You are on page 1of 11

REVIEWS

Zebrafish as tools for drug discovery


Calum A.MacRae14 and Randall T.Peterson36
Abstract | The zebrafish has become a prominent vertebrate model for disease and has
already contributed to several examples of successful phenotype-based drug discovery.
For the zebrafish to become useful in drug development more broadly, key hurdles must be
overcome, including a more comprehensive elucidation of the similarities and differences
between human and zebrafish biology. Recent studies have begun to establish the
capabilities and limitations of zebrafish for disease modelling, drug screening, target
identification, pharmacology, and toxicology. As our understanding increases and as
the technologies for manipulating zebrafish improve, it is hoped that the zebrafish
will have a key role in accelerating the emergence of precision medicine.

Phenotype-based screening
We are currently witnessing a resurgence of interest in As a direct result, compounds advancing from phenotypic
A screen in which the assay phenotype-based screening in drug discovery. Phenotypic screens are often of higher quality than hits from invitro
output is a complex cellular or effects of small molecules were historically the basis of target-based screens.
organismal phenotype that all drug discovery, but over the past 30years this strategy
integrates multiple biochemical
has been largely replaced by target-based approaches. Advantages of screening in zebrafish
pathways and often captures
much of the native biological There is now a growing appreciation that firstinclass Small-molecule screens in zebrafish represent a small
context. drug discovery successes are emerging disproportion- but growing fraction of phenotype-based screens
ately from the remaining phenotype-based efforts. (TABLE1). Zebrafish screens provide the advantages of
For example, a recent analysis of firstinclass drugs phenotype-based screens outlined above, but they also
approved between 1999 and 2008 revealed that 62% were offer unique advantages that come from screening in an
1
Cardiovascular Medicine discovered by phenotype-based screens, despite the fact intactanimal.
and Network Medicine that such screens represented only a small subset of the
Divisions, Brigham and overalltotal1. Broad range of accessible biology. Most phenotypic
Womens Hospital, Boston,
Massachusetts 02115, USA.
Several factors may explain the apparent superiority screens are carried out in cultured cells, which limits
2
Harvard Stem Cell Institute, of phenotype-based approaches over target-based the screens to cell-autonomous phenotypes or those end
Cambridge, Massachusetts screening efforts. First, phenotypic screens can discover points that can be observed in relatively simple culture
02138, USA. efficacious drugs in the absence of a validated target. systems. Zebrafish screens are typically carried out in
3
Department of Medicine,
For example, ezetimibe was discovered based on its living zebrafish embryos or larvae, which exhibit a
Harvard Medical School,
Boston, Massachusetts cholesterol-lowering activity years before NiemannPick diverse repertoire of biological processes and possess
02115, USA. C1like protein 1 (NPC1L1) was validated as a thera- fully integrated vertebrate organ systems. As such, a
4
Broad Institute of MIT peutic target 2,3. Second, phenotypic screens can identify much broader range of phenotypes can be assayed in
and Harvard, Cambridge, compounds that produce a therapeutic effect through zebrafish than in cultured cells. Pain, sedation, tumour
Massachusetts 02142, USA.
5
Cardiovascular Research
simultaneous activity at multiple targets. Thus, amioda metastasis, vascular tone, and gut motility are examples
Center, Massachusetts rone, which remains the definitive antiarrhythmic agent of disease-relevant phenotypes that are observable in
General Hospital, Charlestown, decades after its serendipitous discovery, exhibits activity zebrafish yet simply inaccessible to modelling in cultured
Massachusetts 02129, USA. on multiple ion channels, adrenergic receptors, and cells. The advantages of the intact animal as a focus for
6
Department of Systems
possibly even nuclear hormone pathways4. In addition, screening are particularly evident for neuroscience drug
Biology, Harvard
Medical School, Boston, there is evidence that the acute effects of the intravenous discovery, in which the complexities of cellcell interac-
Massachusetts 02115, USA. form may be attributable in part to the solvent Tween80 tions and endocrine signalling defy even modelling with
e-mails: camacrae@bics.bwh. (REF.5). Third, phenotypic screens often combine screen- patient-derived induced pluripotent stem cells (iPSCs).
harvard.edu; peterson@cvrc. ing and counter-screening (for example, against relevant By contrast, with emerging automated technologies and
mgh.harvard.edu
doi:10.1038/nrd4627
organ-specific toxicity reporter lines) in the same assay, zebrafish screening, it is possible to imagine relatively
Published online discovering compounds that produce a desired effect unbiased capture of a substantial proportion of the
11 September 2015 while parsing out compounds with undesirable qualities. complete phenotypic repertoire.

NATURE REVIEWS | DRUG DISCOVERY VOLUME 14 | O CTOBER 2015 | 721

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

Table 1 | Selected examples of published chemical screens in zebrafish


Screen type Phenotypic readout Refs
Cardiotoxicity screen Heart rate 19
Suppressors of aortic coarctation Blood circulation in the aorta 68
Suppressors of mitotic defects Phospho-histone H3 staining 69
Haematopoiesis Stem cell marker expression 40
Tissue regeneration Extent of fin regeneration 77
Angiogenesis Vascular morphology in transgenic animals 78
Embryogenesis Embryonic morphology 44
Suppressors of ototoxicity Number of surviving hair cells 50
Suppressors of leukaemia Expression of myeloid and erythroid markers 52
Fgf pathway reporter screen Fgf reporter expression 79
Suppressors of polycystic kidney disease Presence of laterality defects/body curvature 70
Behavioural sleep screen Bouts of rest and wakefulness 72
Behavioural screen Response to photic stimuli 73
Dietary lipid absorption Processing of fluorescent lipid analogues 80
Modifiers of copper homeostasis Pigmentation and notochord defects 81
Angiogenesis Vascular morphology in transgenic animals 82
Suppressors of long QT syndrome Atrioventricular heart rhythm 55
Suppressors of melanoma Expression of neural crest markers 83
Learning assay Habituation to acoustic startle 74
Fluorescent reporter Fgf signalling, dusp6 expression 84
Inducers of -cell differentiation Number of fluorescent -cells in pancreas 85
Embryogenesis Pigment cell patterning and number 86
Hair cell regeneration GFP expression in hair cells after ablation 87
Modifiers of hypertrophic cardiomyopathy Natriuretic peptide reporter line 88
Target-based screening Toxicology screen Death and overt structural defects 89
A screen in which the assay
Suppressors of leukaemia Fluorescent Tcells in thymus 90
output is the activity of a
specific molecular target Luciferase reporter Glucocorticoid signalling reporter 91
in a well-defined but often
heterologous context.
Suppressors of organophosphate toxicity Survival 92
Modulators of gluconeogenesis Luciferase reporter of pck1 expression 12
Counter-screening
To screen in parallel for Behavioural screen Motor activity 28
secondary end points that
Suppressors of cyanide toxicity Survival 93
might condition the final
output of the primary screen. Leukocyte migration Migration of leukocytes to wound 94
Together, such screens enable
the incorporation of simple Modifiers of RAS activity dusp6 expression by insitu hybridization 95
logic; for example, a Embryogenesis Morphological defects in the embryo 58
counter-screen might identify
compounds with a specific Embryogenesis Embryo dorsalization 96
form of toxicity, allowing the
Suppressors of Dravet syndrome Inhibition of convulsive behaviours 71
hits from the primary screen to
be weighted appropriately for Embryogenesis Inducers of ectopic tail formation 97
subsequent evaluation.
Suppressors of leukaemia Death of MYC-expressing thymocytes 57
Toxicity reporter lines Craniofacial development Craniofacial morphology 98
Genetically modified zebrafish
lines expressing specific tissue Suppressors of Wnt-activated cancers Presence of eyes in BIO-treated zebrafish 99
damage reporters under an
Suppressors of cardiomyopathy Normalization of natriuretic peptide levels 30
organ-specific promoter. These
lines will release heterologous Stimulators of -cell proliferation Invivo cell cycle indicator technology 100
reporter peptides that can
then be detected using a range Suppressors of cardiomyopathy Rescue of cardiac function 29
of high-throughput detection
Modifiers of glucose homeostasis Biochemical measurement of glucose 101
methods.

722 | O CTOBER 2015 | VOLUME 14 www.nature.com/reviews/drugdisc

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

Table 1 (cont.) | Selected examples of published chemical screens in zebrafish


Screen type Phenotypic readout Refs
Embryogenesis Embryonic morphology 102
Cell migration Migration of labelled lateral line primordium 103
Tumorigenesis Liver size 104
Adult transplantation Imaging of fluorescent stem cell grafts 105
Fluorescent reporter Number of fluorescent -cells in pancreas 106
dusp6, dual specificity phosphatase 6; Fgf, fibroblast growth factor; GFP, green fluorescent protein; pck1, phosphoenolpyruvate
carboxykinase 1.

Early insight into toxicity. Whereas cell-based assays at active sites in the enzymes, channels and receptors that
provide limited information about the absorption, distri- are frequently the targets of drugs because these sites have
bution, metabolism, excretion and toxicity (ADME-Tox) retained the ability, through evolution, to bind to the same
of screening compounds, zebrafish screens often reveal metabolites, ions and other biomolecules. For example,
insights about these pharmacological characteristics. the zebrafish glucocorticoid receptor, which is only about
Zebrafish larvae possess functional livers, kidneys and 50% identical to the human receptor overall, is 74% iden-
bloodbrain barriers68. To produce invivo phenotypes tical in the carboxyterminal half of the protein, which
in zebrafish assays, compounds must exhibit the ability contains the ligand-bindingdomain.
to be absorbed, reach the target tissue, and avoid rapid
metabolism and excretion. This fact may explain the Physiology. Among model systems that are amenable
observation that several compounds that were discov- to screening, zebrafish stand out for their highly con-
ered in zebrafish screens have been rapidly translated to served integrative physiology. Unlike yeast, worms or
invivo mammalian models with minimal optimization flies, zebrafish possess recognizable organ systems
of pharmacological properties. livers, hearts, kidneys, pancreases and so on. Although
there are some major differences resulting from adapta-
Are zebrafish relevant for human drug discovery? tion to aquatic life, most zebrafish organs perform the
Being able to perform high-throughput phenotypic same functions as their human counterparts and exhibit
screens in an invivo context is theoretically very attrac- well-conserved physiology. For example, the zebrafish
tive, but how relevant is the output from zebrafish screens pancreas contains islets comprising , , , and cells
for human biology? Recent studies have cast doubt on the that regulate glucose homeostasis by secreting glucagon,
validity of some well-established rodent disease models, insulin, somatostatin and ghrelin, just as in humans11.
reminding us again of the potential to be misled by models Drugs that modulate glucose homeostasis in humans
of any form9. If it is important to understand the capa- have been shown to have the same effects in zebrafish12.
bilities and limitations of mammalian models, the need The zebrafish haematopoietic system is highly similar
is even greater for zebrafish, which are phylogenetically to the human system and consists of the same cell types
further removed from humans. Over the past few years, erythrocytes, neutrophils, eosinophils, lymphocytes,
we have begun to understand how zebrafish compare to macrophages and so on13. Haematopoietic processes are
humans in terms of targets, physiology, drug metabolism conserved, as are globin switching and iron homeostasis
and pharmacology, in particular during the first few days through the hepcidinferroportin pathway 14,15. In fact,
of life when it is possible to house them in multi-well key components of these processes were first discov-
plates suited for screening (FIG.1). Nevertheless, questions ered in zebrafish16. Drugs affecting haematopoiesis and
about relevance to humans remain of central importance anaemia in humans have similar effects in zebrafish17.
to the use of zebrafish in drug discovery. Cardiovascular physiology is also highly conserved
between humans and zebrafish at anatomical, cellular
Targets. With a high-quality zebrafish genome now avail- and membrane-biology levels. Many human cardiovas-
able, it appears that 71% of human proteins (and 82% of cular drugs have been shown to have identical effects
disease-causing human proteins) have an obvious ortho- on zebrafish physiology, and numerous human cardio-
logue in zebrafish10. Orthologous zebrafish proteins are vascular disorders have been recapitulated in zebrafish
reasonably similar to their human counterparts, particu- genetic models18. Interestingly, the cardiac electrophysi-
larly within functional domains. For example, the protein ology of humans is more similar to that of zebrafish than
targets of the ten most-prescribed drugs have zebrafish it is to that of rodents, adding to a list of cases in which
orthologues with sequence identity ranging from 54% zebrafish physiology may be more relevant than the
(glucocorticoid receptor) to 91% (thyroid receptor). Given rodent counterpart 1921. Many other examples of orthol-
the sequence divergence between zebrafish and human ogy between humans and zebrafish exist; however, per-
proteins, one might expect a modest rate of conservation haps the ultimate consequence of the throughput feasible
of pharmacological effect. In reality, however, the rates of in this organism will be a much more global understand-
conservation are relatively high (see below). One potential ing of how representative zebrafish are of human biology
explanation for this is that the target similarity is greater compared to most other preclinicalmodels.

NATURE REVIEWS | DRUG DISCOVERY VOLUME 14 | O CTOBER 2015 | 723

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

repolarization-related toxicity 19. Similarly, drugs with


specific effects on cardiac contractility and vasomo-
tion in humans consistently recapitulate these effects in
the zebrafish24. Although there may be reporting bias
favouring the publication of cases in which pharmacol-
ogy is conserved, these examples suggest a high prob-
ability of direct correlation between effects in humans
m/z
Conserved targets Conserved drug metabolism and zebrafish. In our laboratories, we have tested ten
compounds discovered via diverse zebrafish screens in
corresponding rodent disease models. Eight of the ten
compounds produced the desired effect in rodents, with
little or no optimization via medicinal chemistry or for-
mulation2531. Although the sample size is small, these
experiences have led us to conclude that conservation of
pharmacological effect is high for the majority of drugs
where the phenotypic correlation is rigorous.

Drug distribution, metabolism and excretion. As small


molecules were systematically tested in the zebrafish,
Conserved it became apparent that not only were the effects of
Conserved physiology pharmacology individual human drugs replicated but so too were the
majority of drugdrug interactions. These findings sug-
Figure 1 | How suitable are zebrafish for discovering
human drugs? Although zebrafish enable invivo studies gested that the distribution, metabolism and excretion
on a large scale, questionsNature
remainReviews | Drug
about how Discovery
relevant the of drugs might also be accessible in zebrafish model-
findings are for drug discovery. Recent studies have begun ling. There is now strong evidence not only of conserved
to elucidate the degree of conservation between humans partitioning of drugs into different passive compart-
and zebrafish, which share 82% of disease-associated ments based on physicochemical characteristics but
targets and a large number of drug metabolism pathways also of the existence of the regulation of drug distribu-
(as illustrated by the mass spectrometry of metabolic tion across active physiological boundaries such as the
by-products). Zebrafish physiology is often well conserved bloodbrain barrier and by conserved tissue-specific
(sometimes more so than rodent physiology); for example,
transporters3235.
cardiac electrophysiology (illustrative electrocardiogram
Although comparative genomics have established
(ECG) tracings from each organism are shown). Several
compounds discovered in zebrafish screens have been that the zebrafish possesses a full complement of
shown to exhibit similar effects in rodent models and cytochrome P450 (CYP) genes, genome duplication
humans, including eight out of ten compounds tested events and functional redundancy have hampered com-
by our groups. m/z, mass to charge ratio. prehensive study of the conservation of the metabolic
processing of small molecules8,36. The conservation of
other drug metabolism pathways is only beginning to
be explored, but to date there is evidence of substan-
Some effort has been made to characterize the par- tial functional parallels across diverse mechanisms,
allels between different stages of zebrafish development including multiple non-CYP enzymes34,37. Not unex-
and those of humans. In these cases, we have observed pectedly, there is also evidence that great care must be
that the key molecular transitions in each of the organs taken in simple extrapolation, as drug metabolism may
relevant for disease modelling or drug metabolism appear vary widely across different developmental stages and,
to take place in sequence, but often much more rapidly as with other models, the presence of one part of any
in zebrafish than in humans. For example, the sequential pathway does not infer the conservation of the remain-
electrophysiological maturation of the zebrafish heart, ing components8,38. Finally, drug excretion is clearly
which has been characterized at the resolution of indi- regulated in the zebrafish, but few studies of drug filtra-
vidual ionic currents, takes place within 96hours post- tion, reabsorption or excretion have been undertaken to
fertilization, whereas some of the parallel events are not date. As mass spectroscopy techniques advance, more
completed until adolescence in humans19,22,23. generalizable approaches to ADME in the zebrafish are
emerging 36.
Pharmacology. How often will a compound discovered
in zebrafish retain its efficacy in humans? This critical Examples of success
question has not yet been answered because only a few To date, more than 65 small-molecule screens in
compounds discovered in zebrafish have been tested in zebrafish have been reported in the literature (TABLE1).
humans. However, it has been possible to test a reason- These screens have targeted diverse phenotypes ranging
able number of human drugs for conserved effects in from embryo morphology to cardiac physiology and
zebrafish. For example, 23 drugs known to exhibit repo- sleep. Some have identified repurposing opportunities
larization cardiotoxicity were tested for their effects for existing drugs, whereas others have discovered novel
on zebrafish, and 22 of the 23 compounds produced compound classes. These screens have been surveyed in

724 | O CTOBER 2015 | VOLUME 14 www.nature.com/reviews/drugdisc

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

detail elsewhere39. Here, we focus on select examples of replicate the dorsalized phenotype in zebrafish embryos,
disease-relevant compounds that have been discovered suggesting that AMPK was not the dorsalizing target.
by zebrafish screens in the hope that they will illustrate Instead, dorsomorphin-treated animals were indistin-
the diverse methodologies accessible for drug discovery guishable from a previously identified genetic mutant
in this organism. called lostafin, which harboured a mutation in Alk8
(REF.44). This phenotypic similarity suggested that dor-
Prohema. Prohema is a stabilized derivative of prosta- somorphin may inhibit Alk8 in zebrafish and ALK2 in
glandin E2 (PGE2) that is currently in PhaseII trials in humans a hypothesis that has since been confirmed.
patients undergoing umbilical cord blood (UCB) trans- As the first small-molecule inhibitor of the BMP path-
plantation for leukaemia or lymphoma. Prohema was way, dorsomorphin and derivatives such as LDN193189
discovered in a zebrafish screen for compounds that have become widely used probes for manipulating BMP
increased or decreased the numbers of haematopoietic signalling 45.
stem cells (HSCs)40. Automated insitu hybridization was Hyperactive BMP signalling causes a variety of dis-
used to stain HSCs in the aortagonadmesonephros ease states, including FOP and anaemia of inflammation.
region of the embryo, and compounds that enhanced The dorsomorphin derivative LDN193189 has proven
PGE2 synthesis were found to increase HSC numbers. to be effective in treating rodent models of FOP and
The ability of PGE2 to enhance production of HSCs was anaemia of inflammation15,25,46. In addition, it has been
later tied to its interaction with WNT signalling 41. shown to be effective in treating mammalian models of
The therapeutic potential of Prohema was first dem- other BMP-linked conditions, including arterial calci-
onstrated in adult zebrafish, in which it enhanced the fication and inflammatory bowel disorder 4749. Because
rate of marrow recovery after sub-lethal irradiation. of the apparent promise of BMP inhibitors for treating
In mice, whole bone marrow exposed to Prohema these conditions, the US National Institutes of Health
exvivo formed more spleen colony-forming units when has selected dorsomorphin and its derivatives for pre-
transplanted into irradiated mice than did untreated clinical development through its TRND (Therapeutics
marrow 40. Similarly, Prohema-treated whole bone mar- for Rare and Neglected Diseases) and BrIDGs (Bridging
row more effectively reconstituted the haematopoietic Interventional Development Gaps) programmes, which
system and contributed more HSCs to short-term will support preclinical development through early
and long-term repopulation than did untreated whole clinical trials in FOP and anaemia of inflammation,
bonemarrow. respectively.
For humans undergoing HSC transplantation as
therapy for leukaemia or lymphoma, UCB has become PROTO1. PROTO1 and its derivatives are benzothio-
an important source of transplantable HSCs. However, phene carboxamides that are currently in development
the number of HSCs present in a typical UCB sample for preventing antibiotic-induced hearing loss. Because
is often suboptimal for treating adults. Therefore, the aminoglycoside antibiotics can cause hearing loss by
apparent ability of Prohema to boost HSC numbers and killing hair cells in the human ear, a zebrafish screen
its repopulating ability may enhance the effectiveness of of 10,960 compounds was conducted to identify those
UCB transplantation. On the basis of this idea, Prohema compounds that could protect zebrafish hair cells from
is currently in a PhaseII clinical trial for the exvivo con- aminoglycoside-induced death50. PROTO1 and a struc-
ditioning of UCB before transplantation into patients turally related benzothiophene carboxamide were found
with leukaemia or lymphoma42,43. to be potent otoprotectants in zebrafish, and to protect
hair cells from aminoglycoside death in cultured murine
Dorsomorphin. Dorsomorphin and its derivatives are utricles.
Fibrodysplasia ossificans inhibitors of the bone morphogenetic protein (BMP) PROTO1 has been licensed to Oricula Therapeutics,
progressiva receptor: in humans this protein is activin receptor-like a company dedicated to developing the benzothio-
(FOP). A rare autosomal
kinase 2 (ALK2, also known as ACVR1) and in zebrafish phene carboxamides as otoprotectants. Oricula reports
dominant condition caused by
gain-of-function mutations in it is Alk8 (also known as Acvr1l). These compounds are that lead optimization has yielded analogues with 100
the gene encoding activin currently in development as therapies for fibrodysplasia times higher potency than PROTO1, a wide safety
receptor-like kinase 2 (ALK2). ossificans progressiva (FOP) and for anaemia of inflammation margin and efficacy in rat models of aminoglycoside-
These mutations result in (FIG.2). Dorsomorphin was discovered in a zebrafish induced hearing loss. Investigational new drug (IND)-
chronic activation of the bone
morphogenetic protein (BMP)
screen that sought to identify compounds that perturb enabling studies are ongoing (Oricula Therapeutics and
pathway with resultant the establishment of the basic body organization during M.Gleser, personal communication).
formation of ectopic bone in early embryogenesis44. In screening a library of a few
muscle tissue. Restriction of thousand compounds, a pyrazolopyrimidine was iden- Repurposing existing drugs
the thoracic skeleton then
tified that causes dorsalization the expansion of dorsal In addition to discovering novel compounds with thera-
leads to respiratory failure,
usually in childhood. tissues at the expense of ventral tissues. Treated embryos peutic potential, zebrafish screens have proven useful for
developed with reduced or missing tails, and the pyra- identifying novel uses for existing drugs. As the follow-
Anaemia of inflammation zolopyrimidine was named dorsomorphin to reflect its ing examples illustrate, repurposing screens can provide
A form of anaemia that is dorsalizing activity. an abbreviated path to clinical investigation because the
characterized by a block in iron
availability for haematopoiesis
The dorsalizing pyrazolopyrimidine had previously compounds they identify have already been approved
and is observed in many been described as an inhibitor of AMP-activated protein for human use or at least have been subjected to prior
chronic inflammatory diseases. kinase (AMPK), but other AMPK inhibitors failed to pharmacokinetic analysis and safety testing.

NATURE REVIEWS | DRUG DISCOVERY VOLUME 14 | O CTOBER 2015 | 725

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

was used to demonstrate that glucocorticoids shorten


action potential duration in LQT hearts 55. Because
glucocorticoids are clinically approved and well tol-
erated in humans, it was possible to quickly test the
clinical hypothesis that glucocorticoids would provide
therapeutic benefit in patients with LQT syndrome.
A clinical trial of the glucocorticoid dexametha-
sone in patients with LQT syndrome was initiated at
HN
Massachusetts General Hospital, USA; early results are
N
consistent with the hypothesis that dexamethasone may
N
shorten QT intervals in humans (D. Milan, personal
N
communication).
N
Challenges in determining mechanisms of action
As with other types of phenotype-based small-molecule
N discovery, determining the mechanism of action (MOA)
remains one of the most substantial hurdles for small
molecules discovered in zebrafish screens. Although
challenging, MOA studies are among the most rewarding
aspects of phenotype-based drug discovery because
they often reveal unexpected and transformative new
biological insights into the disease under investigation.
Figure 2 | BMP pathway inhibitors discovered by invivo screens in zebrafish. A screen Other reviews have described the rich variety of com-
for disruptors of early embryogenesis identified a pyrazolopyrimidine called putational, biochemical and genetic techniques that
dorsomorphin that dorsalized zebrafish embryos by inhibiting the zebrafish
Nature Reviews orthologue
| Drug Discovery can be used to discover MOAs56. Each of these tech-
of the bone morphogenetic protein (BMP) receptor activin receptor-like kinase 2 (ALK2). niques can be used for small molecules discovered in
Dorsomorphin derivatives such as LDN193189 (shown) have since been shown to be zebrafish screens, and several excellent examples have
efficacious in several mammalian models of BMP-associated diseases, including the bone
been reported. For example, affinity chromatography
overgrowth disorder fibrodysplasia ossificans progressiva (FOP). Figure from REF.25 and
coupled with mass spectrometry was used to identify
REF.44, Nature Publishing Group.
the A subunit of protein phosphatase 2A (PP2A) as the
target of perphenazine, an antipsychotic found to kill
MYC-overexpressing thymocytes in a zebrafish model
COX inhibitors in leukaemia. A zebrafish model of of T cell acute lymphoblastic leukaemia (TALL)57.
acute myeloid leukaemia (AML) was generated by Similarly, affinity chromatography was used to determine
transgenic expression of the human AML oncogene that mitochondrial malate dehydrogenase is the target of
AML1ETO51. Transgenic zebrafish accumulated mye- visnagin, a cardioprotective compound discovered in a
loid blast cells reminiscent of those from patients with zebrafish heart-failure screen30.
AML. A zebrafish screen of existing drugs identified Beyond the generic approaches to identify the MOA,
Long QT syndrome cyclooxygenase (COX) inhibitors as potent suppressors there are zebrafish-specific tools that can aid in MOA
A disorder in which restoration of the leukaemia-like phenotype and implicated WNT determination. One of the most effective is the large
of the membrane potential to
equilibrium after a cardiac
catenin signalling in leukemogenesis52. Subsequent collection of zebrafish phenotypes that have been asso-
action potential is delayed as studies in mice confirmed that WNTcatenin sig- ciated with specific gene mutations and knockdowns.
a result of abnormal ion fluxes. nalling is required for self-renewal of leukaemia stem Identifying similarity between a drug-induced phenotype
This delay can be detected cells53; moreover, treatment of murine AML models with and a genetic phenotype can often provide clues as to the
from the simple measurement,
COX inhibitors suppressed leukaemia xenograft forma- principal drug target or targets (FIG.3). In the case of dor-
on a surface electrocardiogram
(ECG), of the time from tion and inhibited invivo progression of transplanted somorphin described above, similarity to the lostafin
depolarization onset until the human leukaemia cells27. Because COX inhibitors are phenotype revealed the BMP pathway as the target of
return to baseline membrane clinically approved and well tolerated in humans, it was dorsomorphin. Embryos treated with dorsomorphin
potential (the QT interval). possible to quickly carry out a PhaseI trial to test the developed missing the ventral side of their tail fins an
2:1 atrioventricular
clinical hypothesis that COX inhibitors would provide unusual phenotype that almost perfectly replicates the
heart block a therapeutic benefit in patients with AML54. phenotype of lostafin mutants44. The ability of dorso-
A cardiac rhythm disorder morphin to phenocopy lostafin revealed that it inhibits
characterized by the failure Glucocorticoids in long QT syndrome. A zebrafish model ALK2, the protein disrupted by the mutation.
ofevery second electrical
of genetic long QT syndrome (LQT syndrome), generated In a more recent example, a zebrafish embryogenesis
impulse to propagate from the
atrium to the ventricle. This can by a mutation in the kcnh2 gene, exhibits prolonged screen identified kalihinol F as causing several distinctive
result from extreme forms of action potential duration and 2:1 atrioventricular heart developmental defects, including undulation of the noto-
the long QT syndrome in block55. A zebrafish screen of existing drugs identified chord and defects in pigmentation, haematopoiesis and
whichthe delay in restoration the glucocorticoid flurandrenolide as a potent suppres- neural development 58. This collection of phenotypes was
of the membrane potential in
the ventricle is such that it is
sor of the LQT-like phenotype and implicated gluco- identified as highly similar to those of calamity, a mutant
refractory to the next electrical corticoid signalling in LQT syndrome. High-resolution with disrupted function of the copper-transporting
impulse from the atrium. physiology, comparable to that in mammalian models, ATPase atp7a. The kalihinolF/calamity phenocopy

726 | O CTOBER 2015 | VOLUME 14 www.nature.com/reviews/drugdisc

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

N
N N

Gene X
N

Figure 3 | Zebrafish phenotypes reveal drug targets. Drug-induced phenotypes in embryonic or larval zebrafish can
be compared with databases containing thousands of mutation-associated zebrafish phenotypes. On several occasions,
the discovery of phenotypic similarity between drug- and mutation-induced phenotypesNature
has revealed the| mechanism
Reviews of
Drug Discovery
action of a poorly understood compound.

led to the hypothesis that kalihinol F chelates copper, toxicology may be the most prevalent use of zebrafish
which was confirmed biochemically. In fact, the addi- in industry, with the majority of large pharmaceutical
tion of exogenous copper was able to rescue the kali- companies reporting some use of zebrafish for toxicology.
hinolFinduced developmental defects in zebrafish, It should be noted that the role of zebrafish in toxicol-
and treatment with kalihinol F was able to rescue ogy is often different from that of mammalian species.
copper overload toxicity in zebrafish embryos and Owing to the cost and effort associated with mamma-
mammaliancells. lian toxicology, it is often performed relatively late in
It should be noted that this type of multidimensional preclinical development, with the goals of determining
phenotype matching, or phenoclustering, in zebrafish the safety of advanced preclinical lead compounds and
can also be used to determine MOAs for compounds obtaining regulatory approval for a clinical trial. By con-
that were not originally discovered using zebrafish trast, zebrafish toxicology is much less expensive and can
screens. Fumagillin, an anti-angiogenic natural product, be performed rapidly on large numbers of compounds
was discovered more than 60years ago, and its binding in parallel. As such, zebrafish can be deployed much
target, methionine aminopeptidase 2 (MetAP2), was earlier in preclinical development. For example, toxicol-
discovered more than 15years ago. However, pheno- ogy testing can be performed on hundreds or thousands
type matching in zebrafish revealed a key insight into of hits from a primary high-throughput screen to eliminate
its MOA. Zebrafish embryos treated with fumagillin toxic compounds at an early stage and to prioritize hits
exhibited a distinctive gastrulation phenotype previously for further development. Indeed, this implicit counter-
observed in zebrafish with mutations in the non-canonical screen for toxicity is an integral part of disease sup-
Wnt5, revealing that MetAP2 inhibition by fumagil- pressor screens in the organism. Therefore, zebrafish
lin ultimately disrupts non-canonical WNT signalling can perform an invaluable function in enabling toxic
downstream of the Frizzled receptor 59. compounds to fail fast, before substantial resources
Because the number and variety of phenotypes have been wasted on their preclinical advancement.
that can be distinguished in a whole organism greatly As suggested earlier, zebrafish screens may generate
exceed the number distinguishable in cultured cells, lead compounds from primary screens that are already
phenotype matching is a powerful way of determin- pre-evaluated for several major forms of toxicity.
ing small-molecule MOAs. As the number of described Zebrafish have been shown to be a good model for
genephenotype pairs continues to rapidly increase, the predicting several types of drug toxicities (FIG.4). Cardio
utility of this approach should continue to grow. We toxicity, for example, appears to be highly conserved
anticipate that phenoclustering will be an important between zebrafish and humans. Systematic studies of
tool for MOA determination, not only for compounds drugs that cause QT prolongation in humans show a
discovered in zebrafish screens but also for drugs discov- >95% conservation of effect in zebrafish19. In the identifi-
ered by any other means. These observations also sug- cation of clinically significant repolarization toxicity, this
gest the potential benefits of more systematic approaches single assay performed as well as the combination of an
to compound annotation in the zebrafish. assay for human ERG (hERG) inhibition and toxicological
experiments on rabbit whole heart and canine Purkinje
Other uses for zebrafish in drug development fibres. Similarly close correlations have been observed
Although much of the academic effort in zebrafish chem- for hepatotoxicity, nephrotoxicity and reproductive tox-
ical biology has been focused on small-molecule discov- icity, in which all of the known toxicants in preclinical
ery, it is by no means the only way in which zebrafish are mammalian models or humans have similar effects in
contributing to drug discovery and development. In fact, zebrafish60,61. Nevertheless, it will require considerable

NATURE REVIEWS | DRUG DISCOVERY VOLUME 14 | O CTOBER 2015 | 727

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

Idgeftrjbavceoykvmv

Idgeftrjbavceoykvmv
Idgeftrjbavceoykvmv

Idgeftrjbavceoykvmv
Idgeftrjbavceoykvmv

Idgeftrjbavceoykvmv

Idgeftrjbavceoykvmv

Idgeftrjbavceoykvmv
Dgeftrjbavceoykvmv
Sgeftrjbavceoykvmv

Pgeftrjbavceoykvmv
Idgefrjbavceoykvmv

Wgeftrjbavceoykvm
Idgeftrjbaceoykvmv
Idgeftrjbavceoykvm

Idgeftrjbavceoykvm

Idgeftrjbavceoykvm

Idgeftrjbavceoyvmv

Idgeftrjbavceoyvmv
Idgftrjbavceoykvmv

Idgeftrjbavcoykvmv
Ideftrjbavceoykvmv

Idgeftrjavceoykvmv

Idgeftrjavceoykvmv
Idgeftrjbaveoykvm
Idgeftrjbavcykvmv
Aeftrjbavcoykvmv

Idgeftrjbavceoykv

Idgeftrjbavcev
Jhdahfda
Jhdahfda
Ahdahfda
Pdahfda
Jhdahfda
Jhdahfda
Jhdahfda
Kdahfda
Jhdahfda
11 12 Thdxfda
10 Jhdahfda
9 Jhdahfda
7 8
Ddahfda
HG
m/z
Jhdahfda
F E
DC 5 6 Jhdahfda
Ohdahfda
B A 3 4 Whdahfda
1 2
Jhdahfda

Organ-specic toxicity Behavioural toxicity Systemic phenotyping Metabolomics Toxicity reporter lines

Figure 4 | Use of zebrafish in toxicology. Zebrafish have been validated for high-throughput screening focused on
specific organ toxicity (liver, kidney and heart) or behavioural toxicity. Alternatively, they have been used to profile
Nature Reviews
compounds systematically for toxicity across multiple systems simultaneously. New technologies, | Drug
including Discovery
metabolomic
profiling and toxicity reporter lines, promise to extend the utility of zebrafish further. m/z, mass to charge ratio.

additional effort, and much higher-resolution assessment Behavioural screens for novel neuroactive drugs. Diseases
of the mechanisms of toxicity in each system, to establish of the nervous system are arguably the area of greatest
the full extent of the homology between zebrafish and challenge for drug discovery. Despite the prevalence of
human toxicology. nervous system disorders, there are generally few effec-
tive pharmacological therapies. Developing new central
Future opportunities for zebrafish nervous system (CNS) drugs is slow, expensive, and
Technologies for working with zebrafish are evolving fraught with failures. Challenges include the fact that
rapidly and opening new opportunities to contribute there are few well-validated targets for most nervous sys-
to improved drug development. Zebrafish are poised to tem disorders. Furthermore, it is difficult to create and
make a difference in a number of areas in the near future, assess faithful models of many CNS disorders. Psychiatric
as outlinedbelow. end points such as depression and psychosis are difficult
to assess in animals, and neurodegenerative end points
Genome engineering to generate sensitized models for are often slow to develop and laborious to quantify.
screening. The advent of transcription activator-like Zebrafish behavioural screens offer an intriguing
effector nucleases (TALENs) and CRISPRCas (clustered alternative approach to CNS drug discovery. Zebrafish
regularly interspaced short palindromic repeatCRISPR- possess a rich repertoire of behaviours, including
associated) nucleases has dramatically changed the simple stimulusresponse behaviours and more com-
landscape for zebrafish-based research. Withinthe past plex behaviours such as learning and sleep. Many of
2years, CRISPRCas technology in particular has made these behaviours have been shown to be amenable to
it possible to rapidly generate targeted lossoffunction assessment in a 96well format, and high-throughput
mutants and knockin lines62,63. For the first time, it is now screens have discovered compounds that alter these
possible to recreate human disease alleles in zebrafish6467. behaviours in specific, reproducible ways7274.
The potential impact of this advance on drug discovery is Although zebrafish and human behaviours differ, the
immense. With thousands of human disease-associated underlying circuits, cells and receptors are frequently
mutations being discovered by human geneticists, the conserved. Therefore, it may be possible to identify
challenge now is to figure out the functions of these muta- compounds that modulate disease-relevant processes by
tions and determine how to translate that knowledge into their ability to modulate zebrafish behaviours controlled
new therapies. Because targeted mutants can be generated by homologous pathways. For example, it may be diffi-
and phenotyped so much more efficiently in zebrafish cult to model schizophrenia in a zebrafish or to assess
than in rodents, zebrafish are likely to be a popular choice schizophrenia-like symptoms directly. However, most of
for testing the relevance of candidate diseasegenes. the proteins and cell types associated with human schizo
As genome editing technologies continue to improve, phrenia are conserved in zebrafish75,76. Antipsychotic
it will be possible to reconstruct tens or hundreds of drugs would be expected to modulate such proteins and
human mutations in zebrafish, followed by rapid assess- cell types, causing specific and distinctive behavioural
ment of their phenotypic effects. Perhaps more impor- changes in zebrafish. With a sufficiently granular under-
tantly, lines that are found to be disease-relevant can standing of the correlations between human disease and
quickly form the basis of large-scale chemical suppressor zebrafish behaviour, it would theoretically be possible to
screens. It has already been demonstrated that zebrafish identify neuroactive drugs simply by screening for com-
screens can identify compounds that reverse the effects pounds that produce the desired behavioural profile in
of genetic mutations30,52,55,6871. Thus, any line that pro- zebrafish.
duces a disease-relevant phenotype could be quickly For such an approach to work, the challenge will be
expanded and used to screen for chemical suppressors to determine what behavioural phenotypes are indica-
of the disease-associated phenotype. tive of a therapeutic candidate. Two approaches seem

728 | O CTOBER 2015 | VOLUME 14 www.nature.com/reviews/drugdisc

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

Cl

NH

Compound X

Figure 5 | Shared behavioural phenotypes reveal shared mechanisms and utilities. Neuroactive compounds often
produce distinctive behavioural changes in zebrafish that can be quantified in high-throughput behavioural assays.
As databases of compound-induced behavioural profiles grow, it is becoming possible Nature
to identify mechanistic
Reviews | Drug Discovery
relationships between compounds and search large chemical libraries for desirable profiles that indicate therapeutic
potential for nervous system disorders.

promising in this regard. The first is to use existing drugs false negatives are not problematic for drug discovery
to establish what the behavioural profile is for a particu- screens, in which the occasional false negative is well
lar therapeutic class. It has been shown that some classes tolerated. In toxicology screening, however, false nega-
of known neuroactive drugs produce distinctive and tives can be more troubling, as they could potentially
reproducible behavioural signatures in zebrafish behav- result in a failure to identify a compounds significant
ioural assays72,73. Establishment of a behavioural signa- toxicity. Improved methods for predicting or quantify-
ture for a therapeutic class would enable subsequent ing drug uptake could dramatically improve the utility
high-throughput screening for novel compounds that of zebrafish for drug discovery.
produce a similar behavioural effect in zebrafish (FIG.5). Of course, several other challenges remain. Natural
A second potential approach is to use genome engineer- variation exists in many invivo phenotypes, especially
ing to recreate human disease-associated mutations in complex phenotypes such as animal behaviours. Creating
zebrafish, then phenotype these animals carefully for robust, high-throughput assays to measure such pheno-
any resulting behavioural deficits. Mutant lines could types is challenging, but has been overcome in a variety
then be subjected to behavioural screening to identify of creative ways by numerous different research groups
compounds that suppress the abnormal behaviour and (TABLE 1) . Another challenge is producing enough
return the behavioural profile tonormal. zebrafish to enable truly large-scale, high-throughput
screens. This can be particularly difficult when work-
Unresolved questions and remaining challenges ing with fragile or inbred transgenic lines. To date,
The zebrafish has filled the niche for a screenable verte- most screens have involved production and screen-
brate in biology and drug discovery, but the emergence ing of a few hundred to tens of thousands of zebrafish
of the model and its rational appraisal as a tool for drug per day 39. Improved methods for mass production and
discovery has also served to highlight several challenges handling of zebrafish embryos would be highly beneficial
that remain. One of the most substantial issues is the to thefield.
challenge of controlling and quantifying drug exposures. Moving forward, substantial innovation and invest-
Although it is easy to control the concentration of a drug ment will be required to generate phenotypes that are
in the water bathing the zebrafish, it is more difficult to both relevant to human disease traits and scalable
predict how much drug will be absorbed. Some drugs for high throughput. Even as genome editing enables
appear to be absorbed poorly, whereas others are taken truly mechanistic modelling in different organisms
up readily. Drug levels are often measured directly in (overcoming some of the limitations inherent to target
serum or tissues of mammalian models, but such meas- choice), the development of adequately characterized
urements are more challenging to make in microscopic phenotypes will be a rate-limiting step in zebrafish
zebrafish, especially for high-throughput applications. chemical screens. It will also be important to ensure
Although some physicochemical attributes of com- that the precise details of each screen are well docu-
pounds, especially logP, have been shown to be predic- mented, as numerous variables can influence screen
tive of drug uptake in zebrafish, predictions are still output. These include the well volume, timing and
imperfect. As a consequence, false negatives are possible mode of exposure, temperature, lighting, pH, assay sen-
in zebrafish screens, whereby a compound that could sitivity and specificity, and linearity of assay response
be a good drug candidate fails to produce a biological characteristics. As experience accumulates, assays
effect owing to poor uptake. Generally speaking, such will hopefully become standardized, and more formal

NATURE REVIEWS | DRUG DISCOVERY VOLUME 14 | O CTOBER 2015 | 729

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

relationships with other preclinical animal models and integration of the zebrafish with traditional approaches
with quantitative metrics of human efficacy or toxicity to drug development will help to realize this vision of
will be established. medicines that empirically balance efficacy and toxicity,
A new scale and efficiency of drug discovery is nec- facilitating the discovery of new classes of therapeutics.
essary to move towards the individualized preventive
therapeutics that medicine will demand in the future. Note added in proof
To achieve these ends, diseases must be stratified and While this review was going to press, several important
specific drugs must be tailored for each new disease zebrafish chemical screens were published, including
entity. As we intervene earlier in disease, we will also screens for modifiers of hedgehog signalling102, cell migra-
require nuanced drug discovery with the goal of pathway tion103, liver tumorigenesis104, marrow engraftment105, and
normalization rather than target inhibition. Successful pancreatic -cell mass106.

1. Swinney,D.C. & Anthony,J. How were new medicines 21. Chi,N.C. etal. Genetic and physiologic dissection of 36. Chng,H.T., Ho,H.K., Yap,C.W., Lam,S.H. &
discovered? Nat. Rev. Drug Discov. 10, 507519 the vertebrate cardiac conduction system. PLoS Biol. Chan,E.C. An investigation of the bioactivation
(2011). 6, e109 (2008). potential and metabolism profile of zebrafish versus
2. Clader,J. W. The discovery of ezetimibe: a view from 22. Milan,D.J., Giokas,A.C., Serluca,F.C., Peterson,R.T. human. J.Biomol. Screen. 17, 974986 (2012).
outside the receptor. J.Med. Chem. 47, 19 & MacRae,C.A. Notch1b and neuregulin are required 37. Reimers,M.J., Flockton,A.R. & Tanguay,R.L.
(2004). for specification of central cardiac conduction tissue. Ethanol- and acetaldehyde-mediated developmental
3. Stitziel,N.O. etal. Inactivating mutations in NPC1L1 Development 133, 11251132 (2006). toxicity in zebrafish. Neurotoxicol. Teratol. 26,
and protection from coronary heart disease. N.Engl. 23. Milan,D.J., Jones,I.L., Ellinor,P.T. & MacRae,C.A. 769781 (2004).
J.Med. 371, 20722082 (2014). Invivo recording of adult zebrafish electrocardiogram 38. Kluver,N. etal. Transient overexpression of adh8a
4. Kodama,I., Kamiya,K. & Toyama,J. Cellular electro and assessment of drug-induced QT prolongation. increases allyl alcohol toxicity in zebrafish embryos.
pharmacology of amiodarone. Cardiovasc. Res. 35, Am. J.Physiol. Heart Circ. Physiol. 291, H269H273 PLoS ONE 9, e90619 (2014).
1329 (1997). (2006). 39. Rennekamp,A.J. & Peterson,R.T. 15years of
5. Path,G.J., Dai,X.Z., Schwartz,J.S., Benditt,D.G. 24. Schwerte,T. & Pelster,B. Digital motion analysis as a zebrafish chemical screening. Curr. Opin. Chem. Biol.
& Bache,R.J. Effects of amiodarone with and without tool for analysing the shape and performance of the 24, 5870 (2014).
polysorbate 80 on myocardial oxygen consumption circulatory system in transparent animals. J.Exp. Biol. 40. North,T.E. etal. Prostaglandin E2 regulates
and coronary blood flow during treadmill exercise 203, 16591669 (2000). vertebrate haematopoietic stem cell homeostasis.
in the dog. J.Cardiovasc. Pharmacol. 18, 1116 25. Yu,P.B. etal. BMP typeI receptor inhibition reduces Nature 447, 10071011 (2007).
(1991). heterotopic [corrected] ossification. Nat. Med. 14, This paper illustrated the power of insitu
6. Li,Z.H. etal. Combined invivo imaging and omics 13631369 (2008). expression screening in zebrafish by discovering
approaches reveal metabolism of icaritin and its 26. Ren,B. etal. ERK1/2Akt1 crosstalk regulates PGE2 as a modulator of hematopoietic stem cell
glycosides in zebrafish larvae. Mol. Biosyst. 7, arteriogenesis in mice and zebrafish. J.Clin. Invest. numbers. Discoveries described here resulted in
21282138 (2011). 120, 12171228 (2010). clinical trials of a PGE2 derivative for improving
7. Jeong,J.Y. etal. Functional and developmental 27. Zhang,Y. etal. AML1ETO mediates hematopoietic HSC transplantation.
analysis of the bloodbrain barrier in zebrafish. self-renewal and leukemogenesis through a COX/- 41. Goessling,W. etal. Genetic interaction of PGE2 and
Brain Res. Bull. 75, 619628 (2008). catenin signaling pathway. Blood 121, 49064916 Wnt signaling regulates developmental specification of
8. Goldstone,J.V. etal. Identification and developmental (2013). stem cells and regeneration. Cell 136, 11361147
expression of the full complement ofcytochrome 28. Kokel,D. etal. Photochemical activation of TRPA1 (2009).
P450 genes in zebrafish. BMC Genomics 11, 643 channels in neurons and animals. Nat. Chem. Biol. 42. Cutler,C. etal. Prostaglandin-modulated umbilical
(2010). 9, 257263 (2013). cord blood hematopoietic stem cell transplantation.
9. Seok,J. etal. Genomic responses in mouse models 29. Liu,Y. etal. Visnagin protects against doxorubicin- Blood 122, 30743081 (2013).
poorly mimic human inflammatory diseases. induced cardiomyopathy through modulation of 43. Hagedorn,E.J., Durand,E.M., Fast,E.M. & Zon,L.I.
Proc. Natl Acad. Sci. USA 110, 35073512 (2013). mitochondrial malate dehydrogenase. Sci. Transl. Med. Getting more for your marrow: boosting hematopoietic
10. Howe,K. etal. The zebrafish reference genome 6, 266ra170 (2014). stem cell numbers with PGE. Exp. Cell Res. 329,
sequence and its relationship to the human genome. This manuscript describes the discovery of 220226 (2014).
Nature 496, 498503 (2013). visnagin, a small molecule that protects the heart 44. Yu,P.B. etal. Dorsomorphin inhibits BMP signals
Completion of the zebrafish reference genome from chemotherapy-induced damage. The effects required for embryogenesis and iron metabolism.
revealed that 82% of disease-associated human of visnagin were conserved in rodent heart failure Nat. Chem. Biol. 4, 3341 (2008).
genes have a zebrafish orthologue. models and were shown to be mediated through The first small-molecule antagonists of the
11. Tiso,N., Moro,E. & Argenton,F. Zebrafish pancreas a novel target: MDH2. BMP pathway were discovered in a screen for
development. Mol. Cell. Endocrinol. 312, 2430 30. Asimaki,A. etal. Identification of a new modulator compounds that perturb zebrafish embryogenesis.
(2009). of the intercalated disc in a zebrafish model of Dorsomorphin derivatives are being developed
12. Gut,P. etal. Whole-organism screening for arrhythmogenic cardiomyopathy. Sci. Transl. Med. as therapeutics for a variety of indications
gluconeogenesis identifies activators of fasting 6, 240ra274 (2014). associated with excessive BMP signalling.
metabolism. Nat. Chem. Biol. 9, 97104 (2013). Screening in a zebrafish model of arrhythmogenic 45. Cuny,G.D. etal. Structureactivity relationship study of
13. Jagannathan-Bogdan,M. & Zon,L.I. Hematopoiesis. cardiomyopathy identified SB216763, a compound bone morphogenetic protein (BMP) signaling inhibitors.
Development 140, 24632467 (2013). capable of reversing the disease phenotype in Bioorg. Med. Chem. Lett. 18, 43884392 (2008).
14. Ganis,J.J. etal. Zebrafish globin switching occurs in zebrafish, rodent cells and cardiac myocytes from 46. Theurl,I. etal. Pharmacologic inhibition of hepcidin
two developmental stages and is controlled by the patient-derived stem cells. The manuscript highlights expression reverses anemia of chronic inflammation
LCR. Dev. Biol. 366, 185194 (2012). the potential of zebrafish screens for repurposing in rats. Blood 118, 49774984 (2011).
15. Steinbicker,A.U. etal. Inhibition of bone existing drugs. 47. Derwall,M. etal. Inhibition of bone morphogenetic
morphogenetic protein signaling attenuates anemia 31. Shin,J.T., Pomerantsev,E.V., Mably,J.D. & protein signaling reduces vascular calcification and
associated with inflammation. Blood 117, MacRae,C.A. High-resolution cardiovascular atherosclerosis. Arterioscler. Thromb. Vasc. Biol.
49154923 (2011). function confirms functional orthology of myocardial 32, 613622 (2012).
16. Donovan,A. etal. Positional cloning of zebrafish contractility pathways in zebrafish. Physiol. Genom. 48. Saeed,O. etal. Pharmacological suppression of
ferroportin1 identifies a conserved vertebrate iron 42, 300309 (2010). hepcidin increases macrophage cholesterol efflux and
exporter. Nature 403, 776781 (2000). 32. Eliceiri,B.P., Gonzalez,A.M. & Baird,A. reduces foam cell formation and atherosclerosis.
17. Paffett-Lugassy,N. etal. Functional conservation of Zebrafish model of the bloodbrain barrier: Arterioscler. Thromb. Vasc. Biol. 32, 299307 (2012).
erythropoietin signaling in zebrafish. Blood 110, morphological and permeability studies. 49. Wang,L. etal. The bone morphogenetic proteinhepcidin
27182726 (2007). Methods Mol. Biol. 686, 371378 (2011). axis as a therapeutic target in inflammatory bowel
18. Asnani,A. & Peterson,R.T. The zebrafish as a tool 33. Fleming,A., Diekmann,H. & Goldsmith,P. Functional disease. Inflamm. Bowel Dis. 18, 112119 (2012).
to identify novel therapies for human cardiovascular characterisation of the maturation of the bloodbrain 50. Owens,K.N. etal. Identification of genetic and
disease. Dis. Model. Mech. 7, 763767 (2014). barrier in larval zebrafish. PLoS ONE 8, e77548 (2013). chemical modulators of zebrafish mechanosensory
19. Milan,D.J., Peterson,T.A., Ruskin,J.N., 34. Farber,S.A. etal. Genetic analysis of digestive hair cell death. PLoS Genet. 4, e1000020 (2008).
Peterson,R.T. & MacRae,C.A. Drugs that induce physiology using fluorescent phospholipid reporters. This paper describes the discovery of compounds
repolarization abnormalities cause bradycardia in Science 292, 13851388 (2001). that protect hair cells from the toxic effects of
zebrafish. Circulation 107, 13551358 (2003). 35. Popovic,M., Zaja,R., Fent,K. & Smital,T. Interaction aminoglycoside antibiotics. The protective effects
20. Burns,C.G. etal. High-throughput assay for small of environmental contaminants with zebrafish organic of these compounds are conserved in mammals,
molecules that modulate zebrafish embryonic heart anion transporting polypeptide, Oatp1d1 (Slco1d1). suggesting their therapeutic potential to mitigate
rate. Nat. Chem. Biol. 1, 263264 (2005). Toxicol. Appl. Pharmacol. 280, 149158 (2014). hearing loss caused by antibiotics and other drugs.

730 | O CTOBER 2015 | VOLUME 14 www.nature.com/reviews/drugdisc

2015 Macmillan Publishers Limited. All rights reserved


REVIEWS

51. Yeh,J.R. etal. AML1ETO reprograms hematopoietic In this paper, a zebrafish model of Dravet 90. Ridges,S. etal. Zebrafish screen identifies novel
cell fate by downregulating scl expression. syndrome is characterized and used to screen compound with selective toxicity against leukemia.
Development 135, 401410 (2008). approved drugs for the ability to attenuate seizure Blood 119, 56215631 (2012).
52. Yeh,J.R. etal. Discovering chemical modifiers of activity. The paper is a significant example of the 91. Weger,B.D., Weger,M., Nusser,M., Brenner-Weiss,G.
oncogene-regulated hematopoietic differentiation. ability to model genetic diseases in zebrafish, & Dickmeis,T. A chemical screening system for
Nat. Chem. Biol. 5, 236243 (2009). and it also highlights the models potential for glucocorticoid stress hormone signaling in an
53. Wang,Y. etal. The Wnt/-catenin pathway is required drug repurposing screens. intact vertebrate. ACS Chem. Biol. 7, 11781183
for the development of leukemia stem cells in AML. 72. Rihel,J. etal. Zebrafish behavioral profiling links (2012).
Science 327, 16501653 (2010). drugs to biological targets and rest/wake regulation. 92. Jin,S. etal. An invivo zebrafish screen identifies
54. Klimek,V.M., Dolezal,E.K., Smith,L., Soff,G. & Science 327, 348351 (2010). organophosphate antidotes with diverse
Nimer,S.D. Phase I trial of sodium salicylate in patients This paper was one of the first to describe mechanisms of action. J.Biomol. Screen 18,
with myelodysplastic syndromes and acute myelogenous high-throughput screening for behaviour-modifying 108115 (2013).
leukemia. Leuk. Res. 36, 570574 (2012). compounds in this case, modifiers of sleep and 93. Nath,A.K. etal. Chemical and metabolomic screens
55. Peal,D.S. etal. Novel chemical suppressors of long wakefulness. The ability to use behaviours as identify novel biomarkers and antidotes for cyanide
QT syndrome identified by an invivo functional readouts for high-throughput screening opens exposure. FASEB J. 27, 19281938 (2013).
screen. Circulation 123, 2330 (2011). new avenues for CNS drug discovery. 94. Liu,Y.J. etal. Cannabinoid receptor 2 suppresses
56. Ziegler,S., Pries,V., Hedberg,C. & Waldmann,H. 73. Kokel,D. etal. Rapid behavior-based identification leukocyte inflammatory migration by modulating
Target identification for small bioactive molecules: of neuroactive small molecules in the zebrafish. the JNK/cJun/Alox5 pathway. J.Biol. Chem. 288,
finding the needle in the haystack. Angew. Chem. Nat. Chem. Biol. 6, 231237 (2010). 1355113562 (2013).
Int. Ed Engl. 52, 27442792 (2013). 74. Wolman,M.A., Jain,R.A., Liss,L. & Granato,M. 95. Le,X. etal. A novel chemical screening strategy
57. Gutierrez,A. etal. Phenothiazines induce Chemical modulation of memory formation in in zebrafish identifies common pathways in
PP2Amediated apoptosis in Tcell acute lymphoblastic larval zebrafish. Proc. Natl Acad. Sci. USA 108, embryogenesis and rhabdomyosarcoma development.
leukemia. J.Clin. Invest. 124, 644655 (2014). 1546815473 (2011). Development 140, 23542364 (2013).
58. Sandoval,I.T. etal. Juxtaposition of chemical and 75. Morris,J. A. Zebrafish: a model system to examine 96. Hao,J. etal. Selective small molecule targeting
mutation-induced developmental defects in zebrafish the neurodevelopmental basis of schizophrenia. -catenin function discovered by invivo chemical
reveal a copper-chelating activity for kalihinol F. Prog. Brain Res. 179, 97106 (2009). genetic screen. Cell Rep. 4, 898904 (2013).
Chem. Biol. 20, 753763 (2013). 76. Singh,K.K. etal. Common DISC1 polymorphisms 97. Gebruers,E. etal. A phenotypic screen in zebrafish
59. Zhang,Y. etal. A chemical and genetic approach to disrupt Wnt/GSK3 signaling and brain development. identifies a novel small-molecule inducer of ectopic
the mode of action of fumagillin. Chem. Biol. 13, Neuron 72, 545558 (2011). tail formation suggestive of alterations in non-
10011009 (2006). 77. Mathew,L.K. etal. Unraveling tissue regeneration canonical Wnt/PCP signaling. PLoS ONE 8, e83293
60. Driessen,M. etal. A transcriptomics-based pathways using chemical genetics. J.Biol. Chem. (2013).
hepatotoxicity comparison between the zebrafish 282, 3520235210 (2007). 98. Kong,Y. etal. Neural crest development and
embryo and established human and rodent invitro 78. Tran,T.C. etal. Automated, quantitative screening craniofacial morphogenesis is coordinated by nitric
and invivo models using cyclosporine A, amiodarone assay for antiangiogenic compounds using transgenic oxide and histone acetylation. Chem. Biol. 21,
and acetaminophen. Toxicol. Lett. 232, 403412 zebrafish. Cancer Res. 67, 1138611392 (2007). 488501 (2014).
(2014). 79. Molina,G. etal. Zebrafish chemical screening reveals 99. Nishiya,N. etal. A zebrafish chemical suppressor
61. Ducharme,N.A., Reif,D.M., Gustafsson,J.A. & an inhibitor of Dusp6 that expands cardiac cell screening identifies small molecule inhibitors of the
Bondesson,M. Comparison of toxicity values across lineages. Nat. Chem. Biol. 5, 680687 (2009). Wnt/-catenin pathway. Chem. Biol. 21, 530540
zebrafish early life stages and mammalian studies: 80. Clifton,J.D. etal. Identification of novel inhibitors of (2014).
implications for chemical testing. Reprod. Toxicol. dietary lipid absorption using zebrafish. PLoS ONE 5, 100. Tsuji,N. etal. Whole organism high content screening
55, 310 (2014). e12386 (2010). identifies stimulators of pancreatic -cell proliferation.
62. Hwang,W.Y. etal. Efficient genome editing in 81. Ishizaki,H. etal. Combined zebrafish-yeast PLoS ONE 9, e104112 (2014).
zebrafish using a CRISPRCas system. Nat. Biotech. chemical-genetic screens reveal gene-copper-nutrition 101. Nath,A.K. etal. PTPMT1 inhibition lowers glucose
31, 227229 (2013). interactions that modulate melanocyte pigmentation. through succinate dehydrogenase phosphorylation.
63. Gonzales,A.P. & Yeh,J.R. Cas9based genome editing Dis. Model. Mech. 3, 639651 (2010). Cell Rep. 10, 694701 (2015).
in zebrafish. Methods Enzymol. 546, 377413 82. Wang,C. etal. Rosuvastatin, identified from a 102. Williams,C.H. etal. An in vivo chemical genetic
(2014). zebrafish chemical genetic screen for antiangiogenic screen identifies phosphodiesterase 4 as a
64. Irion,U., Krauss,J. & Nusslein-Volhard,C. compounds, suppresses the growth of prostate cancer. pharmacological target for Hedgehog signaling
Precise and efficient genome editing in zebrafish Eur. Urol. 58, 418426 (2010). inhibition. Cell Rep. 11, 4350 (2015).
using the CRISPR/Cas9 system. Development 141, 83. White,R.M. etal. DHODH modulates transcriptional 103. Gallardo,V.E. etal. Phenotype-driven chemical
48274830 (2014). elongation in the neural crest and melanoma. Nature screening in zebrafish for compounds that inhibit
65. Gagnon,J.A. etal. Efficient mutagenesis by Cas9 471, 518522 (2011). collective cell migration identifies multiple pathways
protein-mediated oligonucleotide insertion and large- 84. Saydmohammed,M., Vollmer,L.L., Onuoha,E.O., potentially involved in metastatic invasion.
scale assessment of single-guide RNAs. PLoS ONE 9, Vogt,A. & Tsang,M. A high-content screening assay in Dis. Model. Mech. 8, 565576 (2015).
e98186 (2014). transgenic zebrafish identifies two novel activators of 104. Evanson,K.J. etal. Identification of chemical
66. Auer,T.O., Duroure,K., De Cian,A., Concordet,J.P. fgf signaling. Birth Defects Res. C Embryo Today 93, inhibitors of -catenin-driven liver tumorigenesis
& Del Bene,F. Highly efficient CRISPR/Cas9mediated 281287 (2011). in zebrafish. PLoS Genet.11, e1005305 (2015).
knockin in zebrafish by homology-independent DNA 85. Rovira,M. etal. Chemical screen identifies FDA- 105. Li,P. etal. Epoxyeicosatrienoic acids enhance
repair. Genome Res. 24, 142153 (2014). approved drugs and target pathways that induce embryonic haematopoiesis and adult marrow
67. Xiao,A. etal. Chromosomal deletions and inversions precocious pancreatic endocrine differentiation. engraftment. Nature 523, 468471 (2015).
mediated by TALENs and CRISPR/Cas in zebrafish. Proc. Natl Acad. Sci. USA 108, 1926419269 (2011). 106. Wang,G. etal. First quantitative high-throughput
Nucleic Acids Res. 41, e141 (2013). 86. Colanesi,S. etal. Small molecule screening identifies screen in zebrafish identifies novel pathways for
68. Peterson,R.T. etal. Chemical suppression of a genetic targetable zebrafish pigmentation pathways. increasing pancreatic -cell mass. eLife. 4, e08261
mutation in a zebrafish model of aortic coarctation. Pigment Cell. Melanoma Res. 25, 131143 (2012). (2015).
Nat. Biotech. 22, 595599 (2004). 87. Namdaran,P., Reinhart,K.E., Owens,K.N.,
69. Stern,H.M. etal. Small molecules that delay S phase Raible,D.W. & Rubel,E.W. Identification of Acknowledgements
suppress a zebrafish bmyb mutant. Nat. Chem. Biol. modulators of hair cell regeneration in the zebrafish The authors thank A. Rennekamp for his help compiling
1, 366370 (2005). lateral line. J.Neurosci. 32, 35163528 (2012). Table1. R.T.P acknowledges support from the Charles Addison
70. Cao,Y. etal. Chemical modifier screen identifies HDAC 88. Becker,J.R. etal. Invivo natriuretic peptide reporter and Elizabeth Ann Sanders Professorship. The toxicology work
inhibitors as suppressors of PKD models. Proc. Natl assay identifies chemical modifiers of hypertrophic of C.A.M. has been funded by an Innovation in Regulatory
Acad. Sci. USA 106, 2181921824 (2009). cardiomyopathy signalling. Cardiovasc. Res. 93, Science Award from the Burroughs Wellcome Fund.
71. Baraban,S.C., Dinday,M.T. & Hortopan,G.A. 463470 (2012).
Drug screening in Scn1a zebrafish mutant identifies 89. Padilla,S. etal. Zebrafish developmental screening of Competing interests statement
clemizole as a potential Dravet syndrome treatment. the ToxCast Phase I chemical library. Reprod. Toxicol. The authors declare competing interests: see Web version
Nat. Commun. 4, 2410 (2013). 33, 174187 (2012). for details.

NATURE REVIEWS | DRUG DISCOVERY VOLUME 14 | O CTOBER 2015 | 731

2015 Macmillan Publishers Limited. All rights reserved