. Panchal SM, Ennis M, Canon S, Bordeleau LJ. Selecting a BRCA
28. Peto J, Collins N, Barfoot R, et al. Prevalence of BRCA1 and risk assessment model for use in a familial cancer clinic. BMC Med
BRCA2 gene mutations in patients with early-onset breast cancer. J Genet. 2008;9:116.
Natl Cancer Inst. Jun 2, 1999;91(11):943–949. 51. Palma MD, Domchek SM, Stopfer J, et al. The relative contribu-
29. Ford D, Easton DF, Stratton M, et al. Genetic heterogeneity and tion of point mutations and genomic rearrangements in BRCA1
penetrance analysis of the BRCA1 and BRCA2 genes in breast can- and BRCA2 in high-risk breast cancer families. Cancer Res. Sep 1,
cer families. The Breast Cancer Linkage Consortium. Am J Hum 2008;68(17):7006–7014.
Genet. Mar 1998;62(3):676–689. 52. Clarke-Pearson DL. Clinical practice. Screening for ovarian cancer.
30. Offit K. Breast cancer and BRCA1 mutations. N Engl J Med. May 2, N Engl J Med. Jul 9, 2009;361(2):170–177.
1996;334(18):1197–1198; author reply, 1199–1200. 53. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for pre-
31. Brose MS, Rebbeck TR, Calzone KA, Stopfer JE, Nathanson KL, vention of breast cancer: report of the National Surgical Adjuvant
Weber BL. Cancer risk estimates for BRCA1 mutation carriers Breast and Bowel Project P-1 Study. J Natl Cancer Inst. Sep 16,
identified in a risk evaluation program. J Natl Cancer Inst. Sep 18, 1998;90(18):1371–1388.
2002;94(18):1365–1372. 54. King MC, Wieand S, Hale K, et al. Tamoxifen and breast cancer
32. Cancer risks in BRCA2 mutation carriers. The Breast Cancer Linkage incidence among women with inherited mutations in BRCA1 and
Consortium. J Natl Cancer Inst. Aug 4, 1999;91(15):1310–1316. BRCA2: National Surgical Adjuvant Breast and Bowel Project
33. Armstrong K, Eisen A, Weber B. Assessing the risk of breast cancer. (NSABP-P1) Breast Cancer Prevention Trial. JAMA. Nov 14,
N Engl J Med. Feb 24, 2000;342(8):564–571. 2001;286(18):2251–2256.
34. Barnes-Kedar IM, Plon SE. Counseling the at risk patient in
55. Duffy SW, Nixon RM. Estimates of the likely prophylactic effect of
the BRCA1 and BRCA2 Era. Obstet Gynecol Clin North Am. Jun tamoxifen in women with high risk BRCA1 and BRCA2 mutations.
2002;29(2):341–366, vii. Br J Cancer. Jan 21, 2002;86(2):218–221.
35. Easton DF, Steele L, Fields P, et al. Cancer risks in two large breast 56. Narod SA, Brunet JS, Ghadirian P, et al. Tamoxifen and risk of con-
cancer families linked to BRCA2 on chromosome 13q12–13. Am J tralateral breast cancer in BRCA1 and BRCA2 mutation carriers: a
Hum Genet. Jul 1997;61(1):120–128. case-control study. Hereditary Breast Cancer Clinical Study Group.
36. Scully R, Chen J, Plug A, et al. Association of BRCA1 with Rad51 in Lancet. Dec 2, 2000;356(9245):1876–1881.
mitotic and meiotic cells. Cell. Jan 24, 1997;88(2):265–275. 57. Schrag D, Kuntz KM, Garber JE, Weeks JC. Decision analysis—
37. Zhang H, Somasundaram K, Peng Y, et al. BRCA1 physically associ- effects of prophylactic mastectomy and oophorectomy on life expec-
ates with p53 and stimulates its transcriptional activity. Oncogene. tancy among women with BRCA1 or BRCA2 mutations. N Engl J
Apr 2, 1998;16(13):1713–1721. Med. May 15, 1997;336(20):1465–1471.
38. Aprelikova ON, Fang BS, Meissner EG, et al. BRCA1-associated 58. Rebbeck TR, Friebel T, Lynch HT, et al. Bilateral prophylactic
growth arrest is RB-dependent. Proc Natl Acad Sci U S A. Oct 12, mastectomy reduces breast cancer risk in BRCA1 and BRCA2
1999;96(21):11866–11871. mutation carriers: the PROSE Study Group. J Clin Oncol. Mar 15,
39. Wang Q, Zhang H, Kajino K, Greene MI. BRCA1 binds c-Myc 2004;22(6):1055–1062.
and inhibits its transcriptional and transforming activity in cells. 59. Rebbeck TR. Prophylactic oophorectomy in BRCA1 and BRCA2
Oncogene. Oct 15, 1998;17(15):1939–1948. mutation carriers. Eur J Cancer. Nov 2002;38 Suppl 6:S15–17.
C l inic a l C a nc e r G e no m ic s • 5 6 1
60. Kauff ND, Satagopan JM, Robson ME, et al. Risk-reducing
83. Lynch HT, Shaw MW, Magnuson CW, Larsen AL, Krush AJ.
salpingo-oophorectomy in women with a BRCA1 or BRCA2 muta- Hereditary factors in cancer. Study of two large Midwestern kin-
tion. N Engl J Med. May 23, 2002;346(21):1609–1615. dreds. Arch Intern Med. Feb 1966;117(2):206–212.
61. Zhou XP, Woodford-Richens K, Lehtonen R, et al. Germline muta- 84. Salovaara R, Loukola A, Kristo P, et al. Population-based molecu-
tions in BMPR1A/ALK3 cause a subset of cases of juvenile pol- lar detection of hereditary nonpolyposis colorectal cancer. J Clin
yposis syndrome and of Cowden and Bannayan-Riley-Ruvalcaba Oncol. Jun 2000;18(11):2193–2200.
syndromes. Am J Hum Genet. Oct 2001;69(4):704–711. 85. Aaltonen LA, Salovaara R, Kristo P, et al. Incidence of heredi-
62. Waite KA, Eng C. Protean PTEN: form and function. Am J Hum tary nonpolyposis colorectal cancer and the feasibility of
Genet. Apr 2002;70(4):829–844. molecular screening for the disease. N Engl J Med. May 21,
63. Nelen MR, Padberg GW, Peeters EA, et al. Localization of the gene 1998;338(21):1481–1487.
for Cowden disease to chromosome 10q22–23. Nat Genet. May 86. Hampel H, Frankel WL, Martin E, et al. Feasibility of screening
1996;13(1):114–116. for Lynch syndrome among patients with colorectal cancer. J Clin
64. Nelen MR, Kremer H, Konings IB, et al. Novel PTEN mutations Oncol. Dec 10, 2008;26(35):5783–5788.
in patients with Cowden disease: absence of clear genotype–pheno- 87. Hampel H, Frankel WL, Martin E, et al. Screening for the Lynch
type correlations. Eur J Hum Genet. Apr 1999;7(3):267–273. syndrome (hereditary nonpolyposis colorectal cancer). N Engl J
65. Eng C. Will the real Cowden syndrome please stand up: revised Med. May 5, 2005;352(18):1851–1860.
diagnostic criteria. J Med Genet. Nov 2000;37(11):828–830. 88. Aarnio M, Sankila R, Pukkala E, et al. Cancer risk in mutation
66. Tan MH, Mester J, Peterson C, et al. A clinical scoring system for carriers of DNA-mismatch-repair genes. Int J Cancer. Apr 12,
selection of patients for PTEN mutation testing is proposed on the 1999;81(2):214–218.
basis of a prospective study of 3042 probands. Am J Hum Genet. Jan 89. Vasen HF, Wijnen JT, Menko FH, et al. Cancer risk in families
7, 2011;88(1):42–56. with hereditary nonpolyposis colorectal cancer diagnosed by
67. Ngeow J, Mester J, Rybicki LA, Ni Y, Milas M, Eng C. Incidence mutation analysis. Gastroenterology. Apr 1996;110(4):1020–1027.
and clinical characteristics of thyroid cancer in prospective series of 90. Vasen HF, Sanders EA, Taal BG, et al. The risk of brain tumours
individuals with Cowden and Cowden-like syndrome characterized in hereditary non-polyposis colorectal cancer (HNPCC). Int J
by germline PTEN, SDH, or KLLN alterations. J Clin Endocrinol Cancer. Feb 8, 1996;65(4):422–425.
Metab. Dec 2011;96(12):E2063–2071. 91. Millar AL, Pal T, Madlensky L, et al. Mismatch repair gene
68. Tan MH, Mester JL, Ngeow J, Rybicki LA, Orloff MS, Eng C. defects contribute to the genetic basis of double primary can-
Lifetime cancer risks in individuals with germline PTEN mutations. cers of the colorectum and endometrium. Hum Mol Genet. May
Clin Cancer Res. Jan 15, 2012;18(2):400–407. 1999;8(5):823–829.
69. Liaw D, Marsh DJ, Li J, et al. Germline mutations of the PTEN 92. Park YJ, Shin KH, Park JG. Risk of gastric cancer in hereditary
gene in Cowden disease, an inherited breast and thyroid cancer syn- nonpolyposis colorectal cancer in Korea. Clin Cancer Res. Aug
drome. Nat Genet. May 1997;16(1):64–67. 2000;6(8):2994–2998.
70. Fanning AS, Anderson JM. PDZ domains: fundamental building 93. Wagner A, Hendriks Y, Meijers-Heijboer EJ, et al. Atypical
blocks in the organization of protein complexes at the plasma mem- HNPCC owing to MSH6 germline mutations: analysis of a large
brane. J Clin Invest. Mar 1999;103(6):767–772. Dutch pedigree. J Med Genet. May 2001;38(5):318–322.
71. Eng C. PTEN: one gene, many syndromes. Hum Mutat. Sep 94. Wijnen J, de Leeuw W, Vasen H, et al. Familial endometrial cancer
2003;22(3):183–198. in female carriers of MSH6 germline mutations. Nat Genet. Oct
72. Zbuk KM, Eng C. Cancer phenomics: RET and PTEN as illustra- 1999;23(2):142–144.
tive models. Nat Rev Cancer. Jan 2007;7(1):35–45. 95. Nystrom-Lahti M, Kristo P, Nicolaides NC, et al. Founding muta-
73. Ni Y, Zbuk KM, Sadler T, et al. Germline mutations and variants tions and Alu-mediated recombination in hereditary colon cancer.
in the succinate dehydrogenase genes in Cowden and Cowden-like Nat Med. Nov 1995;1(11):1203–1206.
syndromes. Am J Hum Genet. Aug 2008;83(2):261–268. 96. Hutter P, Couturier A, Scott RJ, et al. Complex genetic predisposi-
74. Bennett KL, Mester J, Eng C. Germline epigenetic regulation of tion to cancer in an extended HNPCC family with an ancestral
KILLIN in Cowden and Cowden-like syndrome. JAMA. Dec 22, hMLH1 mutation. J Med Genet. Aug 1996;33(8):636–640.
2010;304(24):2724–2731. 97. Chan TL, Yuen ST, Ho JW, et al. A novel germline 1.8-kb deletion of
75. Orloff MS, He X, Peterson C, et al. Germline PIK3CA and AKT1 hMLH1 mimicking alternative splicing: a founder mutation in the
mutations in Cowden and Cowden-like syndromes. Am J Hum Chinese population. Oncogene. May 24, 2001;20(23):2976–2981.
Genet. Jan 10, 2013;92(1):76–80. 98. Foulkes WD, Thiffault I, Gruber SB, et al. The founder mutation
76. Mester J, Eng C. Estimate of de novo mutation frequency in pro- MSH2*1906G—>C is an important cause of hereditary nonpol-
bands with PTEN hamartoma tumor syndrome. Genet Med. Sep yposis colorectal cancer in the Ashkenazi Jewish population. Am J
2012;14(9):819–822. Hum Genet. Dec 2002;71(6):1395–1412.
77. Hildenbrand C, Burgdorf WH, Lautenschlager S. Cowden syndrome- 99. Wagner A, Barrows A, Wijnen JT, et al. Molecular analysis of hered-
diagnostic skin signs. Dermatology. 2001;202(4):362–366. itary nonpolyposis colorectal cancer in the United States: high
78. Kastrinos F, Syngal S. Recently identified colon cancer pre-
mutation detection rate among clinically selected families and
dispositions: MYH and MSH6 mutations. Semin Oncol. Oct characterization of an American founder genomic deletion of the
2007;34(5):418–424. MSH2 gene. Am J Hum Genet. May 2003;72(5):1088–1100.
79. Lynch HT, de la Chapelle A. Hereditary colorectal cancer. N Engl J 100. Wijnen JT, Vasen HF, Khan PM, et al. Clinical findings with impli-
Med. Mar 6, 2003;348(10):919–932. cations for genetic testing in families with clustering of colorectal
80. Fuchs CS, Giovannucci EL, Colditz GA, Hunter DJ, Speizer cancer. N Engl J Med. Aug 20, 1998;339(8):511–518.
FE, Willett WC. A prospective study of family history and the 101. Gille JJ, Hogervorst FB, Pals G, et al. Genomic deletions of MSH2
risk of colorectal cancer. N Engl J Med. Dec 22, 1994;331(25): and MLH1 in colorectal cancer families detected by a novel muta-
1669–1674. tion detection approach. Br J Cancer. Oct 7, 2002;87(8):892–897.
81. Markowitz SD, Bertagnolli MM. Molecular origins of can-
102. Nakagawa H, Yan H, Lockman J, et al. Allele separation facilitates
cer: molecular basis of colorectal cancer. N Engl J Med. Dec 17, interpretation of potential splicing alterations and genomic rear-
2009;361(25):2449–2460. rangements. Cancer Res. Aug 15, 2002;62(16):4579–4582.
82. Warthin AS. Heredity of carcinoma in man. Ann Intern Med. 103. Vasen HF, Mecklin JP, Khan PM, Lynch HT. The International
1931;4:681–696. Collaborative Group on Hereditary Non-Polyposis Colorectal
5 6 2 • G e no m ic s in C l inic a l P r actic e
Cancer (ICG-HNPCC). Dis Colon Rectum. May 1991;34(5): management of medullary thyroid carcinoma and multiple endo-
424–425. crine neoplasia type 2. Endocrinol Metab Clin North Am. Mar
104. Vasen HF, Watson P, Mecklin JP, Lynch HT. New clinical criteria 1996;25(1):1–25.
for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch 124. Schuffenecker I, Ginet N, Goldgar D, et al. Prevalence and parental
syndrome) proposed by the International Collaborative Group on origin of de novo RET mutations in multiple endocrine neoplasia type
HNPCC. Gastroenterology. Jun 1999;116(6):1453–1456. 2A and familial medullary thyroid carcinoma. Le Groupe d’Etude des
105. Bisgaard ML, Jager AC, Myrhoj T, Bernstein I, Nielsen FC. Tumeurs a Calcitonine. Am J Hum Genet. Jan 1997;60(1):233–237.
Hereditary non-polyposis colorectal cancer (HNPCC): pheno- 125. Wiench M, Wygoda Z, Gubala E, et al. Estimation of risk of inher-
type–genotype correlation between patients with and without ited medullary thyroid carcinoma in apparent sporadic patients. J
identified mutation. Hum Mutat. Jul 2002;20(1):20–27. Clin Oncol. Mar 1, 2001;19(5):1374–1380.
106. Syngal S, Fox EA, Li C, et al. Interpretation of genetic test results for 126. Decker RA, Peacock ML, Borst MJ, Sweet JD, Thompson NW.
hereditary nonpolyposis colorectal cancer: implications for clinical Progress in genetic screening of multiple endocrine neoplasia type
predisposition testing. JAMA. Jul 21, 1999;282(3):247–253. 2A: is calcitonin testing obsolete? Surgery. Aug 1995;118(2):257–
107. Umar A, Boland CR, Terdiman JP, et al. Revised Bethesda 263; discussion 263–254.
Guidelines for hereditary nonpolyposis colorectal cancer (Lynch 127. Brandi ML, Gagel RF, Angeli A, et al. Guidelines for diagnosis and
syndrome) and microsatellite instability. J Natl Cancer Inst. Feb 18, therapy of MEN type 1 and type 2. J Clin Endocrinol Metab. Dec
2004;96(4):261–268. 2001;86(12):5658–5671.
108. Evaluation of Genomic Applications in Practice and Prevention 128. Takahashi M, Ritz J, Cooper GM. Activation of a novel human
(EGAPP) Working Group: Recommendations from the EGAPP transforming gene, RET, by DNA rearrangement. Cell. Sep
Working Group: genetic testing strategies in newly diagnosed 1985;42(2):581–588.
individuals with colorectal cancer aimed at reducing morbidity 129. Takahashi M, Cooper GM. RET transforming gene encodes a
and mortality from Lynch syndrome in relatives. Genet Med. Jan fusion protein homologous to tyrosine kinases. Mol Cell Biol. Apr
2009;11(1):35–41. 1987;7(4):1378–1385.
109. Beamer LC, Grant ML, Espenschied CR, et al. Reflex immu- 130. Takahashi M. Structure and expression of the RET transforming
nohistochemistry and microsatellite instability testing of gene. IARC Sci Publ. 1988(92):189–197.
colorectal tumors for Lynch syndrome among US cancer pro- 131. Pachnis V, Mankoo B, Costantini F. Expression of the c-RET
grams and follow-up of abnormal results. J Clin Oncol. Apr 1, proto-oncogene during mouse embryogenesis. Development. Dec
2012;30(10):1058–1063. 1993;119(4):1005–1017.
110. Heald B, Plesec T, Liu X, et al. Implementation of universal micro- 132. Nakamura T, Ishizaka Y, Nagao M, Hara M, Ishikawa T.
satellite instability and immunohistochemistry screening for diag- Expression of the RET proto-oncogene product in human nor-
nosing Lynch syndrome in a large academic medical center. J Clin mal and neoplastic tissues of neural crest origin. J Pathol. Mar
Oncol. Feb 11, 2013. 1994;172(3):255–260.
111. Burke W, Petersen G, Lynch P, et al. Recommendations for follow- 133. Tsuzuki T, Takahashi M, Asai N, Iwashita T, Matsuyama M, Asai
up care of individuals with an inherited predisposition to cancer. J. Spatial and temporal expression of the RET proto-oncogene
I. Hereditary nonpolyposis colon cancer. Cancer Genetics Studies product in embryonic, infant and adult rat tissues. Oncogene. Jan 5,
Consortium. JAMA. Mar 19, 1997;277(11):915–919. 1995;10(1):191–198.
112. Smith RA, Cokkinides V, Eyre HJ. American Cancer Society 134. Pasini B, Hofstra RM, Yin L, et al. The physical map of the human
guidelines for the early detection of cancer, 2003. CA Cancer J RET proto-oncogene. Oncogene. Nov 2, 1995;11(9):1737–1743.
Clin. Jan–Feb 2003;53(1):27–43. 135. Eng C. RET proto-oncogene in the development of human cancer.
113. Jarvinen HJ, Mecklin JP, Sistonen P. Screening reduces colorectal J Clin Oncol. Jan 1999;17(1):380–393.
cancer rate in families with hereditary nonpolyposis colorectal can- 136. Durbec P, Marcos-Gutierrez CV, Kilkenny C, et al. GDNF signal-
cer. Gastroenterology. May 1995;108(5):1405–1411. ling through the RET receptor tyrosine kinase. Nature. Jun 27,
114. Richards ML. Familial syndromes associated with thyroid cancer in 1996;381(6585):789–793.
the era of personalized medicine. Thyroid. Jul 2010;20(7):707–713. 137. Jing S, Wen D, Yu Y, et al. GDNF-induced activation of the RET
115. Xing M. Molecular pathogenesis and mechanisms of thyroid can- protein tyrosine kinase is mediated by GDNFR-alpha, a novel
cer. Nat Rev Cancer. Mar 2013;13(3):184–199. receptor for GDNF. Cell. Jun 28, 1996;85(7):1113–1124.
116. Gimm O. Multiple endocrine neoplasia type 2: clinical aspects. 138. Kotzbauer PT, Lampe PA, Heuckeroth RO, et al. Neurturin, a
Front Horm Res. 2001;28:103–130. relative of glial-cell-line-derived neurotrophic factor. Nature. Dec
117. Gorlin RJ, Sedano HO, Vickers RA, Cervenka J. Multiple mucosal 5, 1996;384(6608):467–470.
neuromas, pheochromocytoma and medullary carcinoma of the 139. van Weering DH, Bos JL. Signal transduction by the receptor tyro-
thyroid—a syndrome. Cancer. Aug 1968;22(2):293–299 passim. sine kinase RET. Recent Results Cancer Res. 1998;154:271–281.
118. Mathew CG, Chin KS, Easton DF, et al. A linked genetic marker 140. Besset V, Scott RP, Ibanez CF. Signaling complexes and protein–
for multiple endocrine neoplasia type 2A on chromosome 10. protein interactions involved in the activation of the Ras and phos-
Nature. Aug 6–12, 1987;328(6130):527–528. phatidylinositol 3-kinase pathways by the c-RET receptor tyrosine
119. Simpson NE, Kidd KK. Where is the locus for multiple endocrine kinase. J Biol Chem. Dec 15, 2000;275(50):39159–39166.
neoplasia type 2A? Henry Ford Hosp Med J. 1987;35(2–3):168–171. 141. Borrello MG, Pelicci G, Arighi E, et al. The oncogenic versions of
120. Gardner E, Papi L, Easton DF, et al. Genetic linkage studies map the RET and TRK tyrosine kinases bind Shc and Grb2 adaptor
the multiple endocrine neoplasia type 2 loci to a small interval on proteins. Oncogene. Jun 1994;9(6):1661–1668.
chromosome 10q11.2. Hum Mol Genet. Mar 1993;2(3):241–246. 142. Santoro M, Carlomagno F, Romano A, et al. Activation of RET as
121. Mulligan LM, Kwok JB, Healey CS, et al. Germ-line mutations of a dominant transforming gene by germline mutations of MEN2A
the RET proto-oncogene in multiple endocrine neoplasia type 2A. and MEN2B. Science. Jan 20, 1995;267(5196):381–383.
Nature. Jun 3, 1993;363(6428):458–460. 143. Kloos RT, Eng C, Evans DB, et al. Medullary thyroid cancer: man-
122. Eng C, Mulligan LM, Smith DP, et al. Mutation of the RET agement guidelines of the American Thyroid Association. Thyroid.
protooncogene in sporadic medullary thyroid carcinoma. Gene Jun 2009;19(6):565–612.
Chromosome Canc. Mar 1995;12(3):209–212. 144. Farndon JR, Leight GS, Dilley WG, et al. Familial medullary thy-
123. Wohllk N, Cote GJ, Evans DB, Goepfert H, Ordonez NG, roid carcinoma without associated endocrinopathies: a distinct
Gagel RF. Application of genetic screening information to the clinical entity. Br J Surg. Apr 1986;73(4):278–281.
C l inic a l C a nc e r G e no m ic s • 5 6 3
145. Eng C, Clayton D, Schuffenecker I, et al. The relationship between 162. Paez JG, Janne PA, Lee JC, et al. EGFR mutations in lung c ancer:
specific RET proto-oncogene mutations and disease phenotype in correlation with clinical response to gefitinib therapy. Science.
multiple endocrine neoplasia type 2. International RET mutation Jun 4, 2004;304(5676):1497–1500.
consortium analysis. JAMA. Nov 20, 1996;276(19):1575–1579. 163. Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-
146. Gimm O, Marsh DJ, Andrew SD, et al. Germline dinucleotide paclitaxel in pulmonary adenocarcinoma. N Engl J Med. Sep 3,
mutation in codon 883 of the RET proto-oncogene in multiple 2009;361(10):947–957.
endocrine neoplasia type 2B without codon 918 mutation. J Clin 164. Samuels Y, Wang Z, Bardelli A, et al. High frequency of muta-
Endocrinol Metab. Nov 1997;82(11):3902–3904. tions of the PIK3CA gene in human cancers. Science. Apr 23,
147. Smith DP, Houghton C, Ponder BA. Germline mutation of RET 2004;304(5670):554.
codon 883 in two cases of de novo MEN 2B. Oncogene. Sep 4, 165. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene
1997;15(10):1213–1217. in human cancer. Nature. Jun 27, 2002;417(6892):949–954.
148. Mulligan LM, Eng C, Healey CS, et al. Specific mutations of the 166. Stephens P, Hunter C, Bignell G, et al. Lung cancer: intra-
RET proto-oncogene are related to disease phenotype in MEN 2A genic ERBB2 kinase mutations in tumours. Nature. Sep 30,
and FMTC. Nat Genet. Jan 1994;6(1):70–74. 2004;431(7008):525–526.
149. Mulligan LM, Eng C, Healey CS, et al. A de novo mutation of the 167. Bollag G, Hirth P, Tsai J, et al. Clinical efficacy of a RAF inhibitor
RET proto-oncogene in a patient with MEN 2A. Hum Mol Genet. needs broad target blockade in BRAF-mutant melanoma. Nature.
Jun 1994;3(6):1007–1008. Sep 30, 2010;467(7315):596–599.
150. Schuffenecker I, Virally-Monod M, Brohet R, et al. Risk and pen- 168. Tsai J, Lee JT, Wang W, et al. Discovery of a selective inhibitor of
etrance of primary hyperparathyroidism in multiple endocrine oncogenic B-Raf kinase with potent antimelanoma activity. Proc
neoplasia type 2A families with mutations at codon 634 of the Natl Acad Sci U S A. Feb 26, 2008;105(8):3041–3046.
RET proto-oncogene. Groupe D’etude des Tumeurs a Calcitonine. 169. Chapman PB, Hauschild A, Robert C, et al. Improved survival
J Clin Endocrinol Metab. Feb 1998;83(2):487–491. with vemurafenib in melanoma with BRAF V600E mutation.
151. Mulligan LM, Marsh DJ, Robinson BG, et al. Genotype–pheno- N Engl J Med. Jun 30, 2011;364(26):2507–2516.
type correlation in multiple endocrine neoplasia type 2: report of 170. van ‘t Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression pro-
the International RET Mutation Consortium. J Intern Med. Oct filing predicts clinical outcome of breast cancer. Nature. Jan 31,
1995;238(4):343–346. 2002;415(6871):530–536.
152. Eng C, Mulligan LM, Smith DP, et al. Low frequency of germline 171. Paik S, Shak S, Tang G, et al. A multigene assay to predict recur-
mutations in the RET proto-oncogene in patients with apparently rence of tamoxifen-treated, node-negative breast cancer. N Engl J
sporadic medullary thyroid carcinoma. Clin Endocrinol (Oxf ). Jul Med. Dec 30, 2004;351(27):2817–2826.
1995;43(1):123–127. 172. Wang Y, Klijn JG, Zhang Y, et al. Gene-expression profiles to pre-
153. Eng C. Multiple endocrine neoplasia type 2 and the practice dict distant metastasis of lymph-node-negative primary breast can-
of molecular medicine. Rev Endocr Metab Disord. Nov 2000; cer. Lancet. Feb 19–25, 2005;365(9460):671–679.
1(4):283–290. 173. Lo SS, Mumby PB, Norton J, et al. Prospective multicenter study of
154. Frank-Raue K, Rybicki LA, Erlic Z, et al. Risk profiles and pene- the impact of the 21-gene recurrence score assay on medical oncol-
trance estimations in multiple endocrine neoplasia type 2A caused ogist and patient adjuvant breast cancer treatment selection. J Clin
by germline RET mutations located in exon 10. Hum Mutat. Jan Oncol. Apr 1, 2010;28(10):1671–1676.
2011;32(1):51–58. 174. Burton H, Chowdhury S, Dent T, Hall A, Pashayan N, Pharoah
155. Machens A, Dralle H. Genotype–phenotype based surgical con- P. Public health implications from COGS and potential for risk
cept of hereditary medullary thyroid carcinoma. World J Surg. May stratification and screening. Nat Genet. Mar 27, 2013;45(4):
2007;31(5):957–968. 349–351.
156. Yip L, Cote GJ, Shapiro SE, et al. Multiple endocrine neoplasia 175. Sakoda LC, Jorgenson E, Witte JS. Turning of COGS moves for-
type 2: evaluation of the genotype–phenotype relationship. Arch ward findings for hormonally mediated cancers. Nat Genet. Mar
Surg. Apr 2003;138(4):409–416; discussion 416. 27, 2013;45(4):345–348.
157. Machens A, Brauckhoff M, Holzhausen HJ, Thanh PN, Lehnert 176. Michailidou K, Hall P, Gonzalez-Neira A, et al. Large-scale geno-
H, Dralle H. Codon-specific development of pheochromocytoma typing identifies 41 new loci associated with breast cancer risk. Nat
in multiple endocrine neoplasia type 2. J Clin Endocrinol Metab. Genet. Mar 27, 2013;45(4):353–361.
Jul 2005;90(7):3999–4003. 177. Bojesen SE, Pooley KA, Johnatty SE, et al. Multiple indepen-
158. Ogilvie JB, Kebebew E. Indication and timing of thyroid surgery dent variants at the TERT locus are associated with telomere
for patients with hereditary medullary thyroid cancer syndromes. length and risks of breast and ovarian cancer. Nat Genet. Mar 27,
J Natl Compr Canc Netw. Feb 2006;4(2):139–147. 2013;45(4):371–384.
159. Wells SA, Jr., Gosnell JE, Gagel RF, et al. Vandetanib for the treat- 178. Eeles RA, Olama AA, Benlloch S, et al. Identification of 23 new
ment of patients with locally advanced or metastatic hereditary med- prostate cancer susceptibility loci using the iCOGS custom geno-
ullary thyroid cancer. J Clin Oncol. Feb 10, 2010;28(5):767–772. typing array. Nat Genet. Mar 27, 2013;45(4):385–391.
160. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations 179. Feardon ER, Bommer GT. Molecular biology of colorectal cancer.
of c-kit in human gastrointestinal stromal tumors. Science. Jan 23, In: DeVita VT, Lawrence TS, Rosenberg SA, eds. Devita, Hellman
1998;279(5350):577–580. and Rosenberg’s Cancer: Principles and Practice of Oncology.
161. Demetri GD, von Mehren M, Blanke CD, et al. Efficacy and safety Philadelphia, PA: Lippincott Williams & Wilkins; 2008.
of imatinib mesylate in advanced gastrointestinal stromal tumors.
N Engl J Med. Aug 15, 2002;347(7):472–480.
5 6 4 • G e no m ic s in C l inic a l P r actic e
37.
GENOMICS AND INFECTIOUS
DISEASES: SUSCEPTIBILIT Y, RESISTANCE, RESPONSE,
AND ANTIMICROBIAL THERAPY
Michaela Fakiola, Wei Lu, Sarra E. Jamieson, and Christopher S. Peacock
565
14%
22%
7%
9%
6%
4%
2%
30% 6%
Figure 37.1 Leading causes of worldwide deaths (WHO estimates, 2008). Total deaths—56.8 million.
studies; both approaches have proven successful to a degree. a distinct advantage in that it enables the hypothesis-free
For example, GWLSs have been used to identify regions of investigation of the genetic component influencing dis-
the genome carrying disease susceptibility loci (DSL) for ease pathogenesis. GWAS is a powerful tool that com-
schistosomiasis,8 tuberculosis,9,10 leprosy,10,11 and other infec- pares the frequency of several hundred thousand to more
tious diseases.7 However, as most linkage peaks cover a rela- than a million single-nucleotide polymorphisms (SNPs)
tively broad region (10–20 megabases [Mb]) of the genome across the whole genome in hundreds or thousands of unre-
encompassing numerous genes, only one such DSL has sub- lated cases versus ethnically matched controls. Although
sequently been fine-mapped sufficiently to identify the puta- the population-based design is most commonly used,
tive causative gene, that being the PARK/PACRG gene region family-based designs, such as parent/offspring trios, have
in leprosy.12 The candidate-gene approach has potentially also been employed. In both instances, novel and previously
been more successful, with genes identified as putative can- unsuspected genetic risk factors can be identified, provid-
didates on the basis of homology with previously identified ing a broader view of the gene network and pathways impli-
murine-susceptibility genes or on the basis of plausible bio- cated in host responses.
logical function.7 However, the genetic risk factors involved in GWASs have become feasible due to rapid advances in
that susceptibility to infection are expected to have relatively high-throughput array technologies that enable the geno-
small effect sizes (i.e., an odds ratio <1.5), requiring large typing of a dense map of markers across the genome. The
sample sizes to detect them with credible statistical support. successful design of cost-effective genotyping platforms by
The majority of reported candidate-gene studies in infectious Affymetrix and Illumina has been facilitated by the comple-
disease have generally been underpowered (usually based on tion of the Human Genome Project, the International SNP
fewer than 500 cases), mainly due to limitations of budgets Map,13 HapMap,14 and Human Genome Diversity Project,15
and problems of case recruitment, and remain unreplicated in all of which provide a rich resource of known SNPs within
independent studies. the human genome, the frequency at which they are
observed, and the patterns of linkage disequilibrium (LD;
the non-random association of alleles) between them.
G E N O M E -W I D E A S S O C I AT I O N S T U D I E S Furthermore, the 1000 Genome Project16,17 (1KGP;http://
www.1000genomes.org/), which aims to provide next-gen-
Recently there has been a shift from these approaches to eration sequencing data on 2,500 people from 25 popula-
genome-wide association studies (GWASs). The GWAS has tions, has further enriched the public catalogue of known
www.Ebook777.com
confirming their likely role in KD. Novel associations in are discussed here. Contrary to monogenic disorders, in
immune-related genes were also revealed in GWASs of polygenic conditions, genetic polymorphisms may not
KD in Chinese108 and Koreans.109 More recently, a third be instrumental to disease pathogenesis but instead may
KD GWAS highlighted a novel association with a func- modulate host responses to a pathogen. As such, the effect
tional variant in the IgG receptor gene FCGR2A that could of each associated variant on the phenotype is likely to be
potentially advance knowledge of disease pathogenesis weak or modest. GWAS findings for infectious diseases
and response to treatment.110 An infectious trigger has also have certainly expanded our knowledge of the genes and
been proposed for complex diseases, such as Type I diabe- pathways involved but have not always met the initial high
tes (T1D).111 Results of a large GWAS112 in Caucasians fol- expectations for major advances in the field. Although
lowed by large scale resequencing113 to identify genetic risk there is compelling epidemiological evidence for the role
factors underlying T1D identified several rare variants in of genetics in infectious disease susceptibility, power limita-
the interferon induced with helicase C domain 1 (IFIH1) tions due to small sample sizes is one of the major problems
gene that independently decreased T1D risk. This is of of genetic study designs.124 When small genetic effects are
interest, as the IFIH1 gene encodes the cytoplasmic RNA expected (i.e., OR ≤1.5), sample size is the crucial factor
sensor MDA5, which senses and initiates antiviral activity determining the power of the study to detect the effect. As
against RNA viruses. These results lend support to the the- shown by Burgner and colleagues,124 in order for a study to
ory that T1D could be triggered by a viral infection, leading have sufficient power to detect association at common vari-
to the possibility of prevention via vaccination. ants (i.e., minor allele frequency [MAF] >0.1) with mod-
Intriguingly, a number of GWAS variants identified for est effect size (i.e., OR = 1.5), at least 1,500 case-control
infectious diseases coincide with association signals reported pairs or case-parent trios are required. For example, a very
for common noncommunicable diseases, such as psoriasis and large cohort of patients and controls (>11,000 individuals)
inflammatory bowel disease. This might indicate a balance in had to be employed in the Thye and colleagues77 study of
evolutionary genetics in which genetic variants that confer pro- TB to allow identification of common variants with small
tection against pathogens also predispose negatively to exac- effect (OR = 1.19). In contrast, a much smaller GWAS
erbated inflammatory responses.114 For example, it has been (326 individuals) of treatment response to HCV-infection
suggested that people who have psoriasis are more likely to successfully detected association at genome-wide signifi-
carry variants that confer protection against HIV-infection.115 cance due to the much larger effect of the locus detected
Variants associated with HIV infection—including the amino (OR = 18.5–27.2).61 Finally, a GWAS of HIV-infection
acid residues at positions 67, 70, and 97 within HLA-B, focusing on a group of rapid progressors has highlighted the
the causative rs67384697 variant at HLA-C, the HCP5 use of extreme phenotypes as a powerful tool in identifying
rs2395029 variant, and the non-synonymous rs1576 SNP genetic associations, even with a small sample size.41
at PSORS1C3—were found to be major genetic risk factors Many of the reported GWASs for infectious diseases
of susceptibility to the autoimmune diseases of psoriasis and have used cohorts of smaller size (from few hundreds to
psoriatic arthritis.116–118 There have also been cases of remark- a couple of thousand individuals) compared to noncom-
able overlap between the genetic loci underlying communi- municable diseases (some of which have used tens of
cable and noncommunicable disease. For example, results of thousands), partly reflecting difficulties in recruiting large
GWASs for leprosy compared to those for Crohn’s disease sample sizes for infectious diseases in endemic countries.125
(CD) and ulcerative colitis, two common inflammatory This has resulted in association signals that approach but do
bowel diseases, implicate a common immunological pathway not always reach genome-wide significance (P = 5 × 10–7).22
in the development of all these diseases.22,119–122 This is of inter- This strict statistical cutoff aims to filter the plethora of
est, as leprosy and CD are both granulomatous diseases, and observed associations generated by the GWA approach in
the observation of shared genetic risk factors provides support order to minimize the reporting of false discoveries. A direct
to theories proposing a potential causative role for mycobacte- consequence is that truly positive associations with lower P
rial infection in the development of CD.123 values can be missed. Many GWAS have compensated for
the lack of genome-wide significance by employing a com-
bined analysis of the discovery and replication cohorts.
GWA S — C H A L L E N G E S A N D S U C C E S S E S However, further evidence of association with interest-
ing candidates should be sought in independent studies in
Genome-wide association studies for diseases of a poly- order to increase confidence in the findings. Overall, the
genic nature encompass specific challenges, some of which ORs reported in GWASs for infectious diseases range from
3000
$ 100.0
2500
$ 10.0 2000
1500
$ 1.0
$ 0.0 0
99
00
01
02
03
04
05
06
07
08
09
10
11
12
19
20
20
20
20
20
20
20
20
20
20
20
20
20
Figure 37.2
Graphical representation of the rapid decrease in sequencing costs leading to an increased number of completed genome projects. The
availability of new technology and the advent of next-generation sequencing have dramatically reduced sequencing costs. Estimates of costs are
based on the generation of a megabase of raw sequence data sourced from the NHGRI (www.genome.gov/sequencingcosts), and not the multiple
times coverage of finished sequence required for genome assembly. The completed genome data are taken from the Genomes Online database
(http://www.genomesonline. org).
diagnostic laboratories when the U.S. Food and Drug size of a DVD player (GridIon), which can sequence a
Administration (FDA) began testing the practicality of bacterial genome in a day. It is anticipated that availabil-
these machines through placement of MiSeq instruments ity of this technology will lead to more widespread use
in both state and federal laboratories.174 As these bench- in clinical and rural environments (www.nanoporetech.
top sequencers become more robust and easier to use, com/).
there is even the possibility of incorporating this technol-
ogy into rural or mobile laboratories, as has been done
with other molecular equipment. B AC T E R I A L G E N O M E
In addition, there are also new contenders bringing SEQUENCING
novel technologies to the NGS market in the form of
third-generation machines that have potential benefits While the majority of diagnostic services use PCR-based
beyond that provided by next-generation machines. The methods in conjunction with cell culture, we are now
first of these to be commercially available, the PacBio from moving into a new era of genetic diagnostics carried
Pacific Biosciences, has applied itself to the microbial out directly on clinical samples. This move is leading
genomics market by mixing high-quality sequence data to faster results, something that can be critical in deter-
obtained from short, accurate sequencing reads (<1000 mining patient treatment or when dealing with an infec-
base pairs) with the assembly information obtained from tious disease outbreak. The potential of whole-genome
longer (<10,000 bp) reads. The combination of these read sequencing adds another major advantage to the clinical
types leads to improved genome assembly. The develop- arena. Whereas phenotypical and molecular tests have
ment of nanopore-sequencing technology also promises to be targeted to answer one issue, such as identifica-
the possibility of cheap, highly portable, easy-to-use tion, isolate-typing, antibiotic resistance, response to
machines that will deliver results in real time directly vaccination, or pathogen surveillance, whole-genome
to a laptop or mobile device. Oxford Nanopore recently sequencing is an all-encompassing test that can simulta-
unveiled two highly portable machines, one not much neously address all these requirements.175,176 Furthermore,
larger than a USB flash drive (MinIon), and the other the whole-genome sequencing requires no prior knowledge
INTRODUCTION I N N AT E I M MU N I T Y
The analysis of gene expression in tissues, cells, and biologi- The innate immune system is central to the type of the
cal systems has evolved in the last decade from the analysis immune response that is mounted in a given situation. This
of a selected set of genes to an efficient high-throughput, system acts effectively to sense altered signals without pre-
whole-genome screening approach of (potentially) all genes vious exposure to a pathogen such as a bacterium or virus.
expressed in a tissue or cell sample. This development allows The innate immune system controls and assists the acquired
an open-ended survey to identify comprehensively the frac- antigen-specific immune response represented by T and B
tion of genes that defines the samples’ unique biology. This lymphocytes, which both carry an infinite number of recep-
discovery-based research provides the opportunity to char- tor specificities that need to be exploited when necessary.
acterize either new genes with unknown function, or genes The influence of the innate immune system on the adaptive
not previously known to be involved in a biological process. immune response has long been recognized and used for its
The latter category may hold surprises that will sometimes adjuvant effect of (dead) microbes in antibody responses.
urge us to redirect our thinking. Besides a more detailed Since this system is also intimately linked to inflammatory
understanding of basic processes in immunology, microar- and modulatory processes, many human diseases result
ray studies have increased our insight into the immunologi- from an aberrant primary or secondary innate immunity.
cal mechanisms that contribute to chronic immune-related The application of genomics technology provides answers
diseases. Of particular interest is the overlap of genes to basic questions related to the regulation of innate and
identified in these diseases with genes that are induced in adaptive immunity. The extensive experience that immu-
response to a microbial stimulus. Moreover, genome-wide nologists have with cellular aspects of immunology is
studies have provided evidence for the existence of geneti- highly beneficial in the generation of homogeneous cell
cally determined differences in immune activity that might populations for genomics analyses of specific cell subsets.
have consequences for susceptibility and outcome of Within such a framework, we may gain an understanding
immune-related diseases. of immune deregulation in autoimmune diseases, inflam-
Large-scale gene-expression and genome-wide analysis matory responses, and certain tumors.
is of great relevance in the field of immunology to gener-
ating a global view of how the immune systems attacks
R E S P O N S E P RO G R A M S O F T H E
invading microorganisms, maintains tolerance, or creates a
I N NAT E I M MU N E S YS T E M TO
“memory” for past infections. The exciting part of microar-
M I C RO O RG A N I S M S
ray studies is that the many data points that are generated
give rise to unpredictable and unexpected results, which The innate immune system distinguishes between “self ” and
may lead to new insights in immunology. The present chap- “non-self ” by the creation of a variety of pattern-recognition
ter reviews genomic research to characterize gene expres- receptors, called “Toll-like receptors” (TLRs), that rec-
sion in immunology to broaden our understanding of the ognize conserved motifs on pathogens that are not found
biology of immunological processes and immune-related in higher eukaryotes. Each TLR member recognizes and
diseases. responds to different microbial components. Presently,
591
10 different mammalian TLRs have been recognized, response set.” The TNF/NF-kB regulon thus resembles
and their number is likely to increase. For example, lipo- part of the common innate response to infection (Figure
polysaccharide (LPS) from gram-negative bacteria signals 38.1). Interferon-stimulated genes comprise another set of
through TLR4, while the constituents of the cell wall of co-regulated genes included in the common innate response
gram-negative bacteria, lipoteichoic acid (LTA) and mur- program. These genes include OAS1 (2′,5′-oligo-adenylate
amyl dipeptide (MDP), both activate TLR2. Unmethylated synthetase 1), OAS2, OASL, MX1, MX2, G1P2/ISG15,
CpG-containing bacterial DNA sequences are able to IFI16, ISG20/HEM45, IFIM1, IFN-inducible chemo-
activate TLR9. Double-stranded RNA (dsRNA) mol- kines CCL8/MCP2, CXCL9/MIG, CXCL10/IP10, and
ecules that are produced during the replication process of CXCL11/I-TAC, and metallothioneins. Most of the genes
many RNA and DNA viruses mediate an activation signal induced by viruses appear in this cluster. Many cell types
through TLR3. Signaling through different TLRs (TLR9, are able to induce this interferon-stimulated gene program,
2, and 4) leads to activation of the same transcription fac- including macrophages, fibroblasts, PBMC, DCs, B cells,
tors, such as JNK, p38, and NF-kB, which induces the pro- as well as endothelial and epithelial cells. However, it has
duction of cytokines, such as TNFα, which, after binding now become clear that other organisms such as bacteria
to its receptor in turn, induces more NF-κB activation and and fungi are also able to induce an interferon response
induces many other genes as well (Beutler 2004). program, through activation of the Toll-like receptors and
To gain insight into the host response against different induction of interferon β, a concept that has been high-
pathogens and their products, a meta-analysis of 32 stud- lighted by Hertzog and colleagues (Hertzog, O’Neill, and
ies has been performed by Jenner and Young ( Jenner and Hamilton 2003). But unlike stimulation of TLR3 and
Young 2005), in which stimulation of macrophages, den- TLR4 by poly I:C and LPS, respectively, the bacterial prod-
dritic cells, whole blood, peripheral blood mononuclear ucts LTA and MDP from gram-negative bacteria were not
cells (PBMC), T and B cells, and other non-immune cell able to induce an IFN-response program (Nau et al. 2002;
types were compared. Several different activators, such as Jenner and Young 2005). Thus type I interferons, IFNα and
cytokines, parasites, bacteria, viruses, yeast, and microbial IFNβ, are involved in the expression program of many cell
components that bind to specific TLRs, were used to stimu- types, in response to parasites, fungi, and bacterial and viral
late the different cell types. Based on these comparisons, a stimuli that activate different TLRs. Type II interferon,
common host-transcriptional response was defined. Several IFNγ, is produced by T cells, natural killer (NK) cells, and
different pathogens induced a group of genes with a corre- natural killer-T (NKT) cells in response to mitogens and
lated expression profile involved in inflammation, includ- cytokines. Interferons regulate many important activities
ing TNFα, IL-1β, IL-6, IL-8, NF-kB family members, of macrophages and DCs and are required for an effective
and several chemokines (CCL3/MIP1α, CCL4/MIP1β, innate immune response, which in turn directs the adaptive
CCL20/MIP3α, CXCL1/GRO1, CXCL2/MIP2α/GRO2, immune response.
and CXCL3/MIP2β/GRO3). A comparison of the differ-
ent components known to activate distinct TLRs, such as
TLR2 by LTA and MDP, TLR3 by Poly I:C, and TLR4 A DA P T I VE I M MU N I T Y
by LPS, revealed that all three TLRs activated this set of
inflammatory/chemotactic genes in macrophages ( Jenner Both B- and T-lymphocytes can be recognized as the key
and Young 2005; Nau et al. 2002). These genes are expressed, players in the adaptive immune response as a consequence
not only in macrophages, but also in dendritic cells (DCs), of their molecular and functional adaptation to invading
leucocytes, PBMCs, and whole blood during acute infec- pathogens. Within both types of lymphocytes, subtypes
tion with several bacteria. In the meta-analysis it was shown and distinct stages of differentiation exist that are a conse-
that these genes are induced not only by bacteria, but also quence of their functional adaptation to different antigens.
by the protozoan Leishmania chagasi in macrophages, and Differences between T and B cells are reflected by the
by several viruses in peripheral blood cells and other non– differential expression of many genes. Each lymphocyte
immune cell types such as endothelial and epithelial cells. type has its specific gene-expression signature. The T cell
This group of co-regulated genes was previously recog- signature includes CD2, TCR, and genes encoding TCR
nized in PBMC stimulated with several bacteria and LPS signaling proteins (LAT and fyn). The B cell signature con-
(Boldrick et al. 2002). Because many of these genes are tains genes like CD20, immunoglobulin genes, and genes
known to be induced by TNFα, which increases NF-kB encoding BCR components, such as CD79B. Although
activation, it has been referred to as the “TNF/NF-kB these two cell types are clearly distinct, they also share
TLR3
MyD88 DC
IRAK cytoplasm
TRAM
TRIF TRAF6
B cell
? IKKγ
IKKα IKKβ T cell
NF-kB p50
IRF3/7 IkBs
p65 IFNs
–
IL-12
+
P P P
IRF3/7
Nucleus
p50 p65
Figure 38.1
Ligation of TLR3 and TLR4 activate the IFN- and NF-κB pathway. TLR3 stimulation predominantly induces phosphorylation of
interferon regulatory factor 3 (IRF3) and IRF7, which activate type I interferon genes. TLR4 stimulation predominantly induces the release
of inhibitory IκBs from NF-κB, allowing phosphorylation and translocation of NF-κB to the nucleus, where target genes are activated. TLR3
and TLR4 both activate the IFN- and NF-κB pathways; however, TLR4 stimulation leads to a more profound activation of the NF-kB pathway
with production of proinflammatory cytokines ( Jenner and Young 2005). The innate immune response profoundly affects the adaptive/specific
immune response. In particular, the production of IFNs and IL-12 leads to differentiation of T helper cells to Th1 cells, while IL-4 leads to Th2
differentiation. Gene-expression profiling has identified specific gene-expression profiles of these differentiated T cells that explain their function in
autoimmunity and allergy (Rogge et al. 2000, Chtanova et al. 2001, Hamalainen et al. 2001).
remarkable similarities in their activation program. The homing of activated B cells. One of the features of acti-
“activation gene-expression signature” of T lymphocytes vated T cells is the repressed expression of the genes for
and B lymphocytes includes many genes that are either interferon receptors and for STAT-1, which is involved in
induced or repressed upon activation (Alizadeh et al. interferon signal transduction (Teague et al. 1999b).
2000, Teague et al. 1999a, Glynne et al. 2000). Central to
the activation status is the expression of the transcription
T LY M P H O C Y T E S
factor NF-kB. NF-kB factors play a central role in immu-
nity, and members of this family (NF-kB1, NF-kB2, A nice example of the way microarray studies may help
RelB, c-Rel, and IkBα) are included in the activation address basic questions in immunology is revealed by
gene-expression signature together with a prominent a study that aimed to unravel the pathway of CD28
group of target genes, consisting of cytokines, chemo- co-stimulation in T cells. Co-stimulation of T cells via
kines, and anti-apoptotic genes like A1 (an anti-apoptotic both the TCR and CD28 enables sustained activation
bcl-2 family member) and c-IAP2. Differences that exist and cell cycle entry. Diehn and colleagues (Diehn et al.
are at the level of, for instance, CCR7, CXCR5, and 2002) have used human peripheral blood T cells that were
BCL-2 expression, which are expressed at much higher stimulated with antibodies to CD3 and/or CD28. The data
levels in activated B cells. The elevated expression of che- showed that signals induced by anti-CD3 were amplified
mokine receptors is likely to be involved in lymphocyte by co-stimulation with anti-CD28, with the most striking
Healthy
Non-infected infected RA SLE Cancer
Figure 38.2
Immune response in chronic diseases and infections. Gene-expression profiling of immune cells from healthy individuals at baseline and
challenged with different microorganisms identified, at least in part, genetically determined expression programs, which influence the adaptive
immune response. The genetically determined differences in the innate immune response play a key role in the capacity to clear infections. Disease
profiling identifies patient-specific and disease-specific expression programs that partly overlap with expression programs induced by pathogens,
suggesting the involvement of the innate immune system in the susceptibility to chronic diseases.
biology” (Kitano 2002, Hood 2003): a computational MS (Figure 38.1). Because baseline and microbial-induced
modeling approach aimed at understanding the structure alterations of transcript levels of immune-related genes dif-
and dynamics of cellular and organismal functions. Systems fer among individuals, these expression profiles may inform
biology will certainly become a mainstay of drug develop- the prediction of an individual’s response to infectious
ment, allowing the identification of novel drug targets, agents, and his susceptibility to cancer and autoimmune
followed by target-directed drug effects combined with tai- diseases. The activation of the IFN pathway in SLE and RA
lored therapy. It can further be anticipated that this infor- may therefore reflect an inherited genetic capacity for an
mation will allow us to determine how the disease pathway altered activity of the IFN-response program. Analogous
can be manipulated pharmacologically with greater speci- to these findings, the most consistent intrinsic variable
ficity than is achieved by current drugs. gene in healthy donors, DDX17 (Whitney et al. 2003),
Furthermore, differences in responses between indi- was selectively expressed in a subgroup of RA patients who
viduals might provide a lead to an explanation of how showed a high degree of inflammation and high expression
genetic variation affects the immune system. Evidence of IFN-responsive genes (van der Pouw Kraan TC et al.
exists that these inter-individual differences are geneti- 2003a).
cally determined. One of the distinctive gene-expression In the near future, new analysis methods will evolve to
programs that varies among individuals involves the IFN/ further improve the identification of relationships between
STAT signaling pathway (Radich et al. 2004, Whitney et al. the expression levels of many genes, to identify new signal-
2003, Boldrick et al. 2002). These findings make it tempt- ing pathways, determine genetic differences, and further
ing to speculate that inter-individual differences that exist improve prediction of disease susceptibility and subclas-
in healthy controls might explain the heterogeneity that is sification of patients within a disease—finally leading to a
observed in immune-related diseases such as SLE, RA, and more effective and personalized treatment.
603
phenotypical expression. Disorders with well-defined
genetics and common mutations represented in a signifi-
cant proportion of affected individuals are particularly well
suited to this form of analysis. Genomic DNA can be eas-
ily extracted from newborns’ blood specimens by inexpen-
sive, automated methods, further enhancing the feasibility
of this approach[11]. Traditional molecular methodologies,
however, analyze only one gene or mutation at a time,
requiring separate assays for each disease. Thus, while excel-
lent in a research setting, these procedures are not amenable
to population-based, high-throughput newborn-screening
programs. They are primarily used clinically as second-tier
confirmatory assays for such disorders as cystic fibro-
sis, sickle-cell anemia, and other hemoglobinopathies,
and medium-chain acyl-CoA dehydrogenase (MCAD)
deficiency[12].
The recent application of tandem mass spectrometry
(MS/MS) to the specimens on Guthrie cards has further
expanded the ability of newborn-screening programs to
identify congenital disorders. MS/MS recognizes amino
acids and acylcarnitines by their characteristic mass spec-
tra and is thus able to identify more than 20 inborn errors
of metabolism not previously a part of newborn-screening
programs[13,14]. MS/MS also has the advantage of being
able to measure concentrations of multiple compounds,
including amino acids and acylcarnitines, in the same
sample, greatly increasing the utility of this method. MS/
MS has revolutionized the field of newborn-screening by
allowing the detection of multiple metabolic disorders in
a single step. Previously, each screened disease required a
separate assay. This was true not only for older assays, like
the bacterial-inhibition assay for PKU, but also for newer
molecular assays for genetic disorders. Thus, MS/MS is a
Figure 39.1
Example of a Guthrie card for collection of blood specimens
for newborn screening (from the Ohio Department of Health, USA). straightforward, efficient, and cost-effective method for
Blood samples are collected within the dashed circles on filter paper the detection of diseases involving altered amino acid and
attached to the top of the form. (Adapted with permission from acylcarnitine metabolism. Additional work is expanding
‘Applications in Clinical Pediatrics’, by Michael R. Konikof and Michael
D. Bates, p. 425–436. In ‘Genomics and Clinical Medicine’, Eds. Kumar, the range of the utility of MS/MS to include identifica-
D & Weatherall, DJ. Oxford University Press, NY, 2008.) tion of hexose monophosphates in patients with galactose-
mia[14] and of steroids in patients with congenital adrenal
the designs of these programs with respect to legislation, hypoplasia[15].
funding mechanisms, and follow-up also vary. Although The ability to extract DNA from the newborn blood
the cost-effectiveness of newborn-screening for rare con- specimens and recent advances in microarray technology
ditions has been debated in the international community, offer the opportunity for primary screening of a vast num-
the creation of new programs and the expansion of exist- ber of genetic disorders that could not previously be iden-
ing programs continue[10]. tified in the newborn. DNA microarrays, which contain
Work in human genetics has elucidated the under- representations of thousands of genes, are fabricated either
lying molecular defects responsible for many inherited by robotically spotting gene fragments onto a glass slide or
congenital disorders. By applying molecular techniques by synthesizing oligonucleotides onto glass by photolitho-
to newborn-screening, many genetic diseases can be graphic techniques[16]. Microarrays can be used to measure
detected shortly after birth, before their full or irreversible the expression of thousands of genes simultaneously by the
specific hybridization of nucleic acid strands having com- Hybridization signals were detected corresponding to the
plementary nucleotide sequences. This stringent hybridiza- various known mutations, indicating that the microarray
tion can also be applied to the analysis of genomic DNA to correctly identified the tested gene mutations. The entire
detect SNPs, point mutations, insertions, and deletions[17]. process, from DNA extraction to data analysis, can be auto-
Microarrays have been designed to test these concepts mated, and with the addition of multiplex amplification,
using well-characterized genetic disorders such as hemo- several disorders may be assayed on a single microarray. This
globinopathies, α1-antitrypsin deficiency, and factor V will allow population-based first-tier molecular screening
Leiden[18]. Oligonucleotides representing multiple disease for a large number of genetic disorders, which has not pre-
alleles in the above disorders were spotted onto a microar- viously been possible.
ray, and DNA with known mutations was extracted from Primary newborn-screening utilizing microarray tech-
newborn blood specimens and hybridized to the array. nology offers several advantages over traditional methods
Advantages
Screen many genes in single assay.
Identify any mutations rather than only common or previously characterized mutations. Can be automated for high-throughput.
Potential for increased specificity with high sensitivity.
Can detect heterozygotes (disease-gene carriers).
Allows genotyping of modifying genes to improve phenotype prediction of single-gene disorders. Less affected by clinical factors
( prematurity, diet, medications, etc.).
Limitations
Cost.
Amplification of genomic DNA by polymerase chain reaction (PCR) can result in artifacts or contamination. May identify
polymorphisms—mutations that have no effect on gene or protein function.
Does not identify alterations in post-transcriptional or post-translational processes.
Does not identify non-Mendelian genetic events (e.g., gene duplications, epigenetic influences, and non-random X chromosome inactivation).
Not applicable to disorders with no known genetic basis (e.g., congenital hypothyroidism).
SOURCE: Adapted with permission from ‘Applications in Clinical Pediatrics’, by Michael R. Konikof and Michael D. Bates, pages 425–436. In ‘Genomics and
Clinical Medicine’, Eds. Kumar, D & Weatherall, DJ. Oxford University Press, NY, 2008.
(Table 39.2). Just as MS/MS can measure multiple analytes groups where uncommon mutations are over-represented.
from a single sample, DNA microarrays can screen many Microarrays also have the potential to be more specific,
genes for mutations in a single step, with little additional while maintaining sensitivity, than traditional metabolite-
cost for each disorder. The major limiters at present are the or protein-based assays, which commonly use threshold val-
feature density of the microarray and the resolution of laser ues that yield ratios of false-to-true positives ranging from
scanners. The completion of the sequencing of the human 10:1 to 75:1[22]. More-specific tests would reduce the rate of
genome facilitates the design of “resequencing” microar- false positives, thus decreasing the need for follow-up testing
rays to screen for and identify any mutations of a specific and its resultant costs, not to mention the increased anxi-
gene, rather than only common or previously character- ety for parents and medical caregivers. Microarrays are able
ized ones. Because the entire process can be automated with to distinguish disease-gene carriers (heterozygotes) by the
high throughput, microarrays are well suited to population selective detection of hybridization signal intensity and the
screening. This is largely facilitated with the availability of simultaneous probing for wild-type alleles. Carriers could
several next-generation sequencing platforms and com- then be excluded from follow-up testing and laboratory
putational analytical tools. DNA microarrays are able to reporting, which now diverts resources from detecting dis-
screen for certain disorders that are not amenable to tradi- eased individuals, the primary mission of newborn-screening
tional protein-based methods, such as fragile-X syndrome programs. Alternatively, carriers could be easily identified
(through identification of untranslated, triplet-repeat DNA and reported, if this was appropriate from ethical and/or
sequences)[19], the various severe combined immunode- legal perspectives. In addition, microarrays could be used to
ficiencies (through detection of excised T-cell receptor improve phenotype prediction of Mendelian “single-gene”
sequences rather than of various T-cell related proteins)[20], disorders by genotyping known modifying or interacting
and cystic fibrosis. For example, the ΔF508 mutation of genes, which could improve prediction of disease severity
the cystic fibrosis transmembrane conductance regulator that impacts treatment strategies, planning, and counsel-
(CFTR), which accounts for 80% of mutant cystic fibrosis ing[23]. Finally, DNA-based analysis is not affected by clini-
alleles in the Caucasian population (but only 30% of mutant cal factors that may interfere with analysis of blood samples
alleles in Asians or other non-European individuals), is the for metabolites, hormones, or proteins. Factors such as pre-
only mutation assayed in the newborn-screening program maturity, medications of the mother or child, or parenteral
of at least one state in the United States[21]. There are now nutrition will not affect DNA-based analysis.
more than 1,000 known mutations in the CFTR gene, and DNA microarrays do have several limitations that must be
all of these could be assayed using one microarray, which overcome before their widespread use in newborn-screening
would improve detection of cystic fibrosis cases in ethnic programs will become practical (Table 39.2). Cost has
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X
Figure 39.2 Karyotype showing three copies of chromosome 18, consistent with Edwards syndrome.
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X
Figure 39.3 Karyotype showing a single X chromosome, and no Y chromosome, consistent with Turner syndrome.
0.4th centile on the growth chart. She was noticed to have trisomy 21 (mosaic Down syndrome) (47, XY +21/46, XY;
a low posterior hairline, multiple skin nevi, and lack of Figure 39.4). The clinical geneticist suspected a diagnosis
breast development. Cardiac examination was normal. The of mosaic Down syndrome from the history and clinical
pediatric endocrinologist requested a karyotype (45, XO), examination findings[33]. In a “routine” karyotype, it is usual
which confirmed the clinical suspicion of Turner syndrome to analyze up to 10 cells. If mosaicism is suspected, the clini-
(Figure 39.3). The possibility of Turner syndrome should be cian can request the analysis of more cells.
considered in any girl who presents with short stature and
delayed puberty. This child had some additional clues in the
FLU O R E S C E N T I N S IT U
form of low posterior hairline and multiple skin nevi to sug-
H Y B R I D I Z AT I O N—FI S H
gest this diagnosis[32].
Fluorescence in situ hybridization (FISH) is the visualiza-
tion of specific DNA sequences in metaphase chromosomes
CASE 3
and interphase cells using a fluorescent microscope. Both
A six-year-old male child was referred to the clinical geneti- numerical and structural chromosomal rearrangements
cist in view of his developmental delay and dysmorphic fea- can be detected in either dividing or non-dividing inter-
tures. Antenatal period was uneventful. There was history phase nuclei using probes ranging from whole chromo-
of poor feeding in the neonatal period, which improved some “paints” to individual gene-specific probes. The target
with time. He had global developmental delay, especially in sequence can range from one to several hundred kilobases[34].
the area of expressive speech. On examination in the genet-
ics clinic, the child was noticed to have upslanting palpe-
CASE 4
bral fissures, mid-face hypoplasia, unilateral single palmar
crease, and joint laxity. Cardiac examination was normal, A four-year-old girl was referred to the pediatric cardiolo-
and an echocardiogram did not identify any abnormali- gist in view of her cardiac murmur. In the history, she had
ties. The clinical geneticist requested a karyotype with an mild developmental delay. The child was noticed to be quite
extended cell count (30 cells), which showed a mosaic friendly and had some facial dysmorphic features, including
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X Y
→
Figure 39.4 Karyotype showing the cell line with three copies of chromosome 21 consistent with mosaic Down syndrome.
CASE 5
1.40
1.20
1.00
0.80
EHMT1(i)
SMARCB1(ii)
RAB36(ii)
SLC25A19
SMARCB1(i)
EHMT1(ii)
MICAL3
DKFZp566
ATDA1(il)
ATDA1(i)
AI651963
SHANK3
PPP1A3B
USP19
CUGBP2
RAB36(i)
SNAPD3
GATA4
GATA3
APH3AL(i)
KLKB1
GNAZ
BID
0.60
MSRA
NEBL
PPIL2
ARSA
CRK
APH3AL(ii)
TOP3B
FRG1
GEMIN4
HIC2
0.40
TXNRD2
CDC45L
CLTCL1
TBX1(ii)
KLHL22
SNAP29
TBX1(i)
DGCR9
CLDN5
PCQAP
LZTA1
GP1BB
ZNF74
HIRA
0.20
0.00
Figure 39.7
Multiplex Ligation Dependent Probe Amplification (MLPA) using the P250 probe mix showed a deletion at 22q11.2 region consistent
with the 22q11.2 deletion syndrome.
in children with developmental delay and dysmorphic fea- stature. However, dysmorphic features did not suggest any
tures. A number of dedicated systematic databases have specific genetic diagnosis. An MRI brain scan was normal.
now been set up that largely depend on array-CGH; for A karyotype done previously showed an apparently bal-
example, the University of California, Santa Cruz (UCSC; anced de novo translocation between chromosomes 7 and
www.UCSC.ed) and the Wellcome Trust Sanger Institute 13 (46,XX,t(7;13)(q21.2;q12.3); Figure 39.8a). This did
in Hinxton, England (www.sanger.ac.uk). In the United not fully explain developmental delay and mild dysmorphic
Kingdom, the DECIPHER, an acronym of DatabasE of features.
Chromosomal Imbalance and Phenotype in Humans using Following discussion with a clinical geneticist, the
Ensembl Resources, was initiated in 2004 at the Sanger pediatrician requested an array-CGH analysis for the child.
Institute and funded by the Wellcome Trust. It docu- Array-CGH analysis showed that, in the regions of the
ments submicroscopic chromosome abnormalities and breakpoints on chromosomes 7 and 13, there had actually
maps them to the human genome using Ensembl or the been a loss of genetic material (DNA) (Figure 39.8a: loss
UCSC Genome Browser. Its aims are to increase medical of genetic material indicated in red on the diagram of the
and scientific knowledge about chromosomal imbalances, chromosomes); thus the translocation was not balanced. It
improve medical care and genetic advice for affected indi- was agreed that the loss of genetic material from the long
viduals/families, and help facilitate research into the study arms of chromosome 13 and chromosome 7 was causally
of genes that influence human development and health. It is linked with the phenotype and its severity.
a database that is widely utilized to interpret chromosomal Array-CGH analysis of DNA at approximately 1
imbalances detected by array-CGH in routine diagnostic Mb resolution using the BlueGnome CytoChip (v1.01)
laboratories. array showed two interstitial deletions (Figure 39.8b).
An 8.42 Mb deletion of the long arm of chromosome 7
was detected. A 0.2 Mb deletion of the long arm of chro-
CASE 7
mosome 13 was also detected. These deletions appears to
A three-year-old boy was referred to the pediatrician in view be at, or close to, the breakpoints of the apparently bal-
of his developmental delay and some dysmorphic features. anced de novo translocation t(7;13)(q21.2;q12.3) previ-
On examination, the child had microcephaly and short ously reported in this child. It appears therefore that this
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 X Y
B 0
16
32
48
63
79
95
i) Chromosome 7
111
127
143
159
2.00 1.60 1.20 0.80 0.40 0.00 0.40 0.80 1.20 1.60 2.00
Chromosome 7
Log2 Ratio Ch1/Ch2
01100200_top-01100200_bottom.bsn - 14/06/2006
0
11
23
34
46
57
68
ii) Chromosome 13
80
91
103
114
2.00 1.60 1.20 0.80 0.40 0.00 0.40 0.80 1.20 1.60 2.00
Chromosome 13
Log2 Ratio Ch1/Ch2
01100200_top-01100200_bottom.bsn - 14/06/2006
Figure 39.8
(A) Karyotype showing an apparently balanced reciprocal translocation between the long arm of chromosome 7, and the long arm of
chromosome 13, t(7;13)(q21.2;q12.3). (B) Array CGH analysis using the BlueGnome CytoChip 1Mb array chowing a 8.42 Mb deletion of
chromosome 7 and a 0.2 Mb deletion of chromosome 13.
patient has two cryptic deletions at, or near, both break- then be more fully assessed. This child is at 50% risk of pass-
points of the apparently balanced de novo translocation, ing on the abnormal chromosome 15 to any offspring she
t(7;13)(q21.2;q12.3). Submicroscopic imbalances in may have.
patients with de novo apparently balanced translocations
presenting with an abnormal phenotype have recently
CASE 9
been revealed at or close to the breakpoints by array-
CGH in about 50% cases[43, 44]. A four-year-old boy was seen in the pediatric clinic in view
of his global developmental delay, hypotonia, and dysto-
nia. He was born at term, and there were no concerns in
CASE 8
the antenatal or neonatal period. There was no significant
A five-year-old child was referred to the clinical geneticist family history. On examination, the child had mild muscu-
in view of developmental delay and dysmorphic features. lar hypotonia with normal reflexes. There were no signifi-
The child was born at 39 weeks’ gestation, and there were cant dysmorphic features. The pediatrician had previously
no concerns during the antenatal or neonatal period. The arranged a karyotype, which did not identify any abnor-
parents first became concerned when the child was not sit- malities. An array-CGH was requested.
ting without support even by one year of age. On follow-up, Array-CGH analysis of DNA from this child using the
the child had delay in most of her motor milestones, as well BlueGnome CytoChip Oligo ISCA (v2.0) 8x60k array
as speech. An MRI brain scan was arranged when she was showed a copy number gain of the sequences detected by
three years of age, which was normal. A conventional karyo- the oligonucleotide probes located from 130,671,707 bp
type was requested by the pediatrician, which did not iden- to 134,868,378 bp on chromosome 11, and a copy number
tify any abnormality. loss of the sequences detected by the oligonucleotide probes
On examination, the child had normal growth param- located from 230,451 bp to 4,077,584bp on chromosome
eters, including head circumference, for her age. The geneti- 12 (Figure 39.10a,b,c). This equates to a 4.2 Mb terminal
cist noticed facial dysmorphic features, including high copy number gain of the long arm of chromosome 11 from
anterior hairline, prominent forehead, and epicanthic folds. 11q24.3 to 11qter, and a 3.8 Mb terminal copy number
The systemic examination was normal. In view of the devel- loss of the short arm of chromosome 12 from 12pter to
opmental delay and dysmorphic features, and the absence 12p13.32. In DECIPHER, 13 HGNC genes, 10 of which
of any definite clues to suggest a single-gene disorder, the are OMIM genes, are listed in the duplicated region of chro-
geneticist requested array-CGH analysis. mosome 11, and 28 HGNC genes, 23 of which are OMIM
The array-CGH analysis of DNA from the above genes, are listed in the deleted region of chromosome 12.
child using the BlueGnome CytoChip Oligo ISCA The abnormality detected in this child therefore
(v2.0) 8x60k array showed a deletion of the sequences appears to consist of the loss of a small distal segment of
detected by the oligonucleotide probes located from chromosome 12 short-arm material, and its replacement
74,461,225 bp to 78,205,204 bp on chromosome 15 by a small distal segment of chromosome 11 long-arm
(Figure 39.9a,b). This equates to a 3.7 Mb interstitial material, probably as the result of an unbalanced translo-
deletion of the long arm of chromosome 15 at q24.1 to cation. The child is therefore trisomic for the distal seg-
q24.3. In DECIPHER, 47 genes, of which 37 are listed as ment of the long arm of chromosome 11, 11q24.3->qter,
Online Mendelian Inheritance in Man (OMIM) genes, and monosomic for the distal segment of the short arm
map within the deleted region. The 15q24 recurrent of chromosome 12, 12p13.32->pter. The phenotypical
microdeletion syndrome maps within the imbalance. The abnormalities noted in this child are almost certainly the
deletion found in this child overlaps the newly identified result of the chromosome imbalances revealed by aCGH.
recurrent 15q24 recurrent microdeletion syndrome[45, Although there are no patients in DECIPHER with
46]
. Several similar deletions have also been reported in similar copy number gains of chromosome 11, there are
DECIPHER. It appears that deletions of this region are multiple patients with similar deletions of chromosome
associated with clinical problems such as intellectual dis- 12. Phenotypical features noted in these patients include
ability, growth retardation, and dysmorphism. intellectual disability, autism, muscular hypotonia, dysto-
In view of this finding, the next step will be to offer nia, and delayed speech and language development, some
testing to both parents of this child in order to determine of which have also been noted in this patient. This child is
whether the abnormality has arisen de novo or has been at 50% risk of passing on the derived chromosome 12 to
inherited. The clinical significance and recurrence risks can any future offspring.
p13
p12
p11.2
p11.1
q12 q11.2
q14
q21.2 q21.1
q21.3
q22.2
q23
q25.3 q25.2 q25.1
q26.1
q26.2
q26.3
15:102, 531kb −2.00 −1.60 −1.20 −0.80 −0.40 −0.00 0.40 0.80 1.20 1.60 2.00
Log2 ratio Ch1/Ch2
Figure 39.9
(A) Array CGH analysis using the CytoChip Oligo 8x60k (v2.0) array showing a 3.7Mb interstitial deletion at the long arm of
chromosome 15.
(B)
RP11-414J4
Normal chromosome 15
Deleted
chromosome 15
Figure 39.9
(B) Metaphase showing a deletion of the sequences recognized by the RP11-414J4 probe on the long arm of the chromosome 15
homologue using FISH, therefore confirming the array CGH analysis.
In view of this finding, the next step will be to offer in specific panels of microarrays or specific gene mutations
testing to both parents of this child in order to determine in tumors[49]. DNA microarrays can be used to identify
whether one of the parents is a carrier of the balanced mutations and polymorphisms in either single large genes
chromosomal rearrangement. The clinical significance and or panels of genes whose mutation can result in clinically
recurrence risks can then be more fully assessed. similar phenotypes[50]. Such disease-specific microarrays
or gene panels are now commercially available, assisting
in rapid clinical diagnosis of rare disorders and infectious
R E C E N T N EW G E N O M I C diseases[51].
A P P L I C AT I O N S Microarray analyses using genomic DNA may also be
useful for testing panels of polymorphisms or mutations
There are several new genomic techniques that are being that act as modifiers of risk in genetic disorders, or for com-
investigated and rapidly introduced in clinical pediat- binations of genes that result in multigenic disorders. For
ric practice. These include exome-sequencing focusing such analyses, panels of genes might be re-sequenced, or
all exomes of an individual’s genome, deep-sequencing profiles of large numbers of SNPs or copy number varia-
of specific genomic regions, and whole-genome sequenc- tions linked to disease genes or modifier genes could be
ing (see also Chapter 10). Out of these, exome sequenc- analyzed[17]. However, no such specific information could
ing has some clinical utility when the putative gene can be be obtained for use in the promotion of early intervention
located and subsequently identified. This method has led in cases of increased risk.
to the delineation and identification of a few genes for rare
genetic dysmorphic syndromes[47,48]. All these methods are
G E N O M I C C L A S S I FI C AT I O N
now vastly improved at significantly lower costs, with the
O F P E D I AT R I C D I S O R D E R S
availability of new-generation sequencing platforms and
computational analysis tools. (A detailed discussion of the The genomics-based tools developed following the com-
techniques and specific disorders is beyond the scope of pletion of the Human Genome Project are being applied
this chapter.) to determine the genetic basis of pediatric disorders,
Genetic analyses for the identification and character- including congenital malformations. Genomics tech-
ization of genetic disorders are an obvious application of niques are being used to classify many different kinds of
genomics in pediatrics. Most notable development has been diseases (see Chapters 10 & 20) and to determine the
(C)
11q subtelomere probe 12p subtelomere probe
Derivative chromosome 12
Normal chromosome 12
Figure 39.10
(A) & (B) Array CGH analysis using the CytoChip Oligo 8x60k (v2.0) array showing a 4.2Mb copy number gain of the long arm of
chromosome 11, and a 3.8 Mb copy number loss of the short arm of chromosome 12. (C) Metaphase showing a derived chromosome with a partial loss of
the terminal segment of one chromosome 12 long arm material and replaced with a chromosome segment from the short arm of chromosome 11 material.
risk associated with the various classifications in order P H A R M AC O G E N ET I C S
to refine therapeutic approaches. Among the pediatric A N D P H A R M AC O G E N O M I C S
disorders reported to have been studied to date are psy-
An exciting current frontier for genomic medicine is to use
chiatric and developmental disorders, rheumatological
an individual’s genomic information to “personalize” drug
and immune disorders, digestive system disorders (such
therapy through the approaches of pharmacogenetics and
as celiac disease, inflammatory bowel disease, and bili-
pharmacogenomics (see also Chapter 7). Such personaliza-
ary atresia), and renal disorders (Table 39.3). However,
tion will require paying attention to both the patient’s age
most of the work has been done in oncology, in order
and genome, because the targets and toxicities of medica-
to identify prospectively the prognoses associated with
tions change with age[58]. Thus, investigation of developmen-
histologically similar tumors, based on significant differ-
tal pharmacogenetics and pharmacogenomics is needed to
ences in their gene-expression signatures. The first exam-
understand relevant age-related changes in the expression
ple of this approach in pediatric oncology demonstrated
of drug targets and in drug metabolism, so that the most
that medulloblastoma gene-expression profiles could be
efficacious drugs with the least risk for adverse reactions for
used, not only to distinguish between different types of
a patient at a given age can be identified.
tumors, but also to predict survival[52]. Since then, studies
have been reported for a variety of leukemias, lymphomas,
and solid-tumor malignancies, particularly brain tumors.
ET H I C A L , L E G A L , A N D S O C I A L
From the standpoint of cancer biology and pathophysiol-
I M P L I C AT I O N S U N I Q U E TO
ogy, the identification of genes differentially expressed in
P E D I AT R I C S
tumors with a poor prognosis might suggest approaches
to the development of novel and potentially more effec-
No discussion of the use of genomics in clinical pediatrics
tive therapies for patients with such tumors. Clinically,
would be complete without a thorough consideration of
such gene-expression profiling provides a potentially
the ethical, legal, and social implications of this emerg-
more objective approach to disease classification than tra-
ing technology (see also Chapter 17). Traditional medical
ditional histopathological approaches.
ethics has been governed by four principles: beneficence
(doing good), non-maleficence (doing no harm), autonomy
(the right to choose), and justice (being fair and equitable).
I N FEC T I O US D I S E A S E S
These principles also apply to the use of genomic informa-
A key challenge in patients with infectious diseases is to tion in pediatrics, although the “usual” ethics may need
detect and characterize pathogens, particularly for organ- to be modified to deal with the unique ethical quandaries
isms that are difficult to culture, such as Mycobacterium found in genomic medicine[59].
tuberculosis, so that therapy can be initiated in a timely Genomic medicine, by generating vast amounts of
manner. Genomics approaches can also be applied to such personal genetic data, poses significant ethical dilemmas,
challenges once the relevant pathogen genome(s) have both for individuals and for society, and the potential for
been sequenced (see Chapters 11 and 37). This sequence abuse of this information is real. A major concern is that
information is necessary for the preparation of microar- the existence of such detailed genetic or genomic infor-
rays that can detect genomic DNA from the pathogen(s) mation could result in “genetic discrimination” to deny
of interest[53]. The use of microarrays can potentially access to health and life insurance or employment[60]. In
decrease the time necessary to identify a central nervous fact, public pressure in the United States has resulted in
system pathogen from many days currently, to just a few the passage of laws prohibiting genetic discrimination and
hours[54], which would allow targeted therapy to begin regulating the use of genetic information, and U.S. fed-
sooner. These assays can also be used to identify patho- eral legislation has been debated in many recent sessions
gen subtypes and the presence of antimicrobial-resistance of Congress[61]. Although genetic discrimination may
markers[55]. Examples of such approaches include the use be a widespread public concern, there are few examples
of microarrays to detect, subtype, and determine rifampin of such discrimination and, as yet, no evidence that it is
resistance for various Mycobacterium species[56], and for common[62]. Most lawsuits that have been filed involve
the sensitive detection of E. coli O157:H7, an impor- cases of actual disease identified by genetic testing, not
tant food-borne cause of hemorrhagic colitis and the pre-symptomatic or predispositional conditions. There
hemolytic-uremic syndrome[57]. are almost no well-documented cases of health insurers’
Cardiology
Kawasaki disease Nomura et al., 2005
Kawasaki disease (response to therapy) Abe et al., 2005
Right ventricular outflow obstructive lesions Konstantinov et al., 2004
Dentistry
Dental caries Paakkonen et al., 2005
Dermatology
Giant hairy nevi Dasu et al., 2004
Hypertrophic and normal scars Tsou et al., 2000
Gastroenterology/Hepatology
Biliary atresia Bezerra et al., 2002
Celiac disease Diasclado et al., 2004
Eosinophilic esophagitis Blanchard et al., 2006
Inflammatory bowel disease Dieckgraefe et al., 2000
Genetics
Duchenne muscular dystrophy Noguchi et al., 2003
Mitochondrial disorders Crimi et al., 2005
Rett syndrome Colantuoni et al., 2001
X-linked myotubular myopathy Noguchi et al., 2005
Infectious diseases
Neisseria meningitides Sun et al., 2000
Parvo virus B19 infection Kerr et al., 2005
Vaccine-derived polio virus Laassri et al., 2005
Neonatology/Obstetrics
Premature rupture of membranes Tromp et at, 2004
Nephrology
Acute renal allograft rejection Sarwal et al., 2003
Focal segmental glomerulosclerosis Schwab et al., 2004
Neurology/Psychiatry
Brain arteriovenous malformations Hashimoto et al., 2004
Neurofibromatosis type 1 Tang et al., 2004
Neurological diseases Tang et al., 2005
Tardive dyskinesia in therapy of schizophrenia Nikoloff et al., 2002
Oncology
Acute megakaryoblastic leukemia in Down syndrome Lightfoot et al., 2004
Acute myelogenous leukemia Ross, M. E. et al., 2004
Acute myeloid leukemia Yagi et al., 2003
Ependymoma Suarez-Merino et al., 2005
Ewing’s sarcoma Okali et al., 2004
Juvenile pilocytic astrocytomas Wong et al., 2005
Large cell lymphomas Thompson et al., 2005
Leukemias Moos et al., 2002
Medulloblastoma Pomeroy et al, 2002; Fernandez-Teijeiro et al., 2004
(continued)
Table 39.3 CONTINUED
collecting or using pre-symptomatic genetic test results are no treatments or preventative measures available, or the
in their underwriting decisions, and little evidence that condition is adult-onset, testing of children becomes more
such testing affects a person’s ability to obtain health controversial, as is the testing of children for carrier status.
insurance[63]. Despite this apparent lack of improper use Here, the need-to-know of the child and parents must be
of genetic information, as our knowledge of genomics weighed against the adverse effects of a harmful diagnosis
increases, these dilemmas are likely to increase in both fre- on family relationships and the child’s self-esteem and devel-
quency and importance. Deciding what to do about such oping social identity. In addition, mature adolescents who
predispositions will depend on the likelihood of occur- desire genetic testing for future reproductive planning or
rence in at-risk individuals and the natural history of the other reasons must be considered differently than younger
disease. The challenge for clinicians will be to discuss with children who have not yet reached the stage of formal
their patients the possible adverse social consequences of operational thought and thus are unable to give assent. In
testing so that patients can make informed decisions about any case, thorough genetic counseling is a necessity for any
whether or not to proceed with the testing[64]. genetic testing of children, and should include the child, the
Genetic testing of children presents unique ethical con- parents, and the siblings. In most of these situations, genetic
siderations. Such testing may be diagnostic or predictive (see testing of children should be deferred until the child reaches
also Chapter 14). Diagnostic genetic testing refers to testing adulthood or can make reproductive decisions[65].
of a symptomatic child’s genetic material to establish a med-
ical diagnosis. This type of testing, with direct benefit to the
child who is symptomatic, is generally non-controversial. S U M M A RY
Predictive genetic testing, on the other hand, refers to the
testing of asymptomatic children for genetic conditions The application of genomic medicine approaches to the
that may present months to years in the future, if at all. care of children is not as advanced as for adults because
Predictive testing can be pre-symptomatic (i.e., virtually all of the relatively small numbers of affected individu-
children with the specific genotype will develop the disease) als for pediatric as opposed to adult disorders, as well as
or predispositional (i.e., children with the specific genotype because of the unique ethical and social implications for
are at risk of developing the disease, but not certain to do such applications in the care of children. However, clini-
so). Predictive genetic testing can be performed for diseases cal pediatrics offers unique and challenging opportunities
with either childhood- or adult-onset, and this testing may where historically all conventional and new genetic and
offer direct, indirect, or no benefit to the child. If there are genomic laboratory methods are increasingly applied.
treatments or preventative measures available for the con- This chapter has highlighted a few classic case scenarios
dition being tested, predictive genetic testing of children wherein both conventional and newly emerging genomic
should be offered, and, in some cases, strongly advised, simi- methods of array comparative genomic hybridization are
lar to considerations given to neonatal screening. If there demonstrated to have contributed to clinical diagnosis
623
for about half of cataract variability, while twin studies also retinal morphology and function can be improved and
show significant heritability for both cortical (53–58%) that disease progression can also be altered.
and nuclear cataracts (48%) (Heiba et al., 1995; Hammond
et al., 2000, 2001). POAG also has a strong genetic com-
MO N O G E N I C D I S E A S E A N D
ponent, which is known thanks to family and twin concor-
O P HT H A L MO L O GY
dance studies and a range of approaches including the study
of single-gene disorders and association studies. For a long The impact and importance of classical Mendelian genetics
time, the study of common disorders lagged behind that of in ophthalmology has been profound and is well illustrated
monogenic disorders. However, successful genome-wide using the examples of ocular patterning defects, congenital
searches have identified many potential loci—for example, cataract, and retinitis pigmentosa.
for POAG (9q22, 20p12), cataract (6p12-q12), and AMD
(1q31, 10q26 and 17q25) (Wiggs et al., 2004; Iyengar et al.,
Defects of Ocular Patterning
2004; Weeks et al., 2004; Ozel et al., 2014); and have identi-
fied predisposing variations or polymorphisms in a number Defects of early ocular development are often lumped into
of genes—such as for age-related cataract in EPHA2 and a single heterogeneous group termed microphthalmia. This
SLC16A12, and for POAG in Col8A2 (Yang et al., 2013; represents not one, but a spectrum of clinical entities which
Zuercher et al., 2010; Desronvil et al., 2010). The increasing can be referred to as MAC (microphthalmia, anophthal-
impact of genome-wide searches in the study of common mia, coloboma). These may be unilateral or bilateral and
disease is illustrated by the recent success of a combined may range from complete anophthalmia, through microph-
positional/candidate-gene approach that has identified thalmia (defined as an eye with an axial length of <19.5 mm
genes that make an important contribution to AMD (see at the age of one year) to nanophthalmia (defined as a short
Chapter 46). Overall, as a consequence, there is now a real- axial length, high hypermetropia, and shallow anterior seg-
istic prospect that we are beginning to better understand ment, but normal visual function). Such defects result from
groups of conditions whose complex multigenic nature has interruptions in the earliest processes of ocular develop-
previously hindered progress. Identification of the comple- ment, including the formation of the optic vesicle (which
ment system in AMD pathogenesis, for example, now forms as neuroectodermal evaginations of the forebrain),
presents novel therapeutic targets for a disease that was pre- the induction of the lens placode/vesicle, and the differen-
viously difficult to study and treat. tiation of the neural retina and retinal pigment epithelium
The need for a working knowledge of genetics is (RPE). Since, in many cases, there may be an associated fail-
becoming increasingly mainstream as that science ure of the optic fissure to close (resulting in the formation
actively contributes to clinical practice. For conditions of a colobomatous defect), this, too, is seen as a key event.
such as retinoblastoma, inherited retinal dystrophies, The underlying cause of ocular developmental abnor-
and congenital cataract, genetic testing already has a part malities is varied and may be associated with a large num-
to play in the diagnosis, counselling, and management ber of single-gene disorders as well as developmental
of patients and their families. Such services are becom- abnormalities of a broad etiology, including chromosomal
ing more comprehensive and widely available with the abnormalities, copy number variation, maternal infection
advent of novel technologies. As genetic testing con- or toxin ingestion (e.g., anti-epileptic drugs or alcohol),
tributes to the clinician’s ability to diagnose and has a environmental influences, and fetal disruptions. There has
potentially profound impact upon decision-making for been considerable progress in understanding the effects
patients and families, the power to alter disease progres- of copy number variation on ocular developmental dis-
sion or management has seldom been so tangible. This orders (Bardakjian et al., 2009; Raca et al., 2011). An
is exemplified in the progress that has been made in the increasing number of single-gene defects are recognized
development of novel gene-based therapies. Viral-based as underlying these conditions, both in humans and in
gene-delivery approaches to the retina and/or retinal other mammals. CHX10 and RAX (neural retina-specific
pigment epithelium diseases have now been attempted homeodomain-containing transcription factors) were
in rodents and larger animals such as the RPE65- and identified using a candidate-gene approach and have been
RPGRIPL1-deficient dogs (Acland et al., 2001; Lhériteau defined in only a very small number of patients (Voronina
et al., 2014). Most recently, promise has been shown in et al., 2004; Percin et al., 2000). Homozygosity approaches
trials in humans (e.g., Maguire et al., 2009; Bainbridge recognized a number of other contributing genes, including
et al., 2008). Such gene-therapy studies demonstrate that ALDH1A3 (Fares-Taie et al., 2013) and STRA6 (Pasutto
(continued)
key (Khan et al., 2013). The precise delineation of bilat- recognition of isolated autosomal dominant (AD), autoso-
eral congenital cataract in infants therefore requires an mal recessive (AR), X-linked (XL) and digenic forms. RP
interdisciplinary approach involving the ophthalmologist, can also be part of a group of syndromic disorders. Fifty-five
the pediatrician, and the clinical geneticist. In the future loci have been described for non-syndromic RP, of which
diagnosis will be enhanced through the introduction of 48 genes have been identified. A further 158 genes or loci
next-generation sequencing. have been described that are associated with retinal disease
(RetNet, http://www.sph.uth.tmc.edu/RetNet/home.htm,
accessed October 2013).
T H E I N H E R IT E D R ET I NA L
DYS T RO P H I E S , I N C LU D I N G
R ET I N IT I S P I G M E N TO S A Cloning of Single Genes Causing Retinitis
Pigmentosa: A Powerful Approach
In the field of ophthalmology, no one group of conditions
exemplifies the progress that has been made using human Rhodopsin was identified as the first gene mutated in
molecular genetic analyses better than retinitis pigmentosa. retinitis pigmentosa, underlying one form of ADRP.
Retinitis pigmentosa (RP) is a collective description for a Identification relied on a positional candidate-gene
group of inherited retinal degenerations. The condition has approach in 1990 (Dryja et al., 1990), and null mutations
an incidence estimated to be between one in 3,000–5,000 were also found to be responsible for a form of ARRP two
(Heckenlively, 1989; Dryja and Li, 1995). RP presents with years later (Rosenfeld et al., 1992). Since that time, the
an initial, progressive degeneration of the rod photorecep- molecular bases of a vast number of monogenic retinal dys-
tors that starts in the peripheral retina, leading to tunnel trophies have been identified by using techniques that dem-
vision and night blindness (nyctalopia). As the disease pro- onstrate the extraordinary power of genetic approaches for
gresses, the central retina and cone photoreceptors become dissecting genetically heterogeneous disorders.
involved, there is a concomitant reduction in central vision, Positional approaches have been used extremely suc-
loss of visual acuity, and color perception. As the neural cessfully in linked families, initially focusing on standard
retina degenerates, retinal arterioles become attenuated, techniques for autosomal dominant and X-linked families.
and a characteristically waxy optic disc pallor develops. In More recently, the use of homozygosity-mapping in mul-
addition, pigment cells derived from the underlying RPE tiply consanguineous families has successfully identified
migrate into the retina and lead to the formation of bone many mutated genes in recessive forms of retinal disease.
spicules, one of the most characteristic features of the dis- These technologies have collectively identified mutations
ease. An initial reduction of the dark-adapted rod (scotopic) in genes of a wide range of functions but, crucially, have
electroretinogram (ERG) is followed by a reduction of the allowed identification of genes whose function could not
light-adapted, cone-mediated (photopic) ERG. Ultimately, have been identified using hypothesis-driven approaches—
the ERG becomes unrecordable. often these provide the most interesting and unsuspected
Clinically, RP is highly heterogeneous in its onset areas for subsequent investigation. Examples include the
and severity. There is high genetic heterogeneity with the ubiquitous splicing factors PRPF3, 8, 31, and PAP-1;
This is the most common form of glaucoma in the Myocilin is the only gene identified so far in JOAG famil-
Western Hemisphere, with characteristic manifestations of ial and sporadic cases. This gene is located at the GLC1A
636
Table 41.1 GLAUCOMA-RELATED GENES AND LOCI
locus (1q24.3) and is also known as Trabecular meshwork been proposed that MYOC may interact with OPTN,60
Inducible Glucocorticoid Response (TIGR).34 Since the ini- APOE,61 Flotillin-1,62 and α1-Syntrophin.63 Myocilin
tial discovery of MYOC, over 98 different glaucoma-causing transgenic mice show a moderate elevation of intraocular
mutations have been identified throughout the world in pressure, loss of retinal ganglion cells in the peripheral ret-
juvenile- and adult-onset POAG families and sporadic ina, and axonal degeneration in the optic nerve, a pheno-
cases according to the Myocilin Allele-Specific Phenotype type that is similar to those observed in human glaucoma
Database (http://myocilin.com/variants.php). Taken patients.64
together, MYOC mutations are responsible for 2–4% of
glaucoma cases (Table 41.2). At times, Cytochrome P450,
O P T I N EU R I N (O P T I C N EU RO PAT H Y
family 1, subfamily B, polypeptide 1 (CYP1B1), a causative
I N D U C I N G P ROT E I N; O PTN )
primary congenital glaucoma (PCG) gene, has also been
observed to play a role in JOAG patients, thus indicating Optineurin at the GLC1E locus (10p13) was identified
an underlying genetic relationship between the JOAG and using a large British normal tension glaucoma family.65
PCG phenotypes through a digenic interaction between Other groups screened the OPTN gene and reported new
MYOC and CYP1B1 genes.17,34–52 These studies empha- mutations in their NTG, POAG, and JOAG cases.66–81 The
size the genetic heterogeneity of juvenile glaucoma, and most common OPTN-mutation (E50K) is observed in
furthermore suggest that congenital glaucoma and juvenile NTG subjects of a younger age, with more advanced optic
glaucoma may be allelic variants, indicating that the range disc cupping, smaller neuroretinal rim area, and progres-
of expression of MYOC and CYP1B1 mutations may be sive visual fields requiring filtration surgeries (Table 41.2).82
greater than previously appreciated.39,40,43,53–59 It has also Optineurin has also been observed to co-segregate with
MYOC,83–87 APOE,61 and TNF-α88 in several familial and sporadic cases.97–105 Different spectrums of OPTN muta-
sporadic cases. The OPTN protein is known to interact tions have been reported in ALS (autosomal recessive) com-
with at least 14 other proteins, one of which (TBK1) has pared to NTG (autosomal dominant) subjects.
already shown to be equally involved in NTG patients.89
It is anticipated that other OPTN-interacting proteins
WD R E P E AT D O M A I N 36 ( WD R36)
may be equally involved in NTG or other glaucoma phe-
notypes. OPTN transgenic mice and over-expression stud- Gene mutations at the GLC1G locus (5q22.1) were
ies of OPTN-E50K mutations provided evidence for its observed in patients with high IOP, though there were
contribution in neuronal cell processing, impaired protein few NTG individuals who possessed mutations as well.19
trafficking, and retinal ganglion cell apoptosis.90–96 Further Since then, many groups worldwide have screened and
ongoing protein and cellular studies will add new insight confirmed mutations in WDR36 gene.106–112 Four different
into the role of this molecule and the etiology of glaucoma studies suggest that WDR36 may account for 10–17% of
cases. In addition to its role in glaucoma, OPTN is also glaucoma subjects,19,107,110,111,113 while others found no muta-
reported to cause another neurodegenerative condition, tions in their POAG families.32,114–116 It has been shown
amyotrophic lateral sclerosis (ALS), in certain familial and that WDR36 may act cooperatively with certain MYOC
(continued)
INTRODUCTION architecture have all been advanced, but in the last decade,
the field of genetics has offered a truly unique perspective
Age-related macular degeneration (AMD) is the most com- on AMD. Our fundamental knowledge of its pathogenesis
mon cause of irreversible blindness in the developed world. has matured considerably in the past decade, and the model
Currently, the epidemic is estimated to affect between 30 of AMD now serves as a prime example of how to improve
and 50 million people across the globe and is imposing mas- our translational approaches to human genetics to develop
sive health-care expenses while significantly reducing qual- effective therapies. Dozens of genes have now been asso-
ity of life. The growth estimates are startling especially in ciated with this condition, with many relevant pathways
the United States, where an aging Baby Boomer generation being targeted in clinical trials. Therefore, we will focus this
is now in its seventh decade of life; a situation that will most chapter on the polygenic nature of AMD and the charted
certainly be complicated by limited financial resources. development of potential therapeutic approaches that have
Approximately 1.75 million Americans are affected with an been directly related to the robust genetic associations dis-
advanced form of the disease, a prevalence that is expected covered thus far.
nearly double to 3 million by 2020, with another 7 million Dissecting the multifactorial contributions of AMD
at risk.1 Western Europe is estimated to have over 3 million pathogenesis has proven to be a daunting challenge. Several
with advanced disease. large epidemiological studies, including the Beaver Dam
AMD is characterized as a progressive retinal degen- Eye Study, Blue Mountain Eye Study, and the Age-Related
eration with hallmark features such as the accumulation Eye Disease Study (AREDS), have acquired extensive data-
of sub- and intra-retinal lipoproteinaceous deposits called sets that identified significant risk factors such as age and
drusen, abnormal pigment clumping, subretinal fluid or smoking, as well as phenotypical features associated with
hemorrhage, and atrophy of the retinal pigment epithelium disease progression, including drusen size and retinal pig-
(RPE). Severe vision loss from AMD results from choroi- ment abnormalities.2–4 Yet the true significance of genetic
dal neovascularization (CNV), the invasion of the retina susceptibility to this disease was not realized until twin
by abnormal choroidal blood vessels; or from geographic concordance, small family linkage, and segregation analysis
atrophy (GA), the apoptotic loss of RPE, photoreceptors, studies suggested that upwards of 70% of AMD may be due
and choriocapillaris (Figure 42.1). The pathology most to genetic predisposition.5–9 Shortly after the millennium,
often occurs in the central visual field, leading to distortion the application of higher resolution genome-wide associa-
(metamorphopsias), loss of contrast, or black spots (scoto- tion studies to the field resulted in a rush of reports with
mas) experienced by the patient. The disease’s unfortunate multiple single-nucleotide polymorphisms (SNPs) iden-
predilection for the macula, an area that is required for tified as disease cohorts. As more studies continued to be
high-quality binocular visual acuity, has intrigued scien- published, specific pathways began to emerge that were
tists and clinicians since the disease was first identified in subsequently studied in animal models and then targeted
the late nineteenth century. Theories related to the expo- for translation. The strongest genetic signature identified
sure of focused light rays, high metabolic demand, oxida- to date involves SNPs located in immune-related genes
tive stress overload, and spatial heterogeneity of cellular belonging to the complement activation system.
652
(A) (B) (E)
RPE
(F)
200 µm
(C) (D)
(G)
PED
* *
Multimodal imaging of advanced dry or wet (neovascular) age-related macular degeneration. A. Color fundus photograph of patient
Figure 42.1
with advanced dry AMD with large drusen (black arrowheads) and areas of geographic atrophy of the retinal pigment epithelium (RPE, black
arrow). B. Fundus autofluorescent imaging (488nm blue light) clearly revealing focal area of RPE loss (white arrow) and adjacent abnormal
hyperfluorescence indicative of RPE dysfunction in advanced dry AMD. C. Color fundus photograph of patient with acute decrease in vision shows
neovascular AMD with a large CNV membrane, sub-retinal heme, and fluid. D. Fluorescein angiogram of same eye with neovascular AMD at late
time-point (10 minutes) exhibited significant leakage due to CNV formation. E. Spectral domain optical coherence tomography (SD-OCT) image
of a normal retina displaying clear delineation of RPE layer and neural retina. F. SD-OCT of the eye with advanced dry AMD shown in panel A,
revealing severe loss of the RPE and outer retinal layers where photoreceptors are normally located (white arrowheads). G. SD-OCT of the eye
with neovascular AMD shown in panel C, confirming the presence of sub-retinal fluid (asterisks) and significant disruption of the RPE, known as
pigment epithelial detachment (PED).
6 5 4 • G e no m ic s in C l inic a l Pr actic e
researchers corrected for multiple comparisons. Further association studies are warranted, especially in light of
studies of the hypomorphic L412F SNP suggested a pro- expression patterns and newly discovered functions of
tective effect for advanced dry AMD compared to normal specific immune mediators in the retina,59,60 and the per-
controls,48,49 yet this conclusion was similarly contro- sistent speculation on the potential infectious etiology of
versial, as the result was not found in other cohorts.50,51 AMD61 providing a source of potent ligands for this sig-
Importantly, dsRNA-induced retinotoxicity was observed naling pathway.
in mouse models and exhibited many features that reca-
pitulated GA, including large focal areas of isolated RPE
cell loss consistent with the human disease as well as dis- A N G I O G E N ET I C S I N
rupted ZO-1 cell–cell junctions and increased caspase-3 N E O VAS C U LA R AM D
activity.49,52 Stemming from this discovery, downstream
analyses focused on determining whether dsRNAs may Neovascular or “wet” AMD results in upwards of 90% of the
be present and involved in RPE cell death in GA. In vision loss associated with this disease, but it only is found
fact, significantly increased levels of dsRNA sequences in 10% of those affected with this disease. As described, the
were immuno-localized in drusen, Bruch’s membrane, main pathological event leading to retinal dysfunction is the
and surrounding RPE in affected eyes. Using unbiased ingrowth and leakage of choroidal blood vessels (or CNV)
adaptor-based polymerase chain reaction followed by in the sub-retinal and sub-RPE space. This causes leakage of
Sanger sequencing, we identified the accumulation of a fluid and hemorrhage in the macula, and often beneath the
specific ~300-base pair dsRNA transcript derived from fovea, the point of maximum visual acuity in the center of
Alu sequences in eyes with GA. Alu sequences are short, the macula, severely altering the anatomical layering of the
interspersed, repetitive retrotransposons that are spe- chorioretinal tissues as well as the optical properties of the
cific to primates and account for approximately 10% of eye. Early treatment modalities were focused on mechanical
the human genome.53 Interrogation of human tissues for or thermal ablation of the CNV lesion using a sub-retinal
downregulation of RNA-processing enzymes capable of surgical excision or argon-laser photocoagulation, but these
degrading these molecules was performed for a compre- approaches certainly did not improve vision, nor did they
hensive panel of such targets. Eventually, it was discovered prevent further decline in visual acuity compared to con-
that that DICER1, an RNA processing enzyme critical trols. Laser-based approaches were optimized to include the
to micro-RNA biogenesis, was specifically downregu- use of photodynamic therapy, effectively inducing vascular
lated in GA eyes compared to age-matched controls.54 thrombosis and cessation of leakage in abnormal vascula-
Subsequent studies in our laboratory determined that ture, did result in improvements over observation only, but
Alu RNA-induced RPE cell death was dependent on the patients unfortunately still continued to lose their vision.62
immune receptor MyD88, and inflammasome activation In the first few years after the turn of the millennium,
via caspase-1 cleavage.55 Interestingly, MyD88 serves as a investigators gained critical knowledge on essential growth
convergence pathway for most TLR signaling;56 however, factors in CNV pathogenesis63–66 and initiated studies to
to date, Alu RNA has not been shown to interact with evaluate the preclinical efficacy of targeting a key molecule,
TLR3 or with any other innate immune receptors. These vascular endothelial growth factor-A (VEGF-A), with neu-
data highlight the importance of non-coding genetics as tralizing antibodies and RNA aptamers to suppress intra-
an approach for deciphering AMD disease mechanisms. ocular neovascular growth in animal models.67–69 These
With a new understanding of the widespread pathological results prompted a rapid translation to clinical trials; and,
effects of transcribed Alu in the eye and in other disease less than a decade later, the field of wet AMD treatment
models, efforts are currently underway to develop bio- has been transformed by the widespread deployment of
informatics pipelines with massively parallel sequencing antibody-based therapies targeting VEGF and its signal-
data in order to spatially localize Alu inserts in humans ing mechanisms, with unsurpassed gains in visual acuity
and determine sequence composition, as there are many and anatomical reorganization. While early linkage and
subfamilies of these inserts. High-throughput screen- genome-wide association studies (GWAS) did not show
ing assays of these mobile genetic elements have been evidence of VEGF and related polymorphisms in disease
reported57,58 and may be useful if applied to AMD cohorts cohorts,7,70 once VEGFA (6p21) became a candidate gene
for analyses of altered levels or mapping of Alu compared with the success of therapeutic targeting, multiple inde-
to normal, age-matched controls. With these intrigu- pendent studies identified significant association of cer-
ing yet inconclusive reports, further immuno-genetic tain SNPs, specifically in the promoter region, in disease
6 5 6 • G e no m ic s in C l inic a l Pr actic e
CRP
FIB3 CNV
Figure 42.3
Central role of CFH-signaling in neovascular AMD progression. There is a strong genetic association between a specific CFH genotype,
402H, and significantly increased risk for progression to neovascular AMD. Mapping of this SNP on the CFH protein localized the SNP to the
short consensus repeat domain 7 (SCR7), which is a critical binding region for C-reactive protein (CRP), a well-established pro-inflammatory
and angiogenic molecule. CFH is an important auto-regulator of alternative complement-pathway activation and primarily serves as a direct
inhibitor of activated C3b, preventing formation of the C3 convertase. Alternative complement activation also induces the secretion of VEGF-A,
thus accelerating CNV progression. SCR7 dually interacts with TIMP3 and EFEMP1, both of which have been implicated in CNV pathogenesis
and are endogenous inhibitors of angiogenesis. This signaling axis reveals a central role for CFH and SCR7 domain in regulating pathological
angiogenesis during AMD progression.
been contradictory,97,98 but a consensus that suggests lev- targets that interact with other known associated gene
els are unchanged in the retina and RPE is beginning to products, of which ARMS-2 and CFH are likely to be the
emerge.99,100 Currently, over 75% of all cases of AMD are most critical. Undoubtedly, the improved resolution of
attributed to this genotype in combination with CFH signaling networks, coupled with modern bioinformatics
Y402H and smoking (which alone increases the odds ratio analyses that link angiogenesis, immune-mediated inflam-
for AMD risk to 4) with ARMS2 carriers significantly mation, and ECM dysregulation, will considerably aid our
more likely to develop progressive, bilateral, neovascular efforts to develop a complete molecular map of this disease
AMD.101 Subsequent studies have also observed this multi- from which to expand our studies and identify therapeutic
plicative effect, including data from a Finnish AMD cohort targets.
in which carriers of all three major AMD-associated SNP
alleles (CFH, C3, and ARMS2) had an 18-fold higher
risk of disease incidence.102 Further investigations of the G E N ET I C D ET E R M I N A N T S O F ALT E R E D
ARMS2 locus resulted in the discovery of a protein that L I P I D META B O L I SM I N AM D
localized to the outer membrane of mitochondria, which
may interact with specific components of the extracellular Nearly 50 years ago, investigations into disease associations
matrix, including fibulin-1 and -6, COL1A1, COL4A2, and between hyperlipidemia and AMD were first reported. Since
fibronectin,103 but the function of this molecule remains then, significant evidence has accumulated supporting the
unknown and an area of intense investigation. ARMS-2 hypothesis that changes in lipid content and levels may be
affinity for fibulin-6 has attracted considerable interest, involved in early age-related eye diseases. Histological stud-
as the confirmed AMD locus on 1q31 mapped to associ- ies of aged eyes demonstrated increased concentrations of
ated SNPs in fibulin-6 (also known as hemecentin-1).104,105 lipids in Bruch’s membrane, with subsequent studies focused
Further cohorts tested have not revealed significant associa- specifically on cholesterol accumulation.108,109 High-quality
tions with these SNPs,9,106 raising the question of whether ultra-structural imaging on well-preserved eyes suggested
this specific interaction in the ECM is an important con- that lipid particles concentrating in the outer retina were
tributor to ARMS-2 involvement in AMD pathogenesis. derived from membrane debris released by nearby cells.110
Recently, seven new AMD-associated loci were discov- An animal model of CNV-formation has also been devel-
ered in a single study, with regions located near or in other oped based on subretinal lipid injections with pathological
genomic regions also involved in ECM-regulation includ- features similar to the human condition.111 However, there
ing COL8A1, TGFBR1, and ADAMTS9.107 This impres- are conflicting data that increased high-density lipoproteins
sive finding fortifies that notion that ECM dysfunction in are either protective of, or associated with, disease progres-
AMD is probably multifactorial, with an array of potential sion.112 The role of statins in AMD progression has been
6 5 8 • G e no m ic s in C l inic a l Pr actic e
and efficacy of locally delivered injectable small-molecule 11. Dragon-Durey MA, Fremeaux-Bacchi V, Loirat C, et al.
Heterozygous and homozygous factor H deficiencies associated
and biological inhibitors of many of the targets and path- with hemolytic uremic syndrome or membranoproliferative glo-
ways detailed in this chapter are underway, in addition to merulonephritis: report and genetic analysis of 16 cases. J Am Soc
several other approaches. To date, clinical trials have been Nephrol: JASN. Mar 2004;15(3):787–795.
12. Mullins RF, Aptsiauri N, Hageman GS. Structure and composition
completed or are underway, evaluating therapeutics based of drusen associated with glomerulonephritis: implications for the
on antibodies and aptamers to neutralize select comple- role of complement activation in drusen biogenesis. Eye (London,
ment components such as C3, C5aR, and CFD; novel and England). Jun 2001;15(Pt 3):390–395.
13. Hageman GS, Anderson DH, Johnson LV, et al. A common hap-
rediscovered drugs that target the visual cycle to mitigate lotype in the complement regulatory gene factor H (HF1/CFH)
toxic byproducts of retinal metabolism; and stem-cell-based predisposes individuals to age-related macular degeneration. Proc
therapies. While all of these exciting avenues of clinical Natl Acad Sci U S A. May 17 2005;102(20):7227–7232.
14. Jozsi M, Oppermann M, Lambris JD, Zipfel PF. The C-terminus
investigation have immeasurable potential, it is the massive of complement factor H is essential for host cell protection. Mol
acquisition of genomic data in ocular disease, specifically Immunol. Apr 2007;44(10):2697–2706.
with respect to AMD, that has created a unique library, with 15. Klein RJ, Zeiss C, Chew EY, et al. Complement factor H poly-
morphism in age-related macular degeneration. Science. Apr 15
whose aid all future scientists and clinicians will be able to 2005;308(5720):385–389.
improve on and establish new models of age-related biol- 16. Haines JL, Hauser MA, Schmidt S, et al. Complement factor H vari-
ogy and disease. Such precious resources utilized in these ant increases the risk of age-related macular degeneration. Science.
Apr 15 2005;308(5720):419–421.
extreme and time-consuming efforts will provide the neces- 17. Edwards AO, Ritter R 3rd, Abel KJ, Manning A, Panhuysen C,
sary foundation for more detailed studies of inflammation, Farrer LA. Complement factor H polymorphism and age-related
macular degeneration. Science. Apr 15 2005;308(5720):421–424.
cell death, and vascular biology, with a common goal to pro- 18. Clark SJ, Higman VA, Mulloy B, et al. H384 allotypic variant of fac-
vide cures for those at risk of or suffering from age-related tor H associated with age-related macular degeneration has different
forms of blindness. heparin-binding properties from the non-disease-associated form. J
Biol Chem. August 2006;281(34):24713–24720.
19. Weismann D, Hartvigsen K, Lauer N, et al. Complement factor H
binds malondialdehyde epitopes and protects from oxidative stress.
Nature. Oct 6, 2011;478(7367):76–81.
R E F E R E N C ES 20. Giannakis E, Jokiranta TS, Male DA, et al. A common site within fac-
tor H SCR 7 responsible for binding heparin, C-reactive protein and
1. Tomany SC, Wang JJ, Van Leeuwen R, et al. Risk factors for inci- streptococcal M protein. Eur J Immunol. Apr 2003;33(4):962–969.
dent age-related macular degeneration: pooled findings from 3 con- 21. Bhutto IA, Baba T, Merges C, Juriasinghani V, McLeod DS,
tinents. Ophthalmology. Jul 2004;111(7):1280–1287. Lutty GA. C-reactive protein and complement factor H in aged
2. Ferris FL, Davis MD, Clemons TE, et al. A simplified severity scale human eyes and eyes with age-related macular degeneration. Br J
for age-related macular degeneration: AREDS Report No. 18. Arch Ophthalmol. Sep 2011;95(9):1323–1330.
Ophthalmol. Nov 2005;123(11):1570–1574. 22. Wyatt MK, Tsai JY, Mishra S, et al. Interaction of complement fac-
3. Ferris FL 3rd, Wilkinson CP, Bird A, et al. Clinical classifica- tor H and fibulin3 in age-related macular degeneration. PLoS ONE.
tion of age-related macular degeneration. Ophthalmology. Apr 2013;8(6):e68088.
2013;120(4):844–851. 23. Estaller C, Schwaeble W, Dierich M, Weiss EH. Human comple-
4. Klein R, Klein BE, Linton KL. Prevalence of age-related mac- ment factor H: two factor H proteins are derived from alternatively
ulopathy. The Beaver Dam Eye Study. Ophthalmology. Jun spliced transcripts. Eur J Immunol. Mar 1991;21(3):799–802.
1992;99(6):933–943. 24. Ansari M, McKeigue PM, Skerka C, et al. Genetic influences on
5. Seddon JM, Cote J, Page WF, Aggen SH, Neale MC. The US plasma CFH and CFHR1 concentrations and their role in suscep-
twin study of age-related macular degeneration: relative roles of tibility to age-related macular degeneration. Hum Mol Genet. Dec
genetic and environmental influences. Arch Ophthalmol. Mar 2013;22(23):4857–4869.
2005;123(3):321–327. 25. Coffey PJ, Gias C, McDermott CJ, et al. Complement fac-
6. Klein ML, Schultz DW, Edwards A, et al. Age-related macular tor H deficiency in aged mice causes retinal abnormalities
degeneration. Clinical features in a large family and linkage to chro- and visual dysfunction. Proc Natl Acad Sci U S A. Oct 16
mosome 1q. Arch Ophthalmol. Aug 1998;116(8):1082–1088. 2007;104(42):16651–16656.
7. Majewski J, Schultz DW, Weleber RG, et al. Age-related macular 26. Gold B, Merriam JE, Zernant J, et al. Variation in factor B (BF) and
degeneration—a genome scan in extended families. Am J Hum complement component 2 (C2) genes is associated with age-related
Genet. Sep 2003;73(3):540–550. macular degeneration. Nat Genet. Apr 2006;38(4):458–462.
8. Seddon JM, Santangelo SL, Book K, Chong S, Cote J. A genome- 27. Yates JR, Sepp T, Matharu BK, et al. Complement C3 variant and
wide scan for age-related macular degeneration provides evidence the risk of age-related macular degeneration. N Engl J Med. Aug 9,
for linkage to several chromosomal regions. Am J Hum Genet. Oct 2007;357(6):553–561.
2003;73(4):780–790. 28. van de Ven JP, Nilsson SC, Tan PL, et al. A functional variant in the
9. Abecasis GR, Yashar BM, Zhao Y, et al. Age-related macular degen- CFI gene confers a high risk of age-related macular degeneration.
eration: a high-resolution genome scan for susceptibility loci in a Nat Genet. July 2013; 45(7): 813–817.
population enriched for late-stage disease. Am J Hum Genet. Mar 29. Seddon JM, Reynolds R, Maller J, Fagerness JA, Daly MJ, Rosner
2004;74(3):482–494. B. Prediction model for prevalence and incidence of advanced
10. Holers VM, Thurman JM. The alternative pathway of complement age-related macular degeneration based on genetic, demographic,
in disease: opportunities for therapeutic targeting. Mol Immunol. and environmental variables. Invest Ophthalmol Vis Sci. May
Jun 2004;41(2–3):147–152. 2009;50(5):2044–2053.
6 6 0 • G e no m ic s in C l inic a l Pr actic e
70. Weeks DE, Conley YP, Tsai HJ, et al. Age-related maculopathy: a 89. Stone EM, Lotery AJ, Munier FL, et al. A single EFEMP1 muta-
genomewide scan with continued evidence of susceptibility loci tion associated with both Malattia Leventinese and Doyne honey-
within the 1q31, 10q26, and 17q25 regions. Am J Hum Genet. Aug comb retinal dystrophy. Nat Genet. Jun 1999;22(2):199–202.
2004;75(2):174–189. 90. Fu L, Garland D, Yang Z, et al. The R345W mutation in EFEMP1
71. Churchill AJ, Carter JG, Lovell HC, et al. VEGF polymorphisms is pathogenic and causes AMD-like deposits in mice. Hum Mol
are associated with neovascular age-related macular degeneration. Genet. Oct 15 2007;16(20):2411–2422.
Hum Mol Genet. Oct 1, 2006;15(19):2955–2961. 91. Marmorstein LY, McLaughlin PJ, Peachey NS, Sasaki T,
72. Haines JL, Schnetz-Boutaud N, Schmidt S, et al. Functional candi- Marmorstein AD. Formation and progression of sub-retinal pig-
date genes in age-related macular degeneration: significant associa- ment epithelium deposits in Efemp1 mutation knock-in mice: a
tion with VEGF, VLDLR, and LRP6. Invest Ophthalmol Vis Sci. model for the early pathogenic course of macular degeneration.
Jan 2006;47(1):329–335. Hum Mol Genet. Oct 15 2007;16(20):2423–2432.
73. Schick JH, Iyengar SK, Klein BE, et al. A whole-genome screen of a 92. Yang Z, Camp NJ, Sun H, et al. A variant of the HTRA1 gene
quantitative trait of age-related maculopathy in sibships from the Beaver increases susceptibility to age-related macular degeneration.
Dam Eye Study. Am J Hum Genet. Jun 2003;72(6): 1412–1424. Science. Nov 10 2006;314(5801):992–993.
74. Lin JM, Wan L, Tsai YY, et al. Vascular endothelial growth factor 93. Dewan A, Liu M, Hartman S, et al. HTRA1 promoter polymor-
gene polymorphisms in age-related macular degeneration. Am J phism in wet age-related macular degeneration. Science. Nov 10
Ophthalmol. Jun 2008;145(6):1045–1051. 2006;314(5801):989–992.
75. Yu Y, Bhangale TR, Fagerness J, et al. Common variants near 94. Jones A, Kumar S, Zhang N, et al. Increased expression of multi-
FRK/COL10A1 and VEGFA are associated with advanced functional serine protease, HTRA1, in retinal pigment epithelium
age-related macular degeneration. Hum Mol Genet. Sep 15 induces polypoidal choroidal vasculopathy in mice. Proc Natl Acad
2011;20(18):3699–3709. Sci U S A. Aug 30 2011;108(35):14578–14583.
76. Richardson AJ, Islam FM, Guymer RH, Cain M, Baird PN. A 95. Seddon JM, Francis PJ, George S, Schultz DW, Rosner B, Klein
tag-single nucleotide polymorphisms approach to the vascular ML. Association of CFH Y402H and LOC387715 A69S with
endothelial growth factor-A gene in age-related macular degenera- progression of age-related macular degeneration. JAMA. Apr 25
tion. Mol Vis. 2007;13:2148–2152. 2007;297(16):1793–1800.
77. Boekhoorn SS, Isaacs A, Uitterlinden AG, et al. Polymorphisms in 96. Kanda A, Chen W, Othman M, et al. A variant of mitochondrial
the vascular endothelial growth factor gene and risk of age-related protein LOC387715/ARMS2, not HTRA1, is strongly associated
macular degeneration: the Rotterdam Study. Ophthalmology. Nov with age-related macular degeneration. Proc Natl Acad Sci U S A.
2008;115(11):1899–1903. Oct 9, 2007;104(41):16227–16232.
78. Chang W, Noh DH, Sagong M, Kim IT. Pharmacogenetic association 97. Friedrich U, Myers CA, Fritsche LG, et al. Risk- and
with early response to intravitreal ranibizumab for age-related macular non-risk-associated variants at the 10q26 AMD locus influence
degeneration in a Korean population. Mol Vis. 2013;19:702–709. ARMS2 mRNA expression but exclude pathogenic effects due to
79. Zhao L, Grob S, Avery R, et al. Common variant in VEGFA and protein deficiency. Hum Mol Genet. April 1, 2011 2011;20(7):
response to anti-VEGF therapy for neovascular age-related macular 1387–1399.
degeneration. Curr Mol Med. Jul 2013;13(6):929–934. 98. Yang Z, Tong Z, Chen Y, et al. Genetic and functional dissection
80. Green LS, Jellinek D, Jenison R, Ostman A, Heldin CH, Janjic N. of HTRA1 and LOC387715 in age-related macular degeneration.
Inhibitory DNA ligands to platelet-derived growth factor B-chain. PLoS Genet. Feb 2010;6(2):e1000836.
Biochemistry. Nov 12 1996;35(45):14413–14424. 99. Kanda A, Stambolian D, Chen W, Curcio CA, Abecasis GR,
81. Weber BH, Vogt G, Pruett RC, Stohr H, Felbor U. Mutations in the Swaroop A. Age-related macular degeneration-associated vari-
tissue inhibitor of metalloproteinases-3 (TIMP3) in patients with ants at chromosome 10q26 do not significantly alter ARMS2
Sorsby’s fundus dystrophy. Nat Genet. Dec 1994;8(4):352–356. and HTRA1 transcript levels in the human retina. Mol Vis.
82. Jacobson SG, Cideciyan AV, Bennett J, Kingsley RM, Sheffield VC, 2010;16:1317–1323.
Stone EM. Novel mutation in the TIMP3 gene causes Sorsby fun- 100. Wang G, Dubovy SR, Kovach JL, et al. Variants at chromosome
dus dystrophy. Arch Ophthalmol. Mar 2002;120(3):376–379. 10q26 locus and the expression of HTRA1 in the retina. Exp Eye
83. Qi JH, Ebrahem Q, Moore N, et al. A novel function for tissue Res. May 3, 2013;112C:102–105.
inhibitor of metalloproteinases-3 (TIMP3): inhibition of angio- 101. Schmidt S, Hauser MA, Scott WK, et al. Cigarette smoking strongly
genesis by blockage of VEGF binding to VEGF receptor-2. Nat modifies the association of LOC387715 and age-related macular
Med. Apr 2003;9(4):407–415. degeneration. Am J Hum Genet. May 2006;78(5):852–864.
84. Qi JH, Dai G, Luthert P, et al. S156C mutation in tissue inhibitor of 102. Seitsonen SP, Onkamo P, Peng G, et al. Multifactor effects and
metalloproteinases-3 induces increased angiogenesis. J Biol Chem. evidence of potential interaction between complement factor H
Jul 24 2009;284(30):19927–19936. Y402H and LOC387715 A69S in age-related macular degenera-
85. De La Paz MA, Pericak-Vance MA, Lennon F, Haines JL, Seddon JM. tion. PLoS One. 2008;3(12):e3833.
Exclusion of TIMP3 as a candidate locus in age-related macular degen- 103. Kortvely E, Hauck SM, Duetsch G, et al. ARMS2 is a constituent
eration. Invest Ophthalmol Vis Sci. May 1997;38(6):1060–1065. of the extracellular matrix providing a link between familial and
86. Chen W, Stambolian D, Edwards AO, et al. Genetic variants near sporadic age-related macular degenerations. Invest Ophthalmol
TIMP3 and high-density lipoprotein-associated loci influence sus- Vis Sci. Jan 2010;51(1):79–88.
ceptibility to age-related macular degeneration. Proc Natl Acad Sci 104. Thompson CL, Klein BE, Klein R, et al. Complement factor H
U S A. Apr 20 2010;107(16):7401–7406. and hemicentin-1 in age-related macular degeneration and renal
87. Ardeljan D, Meyerle CB, Agron E, et al. Influence of TIMP3/SYN3 phenotypes. Hum Mol Genet. Sep 1, 2007;16(17):2135–2148.
polymorphisms on the phenotypic presentation of age-related mac- 105. Schultz DW, Klein ML, Humpert AJ, et al. Analysis of the
ular degeneration. Eur J Hum Genet. Oct 2013;21(10):1152–1157. ARMD1 locus: evidence that a mutation in HEMICENTIN-1 is
doi:10.1038/ejhg.2013.14. associated with age-related macular degeneration in a large family.
88. Klenotic PA, Munier FL, Marmorstein LY, Anand-Apte B. Tissue Hum Mol Genet. Dec 15 2003;12(24):3315–3323.
inhibitor of metalloproteinases-3 (TIMP-3) is a binding partner of 106. Bojanowski CM, Tuo J, Chew EY, Csaky KG, Chan CC. Analysis
epithelial growth factor-containing fibulin-like extracellular matrix of Hemicentin-1, hOgg1, and E-selectin single nucleotide poly-
protein 1 (EFEMP1). Implications for macular degenerations. J Biol morphisms in age-related macular degeneration. Trans Am
Chem. Jul 16 2004;279(29):30469–30473. Ophthalmol Soc. 2005;103:37–44; discussion 44–35.
6 6 2 • G e no m ic s in C l inic a l Pr actic e
43.
GENOMIC APPLICATIONS IN
AUDIOLOGICAL MEDICINE
Daphne Karfunkel-Doron, Zippora Brownstein, and Karen B. Avraham
663
Middle Inner
(A) ear ear (C) Tectorial membrane (D)
External ear Stria vascularis
marginal cells
Collagen bundles
Vestibule
Kcne1
Tectorin fibrils K+
Kcnq1
Auditory
Temporal nerve K+
bone
Scala vestibuli Endolymph
perilymph e
bran K+
Cochlea
(B)
r ’s mem
Tympanic sne
Reis Endolymph
membrane K+
with high K+
Ossicles
Ear canal (150mM, 85mV)
(E)
Stereocilia
Suppor
nucleus
cell
ting
Ventral Outer
Prestin
cochlear hair cell
nucleus
kcnq4
(G) Stereocilium K+
Whirlin K+
Myosin
Spiral ganglion K+
in
s XVa K+
Myosin armon
c d h 15
Sans H
3/p
cdh2 C×26/C×30 gap junction
3)
ks 2
(F) i n cdh
Actin filaments
p dl
Ti 5 an
h 1
cd
(p Stereocilia
b in
Myosin armon
Ankel links
Cuticular plate
H
vlgr1/usherin
VIIa
My I cldn14
os in V nV Outer
osi
I My hair cell
Figure 43.1
Schematic illustration of the human ear. (A) The ear consists of the outer, middle, and inner ear. (B) Cross-section through one turn of a
cochlea. The sensory epithelium is composed of three rows of outer hair cells, one row of inner hair cells, and supporting cells. Proteins encoding
deafness genes CDH23, CLDN14, GJA1, KCNQ4, MYH9, MYH14, MYO3A, MYO6, MYO7A, MYO15A, PCDH15, POU4F3, PRES, OTOF,
STRC, TFCP2L3, TMC1, TMPRSS3, USH1C, and WFS1 are expressed in inner and outer hair cells. (C) The tectorial membrane is composed
of thin tectorin fibrils and heavy collagen bundles. Proteins encoding deafness genes COL11A2, OTOA, OTOG, and TECTA are expressed in the
tectorial membrane. (D) Marginal cells of the stria vascularis. Proteins encoding deafness genes ATP6B1, BSND, CLCNKA, CLCNKB, EDN3,
EDNRB, GJB2, GJB6, KCNE1, KCNQ1, MITF, MYH14, TFCP2L3, and TMPRSS3 are expressed in the stria vascularis. (E) Single outer hair
cell, surrounded by two Deiters’ supporting cells. Proteins encoding deafness genes CLDN14, GJA1, GJB2, GJB6, KCNQ4, SLC26A4, TFCP2L3,
and TMPRSS3 are expressed in the supporting cells. (F) Enlargement of the hair cell bundle with actin-based stereocilia. (G) A single stereocilium
containing actin filaments. Proteins encoding deafness genes ESPN, PCDH15, STRC, TMIE, and WHRN are expressed in the stereocilia. (H) The
inner hair cells are innervated by the spiral ganglion cells, which in turn produce action potentials that are transmitted via the auditory nerve to
the brain. Proteins encoding deafness genes GJB1, KCNQ1, MPZ, NDP, NDRG1, OTOF, PCDH15, PMP22, SBF2, SLC26A4, TMPRSS3, and
WFS1 are expressed in the spiral ganglion. Modified from[203].
Atopic dermatitis (AD) or atopic eczema is a chronic, Atopic dermatitis, asthma, and hay fever are considered to
itchy, inflammatory skin condition that predominantly be part of a common syndrome of atopic diseases. The word
affects the skin flexures. Approximately 70% of cases start atopy means “strange disease” and was coined in 1923 by
in children younger than five years of age (Williams, 2000). Coca (Coca and Cooke, 1923). The atopic state is recog-
The prevalence has increased three fold over the last three nized by positive skin-prick tests to common allergens, by
decades in developed countries, with a higher prevalence the presence of allergen-specific immunoglobulin E (IgE) in
in urban regions and in higher social classes (Taylor et al., serum, and by elevations of total serum IgE. Eighty percent
1984). The lifetime prevalence in children is 10–20%, of infants with AD have raised total serum IgE levels ( Juhlin
and in adults, 1–3% (Schultz-Larsen, 2002). Though 60% et al., 1969). Two types of AD have been described: an
of children with AD are free of symptoms in adolescence extrinsic type associated with IgE-mediated sensitization,
(Rystedt, 1985), up to 50% may have recurrences in adult- and an intrinsic type, which is clinically identical, with nor-
hood (Lammintausta, 1991). Asthma develops in 30% of mal total serum IgE and no specific IgE responses to com-
children with AD, and allergic rhinitis in 35% (Luoma mon allergens ( Johansson et al., 2001).
et al., 1983). The association between AD and atopy is not clear-cut,
as both forms of AD have the same phenotype clinically.
There is some suggestion from hospital studies that higher
D I AG N O S T I C C R I T E R I A IgE levels are correlated with severe forms of AD, worse
long-term prognosis (Flohr et al., 2004), and an increased
There have been substantial variations in the disease’s defi- likelihood of developing asthma. There are immunological
nitions and diagnoses due to its variable clinical appear- differences between these two types of AD that lie in differ-
ance and distribution patterns, and the intermittent nature ent expression levels in skin-related cytokines that determine
of the disease. Based on the original consensus criteria by T-cell activation patterns ( Jeong et al., 2003). Inflammatory
Hanifin and Rajka (1980), a set of minimum and validated dendritic epidermal cells bear the high-affinity receptor for
discriminators have been determined by the U.K. Working IgE (FcεRI) on their cells’ surface (Novak et al., 2003). This
Party (Williams et al., 1994). The diagnosis requires evi- allows allergens penetrating the impaired skin barrier in AD
dence of itchy skin plus three or more of the following: his- to be focused and efficiently presented by FcεRI-bound IgE
tory of involvement of the skin creases, history of asthma to T cells (Novak et al., 2003; Novak and Bieber, 2003). In
or hay fever, history of generally dry skin, onset in a child intrinsic AD, there is lower surface expression of FcεRI on
under two years of age, and visible flexural dermatitis. These epidermal dendritic cells (Oppel et al., 2000). Interleukin
robust criteria are crucial for effective and reproducible (IL) 13 induces the production of IgE by B cells (Punnonen
genetic studies. et al., 1993). IL13 is decreased in intrinsic AD (Punnonen
683
et al., 1993; Jeong et al., 2003). Intrinsic and extrinsic forms with DNA microarrays has allowed global analyses of the
of AD may, however, be part of a continuum in the natural expression of thousands of genes, and is a high-throughput
history of AD (Novembre et al., 2001). The fact that phe- method for gene identification.
notypically both are identical suggests that IgE sensitization Genomic approaches to understanding AD are new,
is not a prerequisite for developing the skin lesions of AD. and significant inroads have been made only in the last five
Omalizumab, a monoclonal anti-IgE antibody currently years. To date, the following observations have been made
approved for the treatment of asthma, failed to improve in the genetics of AD that have given us new insights into
AD in any patients who were treated (Krathen et al., 2005; the pathogenesis of AD.
Lane et al., 2006). The diagnostic criteria of AD can also be
fulfilled in the absence of elevated IgE (Hanifin and Rajka,
1980; Williams et al., 1994). Subgroup analysis in future POSITIONAL CLONING STUDIES
genetic studies may help dissect these differences better.
Therefore, though atopic mechanisms dominate our under- In positional cloning, families are studied to identify
standing of the pathogenesis of AD, other mechanisms may chromosomal regions that are co-inherited or linked with
predispose to AD independently of atopy. disease. A “genome screen” is a systematic search for such
regions using polymorphic microsatellite markers. The
markers or microsatellites typically contain repeat sequences
G E N ET I C S T U D I E S O F ATO P I C of DNA, such as CA repeats, that vary between individu-
D E R M AT I T I S als. Linkage analyses search regions of the genome for areas
with a higher than expected number of shared alleles among
In genetic-epidemiological and population-based twin affected individuals within a family (Carlson et al., 2004).
studies of AD, the concordance rates for monozygotic Genetic linkage is said to exist if the marker and the phe-
twins were 0.72 to 0.86, and 0.21 to 0.23 for dizygotic twins notype are co-inherited. Genetic linkage identifies broad
(Larsen and Holm, 1986; Schultz Larsen, 1993). Asthma chromosomal regions linked to disease, normally exceeding
shows similar concordance rates of 0.65 in monozygotic 10 cM (~10 million bases). This area has to be saturated
twins and 0.25 in dizygotic twins (Duffy et al., 1990). with closely spaced markers to reduce the interval for local-
Total serum IgE shows heritability of approximately 47.3% ization to between 5–40 Mb (megabases). However, fortu-
(Palmer et al., 2000). Strong genetic factors therefore under- nately there is non-random association between alleles of
lie the development of AD and atopic disorders. However, individual single nucleotide polymorphisms (SNPs) and
inheritance of AD is complex and does not follow Mendel’s neighboring SNPs, known as linkage disequilibrium, which
laws of inheritance. Many genes may be contributory. In extends for shorter distances than genetic linkage. The
rare Mendelian diseases, polymorphisms commonly cause remaining chromosomal region of 0.5–1 Mb can then be
mutations by altering protein-coding sequences. In com- systematically dissected. Because no assumptions need be
mon diseases such as AD, gene functions may be altered by made about the disease’s etiology, positional cloning can
subtler mechanisms. These may affect the initiation of gene identify novel genes and mechanisms. It has been highly
transcription in exons or factors that affect gene expression successful in identifying genes underlying single-gene disor-
and splicing in introns. The genes causing complex disease ders such as cystic fibrosis and Huntington disease. Genome
are likely to show high-frequency allelic variation (fre- screens, however, are difficult to replicate because they are
quency of >1%), which is the basis of the “common disease, sensitive to population stratification, disease definition, the
common variant” hypothesis (Cargill et al., 1999; Halushka panels of markers chosen for genotyping, environmental
et al., 1999; Cheung and Spielman, 2002). There is com- factors, and the frequency of the phenotype in the popu-
plex interplay between environmental factors and the alleles lation. Also, stringent statistical criteria are needed to con-
of many genes and interactions between genes themselves firm whether linkage is real or not that take into account
(epistasis), which play an important role in determining dis- the hundreds of markers tested and the many phenotypes
ease expression. studied (Lander and Kruglyak, 1995).
Two traditional approaches are used to identify There have been four genome screens for AD (Cookson
genes: case-control comparisons of polymorphisms in et al., 2001; Lee et al., 2000; Bradley et al., 2002). The first
known candidate genes, or the identification of new genes screen, carried out in German and Scandinavian families
by positional cloning. Recently, with the sequencing of the with childhood AD, showed linkage to AD on chromo-
human genome, genome-wide gene-expression profiling some 3q21 (Lee et al., 2000). There was also linkage of
G E N ET I C S O F ATO P I C D E R M AT I T I S
D E F E C T I VE I N N AT E I M MU N I T Y C O M PA R E D TO OT H E R
A N D ATO P I C D E R M AT I T I S I N F L A M M ATO RY D I S E A S E S
696
To date, alterations of the MC1R gene on chromosome
16q24.3 have been most comprehensively shown to be
involved in normal variation in human skin and hair color.
MC1R encodes for a 7-pass transmembrane G-protein
coupled receptor, which is expressed by melanocytes. Its
ligand, alpha melanocyte-stimulating hormone (αMSH),
stimulates pigmentation of melanocytes and melanoma
cells in culture and causes skin darkening when injected
into humans (Lerner and McGuire, 1961). Although the
human receptor is only 317 amino acids in length, numer-
ous polymorphisms have been identified in the single 951
base pairs (bp) exon. Alternative splicing of second exons
can lead to a larger receptor with an additional 33 or 65
amino acids (Tan et al., 1999; Rouzaud et al., 2006), and
intergenic splicing of MC1R transcripts with the down-
Figure 45.1
Masson-Fontana stain showing the accumulation of stream tubulin-beta-III gene has also been reported (Dalziel
melanosomes above keratinocyte nuclei, producing a “biological
sun-hat.” et al., 2011). Investigations have indicated that MC1R vari-
ants/polymorphisms in the single-exon-encoded receptor
(the dominant isoform in human skin and hair) are asso-
ranges from blond, to red, to black. In addition, Caucasian ciated with red hair and with fair skin type in Caucasians
people with similarly light skin coloring differ in their abil- (Valverde et al., 1995; Box et al., 1997; Smith et al., 1998;
ity to tan following exposure to UVR, both after single Healy et al., 2000; Flanagan et al., 2000). The variant recep-
doses and after multiple repeated exposures. This variation tors commonly found in red hair and fair skin bind αMSH
in response to UVR was originally used to grade patients with similar affinity as the wild-type receptor, but exhibit
with psoriasis according to their “Fitzpatrick skin type” reduced signaling via cyclic-AMP, (adenosine monophos-
in order to calculate the UVR treatment doses for these phate) yet retain the ability to signal via the ERK (extracel-
patients and avoid UVR-induced burning (Fitzpatrick, lular signal-regulated kinases) pathway (Schioth et al., 1999;
1988). However, Fitzpatrick skin type is based on a combi- Robinson and Healy, 2002; Newton et al., 2005; Herraiz
nation of erythemal and tanning responses, and many sub- et al., 2011). In addition, some variants are expressed at a
jects do not fit neatly into its distinct subgroups (Rampen lower level at the cell surface, accounting for part of their
et al., 1988). Despite this, in the absence of simple alterna- compromised function (Beaumont et al., 2005). Transgenic
tives, the Fitzpatrick scale or modifications of it have been mice have demonstrated that “red-haired” variants fail to
employed in studies investigating the genetics of normal rescue pigmentation to the same extent as wild-type MC1R,
skin color. Previous studies had suggested that the range in indicating that these variants preferentially promote the
skin color from fair-skinned Caucasians to dark-skinned synthesis of phaeomelanin in vivo (Healy et al., 2001). As a
Negroes results from the additive interaction of three to six general rule, individuals with two variant MC1R alleles that
genes (Harrison and Owen, 1964; Stern, 1970). It is known compromise function tend to have red hair, whereas those
that over 350 loci affect pigmentation in mice (Montoliu with a single variant MC1R allele tend to have fair skin.
et al., 2012), but it is thought that significantly fewer of Perhaps predictably, considering the association between
these loci affect pigmentation in humans, with alterations red hair, fair skin, and freckling, MC1R variants have also
in the following genes reported to be associated with varia- been associated with freckling (Smith et al., 1998; Bastiaens
tions in normal human pigmentation: melanocortin 1 et al., 2001a).
receptor (MC1R), solute carrier family 24 (sodium/potas- SLC24A5 on chromosome 15q21.1, whose ortholog is
sium/calcium exchanger) member 5 (SLC24A5), solute responsible for the zebra fish “golden” phenotype, is involved
carrier family 45 member 2 (SLC45A2, also known as in determining variation in skin color between Caucasian
membrane-associated transporter protein MATP), tyrosi- and Asian or Negroid races (Lamason et al., 2005). Several
nase (TYR), tyrosinase-related protein 1 (TYRP1), OCA2 reports of an evolutionarily conserved ancestral allele, Ala111
(also known as the P gene), agouti signaling protein (ASIP), of SLC24A5, in the majority of African and East Asian
proopiomelanocortin (POMC), KIT ligand (KITLG), and populations, but a Thr111 allele in European and American
interferon regulatory factor 4 (IRF4). populations, in addition to the relationship of these alleles
SKIN CANCER
Figure 45.3
Genetic and environmental interactions: basal cell carcinoma
Several different types of cancer affect the skin, including occurring on the medial left cheek in a 43-year-old oil-rig worker with
the fair skin, red hair, and freckling phenotype.
basal cell carcinoma (BCC, from keratinocytes), squamous
cell carcinoma (SCC, from keratinocytes), melanoma
(from melanocytes or melanocytic nevi), Merkel cell car- develop de novo in clinical practice. In addition, a rapidly
cinoma (from Merkel cells), cutaneous T-cell lymphoma growing but self-healing keratinizing tumor, the keratoc-
(from lymphocytes), angiosarcoma (from endothelial cells anthoma, may resemble SCC clinically but is less common
lining blood vessels), and fibrosarcoma (from fibroblasts). than SCC.
In addition, the metastatic spread of internal cancers to the UVR is the most important environmental risk factor
skin can happen occasionally. However, the vast majority of for SCCs and BCCs, with a relationship between cumula-
skin cancer comprises SCC, BCC, and melanoma, with the tive UVR and SCC, while intermittent rather than cumu-
combination of SCC and BCC referred to as non-melanoma lative exposure may be more important for BCC (Vitasa
skin cancer (NMSC). Although “NMSC” is widely used, it et al., 1990; Kricker et al., 1995; Rosso et al., 1996). SCCs
is not an ideal term, and “keratinocyte-derived skin cancer” can also arise from prolonged exposure to infrared radia-
is a more appropriate descriptor. Depending on the depth tion, chronic inflammation, and in certain genodermatoses
of invasion of the primary tumor and other clinical and his- (e.g., dystrophic epidermolysis bullosa, keratitis ichthyosis,
tological features, melanomas and SCCs may metastasize, and deafness [KID] syndrome). Exposure to arsenic and
whereas BCCs metastasize infrequently. photochemotherapy (PUVA) increases the risk of BCC
Skin cancer incidence (melanoma and NMSC) has
risen steadily over the past several decades, and is predicted
to increase further in the next two decades. NMSC is the
most common human cancer, with approximately 13 mil-
lion cases annually worldwide, and its increasing incidence
in aging Western populations now constitutes an important
public health problem (Lucas et al., 2009). BCCs (Figure
45.3) arise in hair-bearing epidermis (most areas of skin
except palms and soles) and may be of follicular or epider-
mal origin (Kruger et al., 1999; Epstein, 2011). Different
clinical/histological subtypes of BCC exist (nodular, scle-
rosing, superficial, and morphoeic types) and are likely to
reflect genetic heterogeneity within the tumors. The inci-
dence of BCCs exceeds that of squamous cell carcinomas
(Figure 45.4) by about 4 to 1. In contrast to BCC, for which
there is no clinically identifiable precursor lesion, premalig-
nant skin lesions (e.g., actinic keratosis and Bowen’s disease) Figure 45.4
A squamous cell carcinoma on the back of the hand of an
may develop into SCC, but it is thought that most SCCs elderly individual with fair skin.
713
the anomalies are isolated and reflect unequal and abnor- techniques (FISH and array CGH) and new genomic tools
mal embryological differentiation of the Müllerian tract. (exome and genome sequencing) are invaluable for eluci-
However, it is likely that abnormal Müllerian canal differ- dating the causative underlying structural genomic abnor-
entiation could be associated with other malformations, mality. It is important that any such case be investigated at
predominantly of mesodermal origin. There are several rec- a dedicated tertiary pediatric gynecology or infertility unit
ognizable multiple malformation syndromes that include equipped with the necessary molecular and imaging diag-
Müllerian and as well as cloacal-developmental anomalies nostic facilities, supported by a skilled multidisciplinary
(Table 46.1). Several genes with regulatory sequences and team (pediatric gynecologist, developmental pediatrician,
polymorphisms are now assigned to some of these dysmor- pediatric surgeon, clinical geneticist, reproductive medicine
phic conditions. Diagnostic applications of the molecular clinician, and clinical psychologist).
pathology in some of these conditions offer reliable tools
for confirmation of the diagnosis and reproductive coun-
D I S O R D E R S O F T H E E N D O M ET R IUM
selling. The scope of this chapter, however, limits the space
for a detailed description of the key dysmorphic conditions The endometrium is a dynamic tissue that responds to mul-
with major female reproductive tract anomalies. tiple stimuli, depending on physiological and environmen-
In addition to the above-listed malformation syndromes, tal conditions, including steroid hormones, an implanting
anomalous development of the female reproductive tract is conceptus, withdrawal of steroid hormones, contraceptive
likely indicate a chromosomal disorder, most notably Turner steroids, selective steroid hormone-receptor modulators,
syndrome. It is essential that detailed cytogenetic analysis infection, transient cell populations, and metaplastic and
be carried out in any such case. New molecular cytogenetic neoplastic agents[5]. Microarray gene-expression profiling
Upregulated genes
Osteopontin AF052124 + + + + +
Apolipoprotein D J02611 + + + +
Dickkopf/DKK1 AB020315 + + + +
Placental protein-14/ glycodelin J04129 + + +
Decay accelerating factor for M31516 + + + +
complement (CD55, Cromer
blood group system)
Adipsin/complement factor D M84526 + + +
Guanylate binding protein 2, M55543 + + +
interferon-inducible
Claudin 4/CEP-R AB000712 + + +
Monoamine oxidase A (MAOA) AA420624/M68840 + + + +
Growth arrest and M60974 + + + +
DNA-damage-inducible protein
(GADD45)
NIP2 AB002365 + + +
Serine (or cysteine) proteinase inhibitor, X54486 + + +
clade G (SERPINGI)
Interleukin 15 (IL15) AF031167 + + + +
Annexin IV (ANXA4) M82809 + + + +
Mitogen-activated protein kinase kinase U67156 + + +
kinase 5 (MAP3K5)
Inhibitor of DNA-binding 4, dominant AL022726 + + +
negative helix-loop-helix protein (ID4)
Complement component 1, r subcom- M14058 + + +
ponent (C1R)
Downregulated genes
Olfactomedin-related ER localized U79299 + + + +
protein
Msh homeobox homolog 1 (MSX1) M97676 + + +
GATA-binding protein 2 (GATA2) M68891 + + +
contribute to the understanding of the endometrial biol- vitro can be induced by number of factors, including pro-
ogy and also improve the accuracy and reproducibility of gesterone primed with estrogen and cyclic adenosine mono
endometrial-dating procedures. phosphate (cAMP). Few transcriptional profiling studies
have been undertaken to try to shed light on the decidual-
ization process.
Endometrial Stromal Cell Differentiation
In another study[14], human endometrial stromal cells
Placental trophoblasts invade into the maternal endome- were decidualized in vitro in response to progesterone
trial stromal compartment during implantation in humans. or cAMP, and the gene-expression profile was compared
Decidualization of stromal cells is essential for successful to non-decidualized stromal cells. Of the 588 genes ana-
implantation and represents differentiation to an excep- lyzed, significant upregulation of cytokines, growth factors,
tional morphological, biosynthetic, and secretory pheno- nuclear transcription factors, cyclin family members, and
type. It is initiated independently of conception during mediators of the cAMP signal transduction pathways were
the late secretory phase of the normal menstrual cycle. The found. Many of these genes have been implicated in decidu-
molecular pathways that are involved in the process of decid- alization in vivo. The group identified new members of the
ualization are very poorly understood. Decidualization in transforming growth factor (TGF) β family upregulated
during decidualization and expression of neuromodulators genes were reprogrammed (i.e., changed expression) within
and neurotransmitter receptors, many of which had not specific functional groups.
been previously documented in the differentiation process.
Temporal expression and regulation of nearly 7,000
Endometriosis
genes during the decidualization process was investigated
by the gene-expression profile of human termed decid- Endometriosis is defined as the presence of endometrial
ual fibroblasts in response to estrogen + progesterone + tissue in ectopic locations outside the uterine cavity, typi-
8Br-cAMP at 0, 2, 4, 6, 9, 12, and 15 days[15]. Of 6,918 gene cally on the pelvic peritoneal surfaces of the uterus and its
analyzed, 121 genes expressed by more than two-fold, 110 ligaments, on the ovaries, and in the rectovaginal septum[16].
were downregulated, and 50 showed biphasic behavior. Between 5% and 10% of women have endometriosis, which
Dynamically regulated genes clustered into nine patterns of causes a range of symptoms, from pain to infertility.
gene expression as a function of time. These genes were bio- The uterine (eutopic) endometrium from women with
logically classified into five categories: cell and tissue func- endometriosis may have endogenous abnormalities that
tion, cell and tissue structure, regulation of gene expression, promote the establishment of the disease. Matrix metal-
expressed sequence tags, and functions unknown. Many loproteinases (MMPs) MMP-7 and MMP-9 are normally
that most Turner syndrome physical features map to Xp, with POF is not known[42]. In one reported series, two of
resulting from reduced dosage of genes on the short arm 52 (3.8%) patients with POF had the triple X syndrome[43].
of the X chromosome. Further investigations narrowed POF has also been reported in a girl with 48XXXX[44].
down the search for the affected chromosomal segment The underlying mechanism could be analogous to that
to the 2.6Mb Xp–Yp pseudoautosomal region. Statistical observed among patients with Klinefelter’s syndrome.
analyses of genotype–phenotype correlations mapped the X-chromosome mosaic individuals (45X/46XX and
Turner syndrome phenotype, including POF, to a critical 45X/47XXX) carry mixed germlines and manifest pheno-
region in Xp11.2–p22.1.X., identical to the correspond- typical abnormalities and POF similar to monosomy X, but
ing Yp region. Eighteen such candidate genes have been 12% are reported to menstruate[45].
reported[39], and more are likely to exist.
It is commonly believed that X trisomy (47,XXX),
Structural X Chromosome Abnormalities
which affects one in 900 women in the general popula-
tion, has no significant effect on fertility; however, associa- X chromosome deletions associated with POF are more
tion with hypergonadotrophic POF has been reported[40]. common than translocations. Deleted X chromosomes
Further documentation of associated ovarian failure in this necessarily leave a portion of the normal X unpaired,
rare sex-chromosome aneuploidy is in the form of occa- and isodicentrics probably interfere with pairing, result-
sional case reports[41]. Its relative prevalence among women ing in atresia of the oocytes. While deletions commonly
including bilateral corneal anesthesia associated with multi- studies on hormone and growth factor responses in fibroid
ple systemic abnormalities[95, 96] and progeria or accelerated smooth muscle cells will be performed in the near future.
aging[97]. Other candidate genes linked to POF are listed in If genetic pathways can be elucidated that differ between
Table 46.4. fibroids and myometrium, it may be possible to create medi-
cal treatments that selectively inhibit fibroid growth with-
out interfering in myometrial function. This is particularly
P O LYC YS T I C OVA R I A N SY N D RO M E
important in premenopausal women, whose families have
(PCOS)
not yet been completed.
Polycystic ovarian syndrome (PCOS) is the most com-
mon endocrine condition afflicting women, occurring in
5% of the population[98]. It is characterized by enlarged M A L E FAC TO R I N F E RT I L I T Y
ovaries with multiple small follicles, anovulation, and an
androgen-secreting stroma, presumably from the theca cells Male factor infertility is implicated in about half of all
of the follicles. Recent genomic studies have improved our infertile couples. This includes wide-ranging potential
understanding of the molecular mechanisms underlying causes that require thorough evaluation for accurate detec-
PCOS[99,100]. Analysis of the isolated and cultured theca tion and management. Couples with a component of male
cells showed an increase in gene transcription for alde- factor infertility need a systematic evaluation directed
hyde dehydrogenase 6 and retinol dehydrogenase 2[101]. at the apparently normal-looking healthy male partner
The gene-expression profiling between normal and PCOS to maximize their reproductive potential. A systematic
whole ovarian tissue involved Wnt signaling, extracellular approach with a complete medical history in conjunction
matrix components, and immunological factors[102]. Using with targeted investigations is necessary in all cases of sus-
cultured smooth muscle cells from fibroids and myome- pected male factor infertility. The physical examination
trium, the autocrine and paracrine action of TGF-beta on should be focused on signs of an associated endocrine dis-
gene expression has been analyzed recently by microar- order; the presence of gynecomastia; examination of the
ray[103,104]. Fibroid and myometrial smooth muscle cells penis, scrotum, and testis; and per rectal examination for
were found to respond differently to TGF-beta treatment, prostatic enlargement. The semen analysis is mandatory
possibly explaining the altered, excessive, extracellular in all cases, but it is by no means sufficient to determine
matrix produced in fibroids compared with myometrium the cause or decide on management plan therapy. Limited
(Figure 46.2). Estrogen responses in cultured fibroid imaging investigations may also assist in evaluating male
smooth-muscle cells have been studied using microarrays, factor infertility.
supporting a role for IGF-1 in mediating estrogen-induced Genetic factors implicated in male factor infertil-
fibroid growth[105]. It is likely that many other genomic ity include chromosomal abnormalities (47,XXY and
6
4
3 2
2
1
1
0 0
Fibroid Myometrium Fibroid Myometrium
1 1
0.5
0 0
Fibroid Myometrium
Fibroid Myometrium
0.5 0.4
mRNA relative units
0.4
0.3
0.3
0.2
0.2
0.1 0.1
0 0.0
Fibroid Myometrium Fibroid Myometrium
Levels of mRNA expression in 18 paired fibroid and myometrical specimens as determined by quantitative RT-PCR for the
three genes CYR 61, CTGF, and COL4A2. In all three, the microarray result was confimed. mRNA levels in arbitary units
shown on the Y-axis of the graphs, with all results corrected against expression of the housekeeping gene β actin.(† p<0.05;
* p<0.01). Graphs on the left are line graphs linking paired specimens, while on the right are box and whiskers plots.
Figure 46.2 Gene-expression profiling in normal and fibroid myometrium showing skewed variation in selected genes. Source: Adapted from Swartz et al., with
permission[115].
structural Y chromosome abnormalities), single-gene clinical diagnosis, but on determining the cause of male fac-
disorders interfering in the sperm’s motility or propaga- tor infertility as well.
tion (cystic fibrosis), spermatogenesis (mutations in SRY
and 5-α reductase genes) and mutations in the DAZ gene
C O N G E N I TA L A N O M A L I E S O F T H E
cluster[106]. The prognosis for any given couple depends, in
M A L E R E P RO D U C T I VE T R AC T
large part, on the cause of the infertility. Without a firm
understanding of the genetics, anatomy, and physiology, Exclusion of a congenital anatomical anomaly of the
and their interactions necessary to permit full functioning male reproductive tract is the essential first step in man-
of the male reproductive system, the evaluation becomes an aging male factor infertility. Most congenital anoma-
inefficient exercise that often fails to elucidate the precise lies of the male reproductive tract are sporadic, without
cause of infertility. Treatment success relies not just on the any appreciable recurrent genetic basis (Table 46.5).
PREVALENCE AND PHENOTYPES OF COMMON CHROMOSOMAL ABNORMALITIES ASSOCIATED WITH MALE INFERTILITY
However, genetic factors play a significant role in the cau- Klinefelter’s Syndrome and Partial X Disomy
sation of congenital anomalies; for example, cystic fibrosis
The mandatory chromosome analysis in the male often
transmembrane-conductance-regulator (CFTR)–related
reveals an additional X chromosome (47,XXY) consistent
congenital bilateral absence of the vas deferens (CBAVD).
with the phenotype of Klinefelter’s syndrome. Amongst
Genetic factors underlying a congenital anomaly of the
most infertile couple, the physical features (for example
male reproductive tract include both chromosomal and
subtle male secondary sexual characteristics and soft tes-
Mendelian conditions (Tables 46.5 and 46.6). There are
tes) were known to the male partner but often overlooked.
limited clinically relevant genomic data or information
A basic clinical examination would confirm the clini-
available on these. It is nevertheless likely that low-risk
cal diagnosis, if supported by X chromosome aneuploidy.
alleles, gene polymorphisms, and genome-wide CNVs
Azoospermia or oligospermia is present in almost all cases.
could have some causal and/or functional roles in male fac-
It is possible to aspirate immature sperms or spermatogonia
tor infertility. Such studies are important, but they might
by direct testicular aspiration or to obtain them at the testic-
be difficult to conduct and interpret due to the pheno-
ular biopsy. Prospects for natural conception are extremely
typical heterogeneity of congenital anomalies of the male
limited; however, it has been reported in rare cases. Most
reproductive tract.
couples would opt for assisted reproduction (in vitro
Image of Y chromosome displaying AZF regions and associated genes. Enlarged portion of AZFc region highlights discussed
microdeletions (A) Normal AZFc region; (B) gr/gr deletion; (C) b1/b3 deletion; (D) g1/g3 deletion; (E) gr/gr duplication.
Yp Yq
DBY (DDX3Y)
USP9Y
RBMY
TPSY
DAZ
PRY
CDY
b1 b2 g1 r1r2 b3 g2 r3r4 g3 b4
A
B gr/gr deletion
C b1/b3 deletion
D g1/g3 deletion
gr/gr duplication
E
Figure 46.3. chromosome abnormality/deleted in azoospermia (DAZ) spectrum. Source: Adapted with permission from O’Flynn et al., 2010 .
[116]
Autosomal
GENE LOCATION DISORDER PATHOLOGY OF MALE INFERTILITY
CFTR 7q Cystic fibrosis Congenital bilateral absent vas deferens (CBAVD)
SHBG 17 - Sex hormone–binding globulin; mutations result in
low level of spermatogenesis and androgen response
ESR1 6 - Estrogen receptor gene 1—abnormal spermatogenesis
ESR2 14 - Estrogen receptor gene 1—abnormal spermatogenesis
FSHR 2 - Follicle stimulating hormone receptor—inefficient spermatogenesis
DAZL 3 - Autosomal homologue of DAZ—azoospermia
MTHFR 1 - Methylene tetrahydrofolate reductase—abnormal DNA
methylation and inefficient spermatogenesis
INSL3 19 Testicular Insulin-like 3—associated with cryptorchidism;
dysgenesis
LGR8 13 Cryptorchidism Receptor for INSL3—associated with testicular
dysgenesis syndrome
Sex Chromosomal
AR Xq Kennedy disease Androgen receptor gene—regulates conversion of
spermatocytes to round spermatids
USP26 Xq Expressed in testes throughout early spermatogenesis
Involved in histone removal and protein breakdown
Reformation during spermatogenesis
TAF7L Xq Related to autosomal TAF7 gene; regulates spermatogenesis
KS1 Xp Kalman syndrome Idiopathic hypogonadotropic
hypogonadism (IHH) combined with anosmia or hyposmia; male infer-
tility is encountered in some cases; autosomal form linked with FGFR1
children[142]. ART could have negative consequences on the disorder, is characterized by large fetal and organ size, hypo-
imprinting of sperm because it may use sperm that are not glycemia, midline abdominal defects, facial moles, and
yet fully mature, and, consequently, whose epigenetic code enlarged tongues. Children with Beckwith-Wiedemann
is not established. If the sperm are too immature or abnor- syndrome are also at risk for developing tumors. One study
mal, it is more likely that the offspring could be born with reports the incidence of Beckwith-Wiedemann syndrome
an imprinting disorder. was almost 5% in children conceived by ART in com-
The correlation between the incidence of imprinting parison with an incidence of less than 1% in the general
disorders and ART in men with abnormal sperm is a con- population[145].
troversial topic. A study found that spermatogonia from
infertile men did not have increased imprinting errors in
M I TO C H O N D R I A L G E N E S I N M A L E
comparison with that of normal men[143]. Several researchers
FAC TO R I N FE RT I L IT Y
have asserted that ART, such as ICSI, causes imprinting dis-
orders like Angelman syndrome and Beckwith-Wiedemann Mitochondrial DNA (mtDNA) inheritance may also have
syndrome[139]. Angelman syndrome is a rare neurologi- an impact on male factor infertility (see also Chapter 9).
cal disorder characterized by cognitive defects, seizures, Abnormal mitochondria are known to interfere with
uncontrolled limb and body movements, spontaneous sperm motility because of aberrations in the mitochondrial
laughter, and difficulties with speech development[144]. Two sheath. There is concern about passing mutated mtDNA
independent groups reported an increased incidence of to offspring using ICSI in ART, because an entire sperm is
Angelman syndrome in offspring from ICSI procedures[145]. injected into the oocyte and the mtDNA is conserved. In
Beckwith-Wiedemann syndrome, also an uncommon normal fertilization, the sperm mitochondria are lost along
Figure: 46.5a
QF-PCR detection of common aneuploidies. QF-PCR electropherogram “trace” showing a result consistent with trisomy for
chromosome 21: Markers D21S1435 and D21S1437 show bi-allelic trisomy, while markers D21S11 and D21S1411 show tri-allelic trisomic results.
Courtesy of Mr. Chris Anderson and Mr. Peter Thompson, Principal Scientists, The Cytogenetic Laboratory, All Wales Genetic Laboratory, University Hospital of Wales, Cardiff, U.K.
Down’s syndrome screening is detecting high-risk pregnan- studies of trisomy 21 (T21) pregnancies include an attempt
cies among younger women (under 35 years of age), who are to identify new maternal serum biochemical markers for
then offered one of the two invasive tests for fetal chromo- Down syndrome pregnancy by cDNA microarrays. Other
some analysis. The invasive tests required (amniocentesis or approaches for faster detection of chromosomal aneuploi-
chorionic villous sampling of amniotic liquor) entail a mis- dies include the use of cell-free fetal DNA from the amni-
carriage risk of 0.5–1.0%. otic fluid to form fluorescent probes for hybridization to
Traditionally, fetal chromosome analysis involves cell CGH arrays[170].
culture and full karyotyping. This is slow and expensive, but
assures a comprehensive analysis, minimizing the risk of fail-
Cell-Free Fetal DNA
ure to detect any other chromosomal abnormality. However,
the rapid introduction of molecular cytogenetic methodol- Most prenatal genetic diagnosis (PND) is based on fetal
ogy led to the increased use of commercially available fluo- tissue obtained by invasive methods involving chorionic
rescent hybridization kits for selective cytogenetic analysis villus biopsy (>11 weeks), amniocentesis (>15 weeks),
for chromosomal aneuploidies of 13, 18, and 21 chromo- and fetoscopy (>18 weeks). All invasive PND methods
somes. The technique has improved to employ QF-PCR. involve additional risk of miscarriage (1–3%). However,
It is claimed that a selective, but fast and less expensive, with high-resolution ultrasound imaging, precise localiza-
cytogenetic analysis using QF-PCR is more cost-effective tion has considerably reduced fetal risks. Efforts to develop
when used in a carefully selected cohort of 17,500 women, non-invasive PND (NIPND) led to the consideration of
using a combined screening strategy of serum testing with testing circulating fetal cells in the maternal circulation. It
increased nuchal translucency (>4 mm)[169]. Microarray has long been recognized that nucleated fetal cells reach the
Figure 46.5b
QF-PCE showing trisomy 18 (Edwards syndrome). QF-PCR electropherogram “trace” showing a result consistent with trisomy for
chromosome 18, Edwards syndrome: Markers D18S535, D18S390, and D19D391 show bi-allelic trisomy; marker D18S386 shows a tri-allelic
trisomic result, and D18S978 is uninformative. Courtesy of Mr. Chris Anderson and Mr. Peter Thompson, Principal Scientists, The Cytogenetic Laboratory, All Wales Genetic Laboratory, University
Hospital of Wales, Cardiff, U.K.
maternal circulatory system, but attempts to isolate these cell-free RNA for PLAC4, a trophoblast-specific gene
rare cells from maternal blood (which typically number located in the Down’s syndrome region of chromosome 21.
1–6 cells per milliliter of maternal blood) and use them The PLAC4 coding sequence has a SNP that allows deter-
for genetic testing have been disappointing because of low mination of allelic ratios when the fetus is heterozygous for
sensitivity. Cell-free fetal DNA is currently the material the SNP[174]. Euploid embryos have an allelic ratio of 1:1.
of choice for NIPND. It represents 3–6% of circulating A ratio of 2:1 indicates a strong likelihood of trisomy 21.
cell-free DNA in maternal plasma, and it can be detected in The analysis of mRNAs encoded by different genes on chro-
the first trimester of pregnancy, increasing in abundance as mosome 21 could improve the sensitivity of this method,
the placenta grows. Cell-free fetal DNA fragments are much but it has not been widely pursued.
smaller than cell-free maternal DNA, which facilitates Since the presence of the Y chromosome defines the
DNA-sequence analysis. Although fetal DNA is detectable male sex, its detection or lack thereof in maternal blood
at 5 weeks of gestation, current methods of analysis are unre- can be used to determine the fetal sex. A recent review and
liable before 7 weeks of gestation[171]. Cell-free fetal RNA meta-analysis of fetal sex determination[175] with the use of
and DNA are not released from the fetus but from apoptotic maternal cell-free fetal DNA reported very good sensitiv-
placental trophoblast cells. These hold greater promise for ity, but the greatest sensitivity and specificity in the use of
genetic testing as a result of advances in DNA-sequencing Y-chromosome sequences to determine sex are obtained
methods and informatics[172;173]. In 2007, Down’s syndrome after 20 weeks of gestation, at which time ultrasonography
was detected by the quantitative assay of maternal blood is the preferred (and inexpensive) method.
INTRODUCTION S T E M C E L L S , T H E I R D E VE L O PM E N TA L
C O M P ET E N C E , A N D I N D U C I B I L I T Y
Stem cells harbor the ability to reconstitute organs and
tissues de novo; therefore, they are the cornerstones of
D EVE L O PM E N TA L C O M P ET E N C E
life: they drive embryonic development, adult homeosta-
A N D E M B RYO G E N E S I S
sis, and regeneration after injury. A common denomina-
tor of stem cells is the quintessential ability to give rise Developmental competence is a common denominator of
to one or more committed cell types, which is other- diverse stem cell populations (Waddington, 1940; Wolff,
wise known as “developmental competence.” Thus stem 1968). Unlike terminally differentiated cells, which have
cells are positioned at a developmental crossroads at stably acquired a given developmental identity, stem cells/
which they can commit to one of multiple fates: under- progenitors are amenable to changes in their cellular iden-
standing how such “decisions” are made is fundamental tity elicited by external instructions, which can instruct
to developmental biology. Nevertheless, myriad unan- them to generate multiple daughter-cell types. These ideas
swered questions continue to surround the potentiality were similarly conceptualized by Waddington over sev-
of stem cells and how downstream lineage decisions are enty years ago, who hypothesized that progenitors have
executed. How is the developmental destiny of stem cells an intrinsic competence, and that extrinsic signals (induc-
(how they know “what” they can become) molecularly ers) acted within this predetermined window of develop-
preconfigured? How do stem cells position themselves in mental opportunity to induce particular embryonic fates
an uncommitted state in a way that they can still access (Waddington, 1940). Several decades on, in modern stem
multiple prospective developmental fates? Upon lineage cell colloquial terminology, the collective of extrinsic fac-
commitment, how is a singular developmental program tors that act on stem cells has been rebranded as the “niche”
executed, and how are other prospective fates rescinded? (Scadden, 2006; Schofield, 1978).
The advent of various genomics platforms provides a win- The developmental competence of a stem cell popula-
dow of opportunity to access the molecular mechanisms tion is delimited by what cell types it can generate (see
underlying these phenomena. Drawing from diverse Box 47.1). One modern approach to assessing develop-
examples, we argue that stem cell transcription factors mental competence is genetic lineage tracing (Blanpain
are lineage specifiers that endow stem cells with the com- and Simons, 2013). Lineage tracing involves the perma-
petence to differentiate into various progeny. During lin- nent genetic labeling of a precursor cell(s) in vivo: upon
eage commitment, extrinsic signals and transcriptional its differentiation into one or more lineages, all its progeny
regulators act within this predetermined range of devel- will inheritably carry the genetic label, therefore defin-
opmental potential to specify various daughter-cell types ing the repertoire of fates the original stem cell popula-
in a mutually exclusive fashion. tion can assume (Blanpain and Simons, 2013). In historic
741
Box 47.1 KEY POINTS progenitors (Kawamoto et al., 2010; Rothenberg, 2011)
and how inductive signals act within (or expand) win-
dows of predetermined competence has attracted renewed
• Developmental competence is the ability to generate
interest.
one or more daughter cell types.
The interaction between inducers and developmentally
• Distinct types of stem cells that exist during competent cells is a recurrent motif occurring throughout
embryonic development and in the adult enable embryonic development for the progressive specification of
organ formation, homeostasis and tissue repair after new cell types. For example, prospective liver progenitors
injury. within the ventral foregut endoderm are directed to com-
mit to the hepatic lineage by signaling molecules Fgf2 ( Jung
• The distinction between stem cells and progenitors is
et al., 1999) and Bmp4 (Rossi et al., 2001), secreted by the
unclear. In certain systems, stem cells and progenitors
surrounding cardiac mesoderm and septum transversum
interconvert.
mesenchyme. Another such interaction is reflected by how
• A ubiquitous “stemness” gene is elusive but the surface ectoderm secretes Bmp4 to induce nearby inter-
emerging evidence indicates the the architecture of mediate mesoderm to form the nephric duct (Obara-Ishihara
transcription factor networks confers developmental et al., 1999). At a deeper level of complexity, juxtaposition
competence. of two reciprocally signaling progenitor populations pro-
•
vides an opportunity for self-organization: for example,
Developmental competence is conferred and
the developing intestinal endoderm (midgut/hindgut)
regulated by extrinsic signaling, lineage specifiers and
provides Shh ligand to the adjacent intestinal mesenchyme;
chromatin regulators.
in turn, the intestinal mesenchyme provides a reciprocal
• High-throughput genomic technologies (including signal (possibly Bmp4) to properly pattern the intestinal
chromatin immunoprecipitation-sequencing and endoderm (Roberts et al., 1995; Roberts et al., 1998). These
RNA-sequencing) enabled the identification interactions between inducers and developmentally compe-
of human genomic elements that regulate tent cells lay the groundwork for embryogenesis.
developmental competence in stem cells and their Sequential steps of embryonic development are driven
lineage specification. by stem cell populations that possess differing ranges of
developmental competence. Typically, developmental com-
petence becomes progressively restricted as body parts and
developmental studies, the developmental competence of organs become increasingly mature. Embryonic develop-
stem cell or progenitor populations was classicially defined ment is initially spearheaded by cells with expansive devel-
by orthotopic or heterotopic transplantation into a recipi- opmental competence. The dozen or so pluripotent cells of
ent organism (Buckingham and Meilhac, 2011; Weissman the peri-implantation epiblast at ~E4.5 in mouse devel-
et al., 1978). Orthotopic grafting of cells back into their opment are the ancestors to all the hundreds of cell types
native environment tested the typical array of lineages these of the future adult (Gardner and Rossant, 1979b). Upon
cells were intrinsically competent to generate (Weissman further developmental progression, the competence of plu-
et al., 1978). By contrast, heterotopic transplantation into ripotent cells is successively relinquished as they commit
other anatomical domains tested whether such progenitors to a number of restricted developmental paths, forming
could be “re-specified” by their extrinsic signaling environ- multipotent stem cell populations harboring the potential
ment to adopt other fates. For example, posterior epiblast to form particular organ domains or body parts and all the
(fated to form mesoderm) may be rediverted to adopt ecto- cell-types within them. Typically, during organ specifica-
dermal fates if heterotopically grafted into the anterior end tion, multipotent organ-progenitor pools proliferate copi-
of the mouse gastrula (Beddington, 1982). Results from ously to expand organ domains, thus delimiting the size
these heterotopic transplantation studies therefore suggest of the future organ in proportion to its surrounding ana-
that the developmental competence of progenitors may be tomical confines. Multipotent organ stem cell pools fur-
broader than initially thought, and that lineage commit- ther differentiate into so-called unipotent progenitors that
ment boundaries are rather labile. Therefore, redefining the are irreversibly committed to form a particular constituent
full scope of developmental potential harbored by various cell-type. Pancreatic development serves as an example of
Trophectoderm Trophectoderm
Proximal epiblast
Primitive endoderm Primitive endoderm
(Hypoblast) Epiblast (Hypoblast)
Epiblast (Post- Epiblast
(Primitive implantation) (Primitive
ectoderm) Distal epiblast ectoderm)
Figure 47.1 The dramatis personae of the stem cell ensemble, including the niches of the adult stem cells (top) and embryonic stem cells (bottom).
A D U LT S T E M C E L L S : T H E G E N E S I S
EXTRINSIC SIGNALING AND
OF CANCER?
D E VE L O PM E N TA L C O M P ET E N C E
Adult stem cells are essential to life and sustain tis-
sue homeostasis and regeneration throughout decades. In above sections, we have presented developmental com-
However, during such extended periods, long-lived stem petence as a fixed characteristic intrinsic to specific stem
cell populations may accrue genetic lesions that fore- cell or progenitor populations (Waddington, 1940; Wolff,
shadow cancer. Particularly during episodes of tissue 1968). However, developmental competence can also be
injury in which stem cell proliferation is extrinsically dic- extrinsically programmed to a certain extent: for example,
tated, one possibility is that chronic injury and repeated upon injury, restricted secretory/enteroendocrine precur-
recall of stem cells may entrap stem cells in a constitu- sors within the intestines gain expanded competency to
tively self-renewing state evocative of the uncontrolled become intestinal stem cells (Buczacki et al., 2013; van Es
proliferation evinced in cancer (Beachy et al., 2004). An et al., 2012), an effect mediated by yet-undiscovered sig-
evolutionary perspective is applicable to the way stem nals. More generally, extrinsic signals play a pivotal role in
cells and cancer are viewed together. It stands to reason altering developmental competence throughout diverse
that, within a stem cell compartment in which mutations contexts.
are being continually inflicted, individual stem cells that
gain mutations that enhance self-renewal and/or block
I N S TA L L AT I O N O F D EVE L O PM E N TA L
differentiation will gradually out-compete their peers,
C O M P ET E N C E BY E X T R I N S I C S I G NA L S
and in doing so, will come to dominate the stem cell pop-
ulation (Rossi et al., 2008). Thus, the stem cell popula- Developmental competence of multipotent stem cells/
tion is continually evolving towards a rapidly proliferative progenitors can be conferred or restricted by extrinsic sig-
state in which stem cell clones gaining successive muta- nals. For example, during renal development in the chick
tions that further enhance self-renewal are being selected embryo, where the intermediate mesoderm is the direct pre-
for, and then such clones then dominate the population cursor to the kidney fate, the dorsal neural tube surrounding
(Rossi et al., 2008). the kidney morphogenetic field secretes an extrinsic TGFβ
Whether stem cells are the genesis of cancer has been signal (Activin) to induce renal competence within these
experimentally tested in recent years. Intestinal adeno- cells (Preger-Ben Noon et al., 2009). This renal competence
mas were effectuated when oncogene β-catenin was sta- is then realized by retinoic acid, which induces the expres-
bilized in Bmi-1+ intestinal stem cells (Sangiorgi and sion of kidney genes like Pax2 and Lhx1 at the posterior
Capecchi, 2008); likewise, when tumor suppressor Apc domain of the intermediate mesoderm in vivo (Preger-Ben
was deleted in Lgr5+ intestinal stem cells (Barker et al., Noon et al., 2009).
2009). Tumor cells that express Lgr5 were multipotent Extrinsic signals also negatively restrict the devel-
cancer stem cells and were able to generate all other cell opmental competence of stem cells/progenitors in
types present in the intestinal adenomas (Schepers et al., vivo. An example can be found during hepatic develop-
2012). Of note, the same oncogenic perturbations failed ment in the zebrafish embryo, wherein the endoderm
to consistently yield adenomas when they were induced is anteriorly-posteriorly patterned to yield either ante-
in transit-amplifying intestinal progenitor cells (Barker rior endoderm (foregut, which gives rise to the liver)
et al., 2009). Therefore, considerable evidence suggests or posterior endoderm (mid/hindgut, which yields
Mesoderm Nanog
Endoderm
Oct4 Tbx3
Sall4
Trophectoderm
Sox2
Ectoderm
Zic3
pluripotency factors concurrently block differentiation to therefore generating a temporary (and precarious) state of
mutually exclusive fates while promoting differentiation developmental indecision. For example, though Sox2 pro-
to other alternative lineages. For example NANOG spe- motes ectoderm formation from ESCs, it is counteracted by
cifically promotes endoderm germ-layer formation from the inhibitory effects of Oct4 and Nanog. Likewise, Nanog
hESCs while reciprocally repressing the alternate ectoderm specifies endoderm from ESCs, yet is held in check by
fate (Teo et al., 2011; Vallier et al., 2009; Wang et al., 2012). endoderm-suppressing Sox2 (Teo et al., 2011; Wang et al.,
Therefore, the previously emphasized differentiation-block- 2012).
ing effects of pluripotency factors ( Jaenisch and Young, In conclusion, coexpression of multiple competing
2008; Silva and Smith, 2008; Young, 2011) may be explained and cross-antagonizing lineage-specifiers in ESCs enables
in the context of their lineage-specifying activities: they are them to simultaneously access diverse differentiation pro-
excluding alternate fates to channel differentiation to their grams while maintaining a temporary state of undifferen-
lineage of choice. Likewise, ectoderm-specifying factor tiated self-renewal (see Box 47.2). Furthermore, stem cell
Sox2 elicits ectoderm differentiation by blocking endoderm transcription factors may generally act as differentiation-
or mesoderm formation, whereas Oct4 specifically induces promoting lineage-specifiers throughout diverse stem cell
mesoderm commitment of ESCs by precluding ectoderm populations, generally directing stem cell differentiation to
induction (Teo et al., 2011; Wang et al., 2012). In each of a particular daughter lineage at the expense of other cell
these examples, each of these pluripotency factors probably types. Concomitant expression of multiple such transcrip-
confer ESCs with the ability to differentiate into a par- tion factors leads to a state of undifferentiated indecision
ticular germ-layer (Loh and Lim, 2011). Coexpression of yet endows individual stem cells with the ability to differ-
diverse lineage-specifying pluripotency factors (Oct4, Sox2 entiate into any of the lineages comprising the stem cell’s
and Nanog) therefore endows ESCs with the capacity to cognate tissue. This model therefore provides an underlying
form all germ layers (endoderm, mesoderm, and ectoderm) molecular explanation for stem cell multilineage potential.
and thus, pluripotency (Loh and Lim, 2011).
If the transcriptional determinants of the pluripotent
regime are each advocating a different path of differentia- S T E M C E L L G E N O M I C S : C H R O M AT I N
tion, this initially appears to be inconsistent with the fact AC C E S S I B I L I T Y E N D OW S T H E
that ESCs may be stably maintained in an uncommitted D E VE L O PM E N TA L C O M P ET E N C E O F
state in culture. We propose a model of transcriptional STEM CELLS
competition in which pluripotency factors inducing dif-
ferent developmental outcomes compete with one another, Developmental competence is the sine qua non of stem cells,
suppressing one anothers’ lineage-specifying activities and yet since the original introduction of the term over seventy
years ago, its molecular origins remain obscure. Of all pos- Chromatin is a complex macrobiomolecule that
sible developmental outcomes, how are a select number consists of genomic DNA wrapped in spools around
of developmental programs made specifically accessible to a protein core, known as histones. The basic unit of
a particular type of stem cell? As aforementioned, extrin- chromatin is the nucleosome core. Each nucleosome
sic signals and transcription factors endow multilineage consists of 147 bp of DNA wrapped around a histone
potency. However, ultimately, competence is regulated octamer core (Olins and Olins, 2003). The structure
at the level of chromatin. Since terminal commitment is and state of chromatin may affect the accessibility of
driven by induction of differentiation genes, the compe- DNA to transcriptional machinery. Chromatin state
tence to differentiate must lie in the facility of uncommit- is associated with post-translational covalent modifi-
ted stem cells to readily upregulate differentiation effectors cations on the N-terminal tails of histone moieties of
in response to developmental signaling. Hence, a following nucleosomes, including post-translational modifica-
proposition is that within stem cells, prospective devel- tions such as acetylation, phosphorylation, methylation
opmental genes must be made uniquely accessible within and ubiquitinylation at key specific residues (Cosgrove
chromatin such that they may be acutely induced and then et al., 2004; Jenuwein, 2001; Strahl and Allis, 2000).
deployed upon extrinsic instruction. Thus, at the crux of Chromatin-immunoprecipitation followed by hybrid-
competence is how developmental genes are “prefigured” ization to DNA-microarray (ChIP-chip) (Heintzman
for activation in stem cells in the undifferentiated state. et al., 2007) or high-throughput-sequencing (ChIP-seq)
• Promoters are DNA sequences that enable the binding of RNA polymerase II preinitiation complexes to
active gene transcription (Kim et al., 2005). Promoters can be active or inactive (Heintzman et al., 2007).
Active promoters are bound by RNA polymerase II preinitiation complexes (Kim et al., 2005) or marked by
trimethylation of Lys4 of histone H3 (H3K4me3) (Heintzman et al., 2007). A subset of inactive promoters
marked by both activating and repressive histone modifications are known as bivalently marked promoters (see
below) (Bernstein et al., 2006).
• Enhancers are cis-acting DNA regulatory elements that enhance the expression of its target gene from varying
orientations and positions, including upstream of, downstream of, or within a target gene(Blackwood, 1998; Bulger
and Groudine, 1999).
• Similar to promoters, enhancers can alternate between active or inactive states. Correlation between enhancer
states and the histone modifications have been studied and histone methylation states (Histone H3 on lysine
4—H3K4me1/2/3) and histone acetylation (Histone H3 Lysine 27—H3K27ac) states were found to reflect the
activity of enhancers (Pekowska et al., 2011).
• Active enhancers are marked by H3K27ac and H3K4me1 (Heintzman et al., 2007). Nevertheless, H3K4me1 and
H3K27ac are also highly enriched over gene bodies (Heintzman et al., 2007).
• Inactive or “poised” enhancers can be marked by H3K4me1 only or H3K4me1 together with H3K27me3
(Heintzman et al., 2007) respectively. The H3K4me1 and H3K27me3 labeled enhancers are labelled as “poised”
enhancers since the genes associated with these enhancers are largely inactive (Rada-Iglesias et al., 2010). It is
noteworthy that H3K4me3 is also found on some active enhancers (Pekowska et al., 2011).
• Inactive enhancers can also be devoid of any mark. Latent enhancers for example are regions of the genome that are
unbound by TFs, and that lack enhancer marks but gain them after stimulation (Ostuni et al., 2013).
• Contrary to enhancers, insulators are DNA elements that reduce transcription by obstructing the spread of the
heterochromatin and block enhancers from activating promoters (Mihaly et al., 1998; Mihaly et al., 1997; Udvardy
et al., 1985; Valenzuela and Kamakaka, 2006).
genes (e.g., Brachyury and Nodal) are paradoxically to access alternate developmental paths, consistent with
co-decorated by both activation-associated (H3K4me3) increasing lineage restriction.
and repression-associated (H3K27me3) histone modifica- Although chromatin preaccessibility of developmental
tions. However, bivalent promoters are thought to repre- genes in anticipation of differentiation has been phenom-
sent some form of transcriptional priming, signaling that enologically described in stem cells, is such chromatin prim-
various developmental genes are being poised for activation ing is functionally essential for sustaining developmental
while being kept temporarily inactive in ESCs. While such competence? For example, in the case of bivalently marked
“bivalently”-marked developmental genes are largely qui- promoters, loss of either the H3K4me3 methyltransferase
escent in ESCs, a subset of such bivalent genes is invoked Mll2 (Hu et al., 2013) or the H3K27me3 methyltransfer-
upon early differentiation. Concomitant with differentia- ase component Eed does not significantly alter mouse ESCs
tion, the bivalent state at these genes becomes unilaterally (Hu et al., 2013). For example, Mll2 depletion (and cor-
resolved: developmental genes that become active exclu- responding loss of H3K4me3 at bivalent promoters in the
sively gain H3K4me3 and lose H3K27me3 (Bernstein undifferentiated state) does not seemingly alter the ability
et al., 2006). By contrast, lineage specifiers associated with of mESCs to upregulate bivalently marked developmental
alternate, unrealized fates inherit H3K27me3 and poten- genes upon differentiation (Hu et al., 2013). Thus, whether
tially lose H3K4me3 (Bernstein et al., 2006; Xie et al., bivalent marking of promoters causally potentiates devel-
2013). Such decommissioning of bivalent domains upon opmental competence, or only correlates with it, remains
lineage commitment could track the loss of competence unclear.
Pre-enhancer states
Pioneer factors
Pioneer Factor
(e.g. FoxA)
Differentiation
Figure 47.4 Models describing chromatin regulation of stem cell developmental competence and lineage commitment.
Furthermore, it remains unclear whether bivalent were separated into vegetal or animal halves, many appar-
domains truly foreshadow the immediate lineage options ently “bivalent” genes were exclusively active (H3K4me3+)
directly available to ESCs. Firstly, many bivalent genes in or repressed (H3K27me3+) in either the vegetal or animal
ESCs are activated only at far later developmental stages domains, such that when whole gastrulae were examined,
(long after germ-layer induction or patterning). Secondly, a these genes superficially appeared bivalent (Akkers et al.,
subset of bivalent domains may be artifactual. For example, 2009). Finally, the increasing prevalence of distal enhancer
an apparently large collection of bivalently marked develop- elements in controlling precise deployment of developmen-
mental genes reside in Xenopus gastrulae; yet, when gastrulae tal genes (whereas proximal promoter elements appear to
766
Local Systemic
Severe
Sterile causes SIRS Shock
SIRS
– Fever
– Tachycardia
Trigger Organ failure Hypotension
– Tachypnea
– Leucocytosis
Severe Septic
Infection Sepsis
sepsis shock
(proven or suspected)
Figure 48.1
Relationship between local and systemic inflammatory responses, the systemic inflammatory response syndrome (SIRS) and sepsis which
may progress to severe sepsis and septic shock. Reproduced with permission from Dejager et al. 2011.114
such interventions have usually been applied unselectively novel biomarkers and disease mechanism. Our understand-
to heterogeneous patient groups, without considering the ing is therefore evolving from considering just single genes
potential influence of their genetic diversity on response. and single-nucleotide polymorphisms (SNPs) into a much
Sepsis has been described as the “graveyard” of pharmaceu- more global and integrated picture that resolves DNA
tical discovery8 because most initially promising drugs have sequence variation, the transcriptome, and the proteome in
proved ineffective. the context of a specific disease such as sepsis, pneumonia
or peritonitis.
G E N ET I C VA R I AT I O N, D I S E A S E
S US C E P T I B I L IT Y, A N D O U TC O M E ET H I C A L C O N S I D E R AT I O N S
Most genomic research in critical illness has focused on For an area of healthcare so resource- and cost-intensive,
analysis of disease susceptibility or outcome based on asso- and for many patients regrettably ultimately futile, it is
ciation with polymorphisms of single “candidate” genes. no surprise that there should be considerable debate sur-
In addition to the inherent problems of adopting such an rounding the ethics of collecting and using genomic data.14
approach discussed elsewhere in this book (see Chapter 1), Genetic studies of critically ill patients raise particular issues
a major challenge remains the heterogeneity of the patient not encountered elsewhere, including the issue of con-
population, making it very difficult to define and study a sent, as typically patients will lack capacity.15 The value of
clear disease phenotype. This is likely to confound analy- genomic approaches is controversial, but the potential ben-
sis and accounts in part for the often conflicting and efits to this group of patients should not be underestimated.
inconclusive literature in this field.9–11 The application Moreover, the opportunities presented by the depth of phe-
of more recently developed genomic approaches, includ- notypical data that can be obtained through already estab-
ing genome-wide association studies (GWAS), may prove lished electronic data-capture and management systems in
to be more informative. Expression quantitative trait loci many ICUs illustrate the scope and complexity that may be
(eQTL) analysis12 may also ameliorate some of these prob- possible in the future.16
lems by viewing expression of a particular mRNA or protein
as the outcome of interest rather than a complex phenotypi-
cal characteristic such as the development of ARDS fol- G E N ET I C A SS O C I AT I O N S I N T H E
lowing traumatic injury. The value of functional genomic C R I T I C A LLY I LL : T H E S TO RY S O FA R
techniques based on high-throughput sequencing for
large-scale analysis of multiple genes including their tran- A comprehensive review of gene association studies in infec-
scription, translation, and epigenetic factors such as histone tious diseases in general, and indeed in critical illness, is out-
acetylation and chromatin remodeling, is highlighted by side the scope of this chapter. What follows aims to give an
recent work from the ENCODE project.13 In the context of overview of the field in recent years, illustrating some of the
sepsis and critical illness, these techniques may help resolve challenges encountered. The disproportionate number of
Cytokines
Complement
Increased
TF and PAI-1
Chemotaxis
Procoagulant effect Lysosomal enzymes
Figure 48.2
The major pathways and inflammatory mediators of sepsis. Lipopolysaccharide (LPS); plasminogen-activator inhibitor-1 (PAI-1); tissue
factor (TF). Reproduced with permission from Cohen 2002.115
studies related to sepsis as opposed to other causes of critical IFN-γ promotes TNF production; studies suggest an
illness, for the reasons already discussed, will be obvious. An effect of the intron 1 (CA)n repeat of IFNG on sepsis
overview of the major pathways and inflammatory media- risk after traumatic injury,19 and intronic SNP rs2430561
tors of sepsis is given (Figure 48.2) to provide context for (IFNG+874T/A) on infectious complications following
the genetic findings presented below. surgery for esophageal cancer.20 These two polymorphisms
are in linkage disequilibrium.
P RO -I N FL A M M ATO RY M E D I ATO R S
Tumor Necrosis Factor Alpha (TNF) and Interferon Interleukins (IL-1, IL-1ra, IL-6, IL-8)
Gamma (IFN-γ) Of three IL-1 genes, two encode the pro-inflammatory
The TNF locus is highly polymorphic, with many SNPs cytokines IL-1α and IL-1β; a third produces the naturally
and microsatellites having been identified and investi- occurring inhibitor IL-1ra. Within IL-1ra intron 2, there
gated using a candidate-gene approach. Studies of the is a variable number tandem repeat (VNTR) that has
promoter SNP rs1800629 (TNF-308G/A) have reported been extensively studied in sepsis (Table 48.1), as has an
conflicting results (Table 48.1). A meta-analysis suggested IL-6 promoter SNP rs1800795 (IL6-174G/C). The SNP
an association with the development of sepsis, but not rs2814778 (DARC-46T/C) in the gene encoding Duffy
mortality.17 In contrast, a U.S. study of over a thousand antigen/receptor for chemokines is associated with worse
patients from 12 ICUs found a significant association clinical outcomes among African Americans with acute
with mortality and duration of mechanical ventilation.18 lung injury (ALI), perhaps related to increased IL-8 levels.21
7 6 8 • G e no m ic s in C l inica l Practic e
Table 48.1 SUMMARY OF SELECTED PRO-INFLAMMATORY MEDIATOR GENE ASSOCIATION STUDIES
STUDY, YEAR , GENE VARIANT OUTCOME POPULATION AND ODDS RATIO NOTES
PMID SAMPLE SIZE (N) (95% CI)
TNF
Stuber TNF -308G/A Development of Severe sepsis (80) Non signifi-
1995 (rs1800629) severe sepsis Healthy controls (153) cant (NS)
8832971
Mira rs1800629 ICU mortality Septic shock (89) 3.7
1999 Healthy blood donors (87) (1.4–10.2)
10450718
Apploni rs1800629 ICU mortality Septic shock (37) 7.4* Excluded 3 patients dying in
2001 (1.2–44.2) first 24 hours; small study
11331061
O’Keefe rs1800629 Severe sepsis Admitted to trauma center 4.6 No association with mortality
2002 (152) (1.9 –10.9)
11988644
Gordon rs1800629 Sepsis Caucasians from eight NS Also examined receptor
2004 Illness severity U.K. and Australian ICUs genes, TNFRSF1A and B
15526005 Outcome (213)
Gill rs1800629 Microchimerism Transfused U.S. trauma 5.7* Immunological consequences
2008 following blood patients (59) (1.2–27.3) uncertain
18199828 transfusion
Shalhub rs1800629 Hospital mortality Severe burns admitted 10.7 Largely accounted for by
2009 U.S. regional center (69) (1.2–95.5) septic shock and multi-organ
19060757 dysfunction (MODS)
Paskulin rs1800629 Sepsis, septic shock Critically ill Caucasian NS
2011 Organ dysfunction general ICU patients from
21670923 Mortality southern Brazil (520)
IL1RN
Arnalich IL-1RN*2 (A2 allele) 30-day mortality Severe sepsis (78) 6.47 Associated with lower PBMC
2002 homozygosity (1.0–41.5) IL-1ra production in vitro
11876758 (VNTR intron 2)
Patwari Absence of Lung injury severity U.S. children of mixed eth- 2.65 Suggestion also of association
2008 IL-1RN*1 (A1 allele) nicity with CAP (847) (1.0– 6.9) with ARDS development
18838927
Hsu IL-1RN*2 Acinetobacter bau- Taiwanese with pneumonia 0.118 Apparent gene dosage effect
2012 homozygosity mannii pneumonia (66) (0.02–0.76) of A1 allele on risk
22558011
IL6
Burzotta IL-6 -174G/C Postoperative death, Italians undergoing elec- 0.24* GG homozygotes had longer
2001 (rs1800795) myocardial infarc- tive coronary artery bypass (0.03–2.2) NS ICU and hospital stays
11703956 tion, or stroke grafting (CABG) (111)
Baier rs1800795 Late bloodstream African-American mechan- 2.6
2006 infections ically ventilated very low (1.1–6.0)
16611358 birthweight infants (233)
Tischendorf rs1800795 Shock German Caucasians with 2.83* -174C conferred low serum
2007 severe sepsis and septic (1.2–7.0) IL6 but high in vitro secre-
18001296 shock (112) tion after LPS stimulation
Martin-Loesches rs1800795 ARDS Caucasian Spanish patients 2.35* Further associations in those
2012 with CAP (1227) (1.2–4.5) with proven S. pneumoniae
22113815
IL8
Hildebrand rs4073 ARDS Germans with blunt mul- NS Increased duration of
2007 IL-8 -251T/A tiple trauma (97) mechanical ventilation
17498967 observed
STUDY, GENE VARIANTS POPULATION AND SAMPLE SIZE (N) SUMMARY FINDINGS
YEAR , PMID
Shu -592C/A (rs1800872) Severe sepsis (116) rs1800896 possibly associated with susceptibility to severe
2003 -819T/C (rs1800871) Chinese sepsis
14642153 -1082G/A (rs1800896) rs1800872 and rs1800871 in complete linkage
disequilibrium and not associated with incidence or
outcome of severe sepsis
Lowe rs1800872 Critically ill (67) rs1800872(A) associated with increased mortality and
2003 Healthy volunteers (132) lower stimulated IL-10 release in vitro
12544990 United Kingdom
Baier rs1800896 Mechanically ventilated very low rs1800896(A) associated with increased incidence of late
2006 birthweight infants (293) bloodstream infection
16611358 Predominantly African-American
Gong rs1800896 Sepsis, trauma, aspiration or massive rs1800896(G) homozygosity associated with lower
2006 transfusion developing ARDS (211) mortality and organ failure
16585075 Controls (429)
Caucasian North Americans
Huebinger rs1800872 Burns (265) rs1800872(A) and/or rs1800871(T) associated with
2010 rs1800871 Healthy volunteers (31) reduced risk of death after burn injury
20863526 North Americans
Patients are adults unless otherwise specified.
PMID = PubMed Identifier.
7 7 0 • G e no m ic s in C l inica l Practic e
TLR4 TLR2 IL-1R/IL-18R
MD2
CD14 CD14
CELL MEM
BRANE
TIR
TIR
TIR
MAL
ENDOSOME
TLR3 TLR4 MyD88 MyD88 TLR7/8 TLR9
IRAK1
TIR
TIR
TIR
TIR
TRAM
IRAK4
TRIF TRIF TRAF6 MyD88 MyD88
TAK1
IKKα IKKβ
IKKγ
TRAF3 MAPKs TRAF3
TBK1 TBK1
IκBγ
NFκB
CYTOPLASM
NUCLEUS
DNA IRFs AP-1 NFκB IRFs
Figure 48.3
Overview of innate imunity signalling pathways examined in critical illness gene association studies. Activating protein (AP)-1; IκB
kinase (IKK); IL-1 receptor-associated kinase (IRAK); interferon regulating factors (IRFs); interleukin (IL)-1; mitogen-activated protein kinase
(MAPK); myeloid differentiation factor 88 (MyD88); MyD88 adaptor-like (MAL); nuclear factor (NF) κB; TANK-binding kinase (TBK); TIR
domain containing adaptor-inducing interferon β (TRIF); TNF-associated factor (TRAF); Toll-like receptor (TLR); transforming growth factor
β-associated kinase (TAK); TRIF-related adaptor molecule (TRAM). Reproduced with permission from Lundberg & Hansson 2010.116
components,33 heat shock proteins,34 and nitric oxide pneumoniae infection. Importantly, however, a low MBL
synthase.35 Figure 48.3 illustrates the ways in which these level appeared to be a more reliable indicator of defi-
molecules interact. ciency than simply carriage of a low-producing MBL2
Mannose-binding lectin (MBL) binds carbohydrates genotype.
on the surface of bacteria, fungi, and viruses to promote Other molecules involved in the innate immune
complement activation and phagocytosis. A combina- response that have received considerable attention
tion of promoter and structural mutations results in a include: CD14, a membrane-anchored protein promot-
deficiency state affecting between a quarter and a third ing the TLR4 response to ligands; LPS-binding protein
of the population. Variant alleles were over-represented (LBP) and bactericidal-permeability–inducing protein
in children developing SIRS requiring ICU admission.36 (BPI), related lipid-transfer proteins that recognize LPS;
A similar finding in adult patients with sepsis has been myeloid-differentiation protein 2 (MD-2), a glycoprotein
reported,37 and lower circulating MBL levels are asso- that confers LPS responsiveness to TLR4-expressing cells;
ciated with a poor outcome.38 The importance of this nucleotide-binding oligomerization domain (NOD) pro-
innate immune-defense protein was highlighted by a teins; and high-mobility group box (HMGB1), a nuclear
meta-analysis39 showing that risk of death was increased and secreted protein. Brief details of studies of interest are
among MBL-deficient patients with Streptococcus given in Table 48.3.
7 7 2 • G e no m ic s in C l inica l Practic e
Procoagulant pathways Anticoagulant pathways Complement
Sepsis
Plasminogen
Activation of complement is a key feature of sepsis; in
Tissue factor PAI-1
+
activators severe sepsis, C5a elevation is associated with increased
Factor X Factor Vlla Protein C mortality.51 In pediatric critical illness, homozygosity for
–
– Plasminogen complement factor H (CFH) 1277T>C (rs1061170) was
aPC
Factor Xa
–
associated with a reduced incidence of early SIRS and
Factor Va
Plasmin sepsis.52 Factor H inhibits complement activation, and
Prothrombin this variant modifies a protein region involved in binding
heparin and CRP. The meningococcus attempts to evade
Fibrin
Thrombin complement-mediated killing by expressing a protein that
FDP
binds CFH. In the specific setting of childhood menin-
Increased
fibrinogen
Impaired
fibrinolysis
gococcal disease, a GWAS has reported susceptibility to
Enhanced formation infection to be significantly associated with two CFH SNPs
of fibrin blood clots
(rs1065489 and rs11582939).53
Thrombosis of small vessels
7 7 4 • G e no m ic s in C l inica l Practic e
complications or systemic procalcitonin levels was observed Results from a murine sepsis model and cDNA microar-
with a variety of different alleles.83 ray system have suggested a collection of specific genes to be
differentially expressed in Gram-positive and Gram-negative
sepsis.91 In a general pediatric hospital population admit-
T R I G G E R I N G R EC E P TO R E X P R E S S E D
ted with acute infections, microarray analysis of peripheral
O N MY E L O I D C E L L S ( T R E M-1)
blood leukocytes appeared able, with a high degree of accu-
Soluble triggering receptor expressed on myeloid cells, racy, to differentiate between influenza A, Escherichia coli,
or (s)TREM-1, a member of the immunoglobulin super- Streptococcus pneumoniae, and Staphylococcus aureus infec-
family, is increased in bronchoalveolar fluid before the tions,92 although blood sampling was performed some days
development of clinical signs of infection in patients into the acute illness. In a small study of critically ill adults,
with ventilator-associated pneumonia.84 In Chinese Han however, no significant differences were demonstrated at
patients, TREM-1 SNPs rs7768162, rs9471535, and the host transcriptome level between Gram-positive and
rs2234237 were not found to be associated with suscep- Gram-negative sepsis.93 In pediatric septic shock, a biologi-
tibility to or outcome of severe sepsis.85 Higher sTREM-1 cally relevant 100-gene expression signature was found to
levels in another group of Chinese ICU patients predicted distinguish three patient groups with important clinical dif-
death, but whilst rs2234237 was significantly associated ferences,94 such as repression of genes involved in adaptive
with sepsis prognosis, no effect of this SNP on sTREM-1 immunity and glucocorticoid-receptor signaling. Such het-
concentrations was evident.86 erogeneity in the genomic response to sepsis might explain
apparent failures of novel therapies in previous studies and
in the future may allow interventions to be targeted to those
FUNCTIONAL GENOMICS most likely to benefit.