ISSN 1294-8322

Dialogues
in
Neuroplasticity

2004

Vo l u m e 6 . N o . 2

clinical neuroscience

Dialogues
Editor-in-chief Jean-Paul MACHER, MD, Rouffach, France Editorial Board Manfred ACKENHEIL, MD, Mnchen, Germany Csar CARVAJAL, MD, Santiago de Chile, Chile Marc-Antoine CROCQ, MD, Rouffach, France Michael DAVIDSON, MD, Tel Hashomer, Israel Margret R. HOEHE, MD, Berlin, Germany Barry D. LEBOWITZ, PhD, Rockville, Md, USA Deborah J. MORRIS-ROSENDAHL, PhD, Johannesburg, South Africa Rajesh M. PARIKH, MD, Bombay, India David RUBINOW, MD, Bethesda, Md, USA Pierre SCHULZ, MD, Chne-Bourg, Switzerland Carol A. TAMMINGA, MD, Baltimore, Md, USA International Consultant Jorge-Alberto COSTA E SILVA, MD, Rio de Janeiro, Brazil Publication Director / Directeur de la Publication Jean-Philippe SETA, MD, Neuilly-sur-Seine, France

Editorial

ear Colleagues,

The known properties of the central nervous system are quite remarkable and we may confidently assume that many more fascinating aspects of the brain remain to be discovered. Until recently, the causation of mental disorders was always explained in terms of abnormalities involving familiar biological concepts, such as monoamine neurotransmission, receptor regulation, and molecular biology. The appearance of a novel explanatory model, accounting for some previously unexplained phenomena, is of tremendous interest. It has long been known that some disorders involve regional modifications that can be evidenced by studying brain structure. Neurotrophic factors preventing cell death have been shown to exist and, more recently, the process of hippocampal neurogenesis has been described. Neuroplasticity is the process that underlies neurogenesis: it leads to protein synthesis and constitutes a defense mechanism against the deleterious effects of stress. Plastic modifications of neurons and synapses have been observed thanks to the development of neuroimaging techniques, which can reach as far as the cellular level. The observations relating to neuroplasticity have led to: New diagnostic markers. A better understanding of certain pathogenetic mechanisms. The proof of activity of certain compounds. We believe that it is important to give a progress report on the concept of neuroplasticity and its influence on the understanding of the mechanisms of depression. We are grateful to Dr David R. Rubinow from the National Institute of Mental Health in Bethesda, Md, for bringing together the most qualified authors in the field to discuss this topic in this issue of Dialogues in Clinical Neuroscience.

Yours sincerely,

Jean-Paul Macher, MD

Marc-Antoine Crocq, MD

113

Dialogues in Clinical Neuroscience is a quarterly publication that aims to serve as an interface between clinical neuropsychiatry and the neurosciences by providing state-of-the-art information and original insights into relevant clinical, biological, and therapeutic aspects. Each issue addresses a specific topic, and also publishes free contributions in the field of neuroscience as well as other nontopic-related material. All contributions are reviewed by members of the Editorial Board and submitted to expert consultants for peer review. Indexed in EMBASE and Elsevier BIOBASE. EDITORIAL OFFICES Editor in Chief Jean-Paul MACHER, MD FORENAP - Institute for Research in Neuroscience and Neuropsychiatry BP29 - 68250 Rouffach - France Tel: + 33 3 89 78 70 18 / Fax: +33 3 89 78 51 24 Secretariat, subscriptions, and submission of manuscripts Marc-Antoine CROCQ, MD FORENAP - Institute for Research in Neuroscience and Neuropsychiatry BP29 - 68250 Rouffach - France Tel: +33 3 89 78 71 20 (direct) or +33 3 89 78 70 18 (secretariat) Fax: +33 3 89 78 51 24 / E-mail: macrocq@forenap.asso.fr Annual subscription rates: Europe 150; Rest of World 170. Production Editor Sarah A. NOVACK, PhD Servier International - Medical Publishing Division 192 avenue Charles-de-Gaulle 92578 Neuilly-sur-Seine Cedex - France Tel: +33 1 55 72 33 10 / Fax: +33 1 55 72 68 88 E-mail: sarah.novack@fr.netgrs.com

PUBLISHER Les Laboratoires Servier 22 rue Garnier - 92578 Neuilly-sur-Seine Cedex - France E-mail: mail.dialneuro@fr.netgrs.com Copyright 2004 by Les Laboratoires Servier All rights reserved throughout the world and in all languages. No part of this publication may be reproduced, transmitted, or stored in any form or by any means either mechanical or electronic, including photocopying, recording, or through an information storage and retrieval system, without the written permission of the copyright holder. Opinions expressed do not necessarily reflect the views of the publisher, editors, or editorial board. The authors, editors, and publisher cannot be held responsible for errors or for any consequences arising from the use of information contained in this journal. ISSN 1294-8322
Design: Christophe Caretti / Layout: Graphie 66 Imprim en France par SIP 1, rue Saint Simon - 95310 Saint-Ouen-lAumne

114

Contents
Page

113 117 119 135 143 157 171 185 199 217

Editorial
Jean-Paul Macher, Marc-Antoine Crocq

In this issue
David R. Rubinow

State of the art

Structural plasticity of the adult brain: how animal models help us understand brain changes in depression and systemic disorders related to depression Bruce S. McEwen

Basic research
Structural plasticity of the adult brain Fred H. Gage Regulation of cellular plasticity and resilience by mood stabilizers: the role of AMPA receptor trafficking Jing Du, Jorge A. Quiroz, Neil A. Gray, Steve T. Szabo, Carlos A. Zarate Jr, Husseini K. Manji

Pharmacological aspects
Neural plasticity: consequences of stress and actions of antidepressant treatment Ronald S. Duman Cellular consequences of stress and depression Eberhard Fuchs, Gabriele Flgge

Clinical research
Cellular abnormalities in depression: evidence from postmortem brain tissue Craig A. Stockmeier, Grazyna Rajkowska Neuroplasticity in mood disorders Wayne C. Drevets Cellular plasticity and resilience and the pathophysiology of severe mood disorders Dennis S. Charney, George DeJesus, Husseini K. Manji

ISSUE COORDINATED BY: David R. RUBINOW

115

Contents
Page

227 235 243

Free papers
Texture analysis of the brain: from animal models to human applications Jean-Franois J. Nedelec, Olivier Yu, Jacques Chambron, Jean-Paul Macher Problems in texture analysis with magnetic resonance imaging Lothar R. Schad Texture analysis methodologies for magnetic resonance imaging Andrzej Materka

ISSUE COORDINATED BY: David R. RUBINOW

116

In this issue...
For many of us, a central tenet of our neurobiological training was that the structure of the brain was fixed, as was the number of neurons in the adult brain. This belief was not only incorrect, but also restricted our understanding of a variety of fundamental processes including learning, adaptation and maladaptation to stress, development of susceptibility to disease, and resilience. This issue of Dialogues in Clinical Neuroscience describes the flexible, adaptive responses of the brainneuroplasticityand the relevance of neuroplastic changes to the pathophysiology of neuropsychiatric illness, the mechanism of action of psychotropic medications, and the transduction of environmental factors to changes in brain function. In the State of the art article, Bruce S. McEwen (page 119) provides an introduction to allostasisadaptation to stressand an overview of the structural and cellular consequences of stress, its molecular mediators in altering brain structure, and the kinetics of stress-induced structural remodeling. The compelling models described make apparent the complexity and dynamic nature of adaptation, as well as the ontogeny of susceptibility to psychiatric illness. In the first Basic research article, Fred H. Gage (page 135) reviews an element of neuroplasticity, neurogenesis, ie, the generation of new neurons, and describes the multiple steps involved in the process of neurogenesis: differentiation, commitment, survival, and functional integration. This capacity for self-repair represents one of the therapeutic frontiers in the treatment of neuropsychiatric illness. In the second Basic research article, Jing Du and colleagues (page 143) review the evidence suggesting the role of glutamatergically mediated synaptic plasticity in both the pathophysiology and treatment of affective illness. While focusing on AMPA (-amino-3-hydroxy-5methyl-4-isoxazolepropionate) and GluR1 (glutamate) receptor trafficking, these authors provide an excellent introduction to glutamate receptor pharmacology and intracellular signaling as a way of demonstrating both the mechanisms of action of mood stabilizers and targets for subsequent drug development. In one of the Pharmacological aspects articles, Ronald S. Duman (page 157) discusses the effects of stress and antidepressants on neuroplasticity, particularly as these effects relate to depression. He focuses on two mediating systems that appear to link, at a molecular level, neuroplasticity, stress, depression, and pharmacotherapy: the neurotrophin BDNF (brain-derived neurotrophic factor), and the cAMP-CREB (cyclic adenosine monophosphate [cAMP]cAMP-response element binding protein) cascade. He suggests that the cellular and molecular underpinnings of structural and functional plasticity offer promising clues to the pathophysiology of depression and targets for drug development. In the other Pharmacological aspects article, Eberhard Fuchs and Gabriele Flgge (page 171) describe the pharmacology of the stress response by focusing on changes in monoamines and monoamine receptors in several animal stress models. They provide a basis for understanding depression as an impairment of synaptic and structural plasticity, with consequent implications for its treatment. In the first Clinical research article, Craig A. Stockmeier and Grazyna Rajkowska (page 185) describe in detail the neural and glial abnormalities identified in several critical brain regions in affective illness. This comprehensive review of postmortem studies discusses the possible functional implications of abnormalities of cell morphology and distribution and introduces the circuitry that is described in more detail in the second article by Wayne C. Drevets (page 199). While focusing on neuroimaging studies, he provides a synthesis of identified neuropathological and imaging abnormalities in affective illness, highlighting those neural circuits strongly implicated as dysfunctional in affective disorder. Elucidation of this circuitry at functional and structural levels will also help illuminate substrates for component processes common to a variety of neuropsychiatric disorders. Indeed, in the last article, Dennis S. Charney and his colleagues (page 217) suggest that studies of neuroplasticity will result in a new psychiatric nosology, as well as new therapeutic targets. Thus, not only will the therapeutic armamentarium be expanded as we better understand the mechanics of neuroplasticity, but this better understanding will also lead to a reconceptualization of how psychiatric illness is acquired, how it is optimally treated (with attention to both structural and functional elements), and, perhaps most importantly, how resilience is expressed at cellular and organismic levels. David R. Rubinow, MD

117

Contributors
Bruce S. McEwen, PhD Wayne C. Drevets, MD Author affiliations: Harold and Margaret Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller University, New York, NY, USA Author affiliations: Mood and Anxiety Disorders Program, NIH NIMH/MIB, Bethesda, Md, USA

Fred H. Gage, PhD

Dennis S. Charney, MD

Author affiliations: Laboratory of Genetics, The Salk Institute, La Jolla, Calif, USA

Author affiliations: National Institute of Mental Health, Bethesda, Md, USA

Husseini K. Manji, MD, FRCPC

Jean-Franois J. Nedelec, PhD

Author affiliations: Laboratory of Molecular Pathophysiology, National Institute of Mental Health, Bethesda, Md, USA

Author affiliations: FORENAP, Rouffach, France

Ronald S. Duman, PhD

Lothar R. Schad, PhD

Author affiliations: Division of Molecular Psychiatry, Departments of Psychiatry and Pharmacology, Yale University School of Medicine, New Haven, CT, USA

Author affiliations: Department of Biophysics and Medical Radiation Physics, German Cancer Research Centre, Heidelberg, Germany

Eberhard Fuchs, PhD

Andrzej Materka, MSEE, PhD, DSc

Author affiliations: Clinical Neurobiology Laboratory, German Primate Center, Gttingen, Germany

Author affiliations: Institute of Electronics, Technical University of Lodz, Lodz, Poland

Craig A. Stockmeier, PhD

Author affiliations: The University of Mississippi Medical Center, Department of Psychiatry and Human Behavior, Jackson, Miss, USA

118

State of the art

Structural plasticity of the adult brain: how animal models help us understand brain changes in depression and systemic disorders related to depression
Bruce S. McEwen, PhD
hen we experience a stressful event, the initial response of the brain, body, and behavior is a protective one, and hormones, cytokines, and other mediators, such as the neurotransmitters, are used to survive and adapt to the challenge. However, repeated stressful experiences have deleterious effects, in part because the very same mechanisms that help protect in the short term are now either mismanaged and/or overused.1 And, over weeks, months, and years, the dysregulation and overactivity of these systems can promote changes that appear to be deleterious, and stressful experiences have been reported to be a major risk factor in the occurrence of depressive disorders. For example, in the brain, the overactivity of stress hormones in the blood and endogenous excitatory The brain interprets experiences and translates them into behavioral and physiological responses. Stressful events are those which are threatening or, at the very least, unexpected and surprising, and the physiological and behavioral responses are intended to promote adaptation via a process called allostasis. Chemical mediators of allostasis include cortisol and adrenalin from the adrenal glands, other hormones, and neurotransmitters, the parasympathetic and sympathetic nervous systems, and cytokines and chemokines from the immune system. Two brain structures, the amygdala and hippocampus, play key roles in interpreting what is stressful and determining appropriate responses. The hippocampus, a key structure for memories of events and contexts, expresses receptors that enable it to respond to glucocorticoid hormones in the blood. It undergoes atrophy in a number of psychiatric disorders; it also responds to stressors with changes in excitability, decreased dendritic branching, and reduction in number of neurons in the dentate gyrus. The amygdala, which is important for emotional memories, becomes hyperactive in posttraumatic stress disorder and depressive illness. In animal models of stress, there is evidence for growth and hypertrophy of nerve cells in the amygdala. Changes in the brain after acute and chronic stressors mirror the pattern seen in the metabolic, cardiovascular, and immune systems, that is, short-term adaptation (allostasis) followed by long-term damage (allostatic load), eg, atherosclerosis, fat deposition obesity, bone demineralization, and impaired immune function. Allostatic load of this kind is seen in major depressive illness and may also be expressed in other chronic anxiety and mood disorders.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:119-133.

Keywords: structural plasticity; brain; allostasis; allostatic load; stress; depression; anxiety Author affiliations: Harold and Margaret Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller University, New York, NY, USA

Address for correspondence: Bruce S. McEwen, PhD, The Rockefeller University, Box 165, 1230 York Avenue, New York, NY 10021, USA (e-mail: mcewen@rockefeller.edu)

Copyright 2004 LLS SAS. All rights reserved

119

www.dialogues-cns.org

State of the art

Selected abbreviations and acronyms
CGRP CRS DG GR IGF-1 MR NMDA PSA-NCAM tPA calcitonin generelated peptide chronic restraint stress dentate gyrus glucocorticoid receptor insulin-like growth factor1 mineralocorticoid receptor N-methyl-D-aspartate polysialated neural cell adhesion molecule tissue plasminogen activator Acute stress induces formation of spine synapses in CA1 region of hippocampus6 and chronic stress also increases spine synapse formation in hippocampus and amygdala.7 The contrasting changes of dendrites in amygdala and hippocampus after chronic restraint stress (CRS) offers an unprecedented opportunity for understanding underlying mechanisms, as will be discussed below. CRS for 21 days or longer impairs hippocampal-dependent cognitive function8,9 and enhances amygdala-dependent unlearned fear and fear conditioning,10 which are consistent with the opposite effects of stress on hippocampal and amygdala structure. CRS also increases aggression between animals living in the same cage (Table II).11 Psychosocial stress suppresses neurogenesis and causes dendritic shrinkage,12-15 and one of these stress models, the tree shrew, is considered to be a model of human depressive illness.16 Indeed, in major depression and a number of other mood and anxiety disorders, there are reports of hippocampal volume loss and enlargement of the amygdala.17,18 Studies in the tree shrew have shown that treatment with anti-

amino acid neurotransmitters in the brain suppress neurogenesis in dentate gyrus (DG) and causes debranching of dendrites in hippocampus and medial prefrontal cortex, whereas chronic stress causes neurons in amygdala to show dendritic growth.2-5 The hippocampus contains receptors for adrenal steroids, which regulate excitability and morphological changes (Figure 1). Along with many other brain regions, the amygdala also contains adrenal steroid receptors, which influence function in this structure as well (Table I).

CA1

Sch Perfusion pathway MF Dentate gyrus

CA3

Figure 1. The hippocampus is a target for adrenal steroids. GR, glucocorticoid receptor; MR, mineralocorticoid receptor; Sch, Schaffer colateral; MF, mossy fiber; CC, corpus callosum.

120

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

depressant, antiseizure, and mood-stabilizing drugs prevents stress-induced hippocampal structural changes.14,15,19 Besides reduced neurogenesis in DG, there is also evidence for reduced size of principal neuron cell bodies in hippocampus, which is consistent with reduced size of the dendritic tree.20 Synaptic reorganization is also a likely consequence of these rather drastic structural changes, and the animal models cited above provide evidence that synapses can be rapidly formed as a result of stress. Taken together, such structural changes seem likely to play a major role in the volume loss in the human hippocampus and the related effects on cognitive function and affect.18 This article will review underlying mechanisms and consider their applicability to furthering our understanding of the pathophysiology of mood and anxiety disorders.

Allostasis and mechanisms for behavioral adaptation

The amygdala and hippocampus are both involved in contextual fear conditioning and in passive avoidance learning. In fear conditioning, glucocorticoids enhance learned fear21 and they play an important role in forming the memory of context in contextual fear conditioning, but not of the actual effect of footshock in rats that are already familiar with the context where the shock is administered.22,23 This suggests that the hippocampal role in contextual fear conditioning is enhanced by moderate levels of glucocorticoids, but the fear conditioning is either not so dependent on glucocorticoids or is so strong that glucocorticoid influences are hard to demonstrate.Yet there is evidence for an influence of glucocorticoids on the flow of information within the amygdala. Glucocorticoids potentiate serotonin inhibition of the processing of excitatory input to the lateral amygdala from the thalamus, suggesting that there is a mechanism for containing, or limiting, the sensory input that is important for
Hippocampus Amygdala Septum Hypothalamus Cerebral cortex Midbrain Brain stem Cerebellum MR and GR GR and some MR GR and some MR GR mostly; low levels of MR GR mostly; low levels of MR GR mostly; low levels of MR GR mostly; patches of MR GR mostly

fear conditioning.24 Thus, adrenal steroids may regulate the nature of the signals that reach the amygdala and allow for greater discrimination of the most salient cues for learning. Moreover, in passive avoidance, both catecholamines and glucocorticoids play a role in facilitating learning.25,26 Catecholamines work outside of the bloodbrain barrier and their effects can be blocked by -adrenergicblocking agents, which do not cross the bloodbrain barrier.26 Glucocorticoids enter the brain, and local implants of exogenous corticosterone into hippocampus, amygdala, and nucleus tractus solitarii are all able to enhance passive avoidance learning.25 Adrenal steroids also play a supporting role in the learning of a spatial navigation task in mice.27 Adrenalectomy impairs the acquisition of the memory of hidden platform location in the Morris water maze, and glucocorticoid administration restores the normal learning curve; however, in mice in which the glucocorticoid receptor (GR) is deleted and replaced with a GR that lacks the DNA binding domain, glucocorticoids do not improve task acquisition.27 This finding illustrates a role for GRs acting upon the genome in a task that is known to depend on the hippocampus. Interestingly, other actions of glucocorticoids via GRs are known to involve the proteinprotein interactions that are not prevented in mice carrying the GR defective in the DNA binding domain.28 Other evidence for glucocorticoid actions supports an inverted U-shaped doseresponse curve in which low to moderate levels of adrenal steroids enhance acquisition of tasks that involve the hippocampus, whereas high levels of glucocorticoids disrupt task acquisition.22,29-31 Adrenal steroids have biphasic effects upon excitability of hippocampal neurons, which may underlie their biphasic actions on memory and recall.30,32-34

Adaptive structural plasticity

One of the ways that stress hormones modulate function within the brain is by changing the structure of neurons. Within the hippocampus, the input from the entorhinal cortex to the DG is ramified by the connections between
Cognitive impairment, spatial recognition memory (hippocampus) Increased anxiety and enhanced fear conditioning (amygdala) Increased aggression (amygdala) Table II. Cumulative effects of restraint stress on behavior.

Table I. Distribution of adrenal steroid receptors in brain regions. GR, glucocorticoid receptor; MR, mineralocorticoid receptor.

121

State of the art

the DG and the CA3 pyramidal neurons. One granule neuron innervates, on average, 12 CA3 neurons; and each CA3 neuron innervates, on the average, 50 other CA3 neurons via axon collaterals, as well as 25 inhibitory cells via other axon collaterals (Figure 2).35 The net result is a 600-fold amplification of excitation as well as a 300-fold amplification of inhibition, which provide some degree of control of the system. As to why this system exists, the DG-CA3 system is believed to play a role in the memory of sequences of events, although long-term storage of memory occurs in other brain regions.36,37 ble of participating in long-term potentiation. In the adult rat, 9000 new neurons are born per day and survive with a half-life of 28 days.41 There are many hormonal and neurochemical modulators of neurogenesis and cell survival in the DG.15,38,42-44 Neurogenesis in the adult DG is enhanced by the hormone insulin-like growth factor1 (IGF-1) and by serotonin and a number of antidepressant drugs. Estradiol accelerates cell proliferation in female rats. IGF-1 is the mediator of the ability of exercise to increase cell proliferation in the DG. Lack of IGF-1 and insulin in diabetes has the opposite effect and decreases cell proliferation. Neurogenesis and/or survival of newly born cells is increased by putting mice in a complex (enriched) environment.45 It is also increased by a form of classical conditioning that activates the hippocampus (trace conditioning) prolongs the survival of newly born DG neurons.46,47 On the other hand, certain types of acute stress and many chronic stressors suppress neurogenesis or cell survival in the DG, and the mediators of these inhibitor effects include excitatory amino acids acting via N-methyl-D-aspartate (NMDA) receptors and endogenous opioids.2,48-50 Chronic stress has even more potent effects on neurogenesis and neuronal survival. CRS for 21 days suppressed neurogenesis and CRS for 42 days causes the number of DG neurons to decrease along with total DG volume (Figure 3).51

Neurogenesis in the DG
There is structural plasticity within the DG-CA3 system, in that new neurons continue to be produced in the DG throughout adult life38 and CA3 pyramidal cells undergo remodeling of their dendrites,2 as will be discussed further below.39 The subgranular layer of the DG contains cells that have properties of astrocytes (eg, expression of glial fibrillary acidic protein) and give rise to granule neurons.40 After administration of bromodeoxyuridine (BrdU) to label DNA of dividing cells, these newly born cells appear as clusters in the inner part of the granule cell layer, where a substantial number of them will go on to differentiate into granule neurons within as little as 7 days. The new granule neurons appear to be quite excitable and capa-

Remodeling of dendrites
Another form of structural plasticity is the remodeling of dendrites in the hippocampus.39 CRS causes retraction and simplification of dendrites in the CA3 region of the hippocampus (Figure 4).2 Such dendritic reorganization can also be seen in rats undergoing adaptation of psychosocial stress in the visible burrow system (VBS). The VBS is an apparatus with an open chamber where there is a food and water supply and several tunnels and chambers.52 Rats can be observed from above by a video camera in this apparatus. In the VBS, male rats housed with several females establish a dominance hierarchy within several days. Over the course of the next week, a few subordinate males may die and others (showing scars from bite marks) will show enlarged adrenals, low testosterone, and many changes in brain chemistry. The dominant shows the fewest scars and has the highest level of testosterone, but also has somewhat larger adrenal glands than cage control rats.

Figure 2. Why is the CA3 so vulnerable? Feed-forward excitability serves memory functions but increases vulnerability for excitotoxicity. DG, dentate gyrus.

122

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Figure 3. A single restraint stress does not suppress cell proliferation. Repeated restraint stress for 21 days suppresses cell proliferation. Repeated restraint stress for 42 days reduces volume of the dentate gyrus (DG) and the number of neurons in the DG.
Reproduced from reference 51 with permission: Pham K, Nacher J, Hof PR, McEwen BS. Repeated restraint stress suppresses neurogenesis and induces biphasic PSA-NCAM expression in the adult rat dentate gyrus. Eur J Neurosci. 2003;17:879-886. Copyright 2003, Blackwell Publishing, Inc.

Regarding changes in brain structure, it was the dominant rats that had a more extensive pattern of debranching of the apical dendrites of the CA3 pyramidal neurons in the hippocampus, compared with the subordinate rats, which showed reduced branching compared with the cage controls.53 What this result emphasizes is that it is not adrenal size or presumed amount of physiological stress per se that determines dendritic remodeling, but a complex set of other factors that modulate neuronal structure. We refer to the phenomenon as dendritic remodeling and we generally find that it is a reversible process. In hibernating hamsters, it occurs in a matter of hours and reverses itself just as quickly when hibernating animals are aroused from torpor (A. M. Magarinos, B. S. McEwen, P. Pevet, unpublished data). Below we consider mechanisms involved in structural remodeling. The role of adrenal steroids in the structural remodeling described above reflects may interactions with neurochemical systems in the hippocampus, including serotonin, -aminobutyric acid (GABA), and excitatory amino acids (Figure 5).2 Probably the most important interactions are those with excitatory amino acids such as glutamate. Excitatory amino acids released by the mossy fiber pathway play a key role in the remodeling of the CA3 region of the hippocampus, and regulation of glutamate release by adrenal steroids may play an important role.54-57 We have found that the glutamate transporter, Glt-1, is elevated by CRS in hippocampus, particularly in the CA3 region, providing another indication that elevated gluta-

Figure 4. Hippocampal CA3 pyramidal neurons are remodeled by 21-d restraint stress. A. Control. B. 21 days chronic restraint stress.

mate levels are an important component of structural plasticity. We previously showed that NMDA receptor blockade and the Na/Ca channel blocker, phenytoin, both block CRS- and glucocorticoid-induced remodeling of dendrites in CA3.58-60 Recent evidence indicates that presynaptic receptors containing kainate receptor subunits such as GluR6 are important for the feed-forward actions of glutamate on mossy fiber terminals,61 and one study showed that a number of kainate receptor subunit mRNAs are regulated biphasically by adrenal steroids.57 In particular, preferential mineralocorticoid receptor (MR) occupancy by low corticosterone (CORT) levels enhanced mRNA levels for KAR2, GluR6, and GluR7.57 This agrees with our finding that MR activation by aldosterone in adrena-

123

State of the art

lectomized (ADX) rats restored levels of [3H]kainate binding in the mossy fiber region of CA3.56 However, further studies are needed. Because excitatory amino acids play a key role along with circulating glucocorticoids, the activation of the CREB (cyclic adenosine monophosphate response elementbinding protein) system is a likely candidate mediator, and recent evidence indicates that phosphorylation of CREB is chronically activated in rats subjected to CRS. CREB has been linked to regulation of synaptic plasticity and particularly neurogenesis.62 It is possible that CREB is involved in activity-dependent synapse formation, which is evident as a result of long-term potentiation.63,64 However, the role of glucocorticoids in activation of the CREB system has not been thoroughly investigated. Nevertheless, treatment with the mood stabilizer lithium prevented CRS-induced structural remodeling of the stress-induced elevation of Glt-1 and CREB phosphorylation (G. E. Wood, L. T. Young, B. S. McEwen, unpublished data), providing further evidence that CRS-induced structural plasticity and the molecular markers Glt-1 and phosphoCREB are useful in study of psychiatric illnesses. Structural changes in dendrites and spine synapses are the result of modifications in the microtubule system of the cytoskeleton,65 and new evidence shows that posttranslational modification of tubulin65 and phosphorylation of the microtubule associated protein tau66 take place along with changes in the actin cytoskeleton,67 under conditions in which reorganization of dendrites and synaptic connections occur. Overall, cytoskeletal changes, such as increased paired-helical-like phosphorylation of tau66 and reduced tyrosinated tubulin,65 are consistent with increased cytoskeletal rigidity. However, this needs much careful study. The Rac/Rho guanosine triphosphatases (GTPases) and related proteins such as the guanosine triphosphate (GTP) exchange factor, kalirin, have been shown to play a key regulatory role in cytoskeletal modifications in developing and adult neurons.67,68 Except for one relevant study on seizures,65 there are no studies thus far of the effects of chronic stress on these pathways or of the modifications of the cytoskeleton itself. Besides glucocorticoids and excitatory amino acids, neurotrophins and gp130 cytokines are implicated in structural plasticity along with extracellular proteases such as tissue plasminogen activator (tPA) and neuropsin. Brainderived neurotrophic factor (BDNF) plays a major role in activity-dependent synaptic and dendritic remodeling,69-73 and is implicated in hippocampal-dependent memory formation.74 BDNF also regulates tPA release from neurons75 and tPA is released from nerve terminals in hippocampus and other brain areas such as amygdala.76-78 It has been suggested that tPA may play a role in the processing of proBDNF into active forms.79 The activity of tPA is associated with structural plasticity and increased fear,77 motor learning,80 and enhancement of long-term potentiation.81 Activity of tPA is an important mediator of structural plasticity and enhanced fear in the amygdala resulting from acute restraint stress. For example, plasminogen (inactive zymogen) leads to plasmin (active serine protease). Using tPA knockout mice, we have found that in medial and central amygdala77: tPA is released under stress and initiates neural remodeling. This release is plasminogen-independent (extracellular signalregulated kinase [ERK1/2]; guanosine triphosphateactivating protein [GAP-43]).

Glutamate levels NMDA receptors Ca++ currents 5-HT turnover 5-HT2 receptors 5-HT1A receptors GABAA receptors

++ ++ ++ ++ ++ -+/-

Figure 5. Glucocorticoids increase glutamate levels, N-methyl-D-aspartate (NMDA) receptors, calcium currents, 5-hydroxytryptamine (5-HT) turnover, and 5-HT2 receptors, decrease 5-HT1A receptors, and alter subunit expression of GABAA receptors. A. Cross-section of dorsal hippocampus. B. Blow-up of CA3 region. C. CA3 neurons highlighting stratum lucium (SL), where mossy fiber terminals form synaptic contacts .GABA, -aminobutyric acid; DG, dentate gyrus; SR, stratum radiatum; SP, stratum pyramidale; SO, stratum oriens.

124

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

tPA induces termination of its own action via plasminogen-activator inhibitor1 (PAI-1). tPA activity is required for increased anxiety in the elevated plus maze. We are presently studying the long-term effects of stress. Neuropsin is another protease that is induced in hippocampus by NMDA-mediated excitation in seizures and leads to proteolysis of the presynaptic adhesion molecule, L1.82 The gp130 cytokines are expressed in hippocampus under stimulation by seizures, along with their receptors, which are constitutively expressed.83 Leukemia inhibitory factor (LIF) is particularly interesting because it interferes with neurotrophin signaling84 and causes dendritic retraction in cell culture.85 However, it has not yet been determined whether acute or chronic stress increases LIF expression, and it is conceivable that increased expression of LIF might play a role in dendritic shortening. The ability of neuronal processes to expand or contract, and newly formed neurons to make connections, is dependent on the extracellular environment in which polysialated neural cell adhesion molecule (PSA-NCAM) plays an important role.86 PSA-NCAM is associated with regions of the brain that show structural plasticity such as the inner granule cell layer of the DG and the mossy fiber terminals of CA3.87 CRS for 21 days causes increased PSA-NCAM expression in the DG proliferative zone even though cell proliferation is suppressed, and these changes have disappeared after CRS for 42 days.51 This raised questions about the role of PSA-NCAM in adaptive structural plasticity, which need to be investigated. Removal of the PSA residue by endoneuraminidase (EndoN)88 is a powerful tool for manipulating this system, since PSA removal abolishes plasticity of suprachiasmatic neurons to environmentally induced phase shifting of the diurnal rhythm.89 We now turn to the important question of whether chronic stress increases or decreases vulnerability of the hippocampus to damage from other insults.

Permanent damage as a result of stress

The remodeling of the hippocampus in response to stress is largely reversible if the CRS is terminated at the end of 3 weeks.10 After 3 weeks of CRS, neurogenesis is reduced in DG and dendrites are shorter and less branched,51,59,60 and there is an increase in PSA-NCAM expression in the DG that is consistent with increased mobility of neuronal processes even in the face of reduced DG neuron pro-

duction. Continuation of CRS for a total of 6 weeks abolishes the upregulation of PSA-NCAM and results in a significant 6% reduction in DG volume and 13% reduction in granule neuron number.51 We do not yet know whether structural changes occurring after 6 weeks of CRS are reversible or whether they can be accelerated by antidepressant or antiepileptic drugs that block the effects of stress and glucocorticoids on remodeling. Nor do we know whether the structural changes occurring with CRS increase or decrease the vulnerability of the hippocampus to damage by excitotoxicity. It is well established that glucocorticoids exacerbate damage to the hippocampus caused by ischemia90 and seizures.91,92 Glucocorticoids exacerbate excitotoxic damage and do so, at least in part, by facilitating trafficking of immune cells to the injury site,93 and, there, cytotoxic T cells are able to produce cytotoxic death of neurons.94 However, the phenomenon of ischemic preconditioning95 reveals that prior stimulation of the hippocampus can induce a protective mechanism that may reduce the damage produced by a full-scale ischemic event. It is not clear whether the same mechanisms might be operative when stress is applied and whether they might affect the response to excitotoxicity in response to seizures, but this possibility needs to be kept in mind if it turns out that prior CRS has a protective effect on subsequent responses to excitotoxic challenge. Protective agents may also involve substances that are upregulated in the brain in response to damage or threat of damage. One of the prominent features of excitotoxic damage or removal of adrenal steroids is the robust induction of calcitonin generelated peptide (CGRP) in terminals and cell bodies in hippocampus and in mossy cells. The increased expression of CGRP in mossy cells is especially prominent after bilateral ADX under conditions in which there is apoptosis of granule cells, and the CGRP immunoreactivity is enhanced within the inner third of the molecular layer of the DG. The neuroimmune peptide, CGRP, is one of the most diverse and influential immunoregulators of the periphery. This important neuropeptide has multiple functions including: its actions as a potent vasodilator96 and an immune modulator,97-102 as well as a neural and immune developmental regulator, a modulator of hormone release involved in growth and development, and a stimulator of sympathetic outflow, which is mediated by CRF and an inducer of apoptosis (reviewed in reference 103). Some of the different functional roles for CGRP may not be independent, but may be part of a

125

State of the art

cascade of events that constitute the healing response to injury. A number of studies have shown that CGRP is expressed following various kinds of trauma and plays an important role in the acute phase response that may be of particular relevance to the outcome of the regional injury response in the central nervous system (CNS).103,104 In recent studies, the expression of CGRP within the hippocampus increases in four separate models of CNS injury: ADX,105 intrahippocampal colchicine injection,105 trimethyltin ingestion,106 and kainic acid injections. In each case, the expression of this peptide was limited to the specific region of damage and in association with the surviving neuronal population. Although the upregulation of CGRP may be associated with neuronal cell survival,107 other studies have shown that both microglia and astrocytes express CGRP receptors and that exposure to physiological levels of CGRP induces c-fos in microglia and astrocytes and increases plasminogen activators.108 The role of CGRP may then not only protect against immune system damage to neurons, but may also participate in plasticity and healing. The brain is the master controller of the three systems noted above and is also a target of these systems, subject to both protection and damage. Allostasis also applies not only to circulating hormones, but also to organs and tissues of the body. In the nervous system, neurotransmitters are released by neuronal activity, and they produce effects locally to either propagate or inhibit further neural activity. Neurotransmitters and hormones are usually released during a discrete period of activation and then are shut off, and the mediators themselves are removed from the intracellular space by reuptake or metabolism in order not to prolong their effects. When that does not happen, however, there is allostatic load and the brain is at increased risk for damage.115,116 The processes of allostasis and allostatic load have been described and measured for metabolic and cardiovascular changes that are associated with obesity, type 2 diabetes, and cardiovascular disease.117 However, the same type of elevated and prolonged secretion of glucocorticoids during aging has also been associated with impairment of cognitive function in rodents118-120 and humans.121-123 Moreover, the endogenous excitatory amino acid neurotransmitters appear to play a major role in these changes,119 even though they are also an essential part of normal synaptic neurotransmission and plasticity.

Protective and damaging effects of the mediators of adaptation

Individual differences in the progression of a number of disorders that accumulate with time can be conceptualized as an accumulation of wear and tear of daily experiences, lifestyle, and major life stressors, which interact with the genetic constitution and predisposing early life experiences.109-111 The neuroendocrine system, autonomic nervous system, and immune system are mediators of adaptation to the challenges of daily life, referred to as allostasis, meaning maintaining stability through change.112 Physiological mediators, such as adrenalin from the adrenal medulla, glucocorticoids from the adrenal cortex, and cytokines from cells of the immune system, act upon receptors in various tissues and organs to produce effects that are adaptive in the short term, but can be damaging if the mediators are not shut off when no longer needed. When release of the mediators is not efficiently terminated, their effects on target cells are prolonged, leading to other consequences that may include receptor desensitization and tissue damage. This process has been named allostatic load,113,114 which refers to the price the tissue or organ pays for an overactive or inefficiently managed allostatic response. Therefore, allostatic load refers to the cost of adaptation.

Allostatic states in depressive illness

Stress hormones are elevated in major depressive illness. In particular the diurnal rhythm is distorted.124 Normally low evening levels of cortisol are increased in a subset of depressed patients125,126 and the stress hormone axis in major depression is resistant to suppression by the synthetic glucocorticoid dexamethasone.127 It is also noteworthy that androgen levels are elevated in women with major depression, which undoubtedly reflects adrenal hyperactivity.128 IGF-1 levels are also reported to be elevated in major depression, and this may reflect elevated growth hormone release as a result of the hypercortisolemia.129 Each of these patterns of elevation constitutes an allostatic state, and represents a pathway for the development of allostatic load in the brain and in other organs throughout the body. Regarding the brain, we already noted the studies showing that hippocampal volume loss in major depressive illness is related to duration of the depression rather than to age per se of the patients.130-132 Not all studies report such changes (see, for example, references 133 and 134); the reasons for these different results are beyond

126

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

the scope of this discussion, but they may be explained by differences in the duration of depression, as well as gender and age. It should be noted that hippocampal size in elderly twins shows only 40% genetic contribution, with the predominant influence being environmental.135 This emphasizes the importance of experimental factors and allostatic load in determining hippocampal volume. Hippocampal atrophy has been found in relation to depression in the elderly,136 with an association detected with presence of the ApoE4 genotype.137 In subjects with a long-term history of depression, Sheline and colleagues described magnetic resonance imaging (MRI) evidence for discontinuities that might represent sites of damage.130 Although some recent postmortem studies on brains from depressed individuals did not show neuron loss in hippocampus,138,139 the duration of the depression and the subtype of depression were not carefully controlled.Thus, the possibility that neural damage may ultimately occur in major depression cannot be disregarded, particularly when depression lasts a long time. However, in a recent study in young depressed subjects, hippocampal volume was not smaller in first-episode depression, but declined rapidly over several years.140 The key, unanswered question is whether such changes can be prevented or even reversed. It is important to note that other brain regions besides hippocampus are affected in depressive illness and undergo structural changes. One region is the prefrontal cortex, and structural imaging141 showed loss of volume in familial pure depressive disorder, whereas autopsy studies142-144 have shown loss of volume and glial cells, as well as neuronal density in both unipolar and bipolar disorder. There is one animal study showing that chronic glucocorticoid treatment induces loss of dendrites in the rat prefrontal cortex.4 However, much more work needs to be done on this brain region. Depressive illness is associated with a hyperactivation of the amygdala,145,146 and more recently, with an actual enlargement of the amygdala in the first episode of major depression.147 This is reminiscent of the increased dendritic branching reported in rats after repeated immoblization stress (see above and reference 148). Since the amygdala integrates information related to fear and strong emotions, and also sends outputs via the central nucleus for autonomic arousal and via the basal nucleus for more active aspects of coping,149 the elevation of amygdala activity may be a first step that leads to overactivation of systems involved in physiological and behavioral coping. The long-term consequences of this may well be a wear

and tear on the body that results in a number of pathophysiological consequences, since the amygdala regulates both autonomic nervous system activity and adrenocorticotropic hormone (ACTH) and cortisol production through outputs of its central nucleus.149,150 It is important to note that there are reports that in recurrent major depression of long duration the amygdala may undergo shrinkage.131,151 It is thus possible that initial hypertrophy gives way to atrophy in this important brain structure. Besides the brain changes in major depression, there are other changes in the body that reflect dysregulated hypothalamopituitary axis (HPA) and autonomic activity, and are slow in developing.These constitute allostatic load that produces cumulative pathophysiology, which may also be reversible if caught in time. Such cumulative, long-term effects include bone mineral loss152-154 and abdominal fat deposition.155-157 Moreover, the combination of long-term allostatic load, together with dysregulation of the autonomic nervous system in major depression,158 is associated with increased blood platelet reactivity159-161 and increased risk for cardiovascular disease.162-165 There are parallels between the story for major depression and what is known about psychiatric and somatic features of Cushings disease involving melancholia, depression, abdominal obesity, bone mineral loss, and increased risk for cardiovascular disease.166-169 In addition, there is evidence for hippocampal atrophy in Cushings disease along with memory impairments.170-172 Interestingly, hippocampal volume loss in Cushings disease is at least partially reversible over several years after correction of the hypercortisolemia.173-175 Finally, a largely unexplored area concerns the effects of antidepressant medication on the brain and body changes associated with depressive illness. On the one hand, certain antidepressants may contribute to some of the associated pathophysiology, such as cardiovascular instability.176 On the other hand, withdrawal from antidpressant treatment may cause imbalances in neurotransmitter systems, with elevations of excitatory amino acid tone,177 and contribute to the allostatic load that occurs as the depressive state continues.178

Conclusion
Translational studies of brain changes in major psychiatric illnesses such as unipolar and bipolar depression and posttraumatic stress disorder are showing that changes in volume of structures such as hippocampus, prefrontal cortex,

127

State of the art

and amygdala must be considered as part of the neurobiological consequences of these illnesses.17,18,140,179,180 Structural remodeling in these brain regions is important for human psychiatric disorders because the altered circuitry is likely to contribute to impaired cognitive function and affect regulation. Moreover, stress is widely acknowledged as a predisposing and precipitating factor in psychiatric illness.181,182 Thus, animal models are relevant to human psychiatric disorders in at least four ways: First, they have led toand continue to contribute basic knowledge to the ongoing studies of how the human brain changes structurally in depression and related psychiatric disorders. Second, the structural changes that occur with chronic stress appear to be reversible as long as the stress is terminated in time. This suggests the hopeful possibility that brain changes in at least some major psychiatric disorders may be treatable if we can find the right agents or therapies and intervene in time. Third, reversible or not, the effects of chronic stress may predispose to greater vulnerability to adverse consequences from other insults. Fourth, the systemic manifestations of the allostatic load generated by chronic psychiatric disorders affects the metabolic, immune, and cardiovascular systems, leading to systemic disorders that add to the costs of health care.
Research support has come from the National Institute of Mental Health Grants MH 41256 and MH58911. The author is also indebted to colleagues in the John D. and Catherine T. MacArthur Foundation Health Program and its Network on Socioeconomic Status and Health (Nancy Adler, PhD, Chair).

REFERENCES
1. McEwen BS. Protective and damaging effects of stress mediators. N Engl J Med. 1998;338:171-179. 2. McEwen BS. Stress and hippocampal plasticity. Annu Rev Neurosci. 1999;22:105-122. 3. Sousa N, Lukoyanov NV, Madeira MD, Almeida OFX, Paula-Barbosa MM. Reorganization of the morphology of hippocampal neurites and synapses after stress-induced damage correlates with behavioral improvement. Neuroscience. 2000;97:253-266. 4. Wellman CL. Dendritic reorganization in pyramidal neurons in medial prefrontal cortex after chronic corticosterone administration. J Neurobiol. 2001;49:245-253. 5. Vyas A, Mitra R, Rao BSS, Chattarji S. Chronic stress induces contrasting patterns of dendritic remodeling in hippocampal and amygdaloid neurons. J Neurosci. 2002;22:6810-6818. 6. Shors TJ, Chua C, Falduto J. Sex differences and opposite effects of stress on dendritic spine density in the male versus female hippocampus. J Neurosci. 2001;21:6292-6297. 7. Sunanda MSR, Raju TR. Effect of chronic restraint stress on dendritic spines and excrescences of hippocampal CA3 pyramidal neuronsa quantitative study. Brain Res. 1995;694:312-317. 8. Luine V, Villegas M, Martinez C, McEwen BS. Repeated stress causes reversible impairments of spatial memory performance. Brain Res. 1994;639:167-170. 9. Conrad CD, Galea LAM, Kuroda Y, McEwen BS. Chronic stress impairs rat spatial memory on the Y-maze and this effect is blocked by tianeptine pretreatment. Behav Neurosci. 1996;110:1321-1334. 10. Conrad CD, Magarinos AM, LeDoux JE, McEwen BS. Repeated restraint stress facilitates fear conditioning independently of causing hippocampal CA3 dendritic atrophy. Behav Neurosci. 1999;113:902-913. 11. Wood GE, Young LT, Reagan LP, McEwen BS. Acute and chronic restraint stress alters the incidence of social conflict in male rats. Horm Behav. 2003;43:205-213. 12. Gould E, Tanapat P, McEwen BS, Flugge G, Fuchs E. Proliferation of granule cell precursors in the dentate gyrus of adult monkeys is diminished by stress. Proc Natl Acad Sci U S A. 1998;95:3168-3171. 13. Gould E, McEwen BS, Tanapat P, Galea LAM, Fuchs E. Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation. J Neurosci. 1997;17:2492-2498.

14. Magarinos AM, McEwen BS, Flugge G, Fuchs E. Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J Neurosci. 1996;16:3534-3540. 15. Czeh B, Michaelis T, Watanabe T, et al. Stress-induced changes in cerebral metabolites, hippocampal volume and cell proliferation are prevented by antidepressant treatment with tianeptine. Proc Natl Acad Sci U S A. 2001;98:12796-12801. 16. Van Kampen M, Kramer M, Hiemke C, Flugge G, Fuchs E. The chronic psychosocial stress paradigm in male tree shrews: evaluation of a novel animal model for depressive disorders. Stress. 2002;5:37-46. 17. McEwen BS. Mood disorders and allostatic load. Biol Psychiatry. 2003;54:200-207. 18. Sheline YI. Neuroimaging studies of mood disorder effects on the brain. Biol Psychiatry. 2003;54:338-352. 19. van der Hart MGC, Czeh B, de Biurrun G, et al. Substance P receptor antagonist and clomipramine prevent stress-induced alterations in cerebral metabolites, cytogenesis in the dentate gyrus and hippocampal volume. Mol Psychiatry. 2002;7:933-941. 20. Stockmeier CA, Mahajan GJ, Konick LC, et al. Preliminary evidence that neuronal and glial density is increased and neuronal size is decreased in hippocampus in major depressive disorder (MDD). Abst Soc Neurosci. 2002;28:497.19. 21. Corodimas KP, LeDoux JE, Gold PW, Schulkin J. Corticosterone potentiation of learned fear. Ann N Y Acad Sci. 1994;746:392. 22. Pugh CR, Tremblay D, Fleshner M, Rudy JW. A selective role for corticosterone in contextual-fear conditioning. Behav Neurosci. 1997;111:503-511. 23. Pugh CR, Fleshner M, Rudy JW. Type II glucocorticoid receptor antagonists impair contextual but not auditory-cue fear conditioning in juvenile rats. Neurobiol Learn Memory. 1997;67:75-79. 24. Stutzmann GE, McEwen BS, LeDoux JE. Serotonin modulation of sensory inputs to the lateral amygdala: dependency on corticosterone. J Neurosci. 1998;18:9529-9538. 25. Roozendaal B. Glucocorticoids and the regulation of memory consolidation. Psychoneuroendocrinology. 2000;25:213-238. 26. Cahill L, Prins B, Weber M, McGaugh JL. -Adrenergic activation and memory for emotional events. Nature. 1994;371:702-704. 27. Oitzl MS, Reichardt HM, Joels M, de Kloet ER. Point mutation in the mouse glucocorticoid receptor preventing DNA binding impairs spatial memory. Proc Natl Acad Sci U S A. 2001;98:12790-12795. 28. Reichardt HM, Schutz G. Glucocorticoid signalingmultiple variations of a common theme. Mol Cell Endocrinol. 1998;146:1-6.

128

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Plasticidad estructural del cerebro adulto: cmo los modelos animales nos ayudan a comprender los cambios cerebrales en la depresin y los trastornos sistmicos relacionados con la depresin
El cerebro interpreta experiencias y las traduce en respuestas conductuales y fisiolgicas. Los acontecimientos estresantes son aquellas situaciones amenazantes, o al menos, inesperadas y sorpresivas; y las respuestas fisiolgicas y conductuales intentan promover una adaptacin a travs de un proceso llamado alostasis. Los mediadores qumicos de la alostasis incluyen el cortisol y la adrenalina de las glndulas adrenales, otras hormonas y neurotransmisores, el sistema nervioso parasimptico y simptico, y citoquinas y quimioquinas del sistema inmune. Dos estructuras cerebrales, la amgdala y el hipocampo, juegan papeles clave en la interpretacin de lo que es estresante y en la determinacin de respuestas apropiadas. El hipocampo, una estructura clave para las memorias de los acontecimientos y del contexto, expresa receptores que lo capacitan para responder a hormonas glucocorticodeas de la sangre. El hipocampo se atrofia en numerosos trastornos psiquitricos y tambin responde a estresores con cambios en la excitabilidad, disminucin de las ramificaciones dendrticas y reduccin del nmero de neuronas del giro dentado. La amgdala, que es importante para las memorias emocionales, aumenta su actividad en el trastorno por estrs postraumtico y en la enfermedad depresiva. En modelos animales de estrs existen evidencias del crecimiento e hipertrofia de clulas nerviosas en la amgdala. Los cambios en el cerebro despus de situaciones de estrs agudo y crnico reflejan el patrn observado en los sistemas metablico, cardiovascular e inmune; esto es, una adaptacin a corto plazo (alostasis) seguida de un dao a largo plazo (carga alosttica), como por ejemplo, ateroesclerosis, obesidad localizada, desmineralizacin del hueso y deterioro de la funcin inmune. La carga alosttica de este tipo se observa en la depresin mayor y tambin se puede expresar en otros trastornos ansiosos y afectivos crnicos.
29. Conrad CD, Lupien SJ, McEwen BS. Support for a bimodal role for type II adrenal steroid receptors in spatial memory. Neurobiol Learn Memory. 1999;72:39-46. 30. Diamond DM, Bennett MC, Fleshner M, Rose GM. Inverted-U relationship between the level of peripheral corticosterone and the magnitude of hippocampal primed burst potentiation. Hippocampus. 1992;2:421-430. 31. Diamond DM, Park CR, Heman KL, Rose GM. Exposing rats to a predator impairs spatial working memory in the radial arm water maze. Hippocampus. 1999;9:542-552. 32. Pavlides C, Kimura A, Magarinos AM, McEwen BS. Type I adrenal steroid receptors prolong hippocampal long-term potentiation. Neuroreport. 1994;5:2673-2677. 33. Pavlides C, Watanabe Y, Magarinos AM, McEwen BS. Opposing role of adrenal steroid type I and type II receptors in hippocampal long-term potentiation. Neuroscience. 1995;68:387-394. 34. Pavlides C, McEwen BS. Effects of mineralocorticoid and glucocorticoid receptors on long-term potentiation in the CA3 hippocampal field. Brain Res. 1999;851:204-214. 35. Feng R, Rampon C, Tang YP, et al. Deficient neurogenesis in forebrainspecific presenilin-1 knockout mice is associated with reduced clearance of hippocampal memory traces. Neuron. 2001;32:911-926. 36. Eichenbaum H, Harris K. Toying with memory in the hippocampus. Nat Neurosci. 2000;3:205-206. 37. Lisman JE. Relating hippocampal circuitry to function: recall of memory sequences by reciprocal dentate-CA3 interactions. Neuron. 1999;22:233-242. 38. Gould E, Tanapat P, Rydel T, Hastings N. Regulation of hippocampal neurogenesis in adulthood. Biol Psychiatry. 2000;48:715-720. 39. McEwen BS, Magarinos AM. Stress and hippocampal plasticity: implications for the pathophysiology of affective disorders. Hum Psychopharmacol. 2001;16:S7-S19. 40. Seri B, Garcia-Verdugo JM, McEwen BS, Alvarez-Buylla A. Astrocytes give rise to new neurons in the adult mammalian hippocampus. J Neurosci. 2001;21:7153-7160. 41. Altman J, Das GD. Autoradiographic and histological studies of postnatal neurogenesis. I. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in neonate rats, with special reference to postnatal neurogenesis in some brain regions. J Comp Neurol. 1965;126:337-390. 42. Malberg JE, Eisch AJ, Nestler EJ, Duman RS. Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J Neurosci. 2000;20:9104-9110. 43. Aberg MA, Aberg ND, Hedbacker H, Oscarsson J, Eriksson PS. Peripheral infusion of IGF-1 selectively induces neurogenesis in the adult rat hippocampus. J Neurosci. 2000;20:2896-2903. 44. Trejo JL, Carro E, Torres-Aleman I. Circulating insulin-like growth factor I mediates exercise-induced increases in the number of new neurons in the adult hippocampus. J Neurosci. 2001;21:1628-1634. 45. Kempermann G, Kuhn HG, Gage FH. More hippocampal neurons in adult mice living in an enriched environment. Nature. 1997;586:493-495. 46. Gould E, Beylin A, Tanapat P, Reeves A, Shors TJ. Learning enhances adult neurogenesis in the hippocampal formation. Nat Neurosci. 1999;2:260265. 47. Shors TJ, Miesegaes G, Beylin A, Zhao M, Rydel T, Gould E. Neurogenesis in the adult is involved in the formation of trace memories. Nature. 2001;410:372-376. 48. Gould E, Tanapat P. Stress and hippocampal neurogenesis. Biol Psychiatry. 1999;46:1472-1479. 49. Cameron HA, Tanapat P, Gould E. Adrenal steroids and N-methyl-Daspartate receptor activation regulate neurogenesis in the dentate gyrus of adult rats through a common pathway. Neuroscience. 1998;82:349-354. 50. Eisch AJ, Barrot M, Schad CA, Self DW, Nestler EJ. Opiates inhibit neurogenesis in the adult rat hippocampus. Proc Natl Acad Sci U S A. 2000;97:7579-7584. 51. Pham K, Nacher J, Hof PR, McEwen BS. Repeated restraint stress suppresses neurogenesis and induces biphasic PSA-NCAM expression in the adult rat dentate gyrus. Eur J Neurosci. 2003;17:879-886.

129

State of the art

Plasticit structurale du cerveau adulte : comment les modles animaux nous aident comprendre les modifications crbrales dans la dpression et les troubles gnraux lis la dpression
Le cerveau interprte les expriences et les traduit en rponses comportementales et physiologiques. Les vnements stressants sont ceux qui sont menaants ou tout au moins inattendus et surprenants et les rponses physiologiques et comportementales ont pour but de promouvoir ladaptation via un processus appel allostasie . Les mdiateurs chimiques de lallostasie incluent le cortisol et ladrnaline scrts par les glandes surrnales, dautres hormones et des neurotransmetteurs, les systmes nerveux sympathique et parasympathique, et les cytokines et chimiokines produites par le systme immunitaire. Deux structures crbrales, lamygdale et lhippocampe, jouent un rle-cl dans lidentification des vnements stressants et llaboration de rponses appropries. Lhippocampe, une structure-cl pour les souvenirs des vnements et contextes, exprime des rcepteurs qui lui permettent de rpondre aux hormones glucocorticodes du sang. Il subit une atrophie au cours de nombreux troubles psychiatriques et ragit galement aux facteurs de stress par des changements de lexcitabilit, une diminution de la ramification dendritique et une baisse du nombre de neurones dans le gyrus dent. Lamygdale, qui joue en rle important dans les souvenirs motionnels , devient hyperactive dans ltat de stress posttraumatique et la dpression. Les modles animaux de stress montrent lexistence dune croissance et dune hypertrophie des cellules nerveuses dans lamygdale. La chronologie des modifications du cerveau la suite de stress aigus ou chroniques (adaptation court terme [allostasie] suivie dune altration long terme [charge allostatique]), reflte celle observe au cours des affections touchant, par ex., les systmes mtabolique, cardiovasculaire et immunitaire, o la phase dadaptation se complique, respectivement, dathrosclrose et obsit localise, de dminralisation osseuse et daltrations de la fonction immunitaire. Une telle charge allostatique se rencontre dans la dpression majeure et peut aussi sexprimer dans lanxit chronique et dautres troubles de lhumeur.
52. Blanchard RJ, McKittrick CR, Blanchard DC. Animal models of social stress: effects on behavior and brain neurochemical systems. Physiol Behav. 2001;73:261-271. 53. McKittrick CR, Magarinos AM, Blanchard DC, Blanchard RJ, McEwen BS, Sakai RR. Chronic social stress reduces dendritic arbors in CA3 of hippocampus and decreases binding to serotonin transporter sites. Synapse. 2000;36:85-94. 54. Lowy MT, Gault L, Yamamoto BK. Adrenalectomy attenuates stressinduced elevations in extracellular glutamate concentrations in the hippocampus. J Neurochem. 1993;61:1957-1960. 55. Lowy MT, Wittenberg L, Novotney S. Adrenalectomy attenuates kainic acidinduced spectrin proteolysis and heat shock protein 70 induction in hippocampus and cortex. J Neurochem. 1994;63:886-894. 56. Watanabe Y, Weiland NG, McEwen BS. Effects of adrenal steroid manipulations and repeated restraint stress on dynorphin mRNA levels and excitatory amino acid receptor binding in hippocampus. Brain Res. 1995;680:217225. 57. Joels M, Bosma A, Hendriksen H, Diegenbach P, Kamphuis W. Corticosteroid actions on the expression of kainate receptor subunit mRNAs in rat hippocampus. Mol Brain Res. 1996;37:15-20. 58. Watanabe Y, Gould E, Cameron HA, Daniels DC, McEwen BS. Phenytoin prevents stress- and corticosterone-induced atrophy of CA3 pyramidal neurons. Hippocampus. 1992;2:431-436. 59. Magarinos AM, McEwen BS. Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: comparison of stressors. Neuroscience. 1995;69:83-88. 60. Magarinos AM, McEwen BS. Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: involvement of glucocorticoid secretion and excitatory amino acid receptors. Neuroscience. 1995;69:89-98. 61. Contractor A, Swanson G, Heinemann SF. Kainate receptors are involved in short- and long-term plasticity at mossy fiber synapses in the hippocampus. Neuron. 2001;29:209-216. 62. Nakagawa S, Kim JE, Lee R, et al. Regulation of neurogenesis in adult mouse hippocampus by cAMP and the cAMP response elementbinding protein. J Neurosci. 2002;22:3673-3682. 63. Andersen P, Soleng AF. Long-term potentiation and spatial training are both associated with the generation of new excitatory synapses. Brain Res Rev. 1998;26:353-359. 64. Toni N, Buchs PA, Nikonenko I, Bron CR, Muller D. LTP promotes formation of multiple spine synapses between a single axon terminal and a dendrite. Nature. 1999;402:421-425. 65. Bianchi M, Heidbreder C, Crespi F. Cytoskeletal changes in the hippocampus following restraint stress: role of serotonin and microtubules. Synapse. 2003;49:188-194. 66. Arendt T, Stieler J, Strijkstra AM, et al. Reversible paired helical filamentlike phosphorylation of tau is an adaptive process associated with neuronal plasticity in hibernating animals. J Neurosci. 2003;23:6972-6981. 67. Luo L. Rho GTPases in neuronal morphogenesis. Nat Rev Neurosci. 2000;1:173-180. 68. Ma XM, Mains RE, Eipper BA. Plasticity in hippocampal peptidergic systems induced by repeated electroconvulsive shock. Neuropsychopharmacology. 2002;27:55-71. 69. Bonhoeffer T. Neurotrophins and activity-dependent development of the neocortex. Curr Opin Neurobiol. 1996;6:119-126. 70. McAllister AK, Katz LC, Lo DC. Neurotrophins and synaptic plasticity. Annu Rev Neurosci. 1999;22:295-318. 71. Poo M. Neurotrophins as synaptic modulators. Nature. 2001;2:24-32. 72. Lee FS, Kim AH, Khursigara G, Chao MV. The uniqueness of being a neurotrophin receptor. Curr Opin Neurobiol. 2001;11:281-286. 73. Gorski JA, Zeiler SR, Tamowski S, Jones KR. Brain-derived neurotrophic factor is required for the maintenance of cortical dendrites. J Neurosci. 2003;23:6856-6865.

130

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

74. Alonso M, Vianna MRM, Depino AM, et al. BDNF-triggered events in the rat hippocampus are required for both short- and long-term memory formation. Hippocampus. 2002;12:551-560. 75. Fiumelli H, Jabaudon D, Magistretti PJ, Martin JL. BDNF stimulates expression, activity and release of tissue-type plasminogen activator in mouse cortical neurons. Eur J Neurosci. 1999;11:1639-1646. 76. Salles FJ, Strickland S. Localization and regulation of the tissue plasminogen activator-plasmin system in the hippocampus. J Neurosci. 2002;22:2125-2134. 77. Pawlak R, Magarinos AM, Melchor J, McEwen B, Strickland S. Tissue plasminogen activator in the amygdala is critical for stress-induced anxiety-like behavior. Nat Neurosci. 2003;6:168-174. 78. Sappino A-P, Madani R, Huarte J, et al. Extracellular proteolysis in the adult murine brain. J Clin Invest. 1993;92:679-685. 79. Lu B. Pro-region of neurotrophins: role in synaptic modulation. Neuron. 2003;39:735-738. 80. Seeds NW, Basham ME, Ferguson JE. Absence of tissue plasminogen activator gene or activity impairs mouse cerebellar motor learning. J Neurosci. 2003;23:7368-7375. 81. Zhuo M, Holtzman DM, Li Y, et al. Role of tissue plasminogen activator receptor LRP in hippocampal long-term potentiation. J Neurosci. 2000;20:542-549. 82. Matsumoto-Miyai K, Ninomiya A, Yamasaki H, Tamura H, Nakamura Y, Shiosaka S. NMDA-dependent proteolysis of presynaptic adhesion molecule L1 in the hippocampus by neuropsin. J Neurosci. 2003;23:7727-7736. 83. Rosell DR, Akama KT, Nacher J, McEwen BS. Differential expression of suppressors of cytokine signaling-1, -2, and -3 in the rat hippocampus after seizure: implications for neuromodulation by gp130 cytokines. Neuroscience. 2003;122:349-358. 84. Ng YP, He W, Ip NY. Leukemia inhibitory factor receptor signaling negatively modulates nerve growth factor-induced neurite outgrowth in PC12 cells and sympathetic neurons. J Biol Chem. 2003;278:38731-38739. 85. Guo X, Chandrasekaran V, Lein P, Kaplan PL, Higgins D. Leukemia inhibitory factor and ciliary neurotrophic factor cause dendritic retraction in cultured rat sympathetic neurons. J Neurosci. 1999;19:2113-2121. 86. Rutishauser U, Landmesser L. Polysialic acid in the vertebrate nervous system: a promoter of plasticity in cell-cell interactions. Trends Neurosci. 1996;19:422-427. 87. Seki T, Arai Y. The persistent expression of a highly polysialylated NCAM in the dentate gyrus of the adult rat. Neurosci Res. 1991;12:503-513. 88. Seki T, Rutishauser U. Removal of polysialic acid-neural cell adhesion molecule induces aberrant mossy fiber innervation and ectopic synaptogenesis in the hippocampus. J Neurosci. 1998;18:3757-3766. 89. Prosser RA, Rutishauser U, Ungers G, Fedorkova L, Glass D. Intrinsic role of polysialylated neural cell adhesion molecule in photic phase resetting of the mammalian circadian clock. J Neurosci. 2003;23:652-658. 90. Sapolsky R, Pulsinelli W. Glucocorticoids potentiate ischemic injury to neurons: therapeutic implications. Science. 1985;229:1397-1399. 91. Stein B, Sapolsky R. Chemical adrenalectomy reduces hippocampal damage induced by kainic acid. Brain Res. 1988;473:175-180. 92. Roozendaal B, Phillips RG, Power AE, Brooke SM, Sapolsky RM, McGaugh JL. Memory retrieval impairment induced by hippocampal CA3 lesions is blocked by adrenocortical suppression. Nat Neurosci. 2001;4:1169-1171. 93. Dinkel K, MacPherson A, Sapolsky RM. Novel glucocorticoid effects on acute inflammation in the CNS. J Neurochem. 2003;84:705-716. 94. Dinkel K, Dhabhar FS, Sapolsky RM. Neurotoxic effects of polymorphonuclear granulocytes on hippocampal primary cultures. Proc Natl Acad Sci U S A. 2004;101:331-336. 95. Dirnag U, Simon RP, Hallenbeck JM. Ischemic tolerance and endogenous neuroprotection. Trends Neurosci. 2003;26:248-254. 96. Drossman DA. Irritable bowel syndrome. Gastroenterologist. 1994;2:315-326. 97. Umeda Y, Arisawa M. Inhibition of natural killer activity by calcitonin gene-related peptide. Immunopharmacol Immunotoxicol. 1989;11:309. 98. Sirinek LP, O'Dorisio MS. Modulation of immune function by intestinal neuropeptides. Acta Oncol. 1991;30:509-517. 99. Lombardi VR, Garcia M, Cacabelos R. Microglial activation induced by factor(s) contained in sera from Alzheimer-related ApoE genotypes. J Neurosci Res. 1998;54:539-553.

100. McGillis JP, Humphreys S, Reid S. Characterization of functional calcitonin gene-related peptide receptors on rat lymphocytes. J Immunol. 1991;147:3482-3489. 101. Nong Y, Titus RG, Ribeiro JMC, Remold HG. Peptides encoded by the calcitonin gene inhibit macrophage function. J Immunol. 1989;143:45-49. 102. Foremen J. Substance P and calcitonin gene-related peptide: effects on mast cells in human skin. Int Arch Allergy Appl Immunol. 1987;82:366. 103. Bulloch K. Regional neural regulation of immunity: anatomy and function. In: McEwen BS, ed. Handbook of Physiology. Coping with the Environment: Neural and Endocrine Mechanisms. New York, NY: Oxford University Press; 2000:353-379. 104. Berczi I, Chalmers IM, Nagy E, Warrington RJ. The immune effects of neuropeptides. Baillieres Clin Rheumatol. 1996;10:227-257. 105. Bulloch K, Prasad A, Conrad CD, McEwen BS, Milner TA. Calcitonin gene-related peptide level in the rat dentate gyrus increases after damage. Neuroreport. 1996;7:1036-1040. 106. Bulloch K, Sadamatsu M, Patel A, McEwen BS. Calcitonin gene-related peptide immunoreactivity in the hippocampus and its relationship to cellular changes following exposure to trimethyltin. J Neurosci Res. 1999;55:441-457. 107. Wang FZ, Feng CH, Liu ZP. The role of Ca in changes of membrane function and the protection of CGRP in hippocampal slice during hypoxia. Abstr Soc Neurosci. 1993;16:479.93. 108. Reddington M, Priller J, Treichel J, Haas C, Kreutzberg GW. Astrocyte and microglia as potential targets for calcitonin gene related peptide in the CNS. Can J Physiol Pharm. 1995;73:1047-1049. 109. Rowe JW, Kahn RL. Successful Aging. New York, NY: Pantheon Books; 1998. 110. Singer BH, Ryff CDE. New Horizons in Health. An Integrative Approach. Washington, DC: National Research Council, National Academy Press; 2001. 111. Geronimus AT. The weathering hypothesis and the health of AfricanAmerican women and infants: evidence and speculations. Ethnic Dis. 1992;2:207-221. 112. Sterling P, Eyer J. Allostasis: a new paradigm to explain arousal pathology. In: Fisher S, Reason J, eds. Handbook of Life Stress, Cognition and Health. New York, NY: John Wiley & Sons; 1988:629-649. 113. McEwen BS, Stellar E. Stress and the individual: mechanisms leading to disease. Arch Intern Med. 1993;153:2093-101. 114. McEwen B. Allostasis and allostatic load: implications for neuropsychopharmacology. Neuropsychopharmacology. 2000;22:108-124. 115. Lowy MT, Wittenberg L, Yamamoto BK. Effect of acute stress on hippocampal glutamate levels and spectrin proteolysis in young and aged rats. J Neurochem. 1995;65:268-274. 116. Moghaddam B, Boliano ML, Stein-Behrens B, Sapolsky R. Glucocorticoids mediate the stress-induced extracellular accumulation of glutamate. Brain Res. 1994;655:251-254. 117. Seeman TE, Singer BH, Rowe JW, Horwitz RI, McEwen BS. Price of adaptationallostatic load and its health consequences: MacArthur studies of successful aging. Arch Intern Med. 1997;157:2259-2268. 118. Landfield P. Modulation of brain aging correlates by long-term alterations of adrenal steroids and neurally active peptides. Prog Brain Res. 1987;72:279-300. 119. Sapolsky R. Stress, the Aging Brain and the Mechanisms of Neuron Death. Cambridge, Mass: MIT Press; 1992;1:423. 120. Nishimura E, Billestrup N, Perrin M, Vale W. Identification and characterization of a pituitary corticotropin-releasing factor binding protein by chemical cross-linking. J Biol Chem. 1987;262:12893-12896. 121. Lupien S, Lecours AR, Lussier I, Schwartz G, Nair NPV, Meaney MJ. Basal cortisol levels and cognitive deficits in human aging. J Neurosci. 1994;14:28932903. 122. Seeman TE, McEwen BS, Singer BH, Albert MS, Rowe JW. Increase in urinary cortisol excretion and memory declines: MacArthur studies of successful aging. J Clin Endocrinol Metab. 1997;82:2458-2465. 123. Lupien SJ, de Leon M, de Santi S, et al. Cortisol levels during human aging predict hippocampal atrophy and memory deficits. Nat Neurosci. 1998;1:69-73. 124. Sachar BJ, Hellman J, Fukushima DK, Gallagher TF. Cortisol production in depressive illness. Arch Gen Psychiatry. 1970;23:289-298. 125. Young EA, Haskett RF, Grunhaus L, et al. Increased evening activation of the hypothalamic-pituitary-adrenal axis in depressed patients. Arch Gen Psychiatry. 1994;51:701-707.

131

State of the art

126. Deuschle M, Weber B, Colla M, Depner M, Heuser I. Effects of major depression, aging and gender upon calculated diurnal free plasma cortisol concentrations: a re-evaluation study. Stress. 1998;2:281-287. 127. Carroll B, Martin F, Davies B. Resistance to suppression by dexamethasone of plasma 11-OHCS levels in severe depressive illness. BMJ. 1968;3:285287. 128. Weber B, Lewicka S, Deuschle M, Colla M, Heuser I. Testosterone, androstenedione and dihydrotestosterone concentrations are elevated in female patients with major depression. Psychoneuroendocrinology. 2000;25:765-771. 129. Deuschle M, Blum WF, Strasburger CJ, et al. Insulin-like growth factorI (IGF-I) plasma concentrations are increased in depressed patients. Psychoneuroendocrinology. 1997;22:493-503. 130. Sheline YI, Wang PW, Gado MH, Csernansky JC, Vannier MW. Hippocampal atrophy in recurrent major depression. Proc Natl Acad Sci U S A. 1996;93:3908-3913. 131. Sheline YI, Sanghavi M, Mintun MA, Gado MH. Depression duration but not age predicts hippocampal volume loss in medically healthy women with recurrent major depression. J Neurosci. 1999;19:5034-5043. 132. Bremner JD, Narayan M, Anderson ER, Staib LH, Miller HL, Charney DS. Hippocampal volume reduction in major depression. Am J Psychiatry. 2000;157:115-117. 133. Vakili K, Pillay SS, Lafer B, et al. Hippocampal volume in primary unipolar major depression. A magnetic resonance imaging study. Biol Psychiatry. 2000;47:1087-1090. 134. Rusch BD, Abercrombie HC, Oakes TR, Schaefer SM, Davidson RJ. Hippocampal morphometry in depressed patients and control subjects: relations to anxiety symptoms. Biol Psychiatry. 2001;50:960-964. 135. Sullivan EV, Pfefferbaum A, Swan GE, Carmelli D. Heritability of hippocampal size in elderly twin men: equivalent influence from genes and environment. Hippocampus. 2001;11:754-762. 136. Steffens DC, Byrum CE, McQuoid DR, et al. Hippocampal volume in geriatric depression. Biol Psychiatry. 2000;48:301-309. 137. Kim DH, Payne ME, Levy RM, MacFall JR, Steffens DC. APOE genotype and hippocampal volume change in geriatric depression. Biol Psychiatry. 2002;51:426-429. 138. Lucassen PJ, Muller MB, Holsboer F, et al. Hippocampal apoptosis in major depression is a minor event and absent from subareas at risk for glucorcorticoid overexposure. Am J Pathol. 2001;158:453-468. 139. Muller MB, Lucassen PJ, Yassouridis A, Hoogendijk WJG, Holsboer F, Swaab DF. Neither major depression nor glucocorticoid treatment affects the cellular integrity of the human hippocampus. Eur J Neurosci. 2001;14:1603-1612. 140. MacQueen GM, Campbell S, McEwen BS, et al. Course of illness, hippocampal function, and hippocampal volume in major depression. Proc Natl Acad Sci U S A. 2003;100:1387-1392. 141. Drevets WC, Price JL, Simpson Jr JR, et al. Subgenual prefrontal cortex abnormalities in mood disorders. Nature. 1997;386:824-827. 142. Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. 143. Rajkowska G. Postmortem studies in mood disorders indicate altered numbers of neurons and glial cells. Biol Psychiatry. 2000;48:766-777. 144. Rajkowska G, Halaris A, Selemon LD. Reductions in neuronal and glial density characterize the dorsolateral prefrontal cortex in bipolar disorder. Biol Psychiatry. 2001;49:741-752. 145. Drevets WC, Videen TO, Price JL, Preskorn SH, Carmichael ST, Raichle ME. A functional anatomical study of unipolar depression. J Neurosci. 1992;12:3628-3641. 146. Sheline YI, Barch DM, Donnelly JM, Ollinger JM, Snyder AZ, Mintun MA. Increased amygdala response to masked emotional faces in depressed subjects resolves with antidepressant treatment: an fMRI study. Biol Psychiatry. 2001;50:651-658. 147. Frodl T, Meisenzahl E, Zetzsche T, et al. Enlargement of the amygdala in patients with a first episode of major depression. Biol Psychiatry. 2002;51:708-714. 148. Chattarji S, Vyas A, Mitra R, Rao BSS. Effects of chronic unpredictable and immobilization stress on neuronal plasticity in the rat amygdala and hippocampus. Soc Neurosci Abs. 2000;26:#571.9, p. 1533. 149. LeDoux JE. The Emotional Brain. New York, NY: Simon and Schuster; 1996. 150. Schulkin J, McEwen BS, Gold PW. Allostasis, amygdala, and anticipatory angst. Neurosci Biobehav Rev. 1994;18:385-396. 151. Sheline YI, Gado MH, Price JL. Amygdala core nuclei volumes are decreased in recurrent major depression. Neuroreport. 1998;9:2023-2028. 152. Michelson D, Stratakis C, Hill L, et al. Bone mineral density in women with depression. N Engl J Med. 1996;335:1176-81. 153. Cizza G, Ravn P, Chrousos GP, Gold PW. Depression: a major, unrecognized risk factor for osteoporosis? Trends Endocrinol Metab. 2001;12:198203. 154. Schweiger U, Weber B, Deuschle M, Heuser I. Lumbar bone mineral density in patients with major depression: evidence of increased bone loss at follow-up. Am J Psychiatry. 2000;157:118-120. 155. Thakore JH, Richards PJ, Reznek RH, Martin A, Dinan TG. Increased intra-abdominal fat deposition in patients with major depressive illness as measured by computed tomography. Biol Psychiatry. 1997;41:1140-1142. 156. Mann JN, Thakore JH. Melancholic depression and abdominal fat distribution: a mini-review. Stress. 1999;3:1-15. 157. Weber-Hamann B, Hentschel F, Kniest A, et al. Hypercortisolemic depression is associated with increased intra-abdominal fat. Psychosom Med. 2002;64:274-277. 158. Thayer JF, Smith M, Rossy LA, Sollers JJ, Friedman BH. Heart period variability and depressive symptoms: gender differences. Biol Psychiatry. 1998;44:304-306. 159. Musselman DL, Tomer A, Manatunga AK, et al. Exaggerated platelet reactivity in major depression. Am J Psychiatry. 1996;153:1313-1317. 160. Lederbogen F, Gilles M, Maras A, et al. Increased platelet aggregability in major depression. Psychiatry Res. 2001;102:255-261. 161. Walsh M-T, Dinan TG, Condren RM, Ryan M, Kenny D. Depression is associated with an increase in the expression of the platelet adhesion receptor glycoprotein Ib. Life Sci. 2002;70:3155-3165. 162. Musselman DL, Evans DL, Nemeroff CB. The relationship of depression to cardiovascular disease. Arch Gen Psychiatry. 1998;55:580-592. 163. Heuser I. Depression, endocrinologically a syndrome of premature aging? Maturitas. 2002;41(suppl 1):S19-S23. 164. Ballenger JC, Davidson JRT, Lecrubier Y, et al. Consensus statement on depression, anxiety, and cardiovascular disease. J Clin Psychiatry. 2001;62: 24-27. 165. Perlmutter JB, Frishman WH, Feinstein RE. Major depression as a risk factor for cardiovascular disease. Therapeutic implications. Heart Dis. 2000;2:75-82. 166. Starkman MN, Schteingart DE. Neuropsychiatric manifestations of patients with Cushing's syndrome. Arch Intern Med. 1981;141:215-219. 167. Starkman MN, Schteingart DE, Schork MA. Depressed mood and other psychiatric manifestations of Cushing's syndrome: relationship to hormone levels. Psychosom Med. 1981;43:3-18. 168. Condren RM, Thakore JH. Cushing's disease and melancholia. Stress. 2001;4:91-119. 169. Gold PW, Loriaux DL, Roy A, et al. Responses to corticotropin-releasing hormone in the hypercortisolism of depression and Cushing's disease. N Engl J Med. 1986;314:1329-1335. 170. Starkman MN, Gebarski SS, Berent S, Schteingart DE. Hippocampal formation volume, memory dysfunction, and cortisol levels in patients with Cushing's syndrome. Biol Psychiatry. 1992;32:756-765. 171. Mauri M, Sinforiani E, Bono G, et al. Memory impairment in Cushing's disease. Acta Neurol Scand. 1993;87:52-55. 172. Forget H, Lacroix A, Somma M, Cohen H. Cognitive decline in patients with Cushing's syndrome. J Int Neuropsychol Soc. 2000;6:20-29. 173. Heinz RE, Martinez J, Haenggeli A. Reversibility of cerebral atrophy in anorexia nervosa and Cushing's syndrome. J Comput Assisted Tomogr. 1977;1:415-418. 174. Starkman MN, Giordani B, Gebrski SS, Berent S, Schork MA, Schteingart DE. Decrease in cortisol reverses human hippocampal atrophy following treatment of Cushing's disease. Biol Psychiatry. 1999;46:1595-1602. 175. Bourdeau I, Bard C, Noel B, et al. Loss of brain volume in endogenous Cushing's syndrome and its reversibility after correction of hypercortisolism. J Clin Endocrinol Metab. 2002;87:1949-1954.

132

Structural plasticity of the adult brain - McEwen

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

176. Lederbogen F, Weber B, Colla M, Heuser I, Dreuschle M, Dempfle CE. Antidepressant treatment and global tests of coagulation and fibrinolysis. J Clin Psychiatry. 2001;62:130. 177. Harvey BH, Jonker LP, Brand L, Heenop M, Stein DJ. NMDA receptor involvement in imipramine withdrawal-associated effects on swim stress, GABA levels and NMDA receptor binding in rat hippocampus. Life Sci. 2002;71:43-54. 178. Harvey B, McEwen BS, Stein D. Neurobiology of antidepressant withdrawal: implications for the longitudinal outcome of depression. Biol Psychiatry. 2003;54:1105-1117.

179. Drevets WC. Neuroimaging studies of mood disorders. Biol Psychiatry. 2000;48:813-829. 180. Bremner JD. Does stress damage the brain? Biol Psychiatry. 1999;45:797805. 181. Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science. 2003;301:386-389. 182. Kendler KS, Karkowski-Shuman L. Stressful life events and genetic liability to major depression: genetic control of exposure to the environment? Psychol Med. 1997;27:539-547.

133

Basic research
Structural plasticity of the adult brain
Fred H. Gage, PhD

The adult brain has long been considered stable and unchanging, except for the inevitable decline that occurs with aging. This view is now being challenged with clear evidence that structural changes occur in the brain throughout life, including the generation of new neurons and other brain cells, and connections between and among neurons. What is as remarkable is that the changes that occur in the adult brain are influenced by the behaviors an individual engages in, as well as the environment in which an individual lives, works, and plays. Learning how behavior and environment regulate brain structure and function will lead to strategies to live more effective lives and perhaps protect from, or repair, brain damage and brain disease.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:135-141.

hose of us who study the nervous system believe that the brain is the organ that controls our behavior. Therefore, what we think and what we do, while obviously influenced by the experience, are results of the brains processing of information and directing our subsequent actions. Given this basic assumption, it is no wonder that the most common model or analogy of how the brain operates is that of a computer. While this analogy may have some heuristic value, it is likely wrong or at least very limiting. The brain is an organ, like the liver, heart, and kidney, and is made of chemicals, cells, and tissue. Communication between brain cells is mediated through neurons with long processes (axons) that connect many cells at once and release small batches of chemical information (neurotransmitters) to a network of other neurons. The neurons receive the signals on their antennae, called dendrites, which protrude, in many cases, quite elaborately from the cell body. The specific site where the chemical signal from one cell makes contact with another cell is called a synapse, which is made up of signaling cells (presynaptic boutons) and receiving cells (postsynaptic spines). The synapse is the structural unit that transmits the majority of information between neurons. Each neuron can have thousands of these synapses on its dendrites and cell body. The real trick for the neuron is to calculate (interpret) the temporal and spatially transmitted information it receives and to send that interpreted message onto the next neurons in a circuit. The aggregation of this information passing and processing results in thought and behavior.

Keywords: neurogenesis; adult stem cell; brain structure; neurological disease; depression Author affiliations: Laboratory of Genetics, The Salk Institute, La Jolla, Calif, USA Address for correspondence: Fred H. Gage, PhD, Laboratory of Genetics, The Salk Institute, 10010 North Torrey Pines Road, La Jolla, CA 92037, USA (e-mail: gage@salk.edu) Copyright 2004 LLS SAS. All rights reserved

Adult neural stability

One of the main reasons for viewing the brain as a stable machine or computer is because this analogy helps explain how we can remember from one instant to the next. If the underlying structure was changing all the
www.dialogues-cns.org

135

Basic research
time, how could we do that? For that matter, if the brain is the seat of consciousness, as proposed by Francis Crick,1 how would we maintain a self identity if the brain were not stable? Well, the dirty little secret is coming out: the brain is not stable and that is a good thing. The structural changes seen in the brain may be required to provide the extra capacity we need for dealing with complexity. It may also provide the underpinning for the adaptability and flexibility, or plasticity as neuroscientists refer to it, that is required for dealing with the variety of challenges that we face throughout life. In addition, and in some ways even more importantly, structural plasticity provides the mechanism for the brain to repair itself. All organs of the body have some capacity to repair themselves following minor injury. Skin, liver, heart, kidney, lung, and blood have some level of repair capacity, and most have the capacity to generate new cells to replace damaged ones, at least to a small extent. Until recently, the brain was considered unique in its lack of ability to repair itself once it had matured to adulthood. Researchers were convinced that Once development was ended, the fonts of growth and regeneration of the axons and dendrites dried up irrevocably. In the adult center the nerve paths are something fixed and immutable, nothing may be regenerated (S. Ramon y Cajal, 1928).2 This dogma even influenced clinical research and the accepted methods for treating brain damage. In general, the therapeutic strategy clinicians would suggest could be summed up as try not to damage your brain, because there is no way to fix it. The dominant strategy for repairing a broken, injured, or damaged brain was to replace the lost neurotransmitters (for example, providing L-dopa for Parkinsons disease [PD], which works pretty well for a while) or, more experimentally, to replace the missing or dead neurons (as in neural transplantation for treating PD, Huntingtons disease [HD],Alzheimers disease, amyotrophic lateral sclerosis, or spinal cord injury). The replacement of dead cells by transplantation of externally derived cells continues both experimentally and clinically and, with the new hope provided by the availability (albeit limited) of the pluripotent human embryonic stem cells, optimism for transplantation therapy has been renewed. The previously accepted dogma of adult neural stability is now being called into question. Pioneering studies by Raisman,3 Bjorklund,4 and Aguayo5 and their colleagues in the 1960s and 1970s revealed that damaged axons could grow under some extraordinary circumstances. These studies have led to a recent stampede of very promising work that could lead to the regeneration of cut or damaged axons due to spinal cord injury.6 A deeper blow to the dogma of adult neural stability has been the recent acceptance of the ability of certain areas of the adult brain to generate new neurons throughout life, known as adult neurogenesis. Early evidence of this ability was generated by Altman and colleagues in the 1960s and 1970s,7 and was beautifully extended to birds by Goldman and Nottebohm in the 1980s,8 and later to nonhuman primates and humans in the 1990s.9 During this same period, it was discovered that adult neurogenesis itself was not stable and predictable, but was, in fact, highly regulated by experience, with stress and aging decreasing neurogenesis and environmental enrichment and exercise increasing neurogenesis.

Stem cells in the adult brain

The surprising observation that neurogenesis continues in the adult nervous system has led to the discovery that there are stem cells in the adult brain that generate the new neurons. A stem cell is an uncommitted cell that, when it divides, can give rise to itself (self-renewal) and can also give rise to any or all of the three main cell lineages of the brain: neurons, astrocytes, and oligodendrocytes. Using a variety of methods, it is now possible to isolate these stem cells from the adult brain and use specific growth factors, like fibroblast growth factor (FGF) and epidermal growth factor (EGF), to induce them to divide indefinitely in culture dishes in the laboratory. Most of the studies that have determined that the cells from the brain are stem cells have done so by studying the cells in vitro; the demonstration of stemness in vivo in the adult brain is difficult. However, the numbers of adult stem cells can be greatly expanded and they can be genetically marked in culture and then transplanted back to the adult nervous system.10 In these studies, the cells survived well and differentiated or matured into authentic neurons in the two areas of the brain where neurogenesis normally occurs, the hippocampus and the olfactory bulb. However, the adult stem cells did not readily differentiate into neurons in any other areas. Interestingly, they did differentiate into astrocytes and oligodendrocytes in other areas. This behavior of adult stem cells that were expanded in culture and transplanted back to the adult brain contrasts with the behavior of fresh tissue derived from the fetal brain that has not been extensively expanded in

136

Structural plasticity of the adult brain - Gage

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

culture. Freshly dissociated cells from the fetal brain, if taken at the appropriate time and from the appropriate location, survive and differentiate quite readily into the types of neurons and glial cells from which they were obtained. In fact, the fetal cells have already matured somewhat and have committed themselves to a particular neuronal type; given minimal local environmental signals, they proceed toward their predetermined fates. These properties of fetal tissue make it more amenable to therapeutic applications. For example, in experimental treatments for PD, committed dopamine cells are being taken from fetal substantia nigra for transplantation; in HD treatment, fetal cells are being taken from fetal basal ganglia and transplanted into patients. The irony then is that fetal tissue grafts are more mature than adult stem cells that have been isolated and expanded in culture. The problem with the adult brain is that, outside of the limited number of stem cells, the adult cells are too mature and will not withstand the isolation and transplantation procedures; they have lost the youthfulness to survive and integrate into the adult brain. Part of the problem with fetal tissue is that there are so few cells available that are at just the right age and in just the right location, which means that either many fetuses must be used for each transplantation or the cells must be put in culture to expand their number. However, once placed in culture, only the primitive fetal stem cells will divide extensively, and, as was seen with adult stem cells, these fetal stem cells are so immature that, unless the adult brain has all the necessary signals to direct them to a particular neural type, ie, a hippocampal neuron, then the cells will either die or become glial cells or merely persist as stem cells. The way to make both fetal and adult stem cells more useful for therapeutic transplantation applications is to determine what the signals are in development that induce the stem cells to become a particular neuronal type, and then induce the stem cells toward that lineage in a culture dish just far enough so that, once they are subsequently transplanted to a particular part of the brain, they will continue toward that cell type and eventually integrate and replace the missing function. At this juncture of stem cell biology and adult neurogenesis, the concept of neural self-repair emerged. The question was posed: if the adult brain has pockets of stem cells that can become neurons, astroglial cells (which play a crucial role in generating and maintaining the health of neurons), and oligodendrocytes (a third type of cell in the

brain that insulates the neuronal axons so that they can transmit their information efficiently), then why cant the brain repair itself after injury or disease? The answer seemed to be that the brain is capable of repairing itself and that it already does, to a limited extent. The current strategy is, therefore, to try to understand how, and perhaps to what end, adult neurogenesis normally occurs, in order to find ways whereby we can enhance it, direct it, and more generally harness the residual elements of neural plasticity that are inherent to neural self-repair as a treatment for brain disorders. Surprisingly, we may not be too far away from this goal. Lets first summarize what we know about the process of adult neurogenesis.

What is adult neurogenesis/cell genesis?

As it turns out, the birth of new brain cells or neurogenesis is not an all-or-nothing event.The multipotent stem cell divides periodically in the brain, giving rise to another stem cell (self-renewal) and some progeny that may grow up to be working cells, but the fate is not guaranteed. The progeny must move away from the influence of the mother stem cell into an area that is permissive for maturation. On average, about 50% of these newborn cells never make it and instead die and disappear. Those that do survive may become a neuron or glial cell, depending on where they end up and what type of activity is going on in that brain area at that time. Even so, it takes over a month from the time the new cell is born until it is functionally integrated in the brain, receiving and sending information.Thus, neurogenesis is a process, not an event, and one thatas I said earlier and will emphasize repeatedlyis highly regulated. The factors that regulate neurogenesis are being intensely investigated and new factors that modulate different components of neurogenesis are being discovered on a regular basis. For example, factors known to be important in development of the nervous system, like Sonic hedgehog11 (which was first discovered in fly brain and called hedgehog), have been shown to regulate the proliferation; BMPs (bone morphogenetic proteins) and Notch12 (which were also first discovered in fly brain) appear to be regulators of whether the newborn cells decide to become glia; and molecules associated with the glial cells that surround the stem cells instruct the newborn cells to become neurons. Once the cells are committed to becoming a neuron or glial cell, other growth factors like brain-derived neurotrophic factor (BDNF)13 and insulin-like growth factor (IGF)14 play important roles in keeping the cells alive and encouraging

137

Basic research
the young cells to mature and become functional. It is the understanding of how these growth factors and cellular environments control neurogenesis in the normal setting that will lead to development of therapies aimed at enhancing and directing neurogenesis in disease states. adaptability to the processing of new information. Since it takes a month from the time the new cells are born until they are integrated into the functional circuits of the brain, the role that the new neurons play in behavior has likely less to do with birth of the cells and more to do with the properties of the newly born functioning neuron.20 Thus, future studies are focusing in part on determining whether the spines and synapses of the newly born neurons have properties that give them advantages over neurons that have been in the circuit for the whole life of an animal. One of the most striking aspects of neurogenesis in the hippocampus is the number of events, experiences, and factors that can regulate either the rate of cell division, the survival of the newly born neurons, or their integration into the neural circuitry. First and foremost, there is a clear genetic underpinning to neurogenesis, with a correlation in mice showing that those strains of mice with higher rates of neurogenesis learn more quickly.21,22 However, as with most things, it is not nature or nurture, but more correctly an interaction or cooperation between the two. For example, movement of adult and even old mice from a rather sterile simplified cage into a large enriched environment with significant complexity and diversity will result in a significant increase in new neurons by decreasing the number of cells that die.This increase in new neurons correlates with increased functioning of the hippocampus, as well as a significant improvement in learning and memory. In an attempt in my laboratory to tease out the elements of the enriched environment that are critical for the increased neurogenesis, van Praag discovered that running on a running wheel alone was sufficient to nearly double the number of dividing cells, resulting in robust increases in new neurons.23,24 In addition to the positive effects of exercise and environmental enrichment, the process of neurogenesis is also negatively regulated by events in the environment, such as stress, injury, and disease. Understanding how neurogenesis is normally regulated will be the key to developing strategies to counteract the misregulations of neurogenesis.

Where does adult neurogenesis/ cell genesis occur?

Neurogenesis, the process of generating new neurons, does not occur spontaneously in every part of the brain. In fact, it only occurs robustly in two areas of the brain, while cell division or cell genesis appears, surprisingly, to occur everywhere in the brain and spinal cord.15,16 In most areas of the brain, cell genesis results in the birth of new glial cells that are likely participating in the microrepair process. Reports that new neurons are born outside of the two well-documented areas of neurogenesis, eg, the frontal cortex, have not been substantiated.17 It is most likely that the complexity of the methods used to prove neurogenesis have led to these anomalous observations, though with new and more sensitive methods, low levels of neurogenesis may be detected in more regions of the adult brain and spinal cord. Certainly, as we learn more about the molecular mechanism that controls neurogenesis, as well as the environmental stimuli that regulate neurogenesis, we anticipate that we will be able to direct neurogenesis anywhere in the brain.10 The most robust cell proliferation occurs in the ventricles of the forebrain, where large numbers of cells migrate forward to the olfactory bulb, a brain structure involved in smell, where the cells differentiate into a variety of different kinds of neurons.We are just now learning about how the olfactory bulb functions normally, and do not have a clear picture as to what role these new cells may play in the function of this brain structure.18,19 The second brain areaand the only structure where neurogenesis has been confirmed in all adult mammals from mice to manis the hippocampus, or more precisely the dentate gyrus of the hippocampus.19 The stem cells of the hippocampus reside in the interior of the densely packed granule cells. Once the stem cells divide and progeny are born, they migrate into the densely packed area and over the next month either die or survive and contribute to the function of the critical brain area. The hippocampus is critical to the formation of new memories, and thus any theory for the functional significance of neurogenesis will likely interpret the value of new neurons in terms of providing flexibility and

How does the process of neurogenesis respond in the damaged, injured, or diseased brain?
In the last 5 years, a striking number of neurological diseases and conditions have been shown to affect neurogenesis, especially in the hippocampus. For example, most forms of experimental epilepsy25,26 result in a robust

138

Structural plasticity of the adult brain - Gage

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

increase in the proliferation of stem cells within the hippocampus. Many of these new cells die, but some survive and, as a result of the epileptic state, these new cells migrate to the wrong place in the hippocampus and appear to differentiate incorrectly. These incorrectly generated new neurons have been speculated to play a role in the persistence of certain types of abnormal behavior and pathology that result from the epileptiform activity. By understanding how neurogenesis normally occurs to generate healthy neurons, it is hoped that this aberrant neurogenesis could be blocked or perhaps the aberrantly generated cells could be trained to wire up correctly (even at a later point in time), given the remarkable structural plasticity of these new brain cells. Cerebral stroke also results in a striking increase in the proliferation of new cells in the hippocampus, but most of these cells die soon thereafter. In addition, in certain types of stroke (like ischemia), there is loss of cells in areas of the brain that do not normally give rise to new neurons, and thus offer little hope for repair.27,28 Quite remarkably, more recent studies have revealed that, in fact, the brain is inducing repair by bringing new cells in from areas of the brain that do have stems cells and directing them to the sites of damage. While with severe strokes, this microrepair is not enough to reverse the damage, it is likely that this microrepair system is adequate to protect, prevent, and repair the brain after small, often-unrecognized strokes. Some of this repair is likely to be behind the often-observed remarkable though quite variable recovery that occurs after many strokes. Growth factors like EGF and FGF are now being used to try to enhance the intrinsic repair process, and with encouraging results.29 One of the most striking correlations between disease and neurogenesis is in depression. As mentioned above, stress reduces the process of neurogenesis leading to fewer newborn cells in the dentate gyrus, and chronic stress is believed to be the most important causal factor in depression aside from genetic predisposition.30-32 Antidepressants (tricyclic antidepressants, selective serotonin reuptake inhibitors, tianeptine, and lithium) augment neurogenesis in the dentate gyrus of experimental animals and, interestingly, the time required to observe therapeutic effects of these drugs corresponds to the time course for neurogenesis. This has led to a hypothesis that depression is in part caused by a decrease in neurogenesis in the dentate gyrus and thus antidepressant therapy and physical therapy (ie, running and exercise)

reverse depression by activating neurogenesis in the dentate gyrus. While this is currently only a working hypothesis, there is converging evidence to support this view, which is leading to the examination of other factors that affect adult neurogenesis and the determination of their effects on depression.

Harnessing the endogenous capacity for self-repair that exists in the adult brain
We now know that the brain does indeed have a pool of residual cells that can divide making new cells that can roam around the brain and spinal cord and, under special conditions, differentiate into new functioning cells. We are also beginning to understand some of the cellular and molecular factors, as well as environment events, that regulate the process of neurogenesis. Importantly, there is a consistent correlation between improved function and increases in neurogenesis. This is particularly the case for hippocampus-associated behaviors and functions; moreover, several neural diseases have been associated with changes in neurogenesis. Now the principle strategy is to learn enough about the factors that regulate each of the components of neurogenesis in order to control cell proliferation (making more cells), migration (getting the cells to places where they are needed), and differentiation (turning the cells into the type of cell that is needed). For diseases of the brain and spinal cord, this will require more knowledge about which cells are affected in a disease, as well as knowing more about the factors that regulate the components of neurogenesis: For depression, epilepsy, and stroke, which are diseases that involve the hippocampus (a structure where neurogenesis does occur), the most straightforward strategy would be to induce more neurogenesis or reroute neurogenesis. In diseases like HD and PD, where very specific cell types die to cause the symptoms, the best strategy would be to induce the local dividing cells to proliferate and then differentiate in small spine neurons, in the case of HD, and dopamine neurons, in the case of PD. In diseases like spinal cord injury or multiple sclerosis, the strategy may not be to make endogenous cells become neurons, but rather to ensheath oligodendrocytes. Since the endogenous cells already have the capacity to make these cells at low frequency in the intact spinal cord, the task will be to enhance the endogenous capacity.

139

Basic research
Conclusion
The task aheadto realize the goals of these strategies is not an easy one, but it is the knowledge that this is a realistic and approachable strategy that heralds a remarkable change in how we even think about brain disease, damage, and repair. I imagine a time when selective drugs will be available to stimulate components of neurogenesis, and this treatment will be combined with very specific physical therapy directed at activating specific brain areas to accept and integrate the new cells in that brain area. The implication of this knowledge is that we will be able to conduct our lives in such a way as to limit brain disease and enhance the natural repair process.
I thank Mary Lynn Gage for her valuable assistance with this manuscript.

Plasticidad estructural del cerebro adulto

Por largo tiempo el cerebro adulto se ha considerado estable e inmodificable, excepto por la declinacin inevitable que ocurre con el envejecimiento. Este punto de vista actualmente se cuestiona debido a claras evidencias acerca de cambios estructurales en el cerebro a lo largo de la vida, incluyendo la generacin de nuevas neuronas y otras clulas cerebrales, como tambin de conexiones entre las neuronas. Es destacable el hecho que los cambios que ocurren en el cerebro adulto son influenciados por las conductas que el individuo adopta, como tambin por el ambiente en que vive, trabaja y se desenvuelve. El conocer cmo la conducta y el ambiente regulan la estructura y funcin cerebral conducir a adoptar modos de vida ms efectivos y tal vez ayudar a protegerse de, o a tratar daos y enfermedades cerebrales.

Plasticit structurale du cerveau adulte

Le cerveau adulte a longtemps t considr comme stable et immuable, sauf en ce qui concerne linvitable dclin survenant avec le vieillissement. Ce point de vue est maintenant remis en question par des arguments manifestes en faveur de changements structuraux apparaissant dans le cerveau au cours de la vie, dont la cration de nouveaux neurones et autres cellules crbrales et de connexions parmi les neurones et entre eux. Le fait remarquable est que les changements apparaissant dans le cerveau adulte sont influencs par les comportements quun individu adopte ainsi que par lenvironnement dans lequel celui-ci vit, travaille et agit. Apprendre comment le comportement et lenvironnement rgulent la structure et la fonction crbrales conduira adopter des modes de vie plus efficaces et peut-tre se prmunir contre des lsions et pathologies crbrales, ou les traiter.

REFERENCES
1. Crick F. The Astonishing Hypothesis. New York, NY: Simon & Schuster; 1994. 2. Ramon y Cajal S. Degeneration and Regeneration of the Nervous System. May RM, trans. New York, NY: Hafner; 1928. 3. Raisman G. Neuronal plasticity in the septal nuclei of the adult brain. Brain Res. 1969;14:25-48. 4. Bjorklund A, Katzman R, Stenevi U, West K. Development and growth of axonal sprouts from noradrenaline and 5-hydroxytryptamine neurons in the rat spinal cord. Brain Res. 1971;31:21-33. 5. Richardson PM, McGuiness UM, Aguya AJ. Axons from CNS neurons regenerate into PNS grafts. Nature. 1980;284:264-265. 6. Horner PJ, Gage FH. Regenerating the damaged central nervous system. Nature. 2000;407:963-970. 7. Altman J, Das GD. Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J Comp Neurol. 1965;124:319-335. 8. Goldman S, Nottebohm F. Neuronal production, migration and differentiation in a vocal nucleus of the adult female canary brain. Proc Natl Acad Sci U S A. 1983;80:2390-2394.

9. Eriksson PS, Perfilieva E, Bjork-Eriksson T, et al. Neurogenesis in the adult human hippocampus. Nat Med. 1998;4:1313-1317. 10. Gage FH. Mammalian neural stem cells. Science. 2000;287:1433-1438. 11. Lai K, Kaspar BK, Gage FH, Schaffer DV. Sonic hedgehog regulates adult neural progenitor proliferation in vitro and in vivo. Nat Neurosci. 2003;6: 21-27. 12. Lim DA, Tramontin AD, Trevejo JM, Herrera DG, Garcia-Verdugo JM, Alverez-Buylla A. Noggin antagonizes BMP signaling to create a niche for adult neurogenesis. Neuron. 1997;28:713-726. 13. Pencea V, Bingaman KD, Wiegand SJ, Luskin MB. Infusion of brainderived neurotrophic factor into the lateral ventricle of the adult rat leads to new neurons in the parenchyma of the striatum, septum, thalamus, and hypothalamus. J Neurosci. 2001;21:6706-6717. 14. Aberg MA, Aberg ND, Hedbacker H, Oscarsson J, Eriksson PS. Peripheral infusion of IGF-I selectively induces neurogenesis in the adult rat hippocampus. J Neurosci. 2000;20:2896-2903. 15. Lie DC, Dziewczapolski G, Willhoite AR, Kaspar BK, Shults CW, Gage FH. The adult substantia nigra contains progenitor cells with neurogenic potential. J Neurosci. 2002;22:6639-6649.

140

Structural plasticity of the adult brain - Gage

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

16. Horner PJ, Power AE, Kempermann G, et al. Proliferation and differentiation of progenitor cells throughout the intact adult rat spinal cord. J Neurosci. 2000;20:2218-2228. 17. Kornack DR, Rakic P. Cell proliferation without neurogenesis in adult primate neocortex. Science. 2001;294:2127-2130. 18. Carleton A, Petreanu LT, Lansford R, Alvarez-Buylla A, Lledo PM. Becoming a new neuron in the adult olfactory bulb. Nat Neurosci. 2003;6:507-518. 19. Carlen M, Cassidy RM, Brismar H, Smith GA, Enquist LW, Frisen J. Functional integration of adult-born neurons. Curr Biol. 2002;12:606-608. 20. van Praag H, Schinder AF, Christie BR, Toni N, Palmer TD, Gage FH. Functional neurogenesis in the adult hippocampus. Nature. 2002;415:1030-1034. 21. Kempermann G, Kuhn HG, Gage FH. Genetic influence on neurogenesis in the dentate gyrus of adult mice. Proc Natl Acad Sci U S A. 1997;94:10409-10414. 22. Kempermann G, Brandon EP, Gage FH. Environmental stimulation of 129/SvJ mice causes increased cell proliferation and neurogenesis in the adult dentate gyrus. Curr Biol. 1998;8:939-942. 23. van Praag H, Christie BR, Sejnowski TJ, Gage FH. Running enhances neurogenesis, learning and long-term potentiation in mice. Proc Natl Acad Sci U S A. 1999;96:13427-13431. 24. van Praag H, Kempermann G, Gage FH. Running increases cell proliferation and neurogenesis in the adult mouse dantate gyrus. Nat Neurosci. 1999;2:266-270.

25. Liu J, Solway K, Messing RO, Sharp FR. Increased neurogenesis in the dentate gyrus after transient global ischemia in gerbils. J Neurosci. 1998;18:7768-7778. 26. Parent JM, Yu TW, Leibowitz RT, Geschwind DH, Sloviter RS, Lowenstein DH. Dentate granule cell neurogenesis is increased by seizures and contributes to aberrant network reorganization in the adult rat hippocampus. J Neurosci. 1997;17:3727-3738. 27. Takagi Y, Nozaki K, Takahashi J, Yodoi J, Ishikawa M, Hashimoto N. Proliferation of neuronal precursor cells in the dentate gyrus is accelerated after transient forebrain ischemia in mice. Brain Res. 1999;831:283-287. 28. Yagita Y, Kitagawa K, Ohtsuki T, et al. Neurogenesis by progenitor cells in the ischemic adult rat hippocampus. Stroke. 2001;32:1890-1896. 29. Nakatomi H, Kuriu, T, Okabe S, et al. Regeneration of hippocampal pyramidal neurons after ischemic brain injury by recruitment of endogenous neural progenitors. Cell. 2002;110:429-441. 30. Jacobs BL, van Praag H, Gage FH. Adult brain neurogenesis and psychiatry: a novel theory of depression. Mol Psychiatry. 2000;5:262-269. 31. DSa C, Duman RS. Antidepressants and neuroplasticity. Bipolar Disord. 2002;4:183-194. 32. Santarelli L, Saxe M, Gross C, et al. Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science. 2003;301: 805-809.

141

Basic research
Regulation of cellular plasticity and resilience by mood stabilizers: the role of AMPA receptor trafficking
Jing Du, MD, PhD; Jorge A. Quiroz, MD; Neil A. Gray, BS; Steve T. Szabo, PhD; Carlos A. Zarate Jr, MD; Husseini K. Manji, MD, FRCPC

espite the devastating impact that mood disorders have on the lives of millions worldwide, there is still a dearth of knowledge concerning their underlying etiology and pathophysiology.The brain systems that have heretofore received the greatest attention in neurobiological studies of mood disorders have been the monoaminergic neurotransmitter systems, which are extensively distributed throughout the network of limbic, striatal, and prefrontal cortical neuronal circuits thought to support the behavioral and visceral manifestations of mood disorders.1-3 Thus, clinical studies over the past 40 years have attempted to uncover the specific defects in these neurotransmitter systems in mood disorders by utilizing a variety of biochemical and neuroendocrine strategies. There is increasing evidence from a variety of sources that severe mood disorders are associated with regional reductions in brain volume, as well as reductions in the number, size, and density of glia and neurons in discrete brain areas. Although the precise pathophysiology underlying these morphometric changes remains to be fully elucidated, the data suggest that severe mood disorders are associated with impairments of structural plasticity and cellular resilience. In this context, it is noteworthy that a growing body of data suggests that the glutamatergic system (which is known to play a major role in neuronal plasticity and cellular resilience) may be involved in the pathophysiology and treatment of mood disorders. Glutamate -amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) GluR1 receptor trafficking plays a critical role in regulating various forms of neural plasticity. It is thus noteworthy that recent studies have shown that structurally dissimilar mood stabilizers lithium and valproate regulate GluR1 receptor subunit trafficking and localization at synapses. These studies suggest that regulation of glutamatergically mediated synaptic plasticity may play a role in the treatment of mood disorders, and raises the possibility that agents more directly affecting synaptic GluR1 represent novel therapies for these devastating illnesses.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:143-155.

Keywords: lithium; valproate; antidepressant; bipolar disorder; glutamate receptor GluR1; phosphorylation Author affiliations: Laboratory of Molecular Pathophysiology, National Institute of Mental Health, Bethesda, Md, USA Copyright 2004 LLS SAS. All rights reserved

Address for correspondence: Husseini K. Manji, MD, Laboratory of Molecular Pathophysiology, Mood and Anxiety Disorders Program, National Institute of Mental Health, 9000 Rockville Pike, Building 10, Unit 3 West, Room 3s250, Bethesda, MD 20892, USA (e-mail: manjih@intra.nimh.nih.gov) www.dialogues-cns.org

143

Basic research
Selected abbreviations and acronyms
-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid CAMKII calcium/calmodulin-dependent protein kinase II EAAT excitatory amino acid transporter LTD long-term depression LTP long-term potentiation NMDA N-methyl-D-aspartate MAPK mitogen-activated protein kinase PCP phencyclidine PKA protein kinase A PP1 protein phosphatase 1 WMH white matter hyperintensities While such investigations have been heuristic over the years, they have been of limited value in elucidating the unique biology of mood disorders, which must include an understanding of the underlying basis for the predilection to episodic and often-profound mood disturbance, which can become progressive over time. These observations have led to the appreciation that, while dysfunction within the monoaminergic neurotransmitter systems is likely to play important roles in mediating some components of the pathophysiology of mood disorders, they do not fully explain all the facets of these complex neuropsychiatric disorders.4,5 In addition to the acknowledgement that investigations into the pathophysiology of complex mood disorders have been excessively focused on monoaminergic systems, there has been a growing appreciation that progress in developing truly novel and improved medications has consequently also been limited.A recognition of the clear need for better treatments and the lack of significant advances in our ability to develop novel, improved therapeutics for these devastating illnesses has led to the investigation of the putative roles of intracellular signaling cascades and nonaminergic systems in the pathophysiology and treatment of mood disorders. Consequently, recent evidence demonstrating that impairments of neuroplasticity may underlie the pathophysiology of mood disorders, and that antidepressants and mood stabilizers exert major effects on the signaling pathways that regulate cellular plasticity and resilience, have generated considerable excitement among the clinical neuroscience community, and are reshaping views about the neurobiological underpinnings of these disorders.1,2,6-8 Somewhat surprisingly, the potential role of the glutamatergic system in the pathophysiology and treatment AMPA of bipolar disorder has only recently begun to be investigated in earnest. Glutamate is the major excitatory synaptic neurotransmitter regulating numerous physiological functions in the mammalian central nervous system (CNS), such as synaptic plasticity, learning, and memory, and represents a major neurotransmitter system in the circuitry thought to subserve many of the symptoms of severe, recurrent mood disorders.3 In this perspectives paper, we review the growing body of data that suggests that severe mood disorders are associated with impairments of cellular plasticity and resilience, effects that may arise from perturbations of neurotrophic signaling cascades and the glutamatergic system. We follow with a discussion of the emerging data that suggests regulating the balance of glutamatergic throughput via N-methyl-D-aspartate (NMDA) and -amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors may play an important role in the actions of our most effective thymoleptic agents, and represent very attractive targets for the development of novel therapeutics for these devastating disorders.

What is the evidence for impairments of cellular plasticity and resilience in severe mood disorders?
Structural imaging studies have demonstrated reduced gray matter volumes in areas of the orbital and medial prefrontal cortex (PFC), ventral striatum, and hippocampus, and enlargement of third ventricle in mooddisordered patients relative to healthy control samples.3,9,10 Postmortem neuropathological studies have shown abnormal reductions in glial cell counts/density, neuron size/density, and cortical volume/thickness in the subgenual PFC, orbital cortex, dorsal anterolateral PFC, amygdala, and in basal ganglia and dorsal raphe nuclei and hippocampus.11-16 Morphometric studies also have reported layer-specific reductions in interneurons in the anterior cingulate cortex (ACC), and reductions in nonpyramidal neurons (~40% lower) in CA2 of the hippocampal formation in bipolar disorder subjects compared with controls.17 Overall, the layer-specific cellular changes observed in several distinct brain regions, including the PFC, ACC, and hippocampus suggest that multiple neuronal circuits underlie the neuropathology of mood disorders. This is not altogether surprising since the behavioral and physiological manifestations of the illnesses are complex and include cognitive, affective,

144

Regulation of cellular plasticity and resilience by mood stabilizers - Du et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

motoric, and neurovegetative symptomatology, as well as alterations of circadian rhythms and neuroendocrine systems, and are thus undoubtedly mediated by networks of interconnected neurotransmitter systems and neural circuits.13 In addition to the accumulating neuroimaging evidence, several postmortem brain studies are now providing direct evidence for reductions in regional CNS volume, cell number, and cell body size. Baumann and associates18 reported reduced volumes of the left nucleus accumbens, the right putamen, and bilateral pallidum externum in postmortem brain samples obtained from patients with unipolar depression or bipolar depression. The abnormal presence of white matter hyperintensities (WMH) has been reported in multiple magnetic resonance imaging (MRI) studies of geriatric patients with affective disorder, particularly those with late-onset depression (ie, elderly depressed patients who experience their first depression after age 60). Elderly adults (>60 years old) with severe WMH are 3 to 5 times more likely to have depressive symptoms as compared with persons with only mild or no white matter lesions.19 Tupler and colleagues20 reported that late-onset depressed patients had more severe hyperintensity ratings in deep white matter than early-onset patients and controls, and that late- and early-onset patients had more severe subcortical gray matter hyperintensities (particularly in the putamen) compared with controls. Recently, Silverstone and colleagues21 reported that bipolar patients showed more severe deep WMH on brain MRI than age-matched unipolar and control subjects. WMH severity has been suggested to predict poorer response to antidepressant therapy.22 In fact, these lesions have been also found to be increased in children with psychiatric disorders, but are highest among bipolar patients, when compared with controls, particularly in the frontal lobes,23 and also early in the course of bipolar illness in adolescent subjects.24 Although the cause of WMH in mood disorders is unknown, their presenceparticularly in the brains of young bipolar patientssuggests importance in the pathophysiology of the disorder.25,26 Together, these results support the contention that WMH indicate damage to the structure of brain tissue, and likely disruption of the neuronal connectivity necessary for normal affective functioning. It is not known whether these structural brain changes seen in patients with severe mood disorders constitute developmental abnormalities that may confer vulnerabil-

ity to abnormal mood episodes, compensatory changes to other pathogenic processes, or the sequelae of recurrent affective episodes per se. Understanding these issues will partly depend upon experiments that delineate the onset of such abnormalities within the illness course and determine whether they antedate depressive episodes in individuals at high familial risk for mood disorders. Nevertheless, these prominent atrophic changes and impairments of plasticity have drawn much attention to the glutamatergic system, sinceas we discuss in detail belowthe glutamatergic system is known to play critical roles in regulating various forms of plasticity. Furthermore, as is discussed extensively in this issue and elsewhere,27 alterations in glutamatergic signaling, mediated by both NMDA and non-NMDA receptors, are known to play important roles in stress-induced morphometric brain changes.14,28,29 Since some clinicians may be less familiar with the intricacies of the regulation of glutamate receptor subtypes, we now present a brief overview of the functioning and regulation of NMDA and AMPA glutamatergic receptors. We follow with a discussion of the exciting emerging data suggesting that glutamatergic signaling represents a very attractive target for the development of novel therapeutics for severe mood disorders.

A primer on glutamatergic signaling: critical roles in cellular plasticity and resilience

As the principal mediator of excitatory synaptic transmission in the mammalian brain, glutamate participates in wide-ranging aspects of both normal and abnormal CNS function. Unlike the monoamines, which require transport of amino acids through the bloodbrain barrier, glutamate and aspartate cannot adequately penetrate into the brain from the periphery and are produced locally by specialized brain machinery.30 The metabolic and synthetic enzymes responsible for the formation of these nonessential amino acids are located in glial cells as well as neurons.30,31 The major metabolic pathway in the production of glutamate is derived from glucose and the transamination of -ketoglutarate; however, a small proportion of glutamate is formed directly from glutamine.The latter is actually synthesized in glia, via an active process (requiring adenosine triphosphate [ATP]), and is then transported to neurons where glutaminase is able to convert this precursor to glutamate (Figure 1). Following release, the concentration of glutamate in the extracellular space is highly

145

Basic research

Figure 1. Glutamatergic system. This figure depicts the various regulatory processes involved in glutamatergic neurotransmission, as described in the text. In astrocytes, glutamine can undergo oxidation to yield -ketoglutarate, which can also be transported to neurons and participate in glutamate synthesis. Glutamate is either metabolized or sequestered and stored into secretory vesicles by vesicular glutamate transporters (VGluT). Glutamate can then be released by a calcium-dependent excitotoxic process. Glutamate has its action terminated in the synapse by reuptake mechanisms utilizing distinct GLU transporters (GLUTs), which exist not only on presynaptic nerve terminals, but also on astrocytes; indeed, current data suggests that astrocytic glutamate uptake may be more important for clearing excess glutamate, raising the possibility that astrocytic loss (as has been documented in mood disorders) may contribute to deleterious GLU signaling, but more so by astrocytes. It is now known that there are a number of important intracellular proteins, which are able to alter the function of glutamate receptors. Gly, glycine; GTn, glutamate transporter; GTg, glutamate transporter (glial); 5-HT1A, 5-hydroxytryptamine (serotonin) receptor 1A; AMPA, -amino-3-hydroxy-5-methyl-4-isoxazole propionic acid; NMDAR, NMDA receptor; PKA and PKC, protein kinase A and C; PP1, PP2A, and PP2B, protein phosphatase 1, 2A, and 2B; Yotiao, NMDA receptor accessory protein; AKAP, protein A kinase anchoring protein; nNOS, nitric oxide synthetase; Src, a family of protein tyrosine kinases; PTP1D, protein-tyrosine phosphatase 1D; SHP2, src homology 2 domaincontaining tyrosine phosphatase; PSD95, postsynaptic density protein 95; CAMKII, calcium/calmodulin-dependent protein kinase II; MyoV, myosin V; SynGAP, Ras guanosine triphosphatase (GTPase)activating protein; GKAP, guanylate kinaseassociated protein; PYK2, proline-rich tyrosine kinase-2; Shank; Shank family of multidomain proteins; Homer, a family of dendritic multidomain proteins; Rap2, a small GTPase; H-ras, Harvey rat sarcoma viral oncogene homologue; Rac1, a Rho family GTPase; ERK, extracellular signalregulated kinase; Raf, MEK, and Rsk, ribosomal S6 kinases; Hsp70, heat-shock protein 70.
Modified and reproduced with permission from reference 30: Szabo ST, Gould TD, Manji HK. Neurotransmitters, receptors, signal transduction, and second messengers in psychiatric disorders. In: Schatzberg A, Nemeroff CB, eds. The American Psychiatric Publishing Textbook of Psychopharmacology. Arlington, VA: American Psychiatric Publishing Inc; 2003:3-52. Copyright 2003; American Psychiatric Publishing Inc.

146

Regulation of cellular plasticity and resilience by mood stabilizers - Du et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

regulated and controlled, primarily by a sodium-dependent reuptake mechanism involving several transporter proteins.The major glutamate transporter proteins found in the CNS include excitatory amino acid transporters (EAATs): EAAT1 (or GLAST-1), EAAT2 (or GLT-1), and EAAT3 (or EAAC1), with EAAT2 being the most predominantly expressed form in the forebrain. Additionally, these transporters are differentially expressed in specific cell types, with EAAT1 and EAAT2 being found primarily in glial cells, EAAT3 localized in neurons, and EAAT4 mainly localized in cerebellum. The physiological events regulating the activity of the glutamate transporters are not well understood, though there is evidence that phosphorylation of the transporters by protein kinases may differentially regulate glutamate transporters and therefore glutamate reuptake (discussed in reference 30). Glutamate concentrations have been shown to rise to excitotoxic levels within minutes following traumatic or ischemic injury, and there is evidence that the function of the glutamate transporters becomes impaired under these excitotoxic conditions.32 Moreover, microdialysis studies have shown that severe stress increases extracellular levels of glutamate in hippocampus, and NMDA glutamate receptor antagonists attenuate stress-induced atrophy of CA3 pyramidal neurons. Glutamate receptor subtypes: a focus on NMDA and AMPA receptors The many subtypes of glutamatergic receptors in the CNS can be classified into two major subtypesthe ionotropic and metabotropic receptors (Table I). The ionotropic glutamate receptor ion channels are assemblies of homooligomeric or hetero-oligomeric subunits integrated into the neurons membrane. Every channel is assembled of (most likely) four subunits associated into a dimer of dimers, as has been observed in crystallographic studies.33,34 Every subunit consists of an extracellular amino terminal and ligand-binding domain, three transmembrane domains and a re-entrant pore loop (located between the first and second transmembrane domains), and an intracellular carboxyl terminal domain.35 The subunits associate through interactions between their amino terminal domains forming a dimer that undergoes a second dimerization mediated by interactions between the ligand-binding domains and/or between transmembrane domains.33,34 Three different subgroups of glutamatergic ion channels have been identified utilizing their pharmacological ability to bind dif-

Ionotropic receptors NMDA NR1 NR2A, NR2B, NR2C, NR2D NR3A, NR3B AMPA GluR1, GluR2, GluR3, GluR4 Kainate GluR5, GluR6, GluR7 KA1, KA2 Metabotropic receptors Group I mGlu1a, mGlu1b, mGlu1c, mGlu1d Group II mGlu2, mGlu3 Group III mGlu4a, mGlu4b, mGlu4c, mGlu4d mGlu6 mGlu7a, mGlu7b, mGlu7c, mGlu7d mGlu8a, mGlu8b, mGlu8c, mGlu8d Table I. Receptor subtype units. Once released from the presynaptic terminal, glutamate is able to bind to numerous excitatory amino acid (EAA) receptors, including both ionotropic (eg, N-methyl-Daspartate [NMDA]) and metabotropic receptors. Presynaptic regulation of glutamate release occurs through metabotropic glutamate receptors (mGluR2 or mGluR3), which subserve the function of autoreceptors. However, these receptors are also located on the postsynaptic element. AMPA, -amino-3-hydroxy5-methyl-4-isoxazole propionic acid.

ferent synthetic ligands, each of which is composed of a different set of subunits. These are the NMDA receptor (NMDAR), the AMPA receptor (AMPAR), and the kainate receptor (KAR). The latter two groups are often referred to together as the non-NMDA receptors, but undoubtedly subserve unique functions (Table I). In the adult mammalian brain, NMDA and AMPA glutamatergic receptors are colocalized in approximately 70% of the synapses.36 By contrast, at early stages of development, synapses are more likely to contain only NMDA receptors. Radioligand binding studies have shown that NMDA and AMPA receptors are found in high density in the cerebral cortex, hippocampus, striatum, septum, and amygdala. NMDA receptors The NMDA receptor is activated by glutamate and requires the presence of a coagonist, namely glycine or D-serine, to be activated. However, the binding of both glutamate and glycine is still not sufficient for the NMDA receptor channel to open, since, at resting membrane potential, the NMDA ion channel is blocked by Mg2+ ions. Only when the membrane is depolarized (eg, by the activation of AMPA or kainate receptors on the same postsynaptic neuron) is the Mg2+ blockade relieved. Under these conditions, the NMDA receptor channel will open and permit the entry of both Na+ and Ca2+ (Figure 1).

147

Basic research
The NMDA receptor channel is composed of combination of NR1, NR2A, NR2B, NR2C, NR2D, NR3A, and NR3B subunits (Table I). The binding site for glutamate has been localized to the NR2 subunit and the site for the coagonist glycine has been localized to the NR1 subunit, which is required for receptor function. Two molecules of glutamate and two of glycine are thought to be necessary to activate the ion channel. Within the ion channel, two other sites have been identified called the sigma () site and the phencyclidine (PCP) site. The hallucinogenic drug PCP, ketamine, and the experimental drug dizocilpine (MK-801), all bind at the latter site and are considered noncompetitive receptor antagonists that inhibit NMDA receptor channel function. In preclinical studies, drugs of this type have been shown to have neuroprotective properties against anoxia and hypoglycemia; these studies await clinical confirmation. In clinical psychiatric studies, ketamine has been shown to transiently induce psychotic symptoms in schizophrenic patients, and to produce rapid antidepressant effects in depressed patients.37 These latter observations have led to the investigation of NMDA antagonists as putative novel antidepressants.29,37 NMDA receptors play a critical role in regulating synaptic plasticity.38 The best-studied forms of synaptic plasticity in the CNS are long-term potentiation (LTP) and longterm depression (LTD) of excitatory synaptic transmission. The molecular mechanisms of LTP and LTD have been extensively characterized and have been proposed to represent cellular models of learning and memory.38 Induction of LTP and LTD in the CA1 region of the hippocampus and in many regions of the brain has now clearly been demonstrated to be dependent on NMDA receptor activation. During NMDA-receptordependent synaptic plasticity, Ca2+ influx through NMDA receptors can activate a wide variety of kinases and/or phosphatases that, in turn, modulate synaptic strength. An important recent development is the finding that two of the primary molecules involvedCa2+/ calmodulin-dependent protein kinase II (CAMKII) and the NMDA subtype of glutamate receptorform a tight complex with each other at the synapse.39 Interestingly, this binding appears to enhance both the autophosphorylation of the kinase and the ability of the entire holoenzyme, which has 12 subunits, to become hyperphosphorylated.39 This hyperphosphorylated state has been postulated to represent a memory switch, which can lead to long-term strengthening of the synapse by multiple mechanisms. One important mechanism involves direct phosphorylation of the glutamate-activated AMPA receptors, which increases their conductance. Furthermore, once CAMKII is bound to the NMDA receptor, it may organize additional anchoring sites for AMPA receptors at the synapse. It is intriguing that activation of synaptic NMDA receptor versus nonsynaptic receptor has an opposite effect on cell survival via differential regulation of CREB (cyclic adenosine monophosphate [cAMP]response element binding protein) function. Calcium entry through synaptic NMDA receptors induced CREB activity and brain-derived neurotrophic factor (BDNF) gene expression as strongly as did stimulation of L-type calcium channels. In contrast, calcium entry through nonsynaptic NMDA receptors, triggered by glutamate exposure or hypoxic/ischemic conditions, activated a general and dominant CREB shut-off pathway that blocked induction of BDNF expression. Synaptic NMDA receptors have antiapoptotic activity, whereas stimulation of extrasynaptic NMDA receptors caused loss of mitochondrial membrane potential (an early marker for glutamate-induced neuronal damage) and cell death.40 AMPA receptor trafficking plays critical roles in the regulation of various forms of neural plasticity The AMPA receptor is stimulated by the presence of glutamate and characteristically produces a fast excitatory synaptic signal that is responsible for the initial reaction to glutamate in the synapse. In fact, as discussed above, it is generally believed that it is the activation of the AMPA receptor that results in neuronal depolarization sufficient to liberate the Mg2+ cation from the NMDA receptor, thereby permitting its activation. The AMPA receptor channel is composed of the combination of GluR1, GluR2, GluR3, and GluR4 subunits, and requires two molecules of glutamate to be activated (Table I). AMPA receptors have a lower affinity for glutamate than the NMDA receptor, thereby allowing for more rapid dissociation of glutamate and therefore a rapid deactivation of the AMPA receptor (reviewed in reference 41). Emerging data suggest that AMPA receptor trafficking, including receptor insertion and internalization, and delivery to synaptic sites, provides an elegant mechanism for activity-dependent regulation of synaptic strength.AMPA receptor subunits undergo constitutive endocytosis and exocytosis; however, the process is highly regulated with a

148

Regulation of cellular plasticity and resilience by mood stabilizers - Du et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

variety of signal transduction cascades being capable of producing short- or long-term changes in synaptic surface expression of AMPA receptor subunits. Indeed, although the mechanisms of LTP and LTD have not been completely elucidated, it is widely accepted that AMPA receptor trafficking is the key player in these phenomena. Most importantly for the present discussion, AMPA receptor trafficking is highly regulated by the protein kinase A (PKA), protein kinase C (PKC), CAMKII, and mitogen-activated protein kinase (MAPK) signaling cascades; these are the very same signaling cascades that mood stabilizers and antidepressants exert major effects on.42-45 These observations have led to an extensive series of studies, which have clearly demonstrated that AMPA receptor trafficking is highly regulated by antidepressants and mood stabilizers46,47 (see below). Regulation of AMPA receptor trafficking by signaling cascades Most vesicle trafficking requires the ordered coating of a donor membrane, budding and fusion to form transport vesicles, transport by passive or active delivery along microtubule, and final fusion with the target membrane.48 AMPA receptors adopted this mechanism to be delivered to the neuronal membrane surface. AMPA receptors are multimeric assemblies of the subunits GluR1 to GluR4. Each subunit is composed of N-terminal extracellular domain, membrane-spanning domain, and C-terminal intracellular domain.49,50 AMPA receptor trafficking is subunit-specific and regulated by phosphorylation of its C-terminal domain, and subsequent alteration of protein-protein interactions. PKA pathway The GluR1 subunit appears to govern the trafficking behavior of heteromeric GluR1/GluR2 receptors, preventing constitutive exchange and conferring inducible delivery of the heteromer.51 Phosphorylation of GluR1 at the PKA site p845 facilitates the insertion of GluR1 onto the membrane and synapses, and is often associated with LTP.52 Dephosphorylation of the GluR1 by protein phosphatases (eg, calcineurin and protein phosphatase 1 [PP1]) target GluR1 to recycling endosomes, where rephosphorylation by PKA may occur and the receptors will be reinserted onto the membrane.53 Phosphorylation of GluR1 at PKA site can be enhanced by synapse-asso-

ciated protein 97 (SAP97)/protein A kinase anchoring protein (AKAP79) complex that direct PKA to GluR1 via a PDZ (PSD95, disk large, ZO1) domain interaction.54 CAMKII pathway Numerous studies have demonstrated that CAMKII is required for the proper formation of LTP in slice preparations, and in regulating learning and memory in rodents.55 In response to stimulation, CAMKII translocates to the postsynaptic site, where it has two major effects on AMPA receptor activity at the postsynaptic site during the formation of LTP.55 First, the AMPA single conductance is directly increased by CAMKII at Ser831 of GluR1 subunit.56 Second, CAMKII is required for the delivery of AMPA receptor to the synapse, which is lacking AMPA receptors.51,57,58 This enhancement of synaptic GluR1 level by activation of CAMKII requires an intact C-terminal domain of GluR1, and is possibly involved in interaction with SAP97.59 PP1, which is also known to be a important modulator for learning and memory, can dephosphorylate the phosphorylation of GluR1 at p831 site by CAMKII.60 Extracellular signalregulated kinase (ERK) MAPK pathway A recent study reported that the small guananine triphosphatases (GTPases) Ras and Rap are involved in AMPA receptor trafficking through a postsynaptic signaling mechanism. Ras mediates activity-evoked increase in GluR1/GluR4containing AMPA receptor surface expression at synapses via a pathway that requires p42/44 MAPK activation. In contrast, Rap mediates NMDA-dependent removal of synaptic GluR2/GluR3containing vesicles via a pathway that involves p38 MAPK. The regulation through Ras and Rap, which work as molecular switches, may in turn control the AMPA receptor level at synapses.61 PKC pathways AMPA GluR2 receptors respond to secondary signals by constitutive receptor recycling. Phosphorylation of Ser880 on GluR2 provides a switch from receptor retention at the membrane by binding to ABP (AMPA receptorbinding protein)/GRIP (glutamate receptorinteracting protein), to receptor internalization by binding to PICK 1 (protein interacting with C kinase-1). Therefore,

149

Basic research
phosphorylation of GluR2 at Ser880 by PKC may release the AMPA receptor from the anchoring proteins and initiate the internalization of receptors.62-65 The mechanism for AMPA receptor trafficking is specific for brain region and neuronal type. For example, the endocytosis of AMPA receptors mediating LTD is triggered by very different signaling cascades in different cell types despite the fact that a conserved cell biological mechanism (ie, clathrin/dynamine-dependent endocytosis) always seems to be involved. Specifically, in CA1 pyramidal cells, protein phosphatases seem to be involved in triggering LTD through dephosphorylation of GluR1 and phosphorylation of PKA site on GluR1 is associated with LTP.53 However, in midbrain dopamine cells, activation of PKA appears to trigger LTD and endocytosis of AMPA receptors.66 ing, we found that an agent that provokes mania, namely the antidepressant imipramine, has an opposite effect as it upregulates AMPA synaptic strength in the hippocampus.47,67 Since chronic administration of mood stabilizers bring about numerous biochemical effects, our laboratory8,68 and others69 have established several criteria that findings should meet in order to maximize the likelihood of their therapeutic importance: This effect of mood stabilizers on GluR1 is a common effect of the structurally dissimilar antimanic agents lithium (a monovalent cation) and valproate (which is an 8-carbon branched fatty acid). This attenuation of synaptic GluR1 by lithium and valproate occurs in the hippocampus, a brain region known to be involved in critical affective neuronal circuits. This effect of lithium and valproate on synaptic GluR1 occurs at therapeutic concentrations both in vivo and in vitro. Similar to the clinical therapeutic effects, the changes in GluR1 were observed only after chronic (and not acute) administration. The effects were specific for antimanic agents, as a promanic antidepressant produced opposite effects. While it is impossible to determine whether synaptic GluR1 attenuation occurs in patients being treated with lithium or valproate, our experimental conditions attempt to mimic this situation as closely as possible. Further supporting our data are recent studies that show that AMPA receptor antagonists attenuate several manic-like behaviors produced by amphetamine administration. Thus, AMPA antagonists have been demonstrated to attenuate psychostimulant-induced development or expression of sensitization and hedonic behavior without affecting spontaneous locomotion; additionally, some studies have demonstrated that AMPA receptor antagonists reduce amphetamine- or cocaine-induced hyperactivity.70-75 The need to use caution in the appropriate application of animal models to complex neuropsychiatric disorders has been well articulated, and in fact it is unlikely we will ever develop rodent models that display the full range of symptomatology clinically expressed in man.76,77 However, one current model of mania, which has been extensively used and has reasonable heuristic value in the study of mood disorders, involves the use of psychostimulants in appropriate paradigms. Thus, psychostimulants like amphetamine and cocaine are known to induce manic-like symp-

AMPA receptor trafficking and mood disorders: implication for development of new medications
In view of the critical role of AMPA receptor trafficking in regulating various forms of plasticity, our laboratory has sought to determine if two structurally highly dissimilar antimanic agents, lithium and valproate, exert effects on AMPA receptor trafficking. Lithium, a monovalent cation, and valproic acid (VPA), an 8-carbon fatty acid, are the two most commonly used agents in the treatment of mania. Because lithium and valproate both require several weeks to exert their therapeutic effects, it is widely believed that adaptive changes in intracellular signaling and/or cellular physiology underlie the beneficial effects; interestingly, these two agents have been shown to exert robust effects on the very same signaling pathways known to regulate AMPA receptor trafficking (vide supra). Thus, we investigated whether lithium and valproate regulate synaptic plasticity and AMPA receptor trafficking in the hippocampus, a brain region presumed to be involved in the circuitry of mood disorders.3 We have found that the structurally highly dissimilar antimanic agents lithium and valproate have a common effect on downregulating AMPA GluR1 synaptic expression in the hippocampus after prolonged treatment with therapeutically relevant concentrations as assessed both in vitro and in vivo. In cultured hippocampal neurons, lithium and valproate attenuated surface GluR1 expression after long-term treatment. Further supporting the therapeutic relevance of the find-

150

Regulation of cellular plasticity and resilience by mood stabilizers - Du et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

toms in healthy volunteers, and trigger frank manic episodes in individuals with bipolar disorder.78 Thus, the best-established animal models mania utilize the administration of amphetamine or cocaine to produce hyperactivity, risk-taking behavior, and increased hedonic driveall very important facets of the human clinical

condition of mania. Moreover, these psychostimulantinduced behavioral changes are attenuated by the administration of chronic lithium in a therapeutically relevant time frame. Thus, the fact that AMPA receptor antagonists are capable of attenuating psychostimulantinduced sensitization, hyperactivity, and hedonic behav-

Glia

Glutamine synthetase gln gln glu

mG mG
EAAT1/2/(3) glu EAAT3 Kainate glu Memantine

luR I

II

luR

mGluRI
gln
glutaminase

mGluRI
? Felbamate ?

-Ketoglutarate TCA

glu (-)

glu

glu

NMDAR

mGluRIII

Riluzole

? ?

AMPAR

mGluRII

KAR

Second-generation mGlu agonist

Presynaptic neuron

Postsynaptic neuron

Figure 2. Thymoleptic agents, which exert major effects on the glutamatergic system. The various glutamate receptors and the presumed antiglutamatergic drug sites of action are presented. Memantine is a noncompetitive antagonist at the N-methyl-D-aspartate (NMDA) receptor. Felbamate is a noncompetitive NMDA receptor antagonist (glycine NR1 and glutamate NR2B), an -amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor antagonist, an mGlu group I receptor (mGluRI) antagonist, and a glutamate release inhibitor. Riluzole is a glutamate release inhibitor (acting through blockade of Na+ voltage dependent channels), a -aminobutyric acid GABAA agonist, and probably an AMPA and kainate (KA) antagonist. The sites for second-generation mGlu group II and III receptor agonists are also depicted. NMDAR, NMDA receptor; AMPAR, AMPA receptor; KAR, KA receptor; glu, glutamate; gln, glutamine; mGluRII (or mGluRIII), mGlu group II (or III) receptor; EAAT, excitatory amino acid transporter; TCA, tricarboxylic acid cycle.
Modified and reproduced with permission from reference 28: Zarate CA, Quiroz J, Payne J, Manji HK. Modulators of the glutamatergic system: implications for the development of improved therapeutics in mood disorders. Psychopharmacol Bull. 2002;36:35-83. Copyright 2002. MedWorks Media LLC.

151

Basic research
ior70-75 provides compelling behavioral support for our contention that AMPA receptors play important roles in regulating affective behavior. As mentioned already, in striking contrast to the effects seen with the antimanic agents lithium and valproate, we found that the chronic administration of the antidepressant imipraminewhich is capable of triggering manic episodes in susceptible individuals78increased hippocampal synaptic expression of GluR1.Very recent studies from other laboratories have also demonstrated that chronic administration of antidepressants enhances membrane expression of GluR1 as well as phosphorylation of GluR1 at the PKA site (p845) and the CAMKII/PKC site (p831).79,80 Furthermore, it is noteworthy that AMPA potentiating agents reportedly have efficacy in preclinical models of depression.81 Additionally, chronic exposure to the psychostimulants amphetamine and cocaine caused an increase in GluR1 level in the ventral tegmental area (VTA), and these effects have been postulated to represent a trigger for sensitization to drug abuse.82 An elegant series of studies has recently provided insights into how dopamine receptors, which are activated during psychostimulant administration, might influence glutamatedependent forms of synaptic plasticity, which are being increasingly recognized as important to drug addiction.83 They showed that surface GluR1 labeling on processes of medium spiny neurons and interneurons was increased by brief incubation with a dopamine D1 agonist.83 Although these studies were designed to investigate the role of GluR1 in mediating the effects of drugs of abuse, it is noteworthy that many of the symptoms of mania resemble the effects of psychostimulants (eg, locomotor hyperactivity, racing thoughts, reduced sleep, and psychosis). Taken together, the biochemical and behavioral studies investigating the effects of antimanic (lithium and valproate) and promanic (antidepressants, cocaine, and amphetamine) agents on GluR1 strongly suggest that AMPA receptor trafficking is an important target in the pathogenesis and treatment of certain facets of bipolar disorder. The mechanisms by which glutamate receptors are actively recruited to synapses have long intrigued the neuroscience community; the information reviewed here suggests that they may also play important roles in the pathophysiology and treatment of complex neuropsychiatric disorders.

Concluding remarks
Regionally selective impairments of structural plasticity and cellular resiliency, which have been postulated to contribute to the development of classical neurodegenerative disorders, may also exist in mood disorders. It remains unclear whether these impairments correlate with the magnitude or duration of the biochemical perturbations in mood disorders, reflect an enhanced vulnerability to the deleterious effects of these perturbations (eg, due to genetic factors and/or early life events), or indeed represent the fundamental etiological process in mood disorders. Nevertheless, it is noteworthy that there is growing evidence from preclinical and clinical research that the glutamatergic system is involved in the pathophysiology and treatment of mood disorders. Over the last few years, an impressive amount of information has been gathered regarding the mechanisms underlying the regulation of AMPA receptor localization at synapses. The findings that mood stabilizersin therapeutically meaningful paradigmsregulate AMPA receptors at synapses opens new potential avenues for new drug development in regards to regulating glutamatergic synaptic strength in critical neuronal circuits (Figure 2). The development of new modulators of AMPA receptor signaling for the treatment of mood disorders may lead to improved therapeutics for these devastating disorders.

REFERENCES
1. Manji HK, Drevets WC, Charney DS. The cellular neurobiology of depression. Nat Med. 2001;7:541-547. 2. Nestler EJ, Barrot M, DiLeone RJ, Eisch AJ, Gold SJ, Monteggia LM. Neurobiology of depression. Neuron. 2002;34:13-25. 3. Drevets WC. Neuroimaging and neuropathological studies of depression: implications for the cognitive-emotional features of mood disorders. Curr Opin Neurobiol. 2001;11:240-249. 4. Manji HK, Lenox RH. Signaling: cellular insights into the pathophysiology of bipolar disorder. Biol Psychiatry. 2000;48:518-530.

5. Payne JL, Quiroz JA, Zarate CA Jr, Manji HK. Timing is everything: does the robust upregulation of noradrenergically regulated plasticity genes underlie the rapid antidepressant effects of sleep deprivation? Biol Psychiatry. 2002;52:921-926. 6. D'Sa C, Duman RS. Antidepressants and neuroplasticity. Bipolar Disord. 2002;4:183-194. 7. Young LT, Bakish D, Beaulieu S. The neurobiology of treatment response to antidepressants and mood-stabilizing medications. J Psychiatry Neurosci. 2002;27:260-265. 8. Manji HK, Lenox RH. The nature of bipolar disorder. J Clin Psychiatry. 2000;61(suppl 13):42-57.

152

Regulation of cellular plasticity and resilience by mood stabilizers - Du et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Regulacin de la plasticidad celular y de la resiliencia mediante estabilizadores del nimo: el papel del trfico del receptor AMPA
Diversas fuentes aportan una evidencia creciente acerca de la asociacin entre los trastornos del nimo severos y reducciones regionales del volumen cerebral, como tambin del nmero, tamao y densidad de la gla y de las neuronas en distintas reas cerebrales. Aunque la fisiopatologa especfica que est a la base de estos cambios morfomtricos no est totalmente aclarada, los datos sugieren que los trastornos del nimo severos estn asociados con un deterioro en la plasticidad estructural y en la resiliencia celular. En este contexto es destacable que una informacin creciente sugiere que el sistema glutamatrgico (que se sabe que juega un papel importante en la plasticidad neuronal y en la resiliencia celular) puede estar involucrado en la fisiopatologa y en el tratamiento de los trastornos del nimo. El trfico de la subunidad GluR1 del receptor AMPA (cido-3-hidroxi-5-metil4-isoxazol propinico) juega un papel decisivo en la regulacin de varias formas de plasticidad neural. Es de destacar que estudios recientes han mostrado que estabilizadores del nimo, estructuralmente dismiles, como el litio y el valproato regulan el trfico de la subunidad GluR1 y su localizacin en las sinapsis. Estos estudios sugieren que la regulacin de la plasticidad sinptica mediada por el sistema glutamatrgico puede jugar un papel en el tratamiento de los trastornos del nimo y se aumenta la posibilidad que agentes que afecten ms directamente la subunidad GluR1 sinptica se transformen en nuevas terapias para estas devastadoras enfermedades.

Rgulation de la plasticit et de la rsilience cellulaires par les stabilisateurs de lhumeur : le rle du trafic neuronal du rcepteur AMPA
Il existe de plus en plus darguments provenant de sources diffrentes en faveur de lassociation des troubles de lhumeur svres avec des rductions rgionales du volume crbral, ainsi quavec des rductions en nombre, taille et densit de la glie et des neurones dans des zones crbrales discrtes. Bien que la physiopathologie exacte sous-tendant ces modifications morphomtriques soit incompltement lucide, les donnes suggrent que les troubles de lhumeur svres sont associs des dficits de la plasticit structurale et de la rsilience cellulaire. Dans ce contexte il faut remarquer quun nombre croissant de donnes est en faveur dune intervention du systme glutamatergique (connu pour jouer un rle majeur dans la plasticit neuronale et la rsilience cellulaire) dans la physiopathologie et le traitement des troubles de lhumeur. Le trafic neuronal de la sous-unit GluR1 du rcepteur AMPA (acide glutamate -amino-3-hydroxy-5mthyl-4-isoxazole propionique) et les changements morphologiques qui en rsultent jouent un rle crucial dans la rgulation des diffrentes formes de plasticit neuronale. ce titre il faut galement noter que de rcentes tudes ont montr que des stabilisateurs de lhumeur structurellement diffrents comme le lithium et le valproate rgulent le trafic neuronal des sous-units GluR1 et leur localisation dans les synapses. Ces tudes suggrent que la rgulation de la plasticit synaptique mdie par le glutamate peut jouer un rle dans le traitement des troubles de lhumeur et voquent la possibilit pour ces maladies invalidantes dun nouveau traitement par lintermdiaire de molcules affectant plus directement le GluR1 synaptique.
14. Manji HK, Quiroz JA, Sporn J, et al. Enhancing neuronal plasticity and cellular resilience to develop novel, improved therapeutics for difficult-totreat depression. Biol Psychiatry. 2003;53:707-742. 15. Manji HK, Duman RS. Impairments of neuroplasticity and cellular resilience in severe mood disorders: implications for the development of novel therapeutics. Psychopharmacol Bull. 2001;35:5-49. 16. Cotter D, Mackay D, Landau S, Kerwin R, Everall I. Reduced glial cell density and neuronal size in the anterior cingulate cortex in major depressive disorder. Arch Gen Psychiatry. 2001;58:545-553. 17. Benes FM, Kwok EW, Vincent SL, Todtenkopf MS. A reduction of nonpyramidal cells in sector CA2 of schizophrenics and manic depressives. Biol Psychiatry. 1998;44:88-97.

9. Strakowski SM, Adler CM, DelBello MP. Volumetric MRI studies of mood disorders: do they distinguish unipolar and bipolar disorder? Bipolar Disord. 2002;4:80-88. 10. Beyer JL, Krishnan KR. Volumetric brain imaging findings in mood disorders. Bipolar Disord. 2002;4:89-104. 11. Rajkowska G. Depression: what we can learn from postmortem studies. Neuroscientist. 2003;9:273-284. 12. Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. 13. Rajkowska G. Postmortem studies in mood disorders indicate altered numbers of neurons and glial cells. Biol Psychiatry. 2000;48:766-777.

153

Basic research
18. Baumann B, Danos P, Krell D, et al. Reduced volume of limbic systemaffiliated basal ganglia in mood disorders: preliminary data from a postmortem study. J Neuropsychiatry Clin Neurosci. 1999;11:71-78. 19. de Groot JC, de Leeuw FE, Oudkerk M, Hofman A, Jolles J, Breteler MM. Cerebral white matter lesions and depressive symptoms in elderly adults. Arch Gen Psychiatry. 2000;57:1071-1076. 20. Tupler LA, Krishnan KR, McDonald WM, Dombeck CB, D'Souza S, Steffens DC. Anatomic location and laterality of MRI signal hyperintensities in late-life depression. J Psychosom Res. 2002;53:665-676. 21. Silverstone T, McPherson H, Li Q, Doyle T. Deep white matter hyperintensities in patients with bipolar depression, unipolar depression and agematched control subjects. Bipolar Disord. 2003;5:53-57. 22. Hickie I, Scott E, Mitchell P, Wilhelm K, Austin MP, Bennett B. Subcortical hyperintensities on magnetic resonance imaging: clinical correlates and prognostic significance in patients with severe depression. Biol Psychiatry. 1995;37:151-160. 23. Lyoo IK, Lee HK, Jung JH, Noam GG, Renshaw PF. White matter hyperintensities on magnetic resonance imaging of the brain in children with psychiatric disorders. Compr Psychiatry. 2002;43:361-368. 24. Pillai JJ, Friedman L, Stuve TA, et al. Increased presence of white matter hyperintensities in adolescent patients with bipolar disorder. Psychiatry Res. 2002;114:51-56. 25. Lenox RH, Gould TD, Manji HK. Endophenotypes in bipolar disorder. Am J Med Genet. 2002;114:391-406. 26. Stoll AL, Renshaw PF, Yurgelun-Todd DA, Cohen BM. Neuroimaging in bipolar disorder: what have we learned? Biol Psychiatry. 2000;48:505-517. 27. McEwen BS, Magarinos AM. Stress and hippocampal plasticity: implications for the pathophysiology of affective disorders. Hum Psychopharmacol. 2001;16:S7-S19. 28. Zarate CA, Quiroz J, Payne J, Manji HK. Modulators of the glutamatergic system: implications for the development of improved therapeutics in mood disorders. Psychopharmacol Bull. 2002;36:35-83. 29. Zarate CA Jr, Du J, Quiroz J, et al. Regulation of cellular plasticity cascades in the pathophysiology and treatment of mood disorders: role of the glutamatergic system. Ann N Y Acad Sci. 2003;1003:273-291. 30. Szabo ST, Gould TD, Manji HK. Neurotransmitters, receptors, signal transduction, and second messengers in psychiatric disorders. In: Schatzberg A, Nemeroff CB, eds. The American Psychiatric Publishing Textbook of Psychopharmacology. Arlington, VA: American Psychiatric Publishing Inc; 2003:3-52. 31. Squire LR, Bloom FE, McConnell SK, Roberts JL, Spitzer NC, Zigmond MJ. Fundamental Neuroscience. New York, NY: Academic Press; 2003. 32. Faden AI, Demediuk P, Panter SS, Vink R. The role of excitatory amino acids and NMDA receptors in traumatic brain injury. Science. 1989;244:798800. 33. Ayalon G, Stern-Bach Y. Functional assembly of AMPA and kainate receptors is mediated by several discrete protein-protein interactions. Neuron. 2001;31:103-113. 34. Madden DR. The structure and function of glutamate receptor ion channels. Nat Rev Neurosci. 2002;3:91-101. 35. Hollmann M, Maron C, Heinemann S. N-Glycosylation site tagging suggests a three transmembrane domain topology for the glutamate receptor GluR1. Neuron. 1994;13:1331-1343. 36. Bekkers JM, Stevens CF. NMDA and non-NMDA receptors are co-localized at individual excitatory synapses in cultured rat hippocampus. Nature. 1989;341:230-233. 37. Krystal JH, Sanacora G, Blumberg H, et al. Glutamate and GABA systems as targets for novel antidepressant and mood-stabilizing treatments. Mol Psychiatry. 2002;7(suppl 1):S71-S80. 38. Malenka RC, Nicoll RA. Long-term potentiationa decade of progress? Science. 1999;285:1870-1874. 39. Lisman JE, McIntyre CC. Synaptic plasticity: a molecular memory switch. Curr Biol. 2001;11:R788-R791. 40. Hardingham GE, Fukunaga Y, Bading H. Extrasynaptic NMDARs oppose synaptic NMDARs by triggering CREB shut-off and cell death pathways. Nat Neurosci. 2002;5:405-414. 41. Dingledine R, Borges K, Bowie D, Traynelis SF. The glutamate receptor ion channels. Pharmacol Rev. 1999;51:7-61. 42. Gould TD, Chen G, Manji HK. In vivo evidence in the brain for lithium inhibition of glycogen synthase kinase-3. Neuropsychopharmacology. 2004;29:32-38. 43. Popoli P, Frank C, Tebano MT, et al. Modulation of glutamate release and excitotoxicity by adenosine A (2A) receptors. Neurology. 2003;61:S69-S71. 44. Gould T, Quiroz J, Singh J, Zarate C, Manji H. Emerging experimental therapeutics for bipolar disorder: novel insights from the molecular and cellular mechanisms of action of mood stabilizers. Mol Psychiatry. In press. 45. Payne JL, Quiroz JA, Gould TG, Zarate CA, Manji HK. Neurobiology of bipolar disorder. In: Charney D, Nestler E, ed. Neurobiology of Mental Illness. In press. 46. Du J, Gray N, Falke C, Yuan P, Szabo S, Manji H. Structurally dissimilar antimanic agents modulate synaptic plasticity by regulating AMPA glutamate receptor subunit GluR1 synaptic expression. Ann N Y Acad Sci. 2003;1003:378-380. 47. Gray N, Du J, Falke C, Yuan P, Manji H. Lithium regulates total and synaptic expression of the AMPA glutamate receptor GluR2 in vitro and in vivo. Ann N Y Acad Sci. 2003;1003:402-404. 48. Antonny B, Schekman R. ER export: public transportation by the COPII coach. Curr Opin Cell Biol. 2001;13:438-443. 49. Bennett JA, Dingledine R. Topology profile for a glutamate receptor: three transmembrane domains and a channel-lining reentrant membrane loop. Neuron. 1995;14:373-384. 50. Wo ZG, Bian ZC, Oswald RE. Asn-265 of frog kainate binding protein is a functional glycosylation site: implications for the transmembrane topology of glutamate receptors. FEBS Lett. 1995;368:230-234. 51. Shi S, Hayashi Y, Esteban JA, Malinow R. Subunit-specific rules governing AMPA receptor trafficking to synapses in hippocampal pyramidal neurons. Cell. 2001;105:331-343. 52. Lee HK, Barbarosie M, Kameyama K, Bear MF, Huganir RL. Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature. 2000;405:955-959. 53. Ehlers MD. Reinsertion or degradation of AMPA receptors determined by activity-dependent endocytic sorting. Neuron. 2000;28:511-525. 54. Colledge M, Dean RA, Scott GK, Langeberg LK, Huganir RL, Scott JD. Targeting of PKA to glutamate receptors through a MAGUK-AKAP complex. Neuron. 2000;27:107-119. 55. Fink CC, Meyer T. Molecular mechanisms of CaMKII activation in neuronal plasticity. Curr Opin Neurobiol. 2002;12:293-299. 56. Derkach V, Barria A, Soderling TR. Ca2+/calmodulin-kinase II enhances channel conductance of -amino-3-hydroxy-5-methyl-4-isoxazolepropionate type glutamate receptors. Proc Natl Acad Sci U S A. 1999;96:3269-3274. 57. Liao D, Scannevin RH, Huganir R. Activation of silent synapses by rapid activity-dependent synaptic recruitment of AMPA receptors. J Neurosci. 2001;21:6008-6017. 58. Shi SH, Hayashi Y, Petralia RS, et al. Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation. Science. 1999;284:1811-1816. 59. Hayashi Y, Shi SH, Esteban JA, Piccini A, Poncer JC, Malinow R. Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domain interaction. Science. 2000;287:2262-2267. 60. Genoux D, Haditsch U, Knobloch M, Michalon A, Storm D, Mansuy IM. Protein phosphatase 1 is a molecular constraint on learning and memory. Nature. 2002;418:970-975. 61. Zhu JJ, Qin Y, Zhao M, Van Aelst L, Malinow R. Ras and Rap control AMPA receptor trafficking during synaptic plasticity. Cell. 2002;110:443-455. 62. Kim CH, Chung HJ, Lee HK, Huganir RL. Interaction of the AMPA receptor subunit GluR2/3 with PDZ domains regulates hippocampal long-term depression. Proc Natl Acad Sci U S A. 2001;98:11725-11730. 63. Matsuda S, Launey T, Mikawa S, Hirai H. Disruption of AMPA receptor GluR2 clusters following long-term depression induction in cerebellar Purkinje neurons. EMBO J. 2000;19:2765-2774. 64. Perez JL, Khatri L, Chang C, Srivastava S, Osten P, Ziff EB. PICK1 targets activated protein kinase Calpha to AMPA receptor clusters in spines of hippocampal neurons and reduces surface levels of the AMPA-type glutamate receptor subunit 2. J Neurosci. 2001;21:5417-5428. 65. Xia J, Chung HJ, Wihler C, Huganir RL, Linden DJ. Cerebellar long-term depression requires PKC-regulated interactions between GluR2/3 and PDZ domaincontaining proteins. Neuron. 2000;28:499-510.

154

Regulation of cellular plasticity and resilience by mood stabilizers - Du et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

66. Gutlerner JL, Penick EC, Snyder EM, Kauer JA. Novel protein kinase Adependent long-term depression of excitatory synapses. Neuron. 2002;36:921-931. 67. Du J, Feng L, Zaitsev E, Je H, Liu X, Lu B. Regulation of TrkB receptor tyrosine kinase and its internalization by neuronal activity and calcium influx. J Cell Biol. 2003;163:385-395. 68. Manji HK, Lenox RH. Protein kinase C signaling in the brain: molecular transduction of mood stabilization in the treatment of manic-depressive illness. Biol Psychiatry. 1999;46:1328-1351. 69. Coyle JT, Duman RS. Finding the intracellular signaling pathways affected by mood disorder treatments. Neuron. 2003;38:157-160. 70. Li Y, Vartanian AJ, White FJ, Xue CJ, Wolf ME. Effects of the AMPA receptor antagonist NBQX on the development and expression of behavioral sensitization to cocaine and amphetamine. Psychopharmacology (Berl). 1997;134:266-276. 71. Mead AN, Stephens DN. AMPA receptors are involved in the expression of amphetamine-induced behavioural sensitisation, but not in the expression of amphetamine-induced conditioned activity in mice. Neuropharmacology. 1998;37:1131-1138. 72. Tzschentke TM. Reassessment of buprenorphine in conditioned place preference: temporal and pharmacological considerations. Psychopharmacology (Berl) 2003;172:58-67. 73. Burns LH, Everitt BJ, Kelley AE, Robbins TW. Glutamate-dopamine interactions in the ventral striatum: role in locomotor activity and responding with conditioned reinforcement. Psychopharmacology (Berl). 1994;115:516-528. 74. Hotsenpiller G, Giorgetti M, Wolf ME. Alterations in behaviour and glutamate transmission following presentation of stimuli previously associated with cocaine exposure. Eur J Neurosci. 2001;14:1843-1855.

75. Backstrom P, Hyytia P. Attenuation of cocaine-seeking behaviour by the AMPA/kainate receptor antagonist CNQX in rats. Psychopharmacology (Berl). 2003;166:69-76. 76. Nestler EJ, Gould E, Manji H, et al. Preclinical models: status of basic research in depression. Biol Psychiatry. 2002;52:503-528. 77. Einat H, Yuan P, Gould TD, et al. The role of the extracellular signal-regulated kinase signaling pathway in mood modulation. J Neurosci. 2003;23:7311-7316. 78. Goodwin FK, Jamison KR. Manic-Depressive Illness. New York, NY: Oxford University Press; 1990. 79. Martinez-Turrillas R, Frechilla D, Del Rio J. Chronic antidepressant treatment increases the membrane expression of AMPA receptors in rat hippocampus. Neuropharmacology. 2002;43:1230-1237. 80. Svenningsson P, Tzavara ET, Witkin JM, Fienberg AA, Nomikos GG, Greengard P. Involvement of striatal and extrastriatal DARPP-32 in biochemical and behavioral effects of fluoxetine (Prozac). Proc Natl Acad Sci U S A. 2002;99:3182-3187. 81. Li X, Tizzano JP, Griffey K, Clay M, Lindstrom T, Skolnick P. Anti-depressant-like actions of an AMPA receptor potentiator (LY392098). Neuropharmacology. 2001;40:1028-1033. 82. Carlezon WA Jr, Nestler EJ. Elevated levels of GluR1 in the midbrain: a trigger for sensitization to drugs of abuse? Trends Neurosci. 2002;25:610-615. 83. Chao SZ, Ariano MA, Peterson DA, Wolf ME. D1 dopamine receptor stimulation increases GluR1 surface expression in nucleus accumbens neurons. J Neurochem. 2002;83:704-712.

155

Pharmacological aspects
Neural plasticity: consequences of stress and actions of antidepressant treatment
Ronald S. Duman, PhD

Neural plasticity is emerging as a fundamental and critical mechanism of neuronal function, which allows the brain to receive information and make the appropriate adaptive responses to subsequent related stimuli. Elucidation of the molecular and cellular mechanisms underlying neural plasticity is a major goal of neuroscience research, and significant advances have been made in recent years. These mechanisms include regulation of signal transduction and gene expression, and also structural alterations of neuronal spines and processes, and even the birth of new neurons in the adult brain. Altered plasticity could thereby contribute to psychiatric and neurological disorders. This article reviews the literature demonstrating altered plasticity in response to stress, and evidence that chronic antidepressant treatment can reverse or block the effects, and even induce neural plasticity-like responses. Continued elucidation of the mechanisms underlying neural plasticity will lead to novel drug targets that could prove to be effective and rapidly acting therapeutic interventions.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:157-169.

eural plasticity is a fundamental process that allows the brain to receive information and form appropriate adaptive responses to the same or similar stimuli. The molecular and cellular adaptations underlying learning and memory are the best-characterized and moststudied examples of neural plasticity. However, many different stimuli can activate neural plasticity processes in different brain structures, including environmental, social, behavioral, and pharmacological stimuli. In fact, it could be argued that neural plasticity is one of the most essential and important processes that the brain performs as it relates to many types of central nervous system functions. Thus, disrupted or abnormal plasticity could lead to maladaptive neuronal responses and abnormal behavior. This could occur in response to genetic abnormalities of the cellular machinery required for plasticity, and abnormal or inappropriate stimuli. For example, exposure to inappropriate or prolonged stress has been reported to alter molecular and cellular markers of neural plasticity, and could contribute to stress-related mood disorders. This review will discuss the literature demonstrating altered neural plasticity in response to stress, and clinical evidence indicating that altered plasticity occurs in depressed patients. The second part of the review will present evidence that antidepressant treatment blocks the effects of stress or produces plasticity-like responses.

General mechanisms of neural plasticity

Neural plasticity encompasses many different types of molecular and cellular responses that occur when cells in the brain are induced to respond to inputs from other
Address for correspondence: Ronald S. Duman, PhD, 34 Park Street, New Haven, CT 06508, USA (e-mail: ronald.duman@yale.edu)

Keywords: signal transduction; gene expression; neurotrophic factor; neurogenesis; neuronal atrophy Author affiliations: Division of Molecular Psychiatry, Departments of Psychiatry and Pharmacology, Yale University School of Medicine, New Haven, CT, USA Copyright 2004 LLS SAS. All rights reserved

157

www.dialogues-cns.org

Pharmacological aspects
Selected abbreviations and acronyms
BDNF cAMP CaRE CREB FGF-2 5-HT LTP NMDA PDE4 PKA SSRI brain-derived neurotrophic factor cyclic adenosine monophosphate cAMP response element cAMP response element binding protein fibroblast growth factor2 5-hydroxytryptamine (serotonin) long-term potentiation N-methyl-D-aspartate phosphodiesterase type IV protein kinase selective serotonin reuptake inhibitor Mechanisms of long-term plasticity: gene expression and protein synthesis The Ca2+/cyclic adenosine monophosphate (cAMP) response element (CaRE) binding protein (CREB) is one of the major transcription factors that mediate the actions of Ca2+, as well as cAMP signaling. CREB has been reported to play a role in both cellular and behavioral models of learning and memory.3 There are a number of gene targets that are influenced by Ca2+, cAMP, and CREB, and the pattern of gene regulation is dependent on the cell type, the length of stimulation, as well as the magnitude of stimulation. Gene targets that have been implicated in learning and memory, and are relevant to the effects of stress and antidepressant treatment, are the neurotrophic factors. Of particular interest is brain-derived neurotrophic factor (BDNF), one of the most abundant neurotrophic factors in the brain.

cells or circulating factors. The systems that have been most extensively studied are cellular and behavioral models of learning and memory, including long-term potentiation (LTP), in slices of brain and rodent models of behavior. The mechanisms identified for learning and memory most likely also subserve plasticity occurring in other regions and for other adaptive functions of the brain. This section will briefly discuss some general mechanisms and concepts of plasticity. Mechanisms of acute neural plasticity: synaptic transmission and protein kinases The effects underlying the rapid responses to neuronal activation are mediated by activation of the excitatory neurotransmitter glutamate and regulation of intracellular signaling cascades (for a review of acute mechanisms underlying LTP, see reference 1). Glutamate causes neuronal depolarization via activation of postsynaptic ionotropic receptors that increase intracellular Na+. This leads to the subsequent activation of N-methyl-D-aspartate (NMDA) receptors and the resulting influx of Ca2+. Ca2+ is a major intracellular signaling molecule that activates a signaling cascade, including activation of Ca2+/ calmodulin-dependent protein kinase. Within minutes to hours, activation of glutamate and Ca2+-dependent pathways can result in structural alterations at the level of dendritic spines. Spines mark the location of glutamate synapses and have been the subject of intensive investigation for understanding synaptic plasticity.2 Changes in the shape and even number of spines can occur very rapidly (minutes to hours) after glutamate stimulation. These alterations are made permanent or long-term when they are stabilized or consolidated, a process that requires gene expression and protein synthesis.

Altered neural plasticity in response to stress

Recent reports have demonstrated altered molecular and cellular responses to stress and have contributed to the hypothesis that altered neural plasticity contributes to stress-related psychiatric illnesses. Some examples of stress responses are discussed in this section. Stress alters learning and memory Stress is known to significantly influence learning and memory, and the effects are dependent on the type, duration, and intensity of the stressor. Emotional arousal can enhance learning and memory via synaptic plasticity of amygdala-dependent pathways, and this is thought to be the basis for intense, long-term memories of traumatic events and posttraumatic stress disorder.4,5 However, stress can also impair subsequent learning and memory and can even lead to amnesia.6 The influence of stress on hippocampal-dependent learning is complex and dependent on the type of learning task. In studies of LTP, a consistent suppression of neural plasticity is observed after exposure to stress or adrenal glucocorticoids.6,7 In one of these studies, the suppression of LTP was observed after exposure to an uncontrollable stressor and correlated with behavioral performance in a learning and memory task. Giving the animals control over the stress (ie, the stress could be terminated) did not lead to reduced LTP or decreased learning and

158

Neural plasticity and depression - Duman

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

memory.8 A role for BDNF in the actions of stress on LTP has also been suggested.9 For additional references and discussion of the effects of stress on learning and memory, see the reviews in references 4 to 7. Stress causes atrophy of hippocampal neurons One of the best-characterized examples of altered structural plasticity in response to stress is the atrophy of hip-

pocampal neurons, which was first described by McEwen and colleagues (Figure 1).10 They found that repeated restraint stress results in atrophy of the dendrites of CA3 pyramidal neurons in the hippocampus, measured as a decrease in the number and length of apical dendrites.11 The reduction in dendritic arborization was found to be dependent on long-term, repeated exposure to restraint stress (3 weeks) and to be reversible when the animals are removed from stress. The atrophy

Hippocampal plasticity

CA1

sc CA3 mf pathway

Granule cell

Dentate gyrus

Stress Glucocorticoid BDNF Antidepressant NE and 5-HT Stress Glucocorticoid BDNF

Antidepressant NE and 5-HT BDNF

BDNF

Atrophy or death

Increased neurogenesis Increased survival and growth Decreased neurogenesis

Increased vulnerability resulting from genetic and environmental factors (eg, hypoxia, hypoglycemia, or viral infections)

Figure 1. Model of hippocampal plasticity showing structural alterations in response to stress: atropy of CA3 pyramidal neurons and decreased neurogenesis of dentate gyrus granule cells. Stress results in powerful effects on the hippocampus, partly because of the high levels of glucocorticoid receptors expressed in this brain region. Stress results in at least two major actions in two different subfields of the hippocampus. Repeated stress causes atrophy or remodeling of CA3 pyramidal neurons, decreasing the number and length of apical dendrites. Administration of glucocorticoids causes a similar effect, and decreased expression of brain-derived neurotrophic factor (BDNF) could contribute to pyramidal cell atrophy. Stress also decreases the proliferation of newborn granule cells in the dentate gyrus, and glucocorticoid administration mimics this effect. Chronic antidepressant administration can reverse the atrophy of CA3 neurons and block the downregulation of neurogenesis in the dentate gyrus. The effects of antidepressant treatment occur via acute regulation of serotonin (5-hydroxytryptamine [5-HT]) and norepinephrine (NE) and the regulation of intracellular signaling and gene expression. mf, mossy fiber; sc, Schaffer colaterals.

159

Pharmacological aspects
of CA3 pyramidal cells appears to result from the elevation of adrenal glucocorticoids that occurs during stress because chronic administration of corticosterone, the active form in rodent, results in a similar decrease in number and length of dendrites.12 The actions of stress and glucocorticoids are blocked by administration of an NMDA receptor antagonist, indicating that this glutamate receptor is required for atrophy of CA3 neurons.10 Atrophy of CA3 pyramidal neurons occurs after 2 to 3 weeks of exposure to restraint stress or more long-term social stress, and has been observed in rodents and tree shrews.11-13 In contrast to the atrophy of hippocampus, recent studies demonstrate that chronic stress causes hypertrophy of neurons in the amygdala.14 This study found chronic immobilization stress increased the dendritic arborization of pyramidal neurons in the basolateral amygdala, but decreased dendrite length and branching in the CA3 pyramidal neurons of the hippocampus. Hypertrophy of the amygdala could underlie increased learning and memory as a result of stressinduced emotional arousal, and may be relevant to the pathophysiology of stress-related disorders, including anxiety, posttraumatic stress, and depression. Increased
Neurogenesis in adult hippocampus

arborization of neurons in the amygdala could thereby enhance emotional states or disrupt normal processing of emotional responses. Stress decreases neurogenesis in the adult hippocampus In addition to regulation of the morphology of neurons in the hippocampus, stress influences the number of newborn neurons or neurogenesis in the adult hippocampus15,16 (Figures 1 and 2). The hippocampus is one of two brain regions where neurogenesis continues to occur in adult organism (the other region is in the subventricular zone). In the hippocampus, neural progenitor cells are found in the subgranular zone, between the granule cell layer and the hilus. These cells give rise to newborn cells that migrate into the granule cell layer and mature into neurons with the morphological and physiological characteristics of adult granule cells.17 Interestingly, the process of neurogenesis is highly regulated by a variety of stimuli and can be considered a form of neural plasticity. For example, enriched environment, exercise, and learning increase neurogenesis, while aging and exposure to drugs of abuse decrease neurogenesis.15,16,18 In addition to these factors, stress also results in a dramatic downregulation of neurogenesis in the hippocampus.10,18 Exposure to just a single stressor is sufficient to significantly decrease neurogenesis in the adult hippocampus. Adult neurogenesis is decreased by different types of stress, including subordination stress,19 predator odor,20 maternal separation,21 and footshock.22 In addition, exposure to inescapable stress in the learned helplessness model of depression decreases adult neurogenesis and this effect correlates with behavioral despair in this model.22 Moreover, the reduction in neurogenesis and the behavioral despair is reversed by antidepressant treatment. Regulation of CREB and decreased expression of BDNF in response to stress Stress results in a wide range of effects that influence many different neurotransmitter and neuropeptide systems, signal transduction pathways, and altered gene expression. The hallmark of the stress response is activation of the hypothalamic-pituitary-adrenal (HPA) axis, which includes increased circulating levels of

GCL SGZ Decreased by: Intruder stress Predator odor Footshock stress Maternal separation Learned helplessness Glucocorticoids Increased by: SSRI NE reuptake inhibitor MAOI ECS Exercise
mfp

Figure 2. Model demonstrating the regulation of adult neurogenesis in the hippocampus. Neural progenitor cells are restricted to the subgranular zone (SGZ) that is located between the granule cell layer (GCL) and hilus. These progenitor cells give rise to newborn neurons that migrate into the granule cell layer and mature into adult neurons. The proliferation and survival of newborn neurons is subject to change and can be considered a form of neural plasticity. Neurogenesis is influenced by a number of different stimuli in either a positive or a negative manner as indicated. SSRI, selective serotonin reuptake inhibitor; NE, noradrenaline; MAOI, monoamine oxidase inhibitor; ECS, electroconvulsive seizures; mfp, mossy fiber pathway.

160

Neural plasticity and depression - Duman

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

adrenal glucocorticoids. The hippocampus contains very high levels of glucocorticoid receptors and is therefore significantly impacted by stress. As mentioned above, studies by McEwen and colleagues have demonstrated that glucocorticoids contribute to the atrophy and decreased neurogenesis of hippocampal neurons resulting from exposure to stress.10 In addition, stress is reported to influence CREB and BDNF in the hippocampus and other brain regions. The transcriptional activity of CREB is regulated by phosphorylation and levels of phospho-CREB are used as an indirect measure of CREB activation and function (Figure 3). The regulation of phospho-CREB is complex and is dependent on the brain region and whether the stress is acute or chronic.23-26 Acute stress increases levels of phospho-CREB in many limbic regions associated with mood disorders and this may represent a normal or appropriate adaptive responsiveness.24 In contrast, chronic stress leads to decreased levels of phosphoCREB in many limbic brain regions, which could lead to decreased plasticity and function.26 Stress has profound effects on the expression of BDNF in the hippocampus. Levels of BDNF expression in hippocampus are dramatically downregulated by both acute and chronic stress, and this effect could contribute to the atrophy and decreased neurogenesis caused by stress (Figure 1).27-29 The role of other factors that could underlie the actions of stress on adult neurogenesis is a subject of interest and could lead to novel targets for drug development.

pocampal volume is also observed in patients with posttraumatic stress disorder (PTSD).32 The reduction in hippocampal volume is directly related to the length of illness.33,34 In addition to hippocampus, atrophy of prefrontal cortex and amygdalabrain regions that control cognition, mood, and anxietyhas also been reported in patients with depression or bipolar disorder.35 Evidence from postmortem studies Atrophy of hippocampus or other brain regions could result from loss of cells (neurons or glia) or decreased size of the cell body or neuronal processes. The most extensive studies have been conducted on prefrontal and cingulate cortex and demonstrate that the neuronal body size and number of glia is decreased in depressed patients.36-38 There is much less known about the hippocampus and additional studies will be required to determine what accounts for the atrophy of hippocampus observed in depressed patients. Postmortem analysis of CREB and BDNF has also provided evidence consistent with a loss of neural plasticity in depression. Levels of CREB are decreased in the cerebral cortex of depressed patients or suicide victims.39,40 Levels of BDNF are also decreased in prefrontal cortex and hippocampus of depressed patients.41 Reduced levels of CREB and BDNF, two molecular markers of neural plasticity, indicate that the ability of limbic brain structures to mount adaptive responses is compromised in depressed patients.

Atrophy of limbic brain structures in depressed patients

Evidence from basic research studies provide strong support for the hypothesis that stress-related illnesses such as depression could include alterations in brain structure and neural plasticity. Indeed, direct evidence to support this hypothesis has been provided by brain imaging and postmortem studies of depressed patients. Evidence from brain imaging studies Magnetic resonance imaging studies have demonstrated that the size of certain brain structures is decreased in mood disorder patients. In particular, these studies demonstrate that the volume of the hippocampus is decreased in patients with depression.30,31 Reduced hip-

Antidepressant treatment increases neural plasticity

In contrast to the effects of stress, antidepressant treatment results in molecular and cellular responses that demonstrate an increase in neural plasticity. Moreover, these studies have paved the way for additional studies that demonstrate that antidepressant treatment results in structural remodeling. In many cases, the effects of antidepressant treatment oppose or reverse the effects of stress. Taken together, these findings provide additional support for the hypothesis that neural plasticity plays a significant role in the treatment, as well as the pathophysiology of mood disorders. The evidence for regulation of neural plasticity at the level of neurogenesis, signal transduction, and gene expression is discussed in the second half of this review.

161

Pharmacological aspects
Antidepressant treatment increases adult neurogenesis
Neurogenesis is increased by chronic antidepressant administration One of the most surprising discoveries of recent times in the field of depression is that antidepressant treatment regulates neurogenesis in the adult hippocampus (Figures 1 and 2). In contrast to the actions of stress, chronic antidepressant treatment increases the number of newborn neurons in the adult hippocampus of rodents or tree shrews.42,43 The upregulation of neurogenesis is dependent on chronic antidepressant treatment, consistent with the time course for the therapeutic action of antidepressants.43 In addition, different classes of antidepressants, including serotonin (5-hydroxytryptamine [5-HT]) and noradrenaline reuptake inhibitors, and electroconvulsive seizures are reported to increase adult neurogenesis.43-45 Antidepressant treatment influences two important aspects of neurogenesis, the rate of cell proliferation (ie, the number of newborn neurons) and the survival of newborn neurons.46 An increase in the number of newborn neurons could contribute to the reversal of hippocampal atrophy observed in depressed patients. Antidepressant treatment blocks the downregulation of neurogenesis caused by stress The influence of antidepressant treatment in the context of stress has also been examined. These studies demonstrate that chronic antidepressant treatment can block or reverse the downregulation of neurogenesis that results from exposure to stress. Several different types of stress have been tested, including blockade of intruder stress,42 maternal separation,47 and learned helplessness.22 In addition, different types of antidepressants have been tested, including an atypical antidepressant, tianeptine,42 a selective serotonin reuptake inhibitor (SSRI),22,47 and a neurokinin-1 receptor antagonist.48 The influence of antidepressant treatment on the atrophy of CA3 pyramidal neurons resulting from chronic exposure to stress has been examined. These studies demonstrate that chronic administration of tianeptine blocks the atrophy of CA3 apical dendrites that is caused by stress.12 Chronic administration of an SSRI antidepressant did not block the atrophy of CA3 neurons in this study. Analysis of dendrite branch number and length is tedious and labor intensive, but additional studies of other antidepressants are necessary to determine the relevance of this effect in the actions of antidepressant treatment. A functional role for neurogenesis in the action of antidepressant treatment A major issue in the field of adult neurogenesis is how to test the function of newborn neurons. A recent study has addressed this question by using a combination of irradiation and mutant mouse approaches.49 This study demonstrates that focused irradiation of hippocampus in the mouse completely blocks neurogenesis and there was a corresponding blockade of the behavioral actions of antidepressant treatment in two behavioral models, novelty suppressed feeding and chronic mild stress. In addition, Santarelli et al49 studied the effects of antidepressants in mice with a null mutation of the 5-HT1A receptor, a subtype that has been implicated in the actions of antidepressant treatment. They found that upregulation of neurogenesis by chronic administration of an SSRI was completely blocked in 5-HT1A null mutant mice, and that the behavioral effects of SSRI treatment were similarly blocked. These results are the first evidence that increased neurogenesis is necessary for an antidepressant response in behavioral models. There are a few limitations to this study. First, although novelty-suppressed feeding is responsive to chronic antidepressant treatmentand this is why it was chosen this paradigm is a better model of anxiety than depression. Second, although the effects of antidepressant treatment were blocked, irradiation and 5-HT1A null mutation alone, in the absence of antidepressant administration, did not produce a depressive phenotype. This is consistent with another report demonstrating that decreased neurogenesis is not correlated with behavior in the learned helplessness model of depression.50 Together these studies indicate that neurogenesis is not required for baseline response. However, it is possible that intact neurons are sufficient to sustain baseline response and that more long-term inhibition of neurogenesis would be required to influence activity.

The cAMP-CREB cascade and depression

Neural plasticity upon antidepressant treatment is likely to involve adaptations of multiple intracellular signaling cascades and even interactions of these pathways. One

162

Neural plasticity and depression - Duman

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

of the pathways that is regulated by antidepressant treatment and has been demonstrated to contribute to the actions of chronic antidepressant responses is the cAMP-CREB cascade, the subject of this section. However, it is likely that other signaling pathways are also regulated byand play a role inthe actions of antidepressants. For reviews covering other signal transduction pathways, see reference 51 and 52.

Antidepressant treatment upregulates the cAMPCREB cascade Several studies have investigated the influence of antidepressant treatment on the cAMP-CREB pathway (Figure 3).53,54 This work demonstrates that chronic antidepressant treatment upregulates the cAMP second-messenger cascade at several different levels. This includes

Antidepressant treatment

5-HT/NE

5-HT/NE

AR 5-HT7

Gs

Adenylyl cyclase

Gi/o

1-AR 5-HT1A

Rolipram

cAMP Ca2+-dependent or MAP kinase cascades

PKA

BDNF/trophic effects: synaptic plasticity, neurogenesis, neuronal survival

P RNA BDNF gene

???

CaRE

Figure 3. Model demonstrating the upregulation of the cyclic adenosine monophosphate (cAMP)cAMP response element binding protein (CREB) cascade and expression of brain-derived neurotrophic factor (BDNF) by antidepressant treatment. Chronic, but not acute, antidepressant treatment upregulates the cAMP-CREB cascade in limbic regions of the brain including the hippocampus and cerebral cortex. This includes increased coupling of stimulatory G protein (Gs) to adenylyl cyclase, increased levels of cAMP-dependent protein kinase (PKA), and increased function and expression of CREB. CREB can also be phosphorylated and activated by other kinases, including Ca2+/calmodulin-dependent kinase and mitogen-activated protein (MAP) kinase. In this way, CREB could serve as a common target for different types of serotonin (5-hydroxytryptamine [5-HT]) and norepinephrine (NE) receptors, including -adrenergic (AR), 5-HT7, 1-adrenergic (1AR), and 5-HT1A receptor subtypes. One downstream target of CREB that has been shown to have antidepressant effects is BDNF. The BDNF promoter has at least one Ca2+/cAMP response element (CaRE) that is regulated by phosphorylation (P) of CREB.

163

Pharmacological aspects
increased coupling of the stimulatory G protein to adenylyl cyclase, increased levels of cAMP-dependent protein kinase (PKA), and increased levels of CREB as well as phospho-CREB.55-57 Upregulation of these components of the cAMP-CREB signaling pathway is dependent on chronic antidepressant treatment, consistent with the time course for the therapeutic action of antidepressants. In addition, upregulation of the cAMP-CREB cascade is observed in response to chronic administration of different classes of antidepressants, indicating that this is a common target of antidepressant treatment. In addition to phosphorylation by PKA, CREB is also phosphorylated by Ca2+-dependent kinases, such as Ca2+/calmodulin-dependent protein kinase, and by mitogen-activated protein kinase pathways (Figure 3). In this way, CREB can serve as a target for multiple signal transduction pathways and neurotransmitter receptors that activate these cascades. Activation of the cAMP-CREB cascade produces an antidepressant response Direct evidence for cAMP-CREB signaling in the action of antidepressant treatment has been tested by pharmacological, viral vector, and mutant mouse approaches. First, drugs that block the breakdown of cAMP produce an antidepressant response in behavioral models of depression.54 The primary target for inhibition of cAMP breakdown is cAMP-specific phosphodiesterase type IV (PDE4), and rolipram was one of the first selective PDE4 inhibitors. In addition, we have found that chronic rolipram administration increases neurogenesis in adult hippocampus.46,58 Second, viral expression of CREB in the hippocampus of rat produces an antidepressant response in the forced swim and learned helplessness models of depression.59 However, further studies demonstrated that the effects of CREB are dependent on the brain region where it is expressed. For example, expression of CREB in the nucleus accumbens produces a prodepressant effect, while expression of a dominant negative mutant of CREB results in an antidepressant response in the forced swim test.60 Transgenic expression of dominant negative CREB in the nucleus accumbens is consistent with this effect.61 The different behavioral effects of CREB can be explained by different target genes in the hippocampus (ie, BDNF) versus the nucleus accumbens (ie, prodynorphin).

Regulation of neurotrophic factors and depression

The regulation of CREB by antidepressant treatment indicates that regulation of gene expression also plays a role in the actions of antidepressants. There have been many gene targets identified for antidepressants,51,52 but BDNF is one that has gained attention and is relevant to neural plasticity responses to antidepressant medications. Studies to identify additional gene targets and gene profiles using gene microarray analysis are currently being conducted. Antidepressant treatment upregulates BDNF Neurotrophic factors were originally identified and studied for their role in development and neuronal survival. However, it is now clear that these factors are expressed in the adult brain, are dynamically regulated by neuronal activity, and are critical for the survival and function of adult neurons. On the basis of these considerations, it is clear why decreased expression of BDNF could have serious consequences for the function of limbic brain structures that control mood and cognition. In contrast, antidepressant treatment results in significant upregulation of BDNF in the hippocampus and cerebral cortex of rodents.28,53,54 Increased expression of BDNF is dependent on chronic treatment, and is observed with different classes of antidepressants, but not other psychotropic drugs. The induction of BDNF would be expected to protect neurons from damage resulting from stress, elevated glucocorticoids, or other types of neuronal insult. BDNF has antidepressant effects in behavioral models of depression The possibility that BDNF contributes to the actions of antidepressant treatment is supported by behavioral studies of recombinant BDNF and transgenic mouse models. Microinfusions of BDNF into the hippocampus produce an antidepressant-like response in the learned helplessness and forced swim models of depression.62 The antidepressant effect of BDNF is observed after a single infusion, compared with repeated administration of a chemical antidepressant, and is relatively long-lasting (up to 10 days after infusion). Transgenic overexpression of a dominant negative mutant of the BDNF receptor, trkB, in the hippocampus and other forebrain

164

Neural plasticity and depression - Duman

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

structures is also reported to block the effect of antidepressant treatment, demonstrating that BDNF signaling is necessary for an antidepressant response.63 Microinfusions of BDNF into the dorsal raphe, a midbrain region where 5-HT cell bodies are localized, also produces an antidepressant response in the learned helplessness model.64 Together, these studies indicate that BDNF could contribute to antidepressant responses in both forebrain and brain stem structures by affecting different populations of neurons. Alternatively, it is possible that microinfusions of BDNF into the hippocampus influence 5-HT neuronal function by acting at presynaptic sites, and could therefore enhance 5-HT signaling as observed after brain stem infusions of BDNF.64 A neurotrophic hypothesis of depression Basic research and clinical studies of BDNF have resulted in a neurotrophic hypothesis of depression and antidepressant action.53,54 This hypothesis is based in part on studies demonstrating that stress decreases BDNF, reduces neurogenesis, and causes atrophy or CA3 pyramidal neurons. Brain imaging and postmortem studies provide additional support, demonstrating atrophy and cell loss of limbic structures, including the hippocampus, prefrontal cortex, and amygdala. In contrast, antidepressant treatment opposes these effects of stress and depression, increasing levels of BDNF, increasing neurogenesis, and reversing or blocking the atrophy and cell loss caused by stress and depression. Additional brain imaging and postmortem studies, as well as basic research approaches will be required to further test this hypothesis. In any case, the studies to date provide compelling evidence that neural plasticity is a critical factor in the pathophysiology and treatment of depression. Antidepressants influence other neurotrophic factor systems Because of the preclinical and clinical evidence implicating neurotrophic factors in the pathophysiology and treatment of depression, studies have been conducted to examine other neurotrophic factor systems. One of the most robust effects identified to date is that antidepressant treatment increases the expression of fibroblast growth factor2 (FGF-2).65 FGF-2 is known to have a potent influence on neurogenesis during development and in the adult brain, and could contribute to antide-

pressant regulation of neurogenesis. Studies are under way to examine the role of FGF-2 in antidepressant regulation of neurogenesis and regulation of behavior in models of depression. Several other growth factors have been identified by microarray analysis and gene expression profiling, including vascular endothelial growth factor, neuritin, and VGF.66 Studies are currently under way to determine the functional significance of these growth factors in models of depression.

Clinical evidence of relevance of neural plasticity to antidepressant treatment

Basic research studies clearly demonstrate that antidepressant treatment regulates signal transduction, gene expression, and the cellular responses that represent neural plasticity. This issue is more difficult to address in clinical studies, but evidence is slowly accumulating. Brain imaging studies have been conducted to examine the influence of antidepressants on the volume of limbic brain regions. One study demonstrates that hippocampal atrophy is inversely proportional to the length of time a patient receives antidepressant medication.67 A longitudinal study of PTSD patients before and after antidepressant treatment has found that there is a partial reversal of hippocampal atrophy in patients receiving medication.68 The latter study demonstrated a corresponding increase in verbal declarative memory in response to antidepressant treatment. Evidence at the molecular level is also provided by postmortem studies. Levels of CREB immunoreactivity are increased in patients receiving antidepressant treatment at the time of death relative to unmedicated patients.39 In addition, levels of BDNF are increased in patients taking an antidepressant at the time of death.59 Although these effects must be replicated and extended (for example, to the regulation of neurogenesis) in additional banks of postmortem tissue, the results are consistent with the hypothesis that neural plasticity is upregulated in patients receiving antidepressant medication.

Novel targets for the treatment of depression

The hypothesis that antidepressant treatment increases neural plasticity provides a number of novel targets for drug development. However, as with any fundamentally important mechanism, care must be taken that the drugs developed for such targets do not interfere with the nor-

165

Pharmacological aspects
mal function of the brain. Nevertheless, regulation of neural plasticity is an exciting area of research for design of new drugs for a variety of indications, including learning, memory, cognition, mood, and neurodegenerative disorders. This section discusses a few of these targets in the context of the pathways regulated by antidepressants and stress. Targets for antidepressant regulation of neurogenesis Identification of the signal transduction and gene expression pathways that are responsible for the actions of antidepressant regulation of neurogenesis is a subject of intense investigation.Activation of the cAMP-CREB signaling cascade using either pharmacological or transgenic approaches is reported to increase both proliferation and survival of newborn neurons in the hippocampus,46,58 supporting the possibility that antidepressants increase neurogenesis via regulation of this intracellular pathway. Gene targets of CREB, as well as other neurotrophic/growth factors that have been shown to regulate adult neurogenesis, include BDNF, FGF-2, and insulin-like growth factor1, to name but a few.18 Because antidepressant treatment increases the expression of both BDNF and FGF-2, these two factors are currently being investigated. This is just a partial listing of the signal transduction cascades and factors that could contribute to antidepressant regulation of adult neurogenesis. Targets for regulation of the cAMP-CREB cascade There are several different sites within the cAMP pathway that could be targeted for drug development. One that has already proven to be effective for antidepressant treatment is blockade of PDE4 and the breakdown of cAMP. Rolipram is a PDE4-selective inhibitor that has been demonstrated to have antidepressant efficacy in early clinical trials and behavioral models of depression.69,70 However, the clinical use of rolipram has been limited by its side effects, primarily nausea. The identification of four different PDE4 isozymes that are equally inhibited by rolipram raises the possibility that one of the isozymes underlies the antidepressant actions of rolipram, while another mediates its side effects. Studies are currently under way to characterize the regional distribution and function of the three PDE4 isozymes expressed in brain (PDE4A, PDE4B, and PDE4D) and the role of these isozymes in the actions of antidepressant treatment.71 Studies of mutant mice demonstrate that null mutation of PDE4D produces an antidepressant-like phenotype indicating a role for this isozyme,72 and similar studies are currently under way for PDE4A and PDE4B. BDNF as a target for drug development The use of BDNF and other neurotrophic factors for the treatment of neurological disorders has been a subject of interest for several years, although problems with delivery, efficacy, and side effects have hampered these efforts. To more directly replicate the in vivo situation, it may be possible to stimulate the expression of endogenous BDNF expression by stimulating signaling pathways known to regulate this neurotrophic factor. First, activation of the cAMP-CREB cascade by inhibition of PDE4 increases the expression of BDNF.56 Small molecular agonists for neurotransmitter receptors have also exhibited some promise. Activation of ionotropic glutamate receptors increases BDNF expression and could be targeted for the treatment of depression.73 One drug that modulates glutamate transmission and increases BDNF expression is memantine.74 Riluzole, a sodium channel blocker, also increases BDNF expression, as well as neurogenesis in adult hippocampus.75 Specific 5-HT and norepinephrine receptor subtypes that activate cAMP (eg, -adrenergic, 5-HT7), Ca2+, or mitogen-activated protein kinase (1-adrenergic, 5-HT1A) pathways could also be targets for development. Characterization of the antidepressant actions of these compounds will be needed, as well as identification of additional neurotransmitter and signal transduction systems that regulate BDNF.

Conclusions
Studies of the molecular and cellular mechanisms underlying neural plasticity responses in learning and memory, as well as fear, anxiety, depression, and drug abuse to name but a few, are some of the most exciting and rapidly advancing areas of research in neuroscience. Progress in our understanding of neural plasticity has profound implications for the treatment of a number of psychiatric and neurodegenerative disorders, and for enhancing performance in what are considered normal subjects. One of the promising aspects of neural plasticity is that it implies that the alterations that occur are

166

Neural plasticity and depression - Duman

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

reversible, even neuronal atrophy and cell loss. Reversibility of structural as well as functional plasticity has already been demonstrated in response to pharmacological treatments or even behavioral therapy. As the fundamental mechanisms of neural plasticity are further elucidated, new targets and paradigms for enhancing

plasticity will be revealed and will lead to more effective and faster-acting therapeutic interventions.
This work is supported by USPHS grants MH45481 and 2 PO1 MH25642, a Veterans Administration National Center Grant for posttraumatic stress disorder, and by the Connecticut Mental Health Center.

Plasticidad neuronal: consecuencias del estrs y efectos del tratamiento antidepresivo

La plasticidad neuronal est resultando un mecanismo fundamental y especfico de la funcin neuronal, lo que permite que el cerebro reciba informacin y ejecute las respuestas de adaptacin apropiadas a los estmulos correspondientes. El esclarecimiento de los mecanismos moleculares y celulares que subyacen a la plasticidad neuronal es uno de los objetivos principales de la investigacin en neurociencias y se han realizado avances significativos en esta rea en los ltimos aos. Estos mecanismos incluyen la regulacin de la transduccin de seales y la expresin gnica, como tambin las alteraciones estructurales de las espinas neuronales y sus procesos e incluso el nacimiento de nuevas neuronas en el cerebro adulto. La alteracin de la plasticidad podra participar en los trastornos psiquitricos y neurolgicos. Este artculo revisa la literatura que demuestra que hay una modificacin en la respuesta de estrs y hay evidencias que el tratamiento antidepresivo crnico puede revertir o bloquear estos efectos, e incluso inducir respuestas neurales semejantes a la plasticidad. El esclarecimiento continuo de los mecanismos que subyacen a la plasticidad neuronal conducir a blancos para nuevos frmacos que podran llegar a constituirse en intervenciones teraputicas efectivas y de rpida accin. REFERENCES
1. Malenka R, Nicoll RA. Long-term potentiationa decade of progress? Science. 1999;285:1870-1874. 2. Lamprecht R, LeDoux J. Structural plasticity and memory. Nat Rev Neurosci. 2004;5:45-54. 3. Silva A, Kogan JH, Frankland PW, Kida S. CREB and memory. Ann Rev Neurosci. 1998;21:127-148. 4. Cahill L, McGaugh JL. Mechanisms of emotional arousal and lasting declarative memory. Trends Neurosci. 1998;21:294-299.

Neuroplasticit : consquence du stress et actions des traitements antidpresseurs

La neuroplasticit se rvle tre un mcanisme fondamental et dterminant de la fonction neuronale, permettant au cerveau de recevoir linformation et dapporter les rponses adaptatives appropries aux stimuli ultrieurs qui sy rattachent. Llucidation des mcanismes molculaires et cellulaires sous-jacents la neuroplasticit est un objectif majeur de la recherche en neurosciences et des progrs significatifs ont t raliss ces dernires annes. Ces mcanismes comprennent la rgulation de la transduction du signal et de lexpression du gne ainsi que les altrations structurales des prolongements et des pines dendritiques des neurones et mme la naissance de nouveaux neurones dans le cerveau adulte. Laltration de ces mcanismes pourrait contribuer aux pathologies psychiatriques et neurologiques. Cet article passe en revue la littrature pour faire le point sur les arguments en faveur de laltration de la plasticit en rponse au stress et de la capacit du traitement antidpresseur au long cours en inverser ou neutraliser les effets voire mme susciter des rponses semblables celles de la neuroplasticit. La poursuite des efforts pour lucider les mcanismes sous-jacents la neuroplasticit permettra de dfinir de nouvelles cibles mdicamenteuses et de dboucher ainsi sur des interventions thrapeutiques efficaces et daction rapide.
5. LeDoux J. Emotion circuits in the brain. Ann Rev Neurosci. 2000;23:155-184. 6. Kim J, Diamond DM. The stressed hippocampus, synaptic plasticity and lost memories. Nat Rev Neurosci. 2002;3:453-462. 7. Pavlides C, Nivon LG, McEwen BS. Effects of chronic stress on hippocampal long-term potentiation. Hippocampus. 2002;12:245-257. 8. Shors T, Seib TB, Levine S, Thompson RF. Inescapable versus escapable shock modulates long-term potentiation in the rat hippocampus. Science. 1989;244:224-226. 9. Zhou J, Zhang F, Zhang Y. Corticosterone inhibits generation of longterm potentiation in rat hippocampus slice: involvement of brain-derived neurotrophic factor. Brain Res. 2000;885:182-191.

167

Pharmacological aspects
10. McEwen B. Stress and hippocampal plasticity. Curr Opin Neurobiol. 1999;5:205-216. 11. Wooley CS, Gould E, McEwen BS. Exposure to excess glucocorticoids alters dendritic morphology of adult hippocampal pyramidal neurons. Brain Res. 1990;531:225-231. 12. Watanabe Y, Gould E, Daniels DC, Cameron H, McEwen BS. Tianeptine attenuates stress-induced morphological changes in the hippocampus. Eur J Pharmacol. 1992;222:157-162. 13. Margarinos A, McEwen BS, Flugge G, Fuchs E. Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J Neurosci. 1996;16:3534-3540. 14. Vyas A, Mitra R, Shankaranarayana Rao BS, Chattarji S. Chronic stress induces contrasting patterns of dendritic remodeling in hippocampal and amygdaloid neurons. J Neurosci. 2002;22:6810-6818. 15. Gage F. Mammalian neural stem cells. Science. 2000;287:1433-1438. 16. Gould E, Beylin A, Tanapat P, Reeves A, Shors TJ. Learning enhances adult neurogenesis in the hippocampal formation. Nat Neurosci. 1999;2:260-265. 17. van Praag H, Schlinder AF, Christie BR, Toni N, Palmer TD, Gage FH. Functional neurogenesis in the adult mouse dentate gyrus. Nature. 2002;415:1030-1034. 18. Duman R, Malberg J, Nakagawa S. Regulation of adult neurogenesis by psychotropic drugs and stress. J Pharmacol Exp Ther. 2001;299:401-407. 19. Gould E, McEwen BS, Tanapat P, Galea LAM, Fuchs E. Neurogenesis in the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA receptor activation. J Neurosci. 1997;17:2492-2498. 20. Tanapat P, Hastings NB, Rydel TA, Galea LAM, Gould E. Exposure to fox odor inhibits cell proliferation in the hippocampus of adult rats via an adrenal hormone-dependent mechanism. J Comp Neurol. 2001;437:496-504. 21. Lee K, Lynch KR, Nguyen T, et al. Cloning and charactization of additional members of the G proteincoupled receptor family. Biochim Biophys Acta. 2000;1490:311-323. 22. Malberg J, Duman RS. Cell proliferation in adult hippocmpus is decreased by inescapable stress: reversal by fluoxetine treatment. Neuropsychopharmacology. 2003;28:1562-1571. 23. Barrot M, Olivier JD, Perrotti LI, et al. CREB activity in the nucleus accumbens shell controls gating of behavioral responses to emotional stimuli. Proc Natl Acad Sci U S A. 2002;99:11435-11440. 24. Bilang-Bleuel A, Rech J, De Carli S, Holsboer F, Reul JMHM. Forced swimming evokes a biphasic response in CREB phosphorylation in extrahypothalamic limbic and neocortical brain structures in the rat. Eur J Neurosci. 2002;15:1048-1060. 25. Bruijnzeel A, Stam R, Compaan JC, Wiegant VM. Stress-induced sensitization of CRH-ir but not P-CREB-ir responsivity in the rat central nervous system. Brain Res. 2001;908:187-196. 26. Trentani A, Kuipers SD, Ter Horst GJ, Den Boer JA. Selective chronic stress-induced in vivo ERK1/2 hyperphosphorylation in medial prefrontocortical dendrites: implications for stress-related cortical pathology? Eur J Neurosci. 2002;15:1681-1691. 27. Duman R. Role of neurotrophic factors in the etiology and treatment of mood disorders. Neuromol Med. 2004;5:11-26. 28. Nibuya M, Morinobu S, Duman RS. Regulation of BDNF and trkB mRNA in rat brain by chronic electroconvulsive seizure and antidepressant drug treatments. J Neurosci. 1995;15:7539-7547. 29. Smith MA, Makino S, Kvetnansky R, Post RM. Stress alters the express of brain-derived neurotrophic factor and neurotrophin-3 mRNAs in the hippocampus. J Neurosci. 1995;15:1768-1777. 30. Bremner J, Narayan M, Anderson ER, Staib LH, Miller H, Charney DS. Smaller hippocampal volume in major depression. Am J Psychiatry. 2000;157:115-117. 31. Sheline Y, Wany P, Gado MH, Csernansky JG, Vannier MW. Hippocampal atrophy in recurrent major depression. Proc Natl Acad Sci U S A. 1996;93:39083913. 32. Bremner JD, Randall P, Scott TM, et al. MRI-based measurement of hippocampal volume in patients with combat-related posttraumatic stress disorder. Am J Psychiatry. 1995;152:973-981. 33. MacQueen G, Campbell S, McEwen BS, et al. Course of illness, hippocampal function, and hippocampal volume in major depression. Proc Natl Acad Sci U S A. 2003;100:1387-1392. 34. Sheline Y, Sanghavi M, Mintun MA, Gado MH. Depression duration but not age predicts hippocampal volume loss in medically healthy wormen with recurrent major depression. J Neurosci. 1999;19:5034-5043. 35. Manji H, Duman RS. Impairments of neuroplasticity and cellular resilience in severe mood disorders: implications for the development of novel therapeutics. Psychopharmacol Bull. 2001;35:5-49. 36. Cotter D, Mackay D, Landau S, Kerwin R, Everall I. Reduced glial cell density and neuronal size in the anterior cingulate cortex in major depressive disorder. Arch Gen Psychiatry. 2001;58:545-553. 37. Ongur D, Drevets WC, Price JL. Glial reduction in the subgenual prefrontal cortex in mood disorders. Proc Natl Acad Sci U S A. 1998;95:13290-13295. 38. Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. 39. Dowlatshahi D, MacQueen GM, Wang JF, Young LT. Increased temporal cortex CREB concentrations and antidepressant treatment in major depression. Lancet. 1998;352:1754-1755. 40. Dwivedi Y, Rizavi HS, Conley RR, Tamminga CA, Pandey GN. Altered gene expression of brain-derived neurotrophic factor and receptor tyrosine kinase B in postmortem brain of suicide subjects. Arch Gen Psychiatry. 2003;60:804-815. 41. Dwivedi Y, Rizavi HS, Roberts RC, Conley RC, Tamminga CA, Pandey GN. Reduced activation and expression of ERK1/2 MAP kinase in the postmortem brain of depressed suicide subjects. J Neurochem. 2001;77:916-928. 42. Czeh B, Michaelis T, Watanabe T, et al. Stress-induced changes in cerebral metabolites, hippocampal volume, and cell proliferation are prevented by antidepressant treatment with tianeptine. Proc Natl Acad Sci U S A. 2001;98:12796-12801. 43. Malberg J, Eisch AJ, Nestler EJ, Duman RS. Chronic antidepressant treatment increases neurogenesis in adult hippocampus. J Neurosci. 2000;20:9104-9110. 44. Madsen T, Treschow A, Bengzon J, Bolwig TG, Lindvall O, Tingstrm A. Increased neurogenesis in a model of electroconvulsive therapy. Biol Psychiatry. 2000;47:1043-1049. 45. Manev H, Uz T, Smalheiser NR, Manev R. Antidepressants alter cell proliferation in the adult brain in vivo and in neural cultures in vitro. Eur J Pharmacol. 2001;411:67-70. 46. Nakagawa S, Kim JE, Lee R, et al. Regulation of neurogenesis in adult mouse hippocampus by cAMP and cAMP response element-binding protein. J Neurosci. 2002;22:9868-9876. 47. Lee H, Kim JW, Yim SV, et al. Fluoxetine enhances cell proliferation and prevents apoptosis in dentate gyrus of maternally separated rats. Mol Psychiatry. 2001;6:725-728. 48. van der Hart M, Czeh B, de Biurrun G, et al. Substance P receptor antagonist and clomipramine prevent stress-induced alterations in cerebral metabolites, cytogenesis in the dentate gyrus and hippocampal volume. Mol Psychiatry. 2002;7:933-941. 49. Santarelli L, Saxe M, Gross C, et al. Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science. 2003;301:805-809. 50. Vollmayr B, Simonis C, Weber S, Gass P, Henn F. Reduced cell proliferation in the dentate gyrus is not correlated with the development of learned helplessness. Biol Psychiatry. 2003;54:1035-1040. 51. Manji H, Drevets WC, Charney DS. The cellular neurobiology of depression. Nat Med. 2001;7:541-547. 52. Nestler E, Barrot M, DiLeone RJ, Eisch AJ, Gold SJ. Monteggia LM. Neurobiology of depression. Neuron. 2002;34:13-25. 53. Duman R, Heninger GR, Nestler EJ. A molecular and cellular theory of depression. Arch Gen Psychiatry. 1997;54:597-606. 54. Duman R, J Malberg, S. Nakagawa, C DSa. Neuronal plasticity and survival in mood disorders. Biol Psychiatry. 2000;48:732-739. 55. Nestler E, Terwilliger RZ, Duman RS. Chronic antidepressant administration alters the subcellular distribution of cAMP-dependent protein kinase in rat frontal cortex. J Neurochem. 1989;53:1644-1647. 56. Nibuya M, Nestler EJ, Duman RS. Chronic antidepressant administration increases the expression of cAMP response element binding protein (CREB) in rat hippocampus. J Neurosci. 1996;16:2365-2372. 57. Thome J, Sakai N, Shin KH, et al. cAMP response element-mediated gene transcription is upregulated by chronic antidepressant treatment. J Neurosci. 2000;20:4030-4036.

168

Neural plasticity and depression - Duman

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

58. Nakagawa S, Kim JE, Lee R, Chen J, Fujioka T, Malberg J. Localization of phosphorylated cAMP response element-binding protein in immature neurons of adult hippocampus. J Neurosci. 2002;22:9868-9876. 59. Chen A-H, Shirayama Y, Shin KH, Neve RL, Duman RS. Expression of the cAMP response element binding protein (CREB) in hippocampus produces antidepressant effect. Biol Psychiatry. 2001;49:753-762. 60. Pliakas A, Carlson RR, Neve RL, Konradi C, Nestler EJ, Carlezon WA. Altered responsiveness to cocaine and increased immobility in the forced swim test associated with elevated CREB expression in the nucleus accumbens. J Neurosci. 2001;21:7397-7403. 61. Newton S, Thome J, Wallace TL, et al. Inhibition of cAMP response element-binding protein or dynorphin in the nucleus accumbens produces an antidepressant-like effect. J Neurosci. 2002;24:10883-10890. 62. Shirayama Y, Chen AC, Nakagawa S, Russell RS, Duman RS. Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. J Neurosci. 2002;22:3251-3261. 63. Saarelainen T, Hendolin P, Lucas G, et al. Activation of the trkB neurotrophin receptor is induced by antidepressant drugs and is required for antidepressant-induced behavioral effects. J Neurosci. 2003;23:349-357. 64. Siuciak JA, Lewis DR, Wiegand SJ, Lindsay R. Antidepressant-like effect of brain-derived neurotrophic factor (BDNF). Pharmacol Biochem Behav. 1997;56:131-137. 65. Mallei A, Shi B, Mocchetti I. Antidepressant treatments induce the expression of basic fibroblast growth factor in cortical and hippocampal neurons. 2002;61:1017-1024. 66. Newton S, Collier E, Hunsberger J, Adams D, Salvanayagam E, Duman RS. Gene profile of electroconvulsive seizures: induction of neurogenic and angiogenic factors. J Neurosci. 2003;23:10841-10851.

67. Sheline Y, Gado MH, Kraemer HC. Untreated depression and hippocampal volume loss. Am J Psychiatry. 2003;160:1-3. 68. Vermetten E, Vythilingam M, Southwick SM, Charney DS, Bremner JD. Long-term treatment with paroxetine increases verbal declarative memory and hippocampal volume in posttraumatic stress disorder. Biol Psychiatry. 2003;54:693-702. 69. Horowski R, Sastre-Y-Hernandez M. Clinical effects of the neurotrophic selective cAMP phosphodiesterase inhibitor rolipram in depressed patients: global evaluation of the preliminary reports. Curr Ther Res. 1985;38:23-29. 70. Wachtel H. Potential antidepressant activity of rolipram and other selective cyclic adenosine 3,5-monophosphate phosphodiesterase inhibitors. Neuropharmacology. 1983;22:267-272. 71. Takahashi M, Terwilliger R, Lane S, Mezes PS, Conti M, Duman RS. Chronic antidepressant administration increases the expression of cAMP phosphodiesterase 4A and 4B isoforms. J Neurosci. 1999;19:610-618. 72. Zhang H-T, Huang Y, Jin SJC, et al. Antidepressant-like profile and reduced sensitivity to rolipram in mice deficient in the PDE4D phosphodiesterase enzyme. Neuropsychopharmacology. 2002;27:587-595. 73. Li X, Tizzano JP, Griffey K, Clay M, Lindstron T, Skolnick P. Antidepressant-like actions of an AMPA receptor potentiator (LY392098). Neuropharmacology. 2001;40:1028-1033. 74. Marvanova M, Lakso M, Pirhonen J, Nawa H, Wong G, Castren E. The neuroprotective agent memantine induces brain-derived neurotrophic factor and trkB receptor expression in rat brain. Mol Cell Neurosci. 2001;18:247258. 75. Katoh-Semba R, Asano T, Ueda H, et al. Riluzole enhances expression of brain-derived neurotrophic factor with consequent proliferation of granule precursor cells in the rat hippocampus. FASEB J. 2001;16:1328-1330.

169

Pharmacological aspects
Cellular consequences of stress and depression
Eberhard Fuchs, PhD; Gabriele Flgge, PhD

Stress is known to activate distinct neuronal circuits in the brain and induce multiple changes on the cellular level, including alterations in neuronal structures. On the basis of clinical observations that stress often precipitates a depressive disease, chronic psychosocial stress serves as an experimental model to evaluate the cellular and molecular alterations associated with the consequences of major depression. Antidepressants are presently believed to exert their primary biochemical effects by readjusting aberrant intrasynaptic concentrations of neurotransmitters, such as serotonin or noradrenaline, suggesting that imbalances within the monoaminergic systems contribute to the disorder (monoaminergic hypothesis of depression). Here, we review the results that comprise our understanding of stressful experience on cellular processes, with particular focus on the monoaminergic systems and structural changes within brain target areas of monoaminergic neurons.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:171-183.

tressful life events are among the most potent factors that trigger or induce depressive episodes in humans. The brain responds to stress experiences in a complex manner related to the activation and inhibition of neurons that are involved in sensory, motor, autonomic, cognitive, and emotional processes. Chronic stress, which is known to be accompanied by hyperactivity in central nervous neurotransmitter systems, induces cellular changes that can be regarded as a form of plasticity. This causes mood alterations in the affected individual and has the potential to reverse the psychopathological processes, thus alleviating the symptoms of depression. Since social stress in animals evokes symptoms that resemble those found in depressed patients, chronic social stress can serve as an experimental paradigm to investigate the neuronal processes that may also occur during depressive disease in humans. Research over past years has led to considerable advances in the understanding of the neural causes of depression and the cellular mechanisms that underlie the beneficial effects of currently available antidepressants. More importantly, such research forms the basis for the future development of more effective antidepressant drugs.

Stress changes the activity of noradrenergic and adrenergic neurons

Stress is known to activate neurohormonal systems, such as the hypothalamo-pituitary-adrenal (HPA) axis, to release the central nervous stress peptide corticotropin-releasing factor,1 and to secrete glucocorticoids from the adrenal gland.2 These corticosteroids have been identified as prominent factors that modify metabolic processes in both the body and the brain during stress as well as depression.3 However, the other group of essential substances in basic and accelerated metabolism includes the monoamines, noradrenaline, adrenaline,
www.dialogues-cns.org

Keywords: noradrenaline; adrenaline; serotonin; dopamine; histamine; neuronal remodeling; 2-adrenoceptor; 5-HT1A receptor; dopamine transporter; tree shrew Author affiliations: Clinical Neurobiology Laboratory, German Primate Center, Gttingen, Germany Address for correspondence: Eberhard Fuchs, PhD, Clinical Neurobiology Laboratory, German Primate Center, Kellnerweg 4, 37077 Gttingen, Germany (e-mail: efuchs@gwdg.de) Copyright 2004 LLS SAS. All rights reserved

171

Pharmacological aspects
Selected abbreviations and acronyms
2-AR -AR DAT GPCR 5-HT LC 2-adrenoceptor -adrenoceptor dopamine transporter G proteincoupled receptor 5-hydroxytryptamine (serotonin) locus ceruleus

dopamine, serotonin (5-hydroxytryptamine [5-HT]), and histamine. The present survey focuses on processes related to stress-mediated activation of monoaminergic neurons in the brain. The noradrenergic and adrenergic neurons are located in the brain stem, where they form groups of cells that project axons to many parts of the brain. The beststudied group of noradrenergic neurons, located in the pontine locus ceruleus (LC), innervate several brain regions including the neocortex and the limbic system. The limbic system is a collection of regions that appear to regulate emotional processes (Figure 1). The noradrenergic LC neurons play an important role in the regulation of mood and emotions as well as of attention span. When stimulated through stressful challenge, for example, noradrenaline is released from the nerve terminals in the target brain region and is bound to adrenergic receptors belonging to the group of G proteincoupled receptors (GPCRs). These membrane-bound proteins convey signals from the extracellular to the intracellular compartment of a cell (Figure 2). GPCR signaling requires several steps for transmission of the signal, lasting from milliseconds to many minutes. The binding of a natural agonist such as noradrenaline or adrenaline to the receptor initiates a cascade of intracellular events that drive the activity of the cell and involve effectors such as enzymes (eg, adenylyl cyclase, phospholipase, kinases, and phosphatases), second messengers (eg, cyclic adenosine monophosphate [cAMP], cyclic guanosine monophosphate [cGMP], calcium ions, and arachidonic acid), as well as ion channels, which modulate the electrical activity of the neuron. A long-term effect occurring minutes after binding GPCR is the regulation of gene transcription and subsequent protein synthesis (Figure 2).5 There are different types of adrenergic receptors in the brain whose activation either stimulates or inhibits the respective target neurons. Noradrenaline and adrenaline bind to the same types of adrenergic receptors, although with slightly different affinities.6 Various experiments have shown that during stress,

noradrenergic and adrenergic neurons release more noradrenaline and adrenaline, respectively, and that the turnover of these neurotransmitters is accelerated so that their concentrations and/or amounts of their metabolites fluctuate in relation to the intensity and duration of the stressor.7-10 Acute stress induces only a transient rise in noradrenaline levels, but chronic stress with recurrent environmental challenges can lead to repetitive increases in concentration. As a consequence, adrenoceptors on the surface of the target neurons are bombarded with noradrenaline, leading to a reduction in adrenoceptor numbers (receptor downregulation).11 On the other hand, low concentrations of noradrenaline induce adrenoceptor upregulation.12

Changes in 2-ARs alter the activity of neurons

The most studied adrenergic receptors, with respect to regulation in chronic stress, are the 2-adrenoceptors (2ARs), of which three subtypes are known (A, B, and C).13 Because of their widespread distribution in the brain, 2ARs are diversely involved in mediating the analgesic and sedative effects of agonists such as dexmedetomidine14 and in modulating the baroreceptor reflex.15 The involvement of 2-ARs in the regulation of attention is suggested by the finding that methylphenidate (the nonamphetamine stimulant used to treat children with attention-deficit hyperactivity disorder) affects neuronal activity in the LC.16 Administration of the antagonist yohimbine (a sympatholytic drug that is used to treat impotence) increases firing of the LC neurons, resulting in anxiety-like behavior in rats and monkeys.17 Brain 2-AR changes have been observed in depressed patients (see below). The 2A-AR autoreceptor in LC noradrenergic neurons, regulates noradrenaline release via a negative feedback loop.14,18 Expression of this autoreceptor is reduced soon after the onset of stress (see below). In addition, 2A-AR is also expressed in neurons that release the excitatory neurotransmitter glutamate.19 In general, 2-AR stimulation leads to a transient inhibition of neuronal firing through hyperpolarization that is related to the modulation of calcium and potassium channels.20,21 There is reduced intracellular formation of the second messenger, cAMP, which itself regulates many cellular functions including gene transcription.22,23 Different forms of stress, such as immobilization or a cold environment, alter 2-AR numbers in distinct brain

172

Cellular consequences of stress and depression - Fuchs and Flgge

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

A. Noradrenaline

B. Dopamine

C. Serotonin

D. Histamine

Figure 1. Monoaminergic neurons innervate almost all brain areas. A. Noradrenaline. The noradrenergic neurons of the locus ceruleus project to the limbic and cortical regions, and to the thalamus, cerebellum, and spinal cord. They play an important role in the regulation of mood and attention. The noradrenergic neurons of cell groups A1, A2, A5, and A7 project to more restricted regions.4 They are important for autonomic function. B. Dopamine. The dopaminergic neurons of the substantia nigra and the adjacent ventral tegmental area (VTA) project to the striatum and to regions in the neocortex. They are important in the initiation of movements and for emotional processes. Furthermore, there is a dopaminergic cell group in the hypothalamus that regulates neuroendocrine processes. C. Serotonin. The serotonergic neurons located in the raphe nuclei project to almost all parts of the brain and are involved in many functions including the regulation of emotional processes. D. Histamine. Histaminergic neurons are located in the tuberomammillary complex of the hypothalamus. They project to all parts of the brain and are important for arousal (the excited brain state). Modulation of neuronal activity by these monoamines is an important factor of well-balanced central nervous activity. Stress leads to hyperactivity of the monoamine neurons and thus to a dysregulation of neuronal activity. Currently available antidepressants are thought to adjust the balance between the different neurotransmitter systems.

173

Pharmacological aspects
regions.24,25 We investigated the consequences of chronic psychosocial stress using a stress paradigm in male tree shrews.26 Our experiments showed that chronic psychosocial stress reduces 2-AR expression in brain regions that regulate autonomic functions and emotional behavior.27 This receptor downregulation is most probaNeurotransmitter

bly related to the stress-mediated rise in noradrenaline concentrations. Regulation of noradrenaline release is impaired soon after the onset of the stress period, as revealed by reduced expression of the 2A-AR in the LC.28 During a stress period lasting several weeks, adrenergic regulation changes, giving an initially high level and

G-protein complex Neuronal membrane

GPCR

Membrane-bound enzyme

cAMP, diacylglycerol, Ca2+ Cytoplasm

Ion channel

Nucleus

Gene transcription Activation of downstream enzymes (protein kinases, phosphatases)

DNA

mRNA

Protein synthesis

Figure 2. Neurotransmission via a G proteincoupled receptor (GPCR): binding of the neurotransmitter to the receptor initiates a cascade of intracellular events that drive the activity of the neuron or cell. The G-protein complex, consisting of subunits , , and , serves as the machinery that transduces the extracellular signal to various effectors at the intracellular side of the plasma membrane, to the enzymes adenylyl cyclase or phospholipase. These enzymes catalyze the synthesis of second messengers, such as cyclic adenosine monophosphate (cAMP) and diacylglycerol, which regulate gene transcription in the nucleus. Transcripts (mRNA) are later translated into protein. Calcium ions released from intracellular stores and other second messengers activate protein kinases and phosphatases. This leads to phosphorylation and/or dephosphorylation of many intracellular proteins as well as ion channels that are located in the plasma membrane of the cell. Phosphorylation/dephosphorylation induces opening and closing of these channels and this modulates the electrical activity of the neuron. These dynamic cellular processes are accelerated during stress when neurotransmitter concentrations are elevated.

174

Cellular consequences of stress and depression - Fuchs and Flgge

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

then finally a low level of noradrenaline. This is the case in the prefrontal cortex, a brain area important for the regulation of mood and behavior.29 Following a chronic stress period, noradrenaline concentrations are obviously low throughout the whole brain, probably due to a gradually acquired deficit in transmitter synthesis, transport, and/or release from the noradrenergic neurons.30 Interestingly, studies on postmortem material from brains of depressed human patients also revealed the upregulation of 2-ARs in several brain regions.31-33 These data therefore support the noradrenaline deficit hypothesis, which assumes there is a reduced noradrenaline concentration in the brains of depressed patients.34 Antidepressants that interact with 2-ARs such as mirtazapine probably counteract this deficit.35

intracellularly sequestered receptor molecules into the plasma membrane. Finally, after 4 weeks of psychosocial stress, the number of 1-ARs was decreased in cells of the hippocampus and parietal cortex.

Stress and 5-HT neurons

It is generally assumed that changes in serotonergic neurons underlie depressive diseases because the most widely used antidepressants are SSRIs, which raise extracellular levels of 5-HT. Several experimental results have confirmed the 5-HT deficit hypothesis of depression. In mammals, the majority of 5-HT-producing neurons are located in the brain stem, most of them on or near the midline, and they innervate almost every area of the brain.46 The serotonergic neurons of the dorsal raphe nucleus that project to the forebrain are autoactive and discharge in a stereotyped pattern that changes during the sleepwakearousal cycle.47,48 Due to its wide distribution, it has been suggested that the 5-HT system is involved in almost every brain function, such as the regulation of neuroendocrine secretion, regulation of cardiovascular and respiratory activity, sleep, nociception, analgesia, and motor output.49-51 5-HT definitely regulates mood, since its transporters and receptors are targets for several psychotropic drugs.52 A polymorphism in the promoter region of the 5-HT transporter (5-HTT) gene is thought to contribute to anxiety in humans,53 and an epidemiological study provides evidence that an allele encoding a short DNA sequence in this promoter region increases the risk of developing a depressive disorder.54 Rhesus monkeys with the short-sequence allele have low concentrations of the 5-HT metabolite 5-hydroxyindoleacetic acid in their cerebrospinal fluid.55 This finding agrees with the view that low brain 5-HT levels (decreased serotonergic activity) have negative effects on emotionality. However, 5-HT concentration per se is probably not the only trigger for mood changes; humans with a genotype conferring high levels of expression of monoamine oxidase A (MAOA, the enzyme that degrades 5-HT) are less likely to develop antisocial problems than individuals with lower MAOA expression.54 Stress elevates the concentrations of 5-HT and its metabolites in several brain regions, indicating increased turnover rates of the neurotransmitter,8,56 although the serotonergic neurons of the dorsal raphe nucleus do not change their discharge rate during stress.46 Nevertheless,

-ARs also change during stress

GPCR -adrenoceptors (-ARs) increase cAMP synthesis.36 They are present in neurons and glial cells.37 When stimulated by agonists (adrenaline or noradrenaline), -ARs are rapidly internalized into the cells. Therefore, high levels of endogenous agonists quickly reduce numbers of -AR molecules in the plasma membrane of target cells, inducing desensitization.11,38 -ARs are first internalized into the cell; they undergo intracellular sequestration with subsequent reinsertion into the plasma membrane, thereby restoring the normal receptor pattern in the membrane. -AR dysfunction is thought to play a role in psychiatric disorders, and -AR blockers have been used to treat depression and anxiety.39 The number of 1-ARs in the temporal and frontal cortex of suicide victims has been found to be significantly lower than in matched controls.40,41 However, the psychotropic role of -AR downregulation is still under discussion since the antidepressant desmethylimipramine also downregulates brain -ARs.42 On the other hand, the treatment of rats with the selective serotonin reuptake inhibitors (SSRIs) citalopram and fluoxetine increased 1-AR radioligand binding in the frontal cortex and striatum.43 Stress downregulates -ARs in the brain.44 Our data from the tree shrew chronic stress model reveal that (i) the effects are dependent on the duration of a stressful event; (ii) 1- and 2-ARs are differentially regulated; and (iii) the effects differ in different brain regions.45 Some of the stress-induced changes are only transient, since normal receptor numbers are restored through the reinsertion of

175

Pharmacological aspects
stress induces alterations in those brain regions that are targets of the serotonergic neurons, so that repeated exposure of rats to forced swimming increased 5-HT concentrations in the striatum, whereas they were reduced in the lateral septum.57 Chronic restraint stress in rats accelerated 5-HT turnover in the hippocampus and produced low amounts of the monoamine.58 Many receptors (>14) are known to mediate the effects of 5-HT.59 The present survey focuses on the 5-HT1A receptor, the best characterized 5-HT receptor. This GPCR inhibits neuronal activity by reducing cAMP formation or phosphoinositide hydrolysis, depending on the type of neuron where it is expressed, and it modulates potassium and calcium channels.60,61 The somatodendritic 5-HT1A autoreceptors located on the serotonergic neurons in the raphe nuclei regulate 5-HT release. Postsynaptic 5-HT1A receptors regulate the activity of neurons in cortical, limbic, and other regions. For example, they affect the activity of pyramidal neurons in the hippocampus.62-64 The 5-HT1A receptor has been implicated in many functions. Like other 5-HT receptors, it is involved in the regulation of mood and emotional behavior,65 and there is evidence that 5-HT1A receptor dysfunction is involved in depressive disorders. The agonists buspirone and gepirone act as anxiolytics and display antidepressantlike effects in clinical trials.66 Human brain studies showed that 5-HT1A receptor binding in depressed patients is lower than in healthy subjects.67,68 However, there are conflicting data on this issue. Brains of nonviolent suicides had increased 5-HT1A receptor binding in the frontal cortex in one report, whereas another report showed no difference between suicides and controls.69,70 Furthermore, other psychiatric diseasesas well as depressionmight cause changes in 5-HT1A receptors of the central nervous system. A variant of the 5HT1A receptor gene was found in Tourettes patients and, in schizophrenics, 5-HT1A receptor binding sites were increased in the ventral prefrontal cortex.71-73 Schizophrenics also displayed some 5-HT1A receptor binding in the cerebellum, a brain region normally devoid of these receptors.74 Restraint stress downregulated 5-HT1A receptors in the hippocampus of rats, and this effect was attributed to a stress-induced rise in plasma glucocorticoids, the adrenal hormones that regulate the transcription of many genes.75,76 The stress-induced downregulation of postsynaptic 5-HT1A receptors in distinct cortical areas and the hippocampal formation, in tree shrews, could also be attributed to high levels of glucocorticoids.64 However, it is interesting to note in relation to postsynaptic 5-HT1A receptor downregulation that the effect is not exclusively due to high glucocorticoid levels, but also to low testosterone. Social stress in male animals lowers testosterone levels, and normal 5-HT1A receptor numbers can be restored by a testosterone substitution (Figure 3).77 It is interesting that the number of somatodendritic 5-HT1A autoreceptors in the dorsal raphe nucleus did not change during chronic stress in male tree shrews, with only their affinity being reduced.64 This agrees with electrophysiological data from the rat brain stem, which showed that stress reduces 5-HT1A autoreceptor functioning.78

Stress affects dopaminergic neurons

Responses of the dopamine system to stress received particular attention because of the potential involvement of this catecholamine in human psychopathologies that are known to be exacerbated by stress, such as schizophrenia. Groups of dopaminergic neurons are located in the midbrain, hypothalamus, and other regions.4 The mesocortical and mesolimbic dopamine pathways, which arise from the ventral tegmental area, have been implicated in emotional and memory processes. Dopaminergic cells of the substantia nigra and the adjacent midbrain tegmentum project to the telencephalon including the striatum, forming the nigrostriatal pathway, which initiates motor responses. Dopamine transporter (DAT) knockout mice with high extracellular dopamine levels were easily aroused by the mild stress of novelty.79 However, in genetically intact animals, the persistent stress-induced activation demonstrated in the noradrenergic system has not been demonstrated in the dopaminergic system. Under restraint stress, an initial increase in mesolimbic dopamine release was later followed by a decline, suggesting that repeated exposure to the same stressor results in inhibition rather than activation of dopaminergic neurons.80 Other data suggest that the effects depend on the severity and controllability of the stressor, the genetic background of the animals, and their life history.81 The mesocortical dopaminergic system is obviously more stress-sensitive than the mesolimbic and the nigrostriatal systems.82,83 In the tree shrew model, 4 weeks of psychosocial stress downregulated DAT in the striatum. We also found a

176

Cellular consequences of stress and depression - Fuchs and Flgge

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

positive correlation between locomotor activity, which is reduced in stressed animals, and the total number of DAT binding sites.84 Low levels of DAT may indicate low extracellular dopamine concentrations. In agreement with these findings, social defeat in male rats also decreased DAT binding sites in the striatum.85 Dopamine was initially considered to convey its cellular actions via two receptor subtypes, D1 and D2; these exert

opposing effects on the adenylate cyclase system. Five distinct dopamine receptors have now been cloned.36 Experiments with various knockouts could not determine where on the neurons these receptor subtypes are located (presynaptic versus postsynaptic location).86 However, there are indications that D1 and D5 receptors are located postsynaptically, whereas D2, D3, and D4 receptors are located presynaptically and postsynapti-

Serotonin receptor Terminal of serotonergic neuron

Release of serotonin

Stress and testosterone

Stress

Normal
Figure 3. Serotonergic nerve endings (schematic drawing, upper left) in the hippocampal formation release the neurotransmitter serotonin (gray balls), which binds to its receptors, the serotonin-1A (5-HT1A) receptors (orange). The three pseudo-color pictures demonstrate receptor binding in normal male tree shrews (left), in tree shrews that were submitted to chronic psychosocial stress (middle), and in stressed tree shrews that had received testosterone as a treatment (right). Colors indicate numbers of receptors in the different regions of the hippocampal formation: orange, high receptor numbers; yellow, moderate numbers; green, low numbers; purple, no receptors. Note that after chronic social stress receptor numbers are decreased, but that the normal receptor number is restored following testosterone treatment.

177

Pharmacological aspects
cally, with the presynaptic receptors acting as inhibitory autoreceptors.87,88 In the tree shrew model, D1 receptors were slightly increased in the striatum after 4 weeks of psychosocial stress (Mijnster et al, unpublished observations), with a reliable increase in the prefrontal cortex, while D2 receptors were upregulated in the hippocampus.89 Taken together, these changes in receptors and DAT indicate impaired dopamine release after stress. Such a deficit in dopamine release might also account for a lack of motivation in depression. Antidepressants that block the D2 receptor (eg, clomipramine and fluvoxamine) might contribute to an improvement in motivation. antagonists and H3 receptor agonists decrease anxiety, and because of the existence of antidepressants that block the H1 receptor (eg, doxepin and amitriptyline).

Stress-induced neuronal remodeling and plasticity

The stress-induced processes described above include changes in different compartments of cells: Alterations in membrane-bound proteins that occur within seconds after the stressful stimulus (eg, conformational changes in receptors, enzymes, ion channels via stimulation of GPCRs). Internalization of receptors and intracellular trafficking as described for -ARs. Changes in large enzyme complexes involved in the intracellular signaling cascade. Changes in gene transcription, which may lead to either increased or decreased synthesis of a given protein (Figure 2). It is possible that these dynamic processes may even lead to morphological changes in the cells; past research has shown that this is indeed the case. The first proof that chronic stress induces a remodeling of brain cells came from morphological studies on dendrites of pyramidal neurons in area CA3 of the hippocampus. Dendrites are the major regions of neuronal synaptic contact with other neurons. Neurons with many or highly arborized dendrites potentially have large receptive fields (Figure 4). The retraction of the dendrites of these neurons was observed after chronic social stress and this effect was attributed to the stress-induced rise in glucocorticoids.96,97 Similar phenomena occur in pyramidal neurons in the prefrontal cortex, where glucocorticoids also induce alterations in the arborization of dendrites.98 In the CA3 pyramidal neurons of the hippocampus, dendritic retraction could be prevented by the antidepressant tianeptine, but not by the SSRIs fluoxetine and fluvoxamine.99 Also, chronic social defeat in male rats induced a shrinkage of the apical dendrites of the CA3 pyramidal neurons, and electrophysiological measurements revealed that this phenomenon was accompanied by a facilitation of action potentials, with reduced thresholds and higher amplitudes.100 In addition, single experiences of social defeat on two consecutive days induced similar changes in the apical dendrites, with these changes persisting over 3 weeks. In contrast to chronic daily social defeat, the arborization

Histaminergic neurons under stress

The central nervous histamine system has been less extensively studied with respect to stress, although it definitely plays an important role in the stress response. In mammalian brains, histaminergic neurons are found exclusively in the posterior ventral hypothalamus, but send their fibers to all brain regions.36,90 The electrophysiological properties of these cells are similar to those of the other aminergic neurons, with slow spontaneous firing, broad action potentials, and pronounced afterhyperpolarization.91,92 Histamine activates three types of receptors whose expression varies between brain regions.36 Histamine modulates glutamatergic neurotransmission. H1 and H2 receptors are mainly postsynaptically located with high densities in limbic brain regions, while H3 is a somatodendritic autoreceptor that regulates release of the bioamine.91 The central histamine system is involved in many functions. Activity in histaminergic neurons correlates closely with the sleepwake cycle, being highest when awake and lowest during rapid-eye movement sleep. Histaminergic neurons are also active in alarm situations and/or during activation of the peripheral sympathetic nervous system.91 H1 and H2 receptors modulate release of the stress hormones corticotropin-releasing factor and vasopressin from hypothalamic neurons,93 while various stressors such as dehydration or hypoglycemia stimulate histamine release. Even handling of rats raised histamine release in the prefrontal cortex of rats.94 Acute restraint stress stimulates histamine turnover throughout the diencephalon, whereas during chronic stress histamine turnover in the striatum and nucleus accumbens is affected.95 A relationship between histaminergic neurotransmission and emotional processes is suggested by the fact that H1 receptor

178

Cellular consequences of stress and depression - Fuchs and Flgge

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

of the dendrites at the basal pole of the pyramidal neurons was increased after the double defeat paradigm.100,101 Therefore, two severely stressful experiences had longlasting consequences on the morphology of neurons that differed from those induced by daily chronic stress. Stress was also shown to prevent long-term potentiation (LTP, a mechanism of synaptic plasticity that is thought to be related to memory formation) of CA neurons in the hippocampus. This inhibition of LTP was observed in male rats after only two exposures to social defeat.101 The antidepressant tianeptine increases the amplitude of excitatory postsynaptic potentials and this mechanism appears to be related to alterations in the phosphorylation of the N-methyl-D-aspartate (NMDA) receptor, one of the most prominent receptors for the excitatory neurotransmitter glutamate.102

Synapses are often located at the tips of the spine protrusions on the dendritic shafts of neurons (Figure 4). The shape of a spine is related to the arrangement of the actin-containing microfilaments, the cytoskeletal fibers.103 Spines may form rapidly under the influence of synaptic activity.104 Activation of the NMDA receptor initiates changes in the actin cytoskeleton that stabilize the synaptic structure.105 Spine formation in the neurons of the prefrontal cortex can be induced by even minor stimuli, such as handling the experimental animals daily.106 In response to an acute stress, spine density was enhanced in the hippocampus of male rats, whereas, in contrast, female rats showed reduced spine density.107 It therefore appears that spine morphology is modulated by stress, although other factors such as sex hormones may also have an effect on their formation.

Chronic stress and neuronal death?

There have been reports that social stress leads to cell death in the hippocampal formation.108 However, recent studies using the optical dissector technique, a reliable method for quantification of neurons within an entire brain region, showed that stress does not affect neuron numbers in the CA1 and CA3 areas of the hippocampus.109 Moreover, experiments using an in situ end-labeling technique to identify apoptotic (dying) cells showed a significant decrease in the number of apoptotic cells when all hippocampal areas were analyzed.110 Although stress-induced death of principal neurons in the hippocampus is questionable, it is clear that stress profoundly affects these neurons.Their nuclear ultrastructure changes as shown in the significant intensification in Nissl staining.111 An electron microscopic analysis indicated that this effect is due to increased heterochromatin formation in the neuronal nuclei.112 The physiological role of these changes is unknown, but one may speculate that they are accompanied by alterations in gene transcription. Recent tree shrew studies showed that chronic psychosocial stress reduced the expression of certain genes that are related to the shape of neurons and other brain cells.113 In the brains of adult rats that had been prenatally stressed through the stressful treatment of the pregnant dams, expression of genes associated with excitatory neurotransmission and mechanisms of neurotransmitter release were significantly altered.114 Furthermore, a large group of genes in the hippocampus has been shown to be differentially expressed after glucocorticoid treatment.76

Apical dentrites

Spine Synapse

Soma

Basal dentrites

Axon

Figure 4. Schematic drawing of a CA3 pyramidal neuron plus its dendrites. Note the small soma in comparison to the highly arborized apical and basal dendrites. Inset: dendritic shafts can build up protrusions (spines) that form synapses with axons or dendrites from other neurons. Synapses are sites of signal transmission between neurons. Formation and disappearance of spines are regulated by many factors such as gonadal hormones. Chronic psychosocial stress reduces the arborization of the apical dendrites, thus reducing the surface area of the neuron with the consequence that the neuron receives less information from other neurons (see text for details).

179

Pharmacological aspects
Conclusions and further directions
Despite extensive preclinical and clinical investigations, the exact neurobiological processes leading to depression and the mechanisms responsible for the therapeutic effects of antidepressant drugs are still not completely understood. Antidepressants are presently believed to exert their primary biochemical effects by readjusting aberrant intrasynaptic concentrations of neuromodulators such as 5-HT. However, the limitations of current antidepressant medications, such as the time delay for a full therapeutic response, the substantial number of nonresponders, and bothersome side effects merit a full exploration of all plausible agents with novel antidepressant mechanisms of action. Recent preclinical and clinical studies suggest that major depressive disorders are associated with cellular resilience and an impairment of synaptic and structural plasticity, and that antidepressant medications may act by correcting this dysfunction. Although this concept is still in its infancy, it has increasingly attracted research efforts that may result in new treatment strategies for the etiopathophysiology of psychiatric disorders, such as major depression.

The contributions of former and current members of the Clinical Neurobiology Laboratory at the German Primate Center are gratefully acknowledged. The work summarized here was in part supported by the German Science Foundation, the DAAD, and the EC.

Consecuencias celulares del estrs y la depresin

Se sabe que el estrs activa distintos circuitos neuronales en el cerebro e induce mltiples cambios a nivel celular incluyendo alteraciones en las estructuras neuronales. A partir de las observaciones clnicas, que indican que a menudo el estrs precipita una enfermedad depresiva, el estrs psicosocial crnico sirve como modelo experimental para evaluar las alteraciones celulares y moleculares asociadas con las consecuencias de la depresin mayor. Actualmente se cree que los antidepresivos ejercen sus efectos bioqumicos primarios mediante un reajuste de las concentraciones intrasinpticas aberrantes de neurotransmisores, como la serotonina y la noradrenalina, lo que sugiere que el desbalance dentro del sistema monoaminrgico contribuye al trastorno (hiptesis monoaminrgica de la depresin). Aqu se revisan los resultados que contribuyen a nuestra comprensin acerca de las consecuencias del estrs sobre los procesos celulares, con particular atencin a los sistemas monoaminrgicos y los cambios estructurales de las reas cerebrales blanco de las neuronas monoaminrgicas.

Consquences cellulaires du stress et dpression

Il est reconnu que le stress met en jeu des circuits neuronaux distincts dans le cerveau et induit de nombreux changements au niveau cellulaire, y compris des altrations des structures neuronales. Sur la base dobservations cliniques ayant montr que le stress dclenchait souvent une maladie dpressive, le stress psychosocial chronique peut tre pris comme modle exprimental pour valuer les altrations molculaires et cellulaires associes aux consquences dune dpression majeure. Selon les conceptions actuelles, les antidpresseurs exercent leur effet biochimique primaire en rajustant les concentrations intrasynaptiques aberrantes des neurotransmetteurs, telles la srotonine et la noradrnaline. Ceci suggre que les dsquilibres des systmes monoaminergiques sont impliqus dans la gense de la dpression (hypothse monoaminergique). Cet article passe en revue les rsultats qui contribuent notre comprhension des consquences du stress sur les processus cellulaires, avec une attention particulire sur les systmes monoaminergiques et les changements structurels des rgions crbrales cibles des neurones monoaminergiques.

180

Cellular consequences of stress and depression - Fuchs and Flgge

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

REFERENCES
1. Koizumi K. The role of the hypothalamus in neuroendocrinology. In: Greger R, Windhorst U, eds. Comprehensive Human Physiology. Heidelberg, Germany: Springer Verlag; 1996;1:379-401. 2. Hierholzer K, Bhler H. Metabolism of cortical steroid hormones and their general mode of action. In: Greger R, Windhorst U, eds. Comprehensive Human Physiology. Heidelberg, Germany: Springer Verlag; 1996;1:403-429. 3. Duman RS. Neural plasticity: consequences of stress and actions of antidepressant treatment. Dialogues Clin Neurosci. 2004;6:157-169. 4. Nieuwenhuys R. Chemoarchitecture of the Brain. Berlin, Germany: Springer Verlag; 1985. 5. Squire LR, Bloom FE, McConnell SK, Roberts JL, Spitzer NC, Zigmond MJ. Fundamental Neuroscience. New York, NY: Academic Press; 2003. 6. Lefkowitz RJ, Hoffman BB, Taylor P. Neurotransmission. The autonomic and somatic motor nervous system. In: Hardman JG, Limbird LE, et al, eds. Goodman & Gilman's The Pharmacologic Basis of Therapeutics. New York, NY: McGraw-Hill; 1996:105-139. 7. Bliss EL, Ailion J, Zwanziger J. Metabolism of norepinephrine, serotonin and dopamine in rat brain with stress. J Pharmacol Exp Ther. 1968;164:122-134. 8. Thierry AM, Javoy F, Glowinski J, Kety S. Effects of stress on the metabolism of norepinephrine, dopamine and serotonin in the central nervous system of the rat. J Pharmacol Exp Ther. 1968;163:163-171. 9. Saavedra JM. Brain epinephrine in hypertension and stress. In: Stolk JM, UPrichard DC, Fuxe K, eds. Epinephrine in the Central Nervous System. Oxford, UK: Oxford University Press; 1988:102-116. 10. Stanford SC. Central noradrenergic neurones. Pharmacol Ther. 1995;68: 297-342. 11. von Zastrow M, Kobilka BK. Antagonist-dependent and -independent steps in the mechanism of adrenergic receptor internalization. J Biol Chem. 1994;269:18448-18452. 12. Ribas C, Miralles A, Busquets X, Garcia-Sevilla JA. Brain 2-adrenoceptors in monoamine-depleted rats: increased receptor density, G-coupling proteins, receptor turnover and receptor mRNA. Br J Pharmacol. 2001;132:1467-1476. 13. MacDonald E, Kobilka BK, Scheinin M. Gene targeting-homing in on 2adrenoceptor-subtype function. Trends Pharmacol Sci. 1997;18:211-219. 14. Kable JW, Murrin LC, Bylund DB. In vivo gene modification elucidates subtype-specific functions of 2-adrenergic receptors. J Pharmacol Exp Ther. 2000;293:1-7. 15. MacMillan LB, Hein L, Smith MS, Piascik MT, Limbird LE. Central hypotensive effects of the 2A-adrenergic receptor subtype. Science. 1996;273:801803. 16. Ishimatsu M, Kidani Y, Tsuda A, Akasu T. Effects of methylphenidate on the membrane potential and current in neurons of the rat locus coeruleus. J Neurophysiol. 2002;87:1206-1212. 17. Redmond AM, Leonard BE. An evaluation of the role of the noradrenergic system in the neurobiology of depression: a review. Hum Psychopharmacol Clin Exp. 1997;12:407-430. 18. Aghajanian GK, Vandermaelen CP. 2-Adrenoceptor-mediated hyperpolarization of locus coeruleus neurons: intracellular studies in vivo. Science. 1982;215:1394-1396. 19. Meyer H, Palchaudhuri M, Scheinin M, Flgge G. Regulation of 2Aadrenoceptor expression by chronic stress in neurons of the brain stem. Brain Res. 2000;880:147-158. 20. Surprenant A, Horstman DA, Akbarali H, Limbird LE. A point mutation of the 2-adrenoceptor that blocks coupling to potassium but not calcium currents. Science. 1992;257:977-980. 21. Boehm S, Huck S, Freissmuth M. Involvement of a phorbol ester-insensitive protein kinase C in the 2-adrenergic inhibition of voltage-gated calcium current in chick sympathetic neurons. J Neurosci. 1996;16:4596-4603. 22. Limbird LE. Receptors linked to inhibition of adenylate cyclase: additional signaling mechanisms. FASEB J. 1988;2:2686-2695. 23. Nestler EJ, Hyman SE. Regulation of gene expression. In: Davis KL, Charney D, Coyle JT, Nemeroff C, eds. Neuropsychopharmacology. The Fifth Generation of Progress. Philadelphia, Pa: Lippincott Williams & Wilkins; 2002:217-228.

24. UPrichard D C, Kvetnansky R. Central and peripheral adrenergic receptors in acute and repeated immobilization stress. In: Usdin E, Kvetnansky R, Kopin IJ, eds. Catecholamine and Stress. Recent Advances, Developments in Neuroscience. Vol 8. Amsterdam, The Netherlands: Elsevier; 1980:299-308. 25. Nukina I, Glavin GB, La Bella FS. Acute cold-restraint stress affects 2adrenoceptors in specific brain regions of the rat. Brain Res. 1987;401:30-33. 26. Fuchs E, Flgge G. Social stress in tree shrews: effects on physiology, brain function, and behavior of subordinate individuals. Pharmacol Biochem Behav. 2002;73:247-258. 27. Flgge G, Jhren O, Fuchs E. [3H]Rauwolscine binding sites in the brains of male tree shrews are related to social status. Brain Res. 1992;597:131-137. 28. Flgge G. Alterations in the central nervous 2-adrenoceptor system under chronic psychosocial stress. Neuroscience. 1996;75:187-196. 29. Flgge G, Ahrens O, Fuchs E. Monoamine receptors in the prefrontal cortex of Tupaia belangeri during chronic psychosocial stress. Cell Tiss Res. 1997;288:1-10. 30. Flgge G, Van Kampen M, Mijnster MJ. Perturbations in brain monoamine systems during stress. Cell Tiss Res. 2003;315:1-14. 31. Meana JJ, Barturen F, Garcia Sevilla JA. 2-Adrenoceptors in the brain of suicide victims: increased receptor density associated with major depression. Biol Psychiatry. 1992;31:471-490. 32. Ordway GA, Widdowson PS, Smith KS, Halaris A. Agonist binding to 2adrenoceptors is elevated in the locus coeruleus from victims of suicide. J Neurochem. 1994;63:617-624. 33. Garcia-Sevilla JA, Escriba PV, Ozaita A, et al. Up-regulation of immunolabeled 2A-adrenoceptors, Gi coupling proteins, and regulatory receptor kinases in the prefrontal cortex of depressed suicides. J Neurochem. 1999;72:282-291. 34. Holsboer F. Molekulare Mechanismen der Depressionstherapie. In: Ganten D, Ruckpaul K, eds. Handbuch der molekularen Medizin. Heidelberg, Germany: Springer Verlag; 1999:273-318. 35. Schatzberg AF. Pharmacological principles of antidepressant efficacy. Hum Psychopharmacol. 2002;17:17-22. 36. Feldman RS, Meyer JS, Quenzer LF. Principles of Neuropsychopharmacology. Sunderland, Mass: Sinauer Associates Inc; 1997. 37. Stone EA, John SM. Further evidence for a glial localization of rat cortical -adrenoceptors: studies of in vivo cyclic AMP responses to catecholamines. Brain Res. 1991;549:78-82. 38. Claing A, Laporte SA, Caron MG, Lefkowitz RJ. Endocytosis of G proteincoupled receptors: roles of G proteincoupled receptor kinases and arrestin proteins. Prog Neurobiol. 2002;66:61-79. 39. Bailly D. The role of -adrenoceptor blockers in the treatment of psychiatric disorders. CNS Drugs. 1996;5:115-136. 40. Depaermentier F, Crompton M R, Katona CLE, Horton RW. Adrenoceptors in brain and pineal from depressed suicide victims. Pharmacol Toxicol. 1993;71:86-95. 41. Little KY, Clark TB, Ranc J, Duncan GE. -Adrenergic receptor binding in frontal cortex from suicide victims. Biol Psychiatry. 1993;34:596-605. 42. Paetsch PR, Greenshaw AJ. Effects of chronic antidepressant treatment on -adrenoceptor subtype binding in the rat cerebral cortex and cerebellum. Mol Chem Neuropathol. 1993;20:21-31. 43. Palvimaki EP, Laakso A, Kuoppamaki M, Syvalahti E, Hietala J. Upregulation of 1-adrenergic receptors in rat brain after chronic citalopram and fluoxetine treatments. Psychopharmacology. 1994;115:543-546. 44. Stone EA, Platt JE. Brain adrenergic receptors and resistance to stress. Brain Res. 1982;237:405-414. 45. Flgge G, Ahrens O, Fuchs E. -Adrenoceptors in the tree shrew brain. II. Time-dependent effects of chronic psychosocial stress on [125I]iodocyanopindolol bindings sites. Cell Mol Neurobiol. 1997;17:417-432. 46. Jacobs BL, Azmitia EC. Structure and function of the brain serotonin system. Physiol Rev. 1992;72:165-229. 47. Aghajanian GK, Vandermaelen CP. Intracellular recordings from serotonergic dorsal raphe neurons: pacemaker potentials and the effect of LSD. Brain Res. 1982;238:463-469. 48. Trulson ME, Jacobs BL. Raphe unit activity in freely moving cats: correlation with level of behavioral arousal. Brain Res. 1979;163:135-150. 49. Whitaker-Azmitia PM, Peroutka SJ. The Neuropharmacology of Serotonin. New York, NY: New York Academy of Sciences; 1990.

181

Pharmacological aspects
50. Jacobs BL, Fornal CA. Serotonin and behavior. A general hypothesis. In: Bloom FE, Kupfer DJ, eds. Psychopharmacology. The Fourth Generation of Progress. New York, NY: Raven Press; 1995:461-469. 51. Lowry CA. Functional subsets of serotonergic neurones: implications for control of the hypothalamic-pituitary-adrenal axis. J Neuroendocrinol. 2002;14:911-923. 52. Nutt DJ. The neuropharmacology of serotonin and noradrenaline in depression. Int Clin Psychopharmacol. 2002;17:1-12. 53. Lesch KP, Bengel D, Heils A, et al. Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science. 1996;274:1527-1531. 54. Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science. 2003;301:386-389. 55. Bennett AJ, Lesch KP, Heils A, et al. Early experience and serotonin transporter gene variation interact to influence primate CNS function. Mol Psychiatry. 2002;7:118-122. 56. Chaouloff F. Physiopharmacological interactions between stress hormones and central serotonergic systems. Brain Res Rev. 1993;18:1-32. 57. Kirby LG, Lucki I. The effect of repeated exposure to forced swimming on extracellular levels of 5-hydroxytryptamine in the rat. Stress. 1998;2:251-263. 58. Torres IL, Gamaro GD, Vasconcellos AP, Silveira R, Dalmaz C. Effects of chronic restraint stress on feeding behavior and on monoamine levels in different brain structures in rats. Neurochem Res. 2002;27:519-25. 59. Aghajanian GK, Sanders-Bush E. Serotonin. In: Davis KL, Charney D, Coyle JT, Nemeroff C, eds. Neuropsychopharmacology. The Fifth Generation of Progress. Philadelphia, Pa: Lippincott Williams & Wilkins; 2002:15-34. 60. Sanders-Bush E, Canton H. Serotonin receptors. Signal transduction pathways. In: Bloom FE, Kupfer DJ, eds. Psychopharmacology. The Fourth Generation of Progress. New York, NY: Raven Press; 1995:431-441. 61. Johnson RG, Fiorella D, Winter JC, Rabin RA. [3H]8-OH-DPAT labels a 5HT site coupled to inhibition of phosphoinositide hydrolysis in the dorsal raphe. Eur J Pharmacol. 1997;329:99-106. 62. Verg D, Daval G, Marcinkiewicz M, et al. Quantitative autoradiography of multiple 5-HT1 receptor subtypes in the brain of control or 5,7-dihydroxytryptamine-treated rats. J Neurosci. 1986;6:3474-3482. 63. Jols M, Hesen W, De Kloet ER. Mineralocorticoid hormones suppress serotonin-induced hyperpolarization of rat hippocampal CA1 neurons. J Neurosci. 1991;11:2288-2294. 64. Flgge G. Dynamics of central nervous 5-HT1A receptors under psychosocial stress. J Neurosci. 1995;15:7132-7140. 65. Lucki I. 5-HT1 receptors and behavior. Neurosci Biobehav Rev. 1992;16: 83-93. 66. Blier P, de Montigny C. A decade of serotonin research: antidepressant mechanisms and therapeutics. Possible serotonergic mechanisms underlying the antidepressant and anti-obsessive-compulsive disorder responses. Biol Psychiatry. 1998;44:313-323. 67. Lopez JF, Chalmers DT, Little KY, Watson SJ. A. E. Bennett Research Award Regulation of serotonin1A, glucocorticoid, and mineralocorticoid receptor in rat and human hippocampus: implications for the neurobiology of depression. Biol Psychiatry. 1998;43:547-573. 68. Drevets WC, Frank E, Price JC, et al. PET imaging of serotonin 1A receptor binding in depression. Biol Psychiatry. 1999;46:1375-1387. 69. Matsubara S, Arora RC, Meltzer HY. Serotonergic measures in suicide brain: 5-HT1A binding sites in frontal cortex of suicide victims. J Neural Transm Gen Sect. 1991;85:181-194. 70. Lowther S, De Paermentier F, Cheetham SC, Crompton MR, Katona CL, Horton RW. 5-HT1A receptor binding sites in post-mortem brain samples from depressed suicides and controls. J Affect Disord. 1997;42:199-207. 71. Lam S, Shen Y, Nguyen T, et al. A serotonin receptor gene (5HT1A) variant found in a Tourette's syndrome patient. Biochem Biophys Res Commun. 1996;219:853-858. 72. Simpson MD, Lubman DI, Slater P, Deakin JF. Autoradiography with [3H]8-OH-DPAT reveals increases in 5-HT1A receptors in ventral prefrontal cortex in schizophrenia. Biol Psychiatry. 1996;39:919-928. 73. Sumiyoshi T, Stockmeier CA, Overholser JC, Dilley GE, Meltzer HY. Serotonin1A receptors are increased in postmortem prefrontal cortex in schizophrenia. Brain Res. 1996;708:209-214. 74. Slater P, Doyle CA, Deakin JF. Abnormal persistence of cerebellar serotonin-1A receptors in schizophrenia suggests failure to regress in neonates. J Neural Transm. 1998;105:305-315. 75. McKittrick CR, Blanchard DC, Blanchard RJ, McEwen BS, Sakai RR. Serotonin receptor binding in a colony model of chronic social stress. Biol Psychiatry. 1995;37:383-393. 76. Datson NA, van der Perk J, de Kloet ER, Vreugdenhil E. Identification of corticosteroid-responsive genes in rat hippocampus using serial analysis of gene expression. Eur J Neurosci. 2001;14:675-689. 77. Flgge G, Kramer M, Rensing S, Fuchs E. 5-HT1A-receptors and behaviour under chronic stress: selective counteraction by testosterone. Eur J Neurosci. 1998;10:2685-2693. 78. Laaris N, Le Poul E, Hamon M, Lanfumey L. Stress-induced alterations of somatodendritic 5-HT1A autoreceptor sensitivity in the rat dorsal raphe nucleusin vitro electrophysiological evidence. Fund Clin Pharmacol. 1997;11:206-214. 79. Spielewoy C, Roubert C, Hamon M, Nosten-Bertrand M, Bncur C, Giros B. Behavioural disturbances associated with hyperdopaminergia in dopamine-transporter knockout mice. Behav Pharmacol. 2000;11:279-290. 80. Imperato A, Cabib S, Puglisi-Allegra S. Repeated stressful experiences differently affect the time-dependent responses of the mesolimbic dopamine system to the stressor. Brain Res. 1993;60:333-336. 81. Cabib S, Puglisi-Allegra S. Stress, depression and the mesolimbic dopamine system. Psychopharmacol. 1996;128:331-342. 82. Abercrombie ED, Keefe KA, DiFrischia DS, Zigmond MJ. Differential effect of stress on in vivo dopamine release in striatum, nucleus accumbens and medial frontal cortex. J Neurochem. 1989;52:1655-1658. 83. Tidey JW, Miczek KA. Social defeat stress selectively alters mesocorticolimbic dopamine release: an in vivo microdialysis study. Brain Res. 1996;721:140-114. 84. Isovich E, Mijnster MJ, Flgge G, Fuchs E. Chronic psychosocial stress reduces the density of dopamine transporters. Eur J Neurosci. 2000;12:10711078. 85. Isovich E, Engelmann M, Landgraf R, Fuchs E. Social isolation after a single defeat reduces striatal dopamine transporter binding in rats. Eur J Neurosci. 2001;13:1254-1256. 86. Picciotto MR. Knock-out mouse models used to study neurobiological systems. Crit Rev Neurobiol. 1999;13:103-149. 87. Meador-Woodruff JH. Update on dopamine receptors. Ann Clin Psychiatry. 1994;6:79-90. 88. Vallone D, Picetti R, Borrelli E. Structure and function of dopamine receptors. Neurosci Biobehav Rev. 2000;24:125-132. 89. Mijnster MJ, Isovich E, Fuchs E. Chronic psychosocial stress alters the density of dopamine D2-like binding sites. Soc Neurosci Abstr. 1998;24:277. 90. Airaksinen MS, Flgge G, Fuchs E, Panula P. Histaminergic system in the tree shrew brain. J Comp Neurol. 1989;286:289-310. 91. Brown RE, Stevens DR, Haas HL. The physiology of brain histamine. Prog Neurobiol. 2001;63:637-672. 92. Watanabe T, Yanai K. Studies on functional roles of the histaminergic neuron system by using pharmacological agents, knockout mice and positron emission tomography. Tohoku J Exp Med. 2001;195:197-217. 93. Knigge U, Warberg J. The role of histamine in the neuroendocrine regulation of pituitary hormone secretion. Acta Endocrinol (Copenh). 1991;124:609-619. 94. Westerink BH, Cremers TI, De Vries JB, Liefers H, Tran N, De Boer P. Evidence for activation of histamine H3 autoreceptors during handling stress in the prefrontal cortex of the rat. Synapse. 2002;43:238-243. 95. Ito C. The role of brain histamine in acute and chronic stresses. Biomed Pharmacother. 2000;54:263-267. 96. Woolley CS, Gould E, Frankfurt M, McEwen BS. Naturally occurring fluctuation in dendritic spine density on adult hippocampal pyramidal neurons. J Neurosci. 1990;10:4035-4039. 97. Magarios AM, McEwen BS, Flgge G, Fuchs E. Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. J Neurosci. 1996;5:3534-3540. 98. Wellman CL. Dendritic reorganization in pyramidal neurons in medial prefrontal cortex after chronic corticosterone administration. J Neurobiol. 2001;49:245-253.

182

Cellular consequences of stress and depression - Fuchs and Flgge

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

99. Magarinos AM, Deslandes A, McEwen BS. Effects of antidepressants and benzodiazepine treatments on the dendritic structure of CA3 pyramidal neurons after chronic stress. Eur J Pharmacol. 1999;371:113-122. 100. Kole MHP, Czh B, Fuchs E. Homeostatic maintenance in tree shrew hippocampal CA3 neuron excitability after chronic stress. Hippocampus. 2004. In press. 101. Kole MHP. CA3 pyramidal neuron correlates of the stress response analyses of form and function. University of Groningen, The Netherlands. PhD Thesis; 2003. 102. Kole MH, Swan L, Fuchs E. The antidepressant tianeptine persistently modulates glutamate receptor currents of the hippocampal CA3 commissural associational synapse in chronically stressed rats. Eur J Neurosci. 2002;16:807-816. 103. Oliver CJ, Terry-Lorenzo RT, Elliott E, et al Targeting protein phosphatase 1 (PP1) to the actin cytoskeleton: the neurabin I/PP1 complex regulates cell morphology. Mol Cell Biol. 2002;22:4690-4701. 104. Muller D, Toni N, Buchs PA. Spine changes associated with long-term potentiation. Hippocampus. 2000;10:596-604. 105. Ackermann M, Matus A. Activity-induced targeting of profilin and stabilization of dendritic spine morphology. Nat Neurosci. 2003;6:1194-1200. 106. Seib LM, Wellman CL. Daily injections alter spine density in rat medial prefrontal cortex. Neurosci Lett. 2003;337:29-32. 107. Shors TJ, Chua C, Falduto J. Sex differences and opposite effects of stress on dendritic spine density in the male versus female hippocampus. J Neurosci. 2001;21:6292-6297.

108. Sapolsky RM, Krey LC, McEwen BS. The neuroendocrinology of stress and aging: the glucocorticoid cascade hypothesis. Endocr Rev. 1986;7:284301. 109. Keuker J, Vollmann-Honsdorf GK, Fuchs E. How to use the optical fractionator: an example based on the estimation of neurons in the hippocampal CA1 and CA3 regions of tree shrews. Brain Res Protocols. 2001;7:211-221. 110. Lucassen PJ, Vollmann-Honsdorf GK, Gleisberg M, Czeh B, De Kloet ER, Fuchs E. Chronic psychosocial stress differentially affects apoptosis in hippocampal subregions and cortex of the adult tree shrew. Eur J Neurosci. 2001;14:161-166. 111. Fuchs E, Uno H, Flgge G. Chronic psychosocial stress induces morphological alterations in hippocampal pyramidal neurons of the tree shrew. Brain Res. 1995;673:275-282. 112. Vollmann-Honsdorf GK, Flgge G, Fuchs E. Cortisol treatment and psychosocial stress differentially alter the nuclear ultrastructure of hippocampal pyramidal neurons. In: Elsner N, Eysel U, eds. Gttingen Neurobiology Report 1999. Stuttgart, Germany: Georg Thieme Verlag; 1999:524. 113. Alfonso J, Pollevick GD, van der Hart MG, Flgge G, Fuchs E, Frasch ACC. Identification of genes regulated by chronic psychosocial stress and antidepressant treatment in the hippocampus. Eur J Neurosci. 2004;19:659666. 114. Kinnunen AK, Koenig JI, Bilbe G. Repeated variable prenatal stress alters pre- and postsynaptic gene expression in the rat frontal pole. J Neurochem. 2003;86:736-748.

183

Clinical research
Cellular abnormalities in depression: evidence from postmortem brain tissue
Craig A. Stockmeier, PhD; Grazyna Rajkowska, PhD

During the past two decades, in vivo neuroimaging studies have permitted significant insights into the general location of dysfunctional brain regions in depression. In parallel and often intersecting ways, neuroanatomical, pharmacological, and biochemical studies of postmortem brain tissue are permitting new insights into the pathophysiology of depression. In addition to long-recognized neurochemical abnormalities in depression, novel studies at the microscopic level support the contention that mood disorders are associated with abnormalities in cell morphology and distribution. In the past 6 years, cell-counting studies have identified changes in the density and size of both neurons and glia in a number of frontolimbic brain regions, including dorsolateral prefrontal, orbitofrontal, and anterior cingulate cortex, and the amygdala and hippocampus. Convergence of cellular changes at the microscopic level with neuroimaging changes detected in vivo provides a compelling integration of clinical and basic research for disentangling the pathophysiology of depression. The ultimate integration of these two research approaches will occur with premortem longitudinal clinical studies on well-characterized patients linked to postmortem studies of the same subjects.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:185-197.

uring the past two decades, anatomical substrates associated with the neuropathology of mood disorders have been revealed through both in vivo neuroimaging studies and morphological and neurochemical studies on postmortem brain tissue. While neuroimaging studies have given significant insight into the gross morphological location of dysfunctional brain regions in depression, the neurochemical, cellular, and molecular features of depression are being unlocked by studies in postmortem brain tissue. Novel studies at the microscopic level are establishing that the mood disorders are associated with abnormalities in cell morphology and distribution, in addition to the long-recognized neurochemical abnormalities. Major depressive disorder (MDD) and bipolar disorder (BPD) have been examined in postmortem brain tissue by several laboratories in the past 6 years. Cell-counting studies report changes in the density and size of both neurons and glia in a number of frontolimbic brain regions, including dorsolateral prefrontal, orbitofrontal, and anterior cingulate cortex, and the amygdala and hippocampus. These studies in postmortem brain tissue confirm and extend structural and functional neuroimaging studies that reveal volumetric and metabolic changes in the same frontolimbic brain regions in the same disorders. Convergence of cellular changes at the microscopic level with neuroimaging changes detected in vivo provides a compelling integration of clinical and basic research for disentangling the pathophysiology of depression. Regionally localized and cell typespecific changes in neuronal and glial cytoarchitecture recently identified in
Author affiliations: The University of Mississippi Medical Center, Department of Psychiatry and Human Behavior, Jackson, Miss, USA Address for correspondence: Craig Stockmeier, PhD, Department of Psychiatry and Human Behavior, University of Mississippi Medical Center, 2500 N State St, Box 127, Jackson, MS 39216, USA (e-mail: cstockmeier@psychiatry.umsmed.edu) www.dialogues-cns.org

Keywords: major depression; bipolar disorder; postmortem brain; glia; neuron; neuropathology Copyright 2004 LLS SAS. All rights reserved

185

Clinical research
Selected abbreviations and acronyms
AVP BDNF BPD CRH GABA GFAP HPA MDD NAA NMDA arginine-vasopressin brain-derived neurotrophic factor bipolar disorder corticotropin-releasing hormone -aminobutyric acid glial fibrillary acidic protein hypothalamic-pituitary-adrenal (axis) major depressive disorder N-acetylaspartate N-methyl-D-aspartate across all cortical laminae.1,3,4 However, when neurons are assessed within individual cortical layers or in subgroups determined by size or immunohistochemistry, marked reductions in neuron density are found in both MDD and BPD. For example, the density of large-sized neuronal cell bodies is reduced in cortical layers II to VI in the dorsolateral prefrontal and rostral orbitofrontal cortex in MDD.5 These reductions in density of large-sized neuronal cell bodies are accompanied by increases in the density of neurons with smaller-sized cell bodies (Figure 1). The concomitant decrease in the density of large neuronal cell bodies and increase in the density of small neuronal cell bodies suggests that neuronal shrinkage/enlargement or perhaps altered neuronal development, rather than outright neuronal loss, is responsible for neuronal abnormalities in mood disorders.
Orbitofrontal cortex Cortical layers
Pia I II II

mood disorders complement and expand hypotheses of dysfunction within the monoaminergic, glutamatergic, and -aminobutyric acid (GABA) neurotransmitter systems in these disorders. While MDD and BPD are clearly not neurodegenerative disorders, impaired neuroplasticity is associated with these mood disorders. The etiology of histopathological changes observed in postmortem brain tissue is unknown. It is not clear how factors such as genetic risk factors, neurodevelopmental abnormalities, the progression of the disease, or exposure to antidepressant or mood-stabilizing medications contribute to the abnormal neuronal and glial observations in mood disorders. It remains to be determined whether the chronic administration of clinically effective therapeutic medications can reverse or even staunch histopathological changes in the mood disorders.

Control
200 I

Depressed

III IV V

II

500 VI

Alterations in neurons and glia in cerebral cortex

In MDD and BPD, reductions in neuronal density and size in some populations of cortical neurons have been independently reported.1-12 These abnormalities have been described in association cortices such as dorsolateral prefrontal, orbitofrontal, and anterior cingulate cortex, but not in the primary sensory cortical regions such as somatosensory1 or visual cortex.2 Thus, neuronal abnormalities at the microscopic level in mood disorders appear to be specific to frontolimbic cortical regions observations in postmortem tissue that are consistent with in vivo neuroimaging studies of volumetric and metabolic alterations in the same frontocortical regions. Neuronal abnormalities in mood disorders are not immediately evident, inasmuch as there is no significant reduction in the density of Nissl-stained neurons measured

White matter

Pia (m)

III

Figure 1. Changes in neuronal size and size-dependent density in layer II of rostral orbitofrontal cortex in a 73-year-old female with MDD as compared to a 71-year-old psychiatrically normal female control subject. For both subjects, the postmortem delay was less than 17 hours and fixation time was less than 10 months. Photomicrograph depicting cell composition across the six cortical layers in rostral orbitofrontal cortex (upper left). Expanded printouts of cortical layers with neuronal cell bodies represented by equivalent diameter circles with the area measured for the individual neuron in its equatorial plane (right panel). Note that neuronal sizes are smaller in layers II and III in the depressed subject than in the control subject. Note especially dramatic increases in the density of small neurons in layer II associated with significant reductions in the density of the largest neurons of this layer.
Reprinted from reference 5: Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. Copyright 1999, Elsevier.

186

Postmortem abnormalities in depression - Stockmeier and Rajkowska

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

In BPD, decreases in laminar neuronal densities have also been reported in the dorsolateral prefrontal cortex4 and anterior cingulate cortex,2,6,7 but not by all studies.1,8 Moreover, in BPD, a decrease in density of pyramidal neurons in cortical layers III and V4 and nonpyramidal neurons in layer II6 has been observed in the same cortical regions. This last observation coincides with reports on reductions in the density of layer II nonpyramidal neurons that are identified with an antibody against the calciumbinding protein, calbindin, in the anterior cingulate cortex7 and dorsolateral prefrontal cortex9 in BPD. Calbindin immunoreactive neurons are known to colocalize GABA. Our recent measurements of the density and size of calbindin-immunoreactive neurons in layer II and the upper part of layer III of the dorsolateral prefrontal cortex revealed a 43% reduction in the density of these neurons in MDD as compared to controls.10 The depression-related decrease in calbindin immunoreactive neurons, which colocalize GABA, may be closely related to in vivo clinical evidence suggesting that MDD is associated with decreased levels of GABA in cerebral cortex.11 Another manifestation of neuronal pathology in cerebral cortex in mood disorders is the reduced size of neuronal cell bodies. Smaller soma sizes have been reported in subjects with MDD, as compared to normal controls, in the dorsolateral prefrontal cortex,3,5 orbitofrontal cortex,5 and anterior cingulate cortex.8,12 Two other studies, however, did not report significant changes in neuronal size in the anterior cingulate cortex.1,2 In a manner more subtle than in MDD, reductions in neuronal soma size have been observed in BPD by some,4,12 but not by all investigators.1,2,8 In another study, a minor increase in the size of small nonpyramidal neurons was noted in the anterior cingulate cortex in BPD subjects.6 Factors leading to a reduction in the size of neuronal soma are not known. Smaller soma size may be related to smaller dendritic trees and/or abnormal morphology of synaptic contacts. However, visualization of neuronal dendritic trees in cerebral cortex using the Golgi silver impregnation method has not yet been conducted in subjects with mood disorders. Studies looking at synaptic proteins in the anterior prefrontal13 and anterior cingulate cortex14 describe reductions14 or no changes13 in synaptic proteins in mood disorders. Systematic studies of dendritic trees and synaptic contacts in prefrontal and cingulate areas are warranted to shed light on the possible etiology of smaller neuronal cell bodies in mood disorders.

The most consistent cell abnormality described in mood disorders has unexpected finding of prominent reductions in the density and number of glial cells. Glial reductions have been reported consistently by independent laboratories in the anterior cingulate cortex, dorsolateral prefrontal cortex, and orbitofrontal cortex in MDD and/or BPD subjects. For example, a 24% to 41% reduction in the number of a general population of Nissl-stained glial cells is reported in the subgenual region of the anterior cingulate cortex (ventral part of Brodmanns area 24) in a small subgroup of patients with familial MDD and familial BPD, as compared to control subjects.1 However, when data from familial and nonfamilial subgroups of patients were combined, the reductions are not found.The estimation of glial cell number in this study is combined across all six cortical layers, and no information is provided on laminar specificity of glial loss. Reductions in glial cell density, however, are reported in specific cortical layers of the anterior cingulate and prefrontal cortices in four other studies. These glial reductions are observed in layer VI of the supragenual anterior cingulate cortex,8 layers III and V of the dorsolateral prefrontal cortex3-5 and in layers III, IV, V, and VI of the caudal orbitofrontal cortex,5 in mood disorder patients. Glial cell size and shape, in addition to density, appears to be affected in mood disorders.The size of glial cell bodies (corresponding to glial cell nuclei in Nissl-stained material) has been estimated in several studies. In three of these investigations, glial size is reported as increased,3-5 whereas two other studies find glial size to be unchanged in MDD or BPD.1,15 Significant increases in glial size are observed in the dorsolateral prefrontal cortex in BPD4 and to a smaller degree in MDD,5 comparing these cohorts to psychiatrically normal control subjects. More recently, similar increases in glial size are noted in the anterior cingulate cortex in MDD.12 In addition, changes in the shape of glial nuclei to a less rounded conformation are detected in the dorsolateral prefrontal cortex in BPD.4 Reductions in glial density, paralleled by an increase in the size of glial nuclei, suggest that some compensatory mechanisms may take place in mood disorders. It can be speculated that a decrease in the density of glial cells is indicative of a decrease in the number of normally functioning glial cells. At the same time, glial cells that survive and are not damaged might be forced to play a larger role in supporting the metabolic needs of the surrounding neurons. As a consequence of increased metabolic demand, the nuclei of these glial cells might enlarge in size and change in shape.

187

Clinical research
Glutamate-induced swelling of astroglia, reported in animal cell cultures,16 may be another factor in the etiology of enlarged glial cells in depression. Glial cell pathology in mood disorders does not appear to be universally noted throughout the cerebral cortex. Changes in glial cell density or number are not found in the sensorimotor cortex in either MDD or BPD.1 Recent reports suggest a lack of marked glial pathology in the supragenual part of the anterior cingulate cortex,12 the entorhinal cortex in BPD and MDD,17 or the most rostral part of the orbitofrontal cortex in MDD (corresponding to the transitional cortex between Brodmann areas 10 and 47).5 Glial pathology in mood disorders has yet to be systematically studied in subcortical structures. Only one report suggests that glial pathology extends to limbic subcortical regions, with a significant reduction in glial number noted in the amygdala in subjects with MDD and unmedicated subjects with BPD.17 studies report no significant cell loss in any hippocampal region in any of the subject groups. In most of the depressives, there was evidence for a slight increase in fragmented DNA associated with apoptosis and necrotic neuron death detected in the dentate gyrus, CA1 and CA4.30 Decreases in astrocytic immunoreactivity for cellular GFAP and the neuron-specific phosphoprotein B50 (or GAP-45) were detected in CA1 and CA2 in depression.31 The authors suggest that apoptosis may only be a minor contributor to volume changes in the hippocampus in depression, while patterns of reactive astrogliosis and synaptic reorganization proteins are significantly altered in only some hippocampal regions in depression. Other reports of hippocampal changes in mood disorders identify a significant decrease in the density of nonpyramidal neurons in the CA2 region and a reduction in reelin-positive cell density in the hilus in subjects with BPD.32,33 Two other studies conducted on the postmortem hippocampal formation in a small sample of subjects with BPD reveal a decrease in the density and size of nonpyramidal neurons in the CA2 region and some disorganization in neuronal clusters in layers II and III of the entorhinal cortex.34,35 Neuronal and glial cell packing density and soma size were estimated recently in Nissl-stained sections including the hippocampal subfields in 16 subjects with MDD and 16 age-matched normal control subjects.36 Representative photomicrographs are presented in Figure 2. Prominent abnormalities in the CA regions and dentate gyrus are found in subjects with MDD.There is a significant increase in the mean density of pyramidal neurons in depressed subjects, as compared with normal control subjects. In the granule cell layer of the dentate gyrus, cell density is significantly increased in MDD. In addition, there is a significant decrease in the mean soma size of pyramidal neurons in depressed subjects, as compared with normal control subjects. On the basis of covariate analyses, the main findings of increased neuronal density and decreased neuron soma size in depression are not significantly altered when taking into consideration such factors as gender, age, postmortem interval, tissue pH, brain weight, smoking, antidepressant drug prescription in the last month of life, or suicide.The substantial increases noted in neuronal packing density and decrease in neuronal soma size detected in postmortem tissue may be related to the decrease in hippocampal volume noted by some in MDD. Glial pathology in depression appears to extend beyond the frontal cortex to the hippocampus. A recent study of the hippocampus in a large number of subjects with MDD

Alterations in neurons and glia in the hippocampus

Preclinical and neuroimaging studies have implicated the hippocampal formation in the pathophysiology of MDD. In addition, plasticity within the hippocampal formation may be involved in neurobiological responses to stress and to antidepressant drug action.18 Evidence for an interaction between the hippocampus and depression comes from magnetic resonance imaging (MRI) studies examining the volume of the hippocampus. Studies using MRI demonstrate reduced volume of the hippocampus in subjects with MDD or a history of MDD,19-26 but not in other studies.27-29 It appears that hippocampal atrophy is preferentially seen in older, recurrently depressed subjects or subjects who are refractory to antidepressant medications. Recently, hippocampal volume and function was assessed over the course of illness in younger patients with MDD.26 Recollection memory is diminished in subjects with either a first episode or multiple episodes of depression. However, hippocampal volume is generally considered to be significantly decreased only in older depressed subjects with multiple episodes of depression.25,26 Few studies have structurally examined the postmortem human hippocampus in depression. Cellular integrity and apoptosis have been examined in the hippocampus in subjects with depression, steroid-treated subjects, and normal control subjects.30,31 Using semiquantitative methods, these

188

Postmortem abnormalities in depression - Stockmeier and Rajkowska

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

and aged-matched normal control subjects reports a significant increase in the density of glial cells in all hippocampal CA subfields and the granule cell layer of the dentate gyrus.36 In MDD, increases in the packing density of glial cells detected in postmortem tissue suggests a potential reduction in surrounding neuropil (see above), and may be related to decreases in hippocampal volume noted by neuroimaging studies in MDD (see above). The different pattern of density change noted in depression in the hippocampus in contrast to frontal cortical areas may be related to a unique reduction in neuropil in the hippocampus in depression. Neuropil consists of the lattice of glial cells and their processes, dendrites, and proximal axons surrounding neuron cell bodies. The hypothesis of neuropil reduction in the hippocampus in MDD is supported by other postmortem studies revealing a decrease in dendritic spine density on neurons and diminished arborization of apical dendrites in the subiculum in a small group of mixed subjects with bipolar disorder or depression37 and decreased levels of synaptic proteins

found in CA4 in BPD.38 Thus, the diminished volume of the hippocampus noted by some in depression may be critically determined by a loss in neuropil including dendritic branching, dendritic spine complexity, and glial processes. The expression of brain-derived neurotrophic factor (BDNF) has been measured in the hippocampus of subjects with depression, and alterations in these factors might be related to changes in cell density and volume in depression. There is preliminary evidence that BDNF in the human hippocampus may be regulated by chronic treatment with antidepressant medications. In an immunohistochemical study of subjects with MDD and others with BPD or schizophrenia, the immunoreactivity of BDNF, as measured by optical density, is upregulated in the dentate gyrus and hilus only in subjects taking antidepressant medications (regardless of psychiatric diagnosis).39 Chen et al39 provide the first evidence beyond rodent studies that chronic antidepressant drugs upregulate the expression of BDNF in the human hippocampus. In a recent study, Dwivedi et al40 observed a significant reduction in mRNA and protein levels of BDNF in hippocampus as well as dorsolateral prefrontal cortex in suicide victims with either MDD or other psychiatric disorders. In the Dwivedi et al40 study, the decrease in expression of BDNF occurred regardless of antidepressant treatment. It remains to be determined whether alterations in BDNF are related to increases in the packing density of neurons in the hippocampal formation or prefrontal cortex. The different pattern of neuronal pathology in the frontal cortex (decrease in density) and hippocampus (increase in density) suggests unique involvement of these brain regions in the neuropathology of depression. Other evidence of dissimilarities between prefrontal cortex and hippocampus has been reported in MDD.41-43 Successful clinical treatment (or even the use of placebo) in depression was associated with an increase in metabolism in prefrontal cortex and a decrease in metabolism in hippocampus.

Figure 2. Brightfield photomicrographs of coronal sections of the postmortem human hippocampal formation. A. Cresyl violetstained coronal section from a 54-year-old male (23-h postmortem interval). B. An adjacent coronal section processed by Timm staining. Note the intensely stained granule cell layer of the dentate gyrus (DGgr) in A and B, and the clear demarcation in B between hippocampal subfields CA2 and CA3 afforded by the Timm staining. Pyramidal neurons and glial nuclei of CA3 are highlighted in C by the large white arrows and white arrowheads, respectively. Neurons and glial nuclei of the granule cell layer of the dentate gyrus are depicted in D by the large black arrows and black arrowheads, respectively. The scale bars in A and C are 750 m and 25 m, respectively.

Alterations in neurons and glia in subcortical structures

The search for morphological abnormalities in subjects with mood disorders has been less intense in subcortical structures than in cerebral cortical regions. Only a few studies in postmortem brain tissue on a relatively small number of subjects have attempted to estimate the number of neurons in such subcortical structures as hypothalamus, dorsal raphe nucleus, locus ceruleus, and amyg-

189

Clinical research
dala.44-52 Results of these subcortical histopathological studies are somewhat inconsistent. Increases, decreases, or no change in the cell number or density are reported in the hypothalamus and brain stem nuclei in depressed subjects. Stereological investigation of specific types of hypothalamic neurons reveals an increase in the numbers of arginine-vasopressin (AVP)immunoreactive neurons, oxytocin-expressing neurons, and corticotropin-releasing hormone (CRH) neurons in the paraventricular nucleus in subjects with BPD or MDD, compared to normal controls.44,45 Moreover, increases in CRH mRNA, and in the number of CRH neurons colocalizing AVP are also found in depressed patients.46,47 These findings of increases in specific immunoreactive neurons are consistent with the evidence of activation of the hypothalamicpituitary-adrenal (HPA) axis in some subsets of depressed patients.48 On the other hand, decreased number and density of nitric oxide synthasecontaining neurons in the paraventricular hypothalamic nucleus are described in a small group of subjects with either MDD or BPD.49 Subtle structural abnormalities have been reported in mood disorders in the monoaminergic brain stem nuclei, the major sources of serotonin (dorsal raphe nucleus) and norepinephrine (locus ceruleus) projections to the cerebral cortex. An increased number and density of tryptophan hydroxylase immunoreactive neurons is observed in the dorsal raphe nucleus of suicide victims with MDD compared with controls.50 In suicide victims, Arango et al51 report fewer pigmented neurons within the rostral locus ceruleus. Another study in a larger number of subjects found no differences in the number of pigmented neurons in the locus ceruleus between subjects with MDD (most were suicides) and control subjects.52 Although the number of neurons in the locus ceruleus does not appear altered in MDD, CRH immunoreactivity is increased in the locus ceruleus and pontine dorsal and median raphe nuclei.53,54 No changes in neuronal densities were detected in amygdala in subjects with either MDD or BPD, as compared to normal controls.17 These postmortem findings suggest that some changes in the morphology of hypothalamic neurons and brain stem neurons may take place in mood disorders. However, future studies employing stereological techniques and a larger number of subjects are required to determine the exact pathology in these regions in depression.

Functional implications of pathological changes in neural circuits

Morphological abnormalities detected postmortem in mood disorders are most likely related to dysfunction of neural circuits regulating emotional, cognitive, and somatic symptoms exhibited by subjects with MDD or BPD. In fact, alterations in neuronal density and size have been found in the dorsolateral prefrontal, orbitofrontal, and anterior cingulate cortex, the neurons of which give rise to the frontal circuits critical for higher cognitive and limbic functioning.55 Subtle neuronal alterations are also reported in the hypothalamus and hippocampus, further evidence of dysfunction in limbic circuits in depression. Some of the cellular abnormalities detected postmortem in cortical and subcortical structures in MDD and BPD may be related to disruption of monoaminergic transmission in depression. Studies in postmortem brain tissue identify alterations in serotonin and norepinephrine receptors and transporters in the dorsolateral prefrontal cortex and ventrolateral/orbitofrontal cortex in brains from suicide victims with or without clinical depression.56 These cortical regions also exhibit abnormal cell density and size in cell-counting studies of postmortem tissue. For example, cellular changes found in superficial layers of the prefrontal cortex in depressed subjects may be related to alterations in serotonin-1A receptors in superficial layers of cortex in suicide victims.57 In a neuroimaging study, the authors find that radioligand binding to serotonin-1A receptors is decreased in medication-free subjects with MDD in several cortical regions, including medial temporal cortex, the temporal pole, orbitofrontal cortex, anterior cingulate cortex, insula and dorsolateral prefrontal cortex.58 Expression of another component of serotonin neurotransmission, the serotonin transporter, is also decreased in the dorsolateral prefrontal and ventral/orbitofrontal cortex in postmortem brains from depressed suicide victims.59,60 Detailed laminar analysis of the density of serotonin transporterimmunoreactive axons reveals that this deficit in depression is localized in cortical layer VI of the dorsolateral prefrontal cortex.59 The serotonin-transporter deficit may be related to the pathology of layer VI neurons reported in the same cortical layer by postmortem cell-counting studies in depression. Moreover, subtle neuronal abnormalities reported by some studies in the monoaminergic brain stem nuclei

190

Postmortem abnormalities in depression - Stockmeier and Rajkowska

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

suggest dysfunction of monoaminergic projections originating from the brain stem neurons and terminating in frontolimbic cortical regions. It is likely that the functions and morphology of cortical neurons are affected by alterations in the functional state of noradrenergic, serotonergic, and dopaminergic neurons that project axons to prefrontal and anterior cingulate cortex. Postmortem neurochemical studies in MDD report alterations in noradrenergic 2-adrenergic receptors and the norepinephrine transporter, as well as levels of tyrosine hydroxylase in the locus ceruleus,52,61,62 serotonin-1A receptors in the midbrain dorsal raphe nucleus,63 and dopaminergic receptors and transporters in the amygdala.64 The layer-specific changes in neuronal density and size identified in mood disorders implies that both inhibitory local circuit neurons and excitatory projection types of cortical neurons may be involved in the neuropathology of mood disorders. Nonpyramidal inhibitory neurons using GABA are localized mainly in cortical layer II and establish local corticocortical connections within or between adjacent functional columns of cortical cells. In contrast, pyramidal glutamatergic excitatory neurons reside predominantly in cortical layers III, V, and VI and give rise to long projections to other cortical associational regions (layer III), striatum (layer V), and thalamus (layer VI). Neuronal pathology detected in cortical layers III, V, and VI of the dorsolateral prefrontal cortex and anterior cingulate cortex in MDD may be associated with the pathology of excitatory pyramidal neurons within these laminae that use glutamate as their neurotransmitter. Moreover, the density of pyramidal neurons is selectively reduced in the dorsolateral prefrontal cortex in subjects with BPD,4 further confirming the pathology of glutamatergic neurons in mood disorders. These findings in postmortem brain tissue coincide with an in vivo proton magnetic resonance spectroscopy study in the anterior cingulate cortex revealing a reduction in glutamate levels in depression.65 There is increasing preclinical and clinical evidence that antidepressant drugs directly or indirectly reduce the function of N-methyl-D-aspartate (NMDA) glutamate receptors.66 Depression-related decreases in glutamate levels or the density of glutamatergic pyramidal neurons may alter in cortex and elsewhere the glutamatergic recognition site and its coupling to the NMDA receptor complex. One study of suicide victims, some of whom were diagnosed with MDD, reveals changes in the glutamatergic recognition site and

its coupling to the NMDA receptor complex in the anterior prefrontal cortex.67 Interestingly, drugs that reduce glutamatergic activity or glutamate receptorrelated signal transduction may also have antimanic effects.66 Reductions in size and density of layer II neurons in the orbitofrontal and dorsolateral prefrontal cortex, as well as reductions in the density of nonpyramidal neurons in layer II of the anterior cingulate cortex suggest deficient GABAergic neurotransmission. Most nonpyramidal neurons in cortical layer II colocalize GABA and recent clinical evidence suggests that MDD is associated with decreased levels of cortical GABA.11 In summary, the localization of morphological abnormalities in the mood disorders occurs in prefrontolimbic circuits that are likely to regulate emotional, cognitive, and somatic symptoms in depression. The observation in the mood disorders of neuronal pathology in specific cortical layers gives support to the hypotheses that the monoamine, glutamate, and GABA neurotransmitter systems are involved in the pathophysiology of these disorders. It remains to be determined whether the cellular pathology is the reason for, or the consequence of, depression.

Functional implications of glial abnormalities in depression

The glial cells analyzed in the above studies do not represent a homogeneous population of cells. Glial cells are composed of distinct populations of oligodendrocytes, microglia, and astrocytes. The crucial role of glial cell types in brain function is currently being reevaluated. In addition to their traditional roles in neuronal migration (radial glia), myelin formation (oligodendrocytes), and inflammatory processes (astrocytes and microglia), glia (predominantly astrocytes) are now thought to provide trophic support to neurons, neuronal metabolism, and the formation of synapses and neurotransmission.15 The three distinct glial cell types cannot be identified in the previously mentioned studies as those tissues were stained for Nissl substance and such staining does not distinguish reliably between types of glial cells. Nissl staining only reveals morphological features of glial cell bodies and not glial cell processes. On the other hand, recent immunohistochemical examination of glial fibrillary acidic protein (GFAP), a marker of reactive astroglia, in the dorsolateral prefrontal cortex implicates astrocytes in the overall glial pathology in MDD.68 Although no significant group dif-

191

Clinical research
ferences in the packing density of GFAP-reactive astrocytes are present in this study, there is a significant correlation between age and GFAP immunoreactivity among subjects with MDD, when the entire group of MDD (young and old) is compared with normal controls. A significant reduction in the population of reactive astroglia is found in a small subgroup of young (30 to 45 years old) subjects with MDD, as compared to young control subjects and older (46 to 86 years old) subjects with MDD (Figures 3A and 3B). This subgroup of younger adults with MDD also had a shorter duration of depression and most of these subjects were suicide victims. Recent observations from our laboratory confirm that the levels of GFAP protein are also reduced in these young adults with MDD as compared to age-matched control subjects (Figures 3C and 3D), and that GFAP levels are positively correlated with age at the time of death and with the age of onset of depression.69 Thus, the involvement of GFAP expression in early- versus late-life depression differs because the underlying pathophysiology in early-life depression is different from that in late-life depression. Clinical evidence confirms that late-onset depression (first depressive episode when older than 50 years) differs from early-onset depression by its etiology, phenomenology, and cerebrovascular pathology.70-72 Alterations in GFAP in both BPD and MDD are also suggested by a proteomic study in which different forms of GFAP proteins displayed disease-specific abnormalities.73 Oligodendrocytes may also be involved in the cellular pathology of depression. In both the dorsolateral prefrontal and anterior frontal cortex in subjects with BPD or MDD, there are ultrastructural changes in oligodendrocytes and there is a reduction in the density and immunoreactivity of these cells.74,75 Moreover, key oligodendrocyte-related and myelin-related gene expression is reduced in the dorsolateral prefrontal cortex in BPD.76 While these results are intriguing, further immunohistochemical and molecular studies are needed to definitively determine which specific glial cell types are compromised in BPD and whether the same or different types of glial cells are involved in the pathology reported in MDD. Reductions in glial number and density, in addition to changes in size and shape, might be related to the dysfunction of monoamine and glutamate systems reported extensively in depression. For example, astrocytes express virtually all of the receptor systems, ion channels, and

GFAP immunohistochemistry A
100

GFAP protein level C

B
100 r=0.890 P<0.0005

Area fraction (%)

Area fraction (%)

80

80

60

60

Controls (70) Depressed (73)

40

40

D
GFAP

20

20

0 20

Actin
30 40 50 60 70 80 90

Controls

Depressed

Depressed age (years)

Controls (27) Depressed (34)

Figure 3. An illustration of the pathology of glial cells found in the dorsolateral prefrontal cortex in MDD.5,68 Reductions in the glial fibrillary acidic protein (GFAP) immunoreactive astroglia are found in a subgroup of young adults with major depression as compared to aged-matched control subjects and older subjects with major depression (A) and these reductions are correlated with the age of the subjects at the time of death (B). Recent preliminary observations (Si et al, unpublished observation) indicate that the levels of GFAP protein in the same area of the dorsolateral prefrontal cortex are also reduced in these young (D) but not old (C) subjects with major depression as compared to age-matched control subjects. Note that the level of actin, another protein in brain, is unchanged in a depressed subject as compared to the control.
Reproduced (A and B) from reference 68: Miguel-Hidalgo JJ, Baucom C, Dilley G, et al. Glial fibrillary acidic protein immunoreactivity in the prefrontal cortex distinguishes younger from older adults in major depressive disorder. Biol Psychiatry. 2000;48:861-873. Copyright 2000, Elsevier.

192

Postmortem abnormalities in depression - Stockmeier and Rajkowska

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

transporters found in neurons.15 Thus, the postsynaptic monoaminergic receptors distributed on glial cell bodies and processes may play a role in serotonin, norepinephrine, or dopamine neurotransmission. Moreover, astroglia are the primary sites of glutamate uptake by glial transporters and are important in regulating NMDA receptor activity. Astroglia regulate the levels of extracellular glutamate and thereby protect neurons in vitro from cell death and provide energy for neurons. Astrocytic pathology in MDD may indirectly promote glutamate-mediated neuronal excitotoxicity, with consequences that may be detected by functional neuroimaging. A mounting body of data suggests that treatment with antidepressant or mood-stabilizing medications regulates neuronal survival and also influences neurogenesis. Pharmacologically induced increases in neurogenesis in adult rodent brain have been reported in two independent studies.77,78 Moreover, there is evidence that treatment with lithium induces an increase in the astrocytic protein GFAP in rodent hippocampus79,80 and the neural lobe of the pituitary.81 However, whether these increases represent a protective or compensatory effect of these medications, and the mechanisms underlying the regulation of neurogenesis and glial proliferation have to be further investigated. Furthermore, a precise link between cell loss and atrophy, observed in the postmortem human brain, and medication-induced production of new cells, observed in the animal brain, has yet to be established.

Limitations in postmortem pathology studies in mood disorders

Postmortem studies cannot yet clearly define whether a true loss of cells underlies prominent reductions in cell density and size detected in mood disorders. For the estimation of a total number of neurons or glia in a particular brain region, it is essential that the total volume of a studied area be calculated. To measure the entire volume, the exact borders of the studied region have to be established,82,83 so that sampling is confined to the region within these borders. Unfortunately, in most studies of mood disorders in postmortem tissue, limited availability of the complete tissue region, as well as limitations in reliably distinguishing cytoarchitectonic borders of a studied region, have prevented the estimation of a total tissue volume and, consequently, total cell number. In one study where the total cell number was estimated in the subgenual cortex, a loss of glial but not neuronal cells has been demon-

strated in familial mood disorders.1 Glial reductions reported in this study may in fact reflect a true loss of glial cells since the neuroimaging studies in the same cortical region show a reduction in the volume of gray matter.84 There are unquestionable limitations to the use of postmortem brain tissue in studying the mood disorders.56 Some of the critical issues to be considered when interpreting the studies of postmortem brain tissue include the psychiatric status of the subject at the time of death and the underlying psychiatric disorder, whether control subjects were psychiatrically normal, the cause of death of the subjects (suicide or by other means), evolving criteria used to establish psychiatric diagnoses, the possible inclusion of subjects with concurrent psychoactive substance use disorders, the regional and hemispheric localization of the brain regions being studied, and the presence and duration of treatment with a psychotropic medication. Other frequent drawbacks to studies of postmortem brain tissue include low numbers of subjects per cohort, or inadequate expertise in cytoarchitectonic delineation of individual brain regions. Ideally, longitudinal clinical studies on wellcharacterized patients should be linked to subsequent postmortem studies of the same subjects. It is important to seek to control for the potential effects of suicide on postmortem biological observations in depression. In two of our studies,5,36 enough depressed nonsuicide subjects were available to tentatively determine that the main findings of these studies appear to persist regardless of whether the depressed subjects died by suicide or natural causes.While suicide makes tissues available for most postmortem studies of depression, the results obtained with this cohort must be cautiously interpreted since the majority of living individuals with depression do not attempt or commit suicide. In the mood disorders, the alterations in cell density and size are likely to be related to the disorder itself and not to the age of subjects at the time of death, postmortem delay, or the time of fixation of the tissue. Statistical analyses conducted in all of the above morphometric studies yielded no significant correlation between cell density or size and any of these confounding variables. It cannot be ruled out, however, that some of the cellular alterations in mood disorders are related to prior treatment with antidepressants and lithium (for further discussion see reference 85). The question of whether cell abnormalities can be attributed to the effect of therapeutic medications is open to debate. There have been no systematic studies on the effect of antidepressant and mood-stabilizing medications

193

Clinical research
on cell number and morphology in the postmortem human brain, most likely due to an insufficient number of treated versus untreated subjects. in glutamate-glutamine-GABA metabolites in the adult anterior cingulate cortex of these animals were also observed 10 years after the stressors.These NAA measures reflect neuronal integrity and metabolism whereas changes in glutamate-glutamine-GABA metabolites may reflect changes in membrane structure, glial functions, and glutamate content. Together, the above data suggest that structural and metabolic alterations observed in vivo may be related to alterations in cell viability, which, itself, may be related to alterations in cell number, density, and size observed in postmortem tissues at the microscopic level. The studies reviewed above undeniably prove the usefulness of postmortem tissue in unraveling the microscopic anatomical substrate of depression. For the first time, postmortem cell-counting studies in mood disorders have established that MDD and BPD are brain diseases with unique pathological alterations in neuronal and glial cells. The precise region- and layer-specific alterations in neuronal and glial architecture observed in mood disorders are consistent with the hypotheses of specific dysfunction in monoamine, glutamate, and GABA neurotransmitter systems in these disorders. Moreover, colocalization of cellular changes detected in postmortem tissues with in vivo neuroimaging findings proves that postmortem studies provide an important interface between clinical and basic research in unraveling the neuroanatomical substrates of depression. Postmortem studies in depression also indicate that while MDD and BPD are clearly not neurodegenerative disorders, these disorders are associated with impaired cellular neuroplasticity and resilience. It remains to be fully elucidated to what extent these findings represent neurodevelopmental abnormalities, progression of the disorder, biochemical changes (in glucocorticoid or trophic factors levels) accompanying repeated disease episodes, or the results of treatment with therapeutic medications. It is unknown whether the cellular changes observed postmortem in mood disorders can be reversed by antidepressant and mood-stabilizing medications. Although molecular and genetic mechanisms associated with depression are yet to be unraveled, preliminary microarray studies of gene expression in postmortem brain tissues from subjects with mood disorders confirm that the dorsolateral prefrontal and anterior cingulate cortex are sites of pathology in mood disorders.93,94
The authors acknowledge the support of the National Alliance for Research on Schizophrenia and Depression, and Public Health Service Grants MH60451, MH61578, MH63187, MH67996, and P20 RR17701.

Conclusion
Cellular abnormalities in mood disorders are observed in the dorsolateral prefrontal cortex, anterior cingulate cortex, orbitofrontal cortex, hippocampus, and amygdala. In these same brain regions, neuroimaging studies reveal volumetric, metabolic, and neurochemical alterations in subjects with mood disorders. Structural neuroimaging studies in mood disorders provide evidence of modest but intriguing volumetric changes that suggest cell loss and/or atrophy.86 Some studies, but not all, report enlargement of the lateral and third ventricles in mood disorders87 that may be indicative of atrophy of surrounding cortical and subcortical regions. Functional neuroimaging studies in MDD and BPD lend further support to physiological abnormalities in cortical and subcortical frontolimbic regions. Abnormal regulation of glucose metabolism, regional cerebral blood flow, and high-energy phosphate metabolism are observed in the prefrontal and temporal cortex, basal ganglia, and amygdala in mood disorders.88 Neuroimaging studies that examine neurochemical changes in the living brain provide further support for the hypothesis that mood disorders are associated with changes in cell viability and function. For example, high-resolution magnetic resonance spectroscopy in unmedicated subjects with BPD report decreased N-acetylaspartate (NAA) levels bilaterally in the hippocampus89 and in the dorsolateral prefrontal cortex,90 as compared to healthy controls. In contrast, therapeutic doses of lithium increase levels of NAA in the brain of subjects with BPD.91 Such increases in NAA are found in a number of regions including frontal cortex, and are localized almost exclusively in the gray matter. NAA is regarded as a measure of neuronal viability and function, and therefore the changes in NAA levels seen in BPD strongly implicate alterations in neuronal viability, which may be related to alterations in cell number, cell density, and size, and related volumetric changes. Interestingly, recent magnetic resonance spectroscopic studies of nonhuman primates exposed to early life stressors or repeated stressors also reveal a significant decrease in NAA. The NAA decrease in the animals exposed to repeated stressors was normalized by chronic treatment with the antidepressant tianeptine.92 Increases

194

Postmortem abnormalities in depression - Stockmeier and Rajkowska

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Alteraciones celulares en la depresin: evidencia proveniente de tejido cerebral postmortem

Durante los ltimos veinte aos, los estudios de neuroimgenes in vivo han permitido un importante conocimiento acerca de la ubicacin general de regiones cerebrales disfuncionales en la depresin. En paralelo y a menudo intersectndose, los estudios neuroanatmicos, farmacolgicos y bioqumicos del tejido cerebral postmortem estn facilitando nuevos conocimientos sobre la fisiopatologa de la depresin. Adems de las alteraciones neuroqumicas reconocidas desde hace bastante tiempo, nuevos estudios a nivel microscpico permiten sostener que los trastornos afectivos estn asociados con alteraciones en la morfologa y en la distribucin celular. En los ltimos sis aos, estudios de recuento celular han identificado cambios en la densidad y el tamao tanto de las neuronas como de la gla en varias regiones cerebrales frontolmbicas que incluyen las cortezas dorso-lateral prefrontal, rbito-frontal y cingulada anterior, y la amgdala y el hipocampo. La convergencia de cambios celulares a nivel microscpico con cambios en las neuroimgenes detectados in vivo provee una integracin forzada de la investigacin clnica y bsica para desentraar la fisiopatologa de la depresin. La integracin definitiva de estas dos aproximaciones de investigacin ocurrir cuando se puedan relacionar los estudios clnicos longitudinales premortem en pacientes bien caracterizados con los estudios postmortem de los mismos sujetos.

Anomalies cellulaires observes dans le tissu crbral post mortem au cours de la dpression
Ces 20 dernires annes, les tudes de neuro-imagerie in vivo ont permis des avances significatives dans la comprhension de la localisation gnrale des rgions crbrales dysfonctionnelles au cours de la dpression. Les tudes neuroanatomiques, pharmacologiques et biochimiques des tissus crbraux post mortem offrent ainsi, en parallle et souvent de faon croise, un aperu renouvel de la physiopathologie de la dpression. De nouvelles tudes au niveau microscopique ont confort lhypothse selon laquelle les troubles de lhumeur sont associs des anomalies de la distribution et de la morphologie cellulaires, tout en confirmant lexistence danomalies neurochimiques connues depuis longtemps dans la dpression. Ces 6 dernires annes, des tudes de comptage cellulaire ont identifi des modifications dans la densit et la taille des neurones et de la nvroglie au niveau dun certain nombre de rgions crbrales frontolimbiques comprenant les cortex dorsolateral prfrontal, orbitofrontal et cingulaire antrieur, ainsi quau niveau de lamygdale et de lhippocampe. La convergence entre les modifications cellulaires observes au niveau microscopique et les changements in vivo dtects par neuro-imagerie tmoigne avec loquence de lutilit dassocier la recherche clinique et fondamentale en vue dlucider la physiopathologie de la dpression. Lultime intgration de ces deux approches de recherche consistera rapprocher les donnes issues dtudes cliniques longitudinales pre mortem sur des patients bien dfinis avec celles provenant dtudes post mortem sur ces mmes patients.

REFERENCES
1. Ongur D, Drevets WC, Price JL. Glial reduction in the subgenual prefrontal cortex in mood disorders. Proc Natl Acad Sci U S A. 1998;95:13290-13295. 2. Bouras C, Kovari E, Hof PR, et al. Anterior cingulate cortex pathology in schizophrenia and bipolar disorder. Acta Neuropathol (Berl). 2001;102:373379. 3. Cotter D, Mackay D, Chana G, et al. Reduced neuronal size and glial cell density in area 9 of the dorsolateral prefrontal cortex in subjects with major depressive disorder. Cereb Cortex. 2002;12:386-394. 4. Rajkowska G, Halaris A, Selemon LD. Reductions in neuronal and glial density characterize the dorsolateral prefrontal cortex in bipolar disorder. Biol Psychiatry. 2001;49:741-752.

5. Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. 6. Benes FM, Vincent SL, Todtenkopf M. The density of pyramidal and nonpyramidal neurons in anterior cingulate cortex of schizophrenic and bipolar subjects. Biol Psychiatry. 2001;50:395-406. 7. Cotter D, Landau S, Beasley C, et al. The density and spatial distribution of GABAergic neurons, labelled using calcium binding proteins, in the anterior cingulate cortex in major depressive disorder, bipolar disorder, and schizophrenia. Biol Psychiatry. 2002;51:377-386. 8. Cotter D, Mackay D, Landau S, et al. Reduced glial cell density and neuronal size in the anterior cingulate cortex in major depressive disorder. Arch Gen Psychiatry. 2001;58:545-553.

195

Clinical research
9. Reynolds GP, Zhang ZJ, Patten I, et al. Selective deficits of frontal cortical GABAergic neuronal subtypes defined by calcium binding proteins in psychotic illness. Schizophr Res. 2000;41:255. 10. Rajkowska G, O'Dwyer G, Shao Q, et al. Calbindin immunoreactive nonpyramidal neurons are reduced in the dorsolateral prefrontal cortex in major depression and schizophrenia. Society for Neuroscience, 32nd Annual Meeting, Orlando, Fla. Program No. 497.20. 2002 Abstract Viewer/ Itinerary Planner. Washington, DC: Society for Neuroscience; 2002. Available at http://sfn.scholarone.com/itin2002/. Accessed 5 May 2004. 11. Sanacora G, Mason GF, Krystal JH. Impairment of GABAergic transmission in depression: new insights from neuroimaging studies. Crit Rev Neurobiol. 2000;14:23-45. 12. Chana G, Landau S, Beasley C, et al. Two-dimensional assessment of cytoarchitecture in the anterior cingulate cortex in major depressive disorder, bipolar disorder, and schizophrenia: evidence for decreased neuronal somal size and increased neuronal density. Biol Psychiatry. 2003;53:1086-1098. 13. Honer WG, Falkai P, Chen C, et al. Synaptic and plasticity-associated proteins in anterior frontal cortex in severe mental illness. Neuroscience. 1999;91:1247-1255. 14. Eastwood SL, Harrison PJ. Synaptic pathology in the anterior cingulate cortex in schizophrenia and mood disorders. A review and a Western blot study of synaptophysin, GAP-43 and the complexins. Brain Res Bull. 2001;55:569-578. 15. Cotter DR, Pariante CM, Rajkowska G. Glial pathology in major psychiatric disorders. In: Agam G, Everall IP, Belmaker RH, eds. The Postmortem Brain in Psychiatric Research. Boston, Mass: Kluwer Academic Publishers; 2002:49-73. 16. Hansson E, Muyderman H, Leonova J, et al. Astroglia and glutamate in physiology and pathology: aspects on glutamate transport, glutamateinduced cell swelling and gap-junction communication. Neurochem Int. 2000;37:317-329. 17. Bowley MP, Drevets WC, Ongur D, et al. Low glial numbers in the amygdala in major depressive disorder. Biol Psychiatry. 2002;52:404-412. 18. Duman RS, Malberg J, Thome J. Neural plasticity to stress and antidepressant treatment. Biol Psychiatry. 1999;46:1181-1191. 19. Sheline YI, Wang PW, Gado MH, Csernansky JG, Vannier MW. Hippocampal atrophy in recurrent major depression. Proc Natl Acad Sci U S A. 1996;93:3908-3913. 20. Shah PJ, Ebmeier KP, Glabus MF, Goodwin GM. Cortical grey matter reductions associated with treatment-resistant chronic unipolar depression. Controlled magnetic resonance imaging study. Br J Psychiatry. 1998;172:527532. 21. Sheline YI, Sanghavi M, Mintun MA, et al. Depression duration but not age predicts hippocampal volume loss in medically healthy women with recurrent major depression. J Neurosci. 1999;19:5034-5043. 22. Bremner JD, Narayan M, Anderson ER, et al. Hippocampal volume reduction in major depression. Am J Psychiatry. 2000;157:115-118. 23. Mervaala E, Fohr J, Kononen M, et al. Quantitative MRI of the hippocampus and amygdala in severe depression. Psychol Med. 2000;30:117125. 24. Steffens DC, Byrum CE, McQuoid DR, et al. Hippocampal volume in geriatric depression. Biol Psychiatry. 2000;48:301-309. 25. Frodl T, Meisenzahl EM, Zetzsche T, et al. Hippocampal changes in patients with a first episode of major depression. Am J Psychiatry. 2002;159:1112-1118. 26. MacQueen GM, Campbell S, McEwen BS, et al. Course of illness, hippocampal function, and hippocampal volume in major depression. Proc Natl Acad Sci U S A. 2003;100:1387-1392. 27. Vakili K, Pillay SS, Lafer B, et al. Hippocampal volume in primary unipolar major depression: a magnetic resonance imaging study. Biol Psychiatry. 2000;47:1087-1090. 28. Rusch BD, Abercrombie HC, Oakes TR, Schaefer SM, Davidson RJ. Hippocampal morphometry in depressed patients and control subjects: relations to anxiety symptoms. Biol Psychiatry. 2001;50:960-964. 29. Posener JA, Wang L, Price JL, et al. High-dimensional mapping of the hippocampus in depression. Am J Psychiatry. 2003;160:83-89. 30. Lucassen PJ, Muller MB, Holsboer F, et al. Hippocampal apoptosis in major depression is a minor event and absent from subareas at risk for glucocorticoid overexposure. Am J Pathol. 2001;158:453-468. 31. Mller MB, Lucassen PJ, Yassouridis A, et al. Neither major depression nor glucocorticoid treatment affects the cellular integrity of the human hippocampus. Eur J Neurosci. 2001;14:1603-1612. 32. Benes FM, Kwok EW, Vincent SL, et al. A reduction of nonpyramidal cells in sector CA2 of schizophrenics and manic depressives. Biol Psychiatry. 1998;44:88-97. 33. Fatemi SH, Earle JA, McMenomy T. Reduction in reelin immunoreactivity in hippocampus of subjects with schizophrenia, bipolar disorder and major depression. Mol Psychiatry. 2000;56:654-663. 34. Beckmann H, Jakob H. Prenatal disturbances of nerve cell migration in the entorhinal region: a common vulnerability factor in functional psychoses? J Neural Transm. 1991;84:155-164. 35. Bernstein HG, Krell D, Baumann B, et al. Morphometric studies of the entorhinal cortex in neuropsychiatric patients and controls: clusters of heterotopically displaced lamina II neurons are not indicative of schizophrenia. Schizophr Res. 1998;33:125-132. 36. Stockmeier CA, Mahajan GJ, Konick L, et al. Neural and glial density is increased and neural soma size is decreased in hippocampus in major depressive disorder. Biol Psychiatry. 2003;53(suppl 8):198. 37. Rosoklija G, Toomayan G, Ellis SP, et al. Structural abnormalities of subicular dendrites in subjects with schizophrenia and mood disorders: preliminary findings. Arch Gen Psychiatry. 2000;57:349-356. 38. Harrison PJ, Eastwood SL. Neuropathological studies of synaptic connectivity in the hippocampal formation in schizophrenia. Hippocampus. 2001;11:508-519. 39. Chen B, Dowlatshahi D, MacQueen GM, et al. Increased hippocampal BDNF immunoreactivity in subjects treated with antidepressant medication. Biol Psychiatry. 2001;50:260-265. 40. Dwivedi Y, Rizavi HS, Conley RR, Roberts RC, Tamminga CA, Pandey GN. Altered gene expression of brain-derived neurotrophic factor and receptor tyrosine kinase B in postmortem brain of suicide subjects. Arch Gen Psychiatry. 2003;60:804-815. 41. Mayberg HS, Brannan SK, Tekell JL, et al. Regional metabolic effects of fluoxetine in major depression: serial changes and relationship to clinical response. Biol Psychiatry. 2000;48:830-843. 42. Kennedy SH, Evans KR, Kruger S, et al. Changes in regional brain glucose metabolism measured with positron emission tomography after paroxetine treatment of major depression. Am J Psychiatry. 2001;158:899-905. 43. Mayberg HS, Silva JA, Brannan SK, et al. The functional neuroanatomy of the placebo effect. Am J Psychiatry. 2002;159:728-737. 44. Purba JS, Hoogendijk WJ, Hofman MA, et al. Increased number of vasopressin- and oxytocin-expressing neurons in the paraventricular nucleus of the hypothalamus in depression. Arch Gen Psychiatry. 1996;53:137-143. 45. Raadsheer FC, Hoogendijk WJ, Stam FC, et al. Increased numbers of corticotropin-releasing hormone expressing neurons in the hypothalamic paraventricular nucleus of depressed patients. Neuroendocrinology. 1994;60:436444. 46. Raadsheer FC, van Heerikhuize JJ, Lucassen PJ, et al. Corticotropin-releasing hormone mRNA levels in the paraventricular nucleus of patients with Alzheimers disease and depression. Am J Psychiatry. 1995;152:1372-1376. 47. Swaab DF, Hofman MA, Lucassen PJ, et al. Functional neuroanatomy and neuropathology of the human hypothalamus. Anat Embryol (Berl). 1993;187:317-330. 48. Holsboer F, Spengler D, Heuser I. The role of corticotropin-releasing hormone in the pathogenesis of Cushing's disease, anorexia nervosa, alcoholism, affective disorders and dementia. Prog Brain Res. 1992;93:385-417. 49. Bernstein HG, Stanarius A, Baumann B, et al. Nitric oxide synthase-containing neurons in the human hypothalamus: reduced number of immunoreactive cells in the paraventricular nucleus of depressive patients and schizophrenics. Neuroscience. 1998;83:867-875. 50. Underwood MD, Khaibulina AA, Ellis SP, et al. Morphometry of the dorsal raphe nucleus serotonergic neurons in suicide victims. Biol Psychiatry. 1999;46:473-483. 51. Arango V, Underwood MD, Mann JJ. Fewer pigmented locus coeruleus neurons in suicide victims: preliminary results. Biol Psychiatry. 1996;39:112-120. 52. Klimek V, Stockmeier C, Overholser J, et al. Reduced levels of norepinephrine transporters in the locus coeruleus in major depression. J Neurosci. 1997;17:8451-8458.

196

Postmortem abnormalities in depression - Stockmeier and Rajkowska

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

53. Austin MC, Janosky JE, Murphy HA. Increased corticotropin-releasing hormone immunoreactivity in monoamine-containing pontine nuclei of depressed suicide men. Mol Psychiatry. 2003;8:324-332. 54. Bissette G, Klimek V, Pan J, Stockmeier C, Ordway G. Elevated concentrations of CRF in the locus coeruleus of depressed subjects. Neuropsychopharmacology. 2003;28:1328-1335. 55. Alexander GE, Crutcher MD, DeLong MR. Basal ganglia-thalamocortical circuits: parallel substrates for motor, oculomotor, prefrontal and limbic functions. Prog Brain Res. 1990;85:119-146. 56. Stockmeier CA, Jurjus G. Monoamine receptors in postmortem brain: do postmortem brain studies cloud or clarify our understanding of the affective disorders. In: Agam G, Everall IP, Belmaker RH, eds. The Postmortem Brain in Psychiatric Research. Boston, Mass: Kluwer Academic Publishers; 2002:363. 57. Arango V, Underwoood MD, Gubbi AV, et al. Localized alterations in preand postsynaptic serotonin binding sites in the ventrolateral prefrontal cortex of suicide victims. Brain Res. 1995;688:121-133. 58. Sargent PA, Kjaer KH, Bench CJ, et al. Brain serotonin1A receptor binding measured by positron emission tomography with [11C]WAY-100635: effects of depression and antidepressant treatment. Arch Gen Psychiatry. 2000;57:174180. 59. Austin M, Whitehead R, Edgar C, et al. Localized decrease in serotonin transporter-immunoreactive axons in the prefrontal cortex of depressed subjects committing suicide. Neuroscience. 2002;114:807. 60. Mann JJ, Huang YY, Underwood MD, et al. A serotonin transporter gene promoter polymorphism (5-HTTLPR) and prefrontal cortical binding in major depression and suicide. Arch Gen Psychiatry. 2000;57:729-738. 61. Ordway GA, Widdowson PS, Smith KS, et al. Agonist binding to 2-adrenoceptors is elevated in the locus coeruleus from victims of suicide. J Neurochem. 1994;63:617-624. 62. Zhu M-Y, Klimek V, Dilley GE, et al. Elevated levels of tyrosine hydroxylase in the locus coerleus in major depression. Biol Psychiatry. 1999;46;1275-1286. 63. Stockmeier CA, Shapiro LA, Dilley GE, et al. Increase in serotonin-1A autoreceptors in the midbrain of suicide victims with major depression-postmortem evidence for decreased serotonin activity. J Neurosci. 1998;18:7394-7401. 64. Klimek V, Schenck JE, Han H, et al. Dopaminergic abnormalities in amygdaloid nuclei in major depression: a postmortem study. Biol Psychiatry. 2002;52:740-748. 65. Auer DP, Putz B, Kraft E, et al. Reduced glutamate in the anterior cingulate cortex in depression: an in vivo proton magnetic resonance spectroscopy study. Biol Psychiatry. 2000;47:305-313. 66. Krystal JH, Sanacora G, Blumberg H, et al. Glutamate and GABA systems as targets for novel antidepressant and mood- stabilizing treatments. Mol Psychiatry. 2002;7:S71-S80. 67. Nowak G, Ordway GA, Paul IA. Alterations in the N-methyl-D-aspartate (NMDA) receptor complex in the frontal cortex of suicide victims. Brain Res. 1995;675:157-164. 68. Miguel-Hidalgo JJ, Baucom C, Dilley G, et al. Glial fibrillary acidic protein immunoreactivity in the prefrontal cortex distinguishes younger from older adults in major depressive disorder. Biol Psychiatry. 2000;48:861-873. 69. Si X, Miguel-Hidalgo JJ, Rajkowska G. GFAP expression is reduced in the dorsolateral prefrontal cortex in depression. Society for Neuroscience, 33rd Annual Meeting, New Orleans, La. Program No. 640.8. 2003 Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience, 2003. Online. Available at http://sfn.scholarone.com/itin2003/. Accessed 5 May 2004. 70. Heun R, Kockler M, Papassotiropoulos A. Distinction of early- and late-onset depression in the elderly by their lifetime symptomatology. Int J Geriatr Psychiatry. 2000;15:1138-1142. 71. Krishnan K, Hays J, Tupler L, et al. Clinical and phenomenological comparisons of late-onset and early-onset depression. Am J Psychiatry. 1995; 152:785-788. 72. Lavretsky H, Lesser IM, Wohl M, et al. Relationship of age, age at onset, and sex to depression in older adults. Am J Geriatr Psychiatry. 1998;6:248-256.

73. Johnston-Wilson NL, Sims CD, Hofmann JP, et al. Disease-specific alterations in frontal cortex brain proteins in schizophrenia, bipolar disorder, and major depressive disorder. The Stanley Neuropathology Consortium. Mol Psychiatry. 2000;5:142-149. 74. Orlovskaya DD, Vostrikov VM, Rachmanova NA, et al. Decreased numerical density of oligodendroglial cells in postmortem prefrontal cortex in schizophrenia, bipolar affective disorder and major depression. Schizophr Res. 2000;41:105. 75. Uranova N, Orlovskaya D, Vikhreva O, et al. Electron microscopy of oligodendroglia in severe mental illness. Brain Res Bull. 2001;55:597-610. 76. Tkachev D, Mimmack ML, Ryan MM, et al. Oligodendrocyte dysfunction in schizophrenia and bipolar disorder. Lancet. 2003;362:798-805. 77. Chen G, Rajkowska G, Du F, et al. Enhancement of hippocampal neurogenesis by lithium. J Neurochem. 2000;75:1729-1734. 78. Malberg JE, Eisch AJ, Nestler EJ, et al. Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J Neurosci. 2000;20:91049110. 79. Rocha E, Achaval M, Santos P, et al. Lithium treatment causes gliosis and modifies the morphology of hippocampal astrocytes in rats. Neuroreport. 1998;9:3971-3974. 80. Rocha E, Rodnight R. Chronic administration of lithium chloride increases immunodetectable glial fibrillary acidic protein in the rat hippocampus. J Neurochem. 1994;63:1582-1584. 81. Levine S, Saltzman A, Klein AW. Proliferation of glial cells in vivo induced in the neural lobe of the rat pituitary by lithium. Cell Prolif. 2000;33:203-207. 82. Rajkowska G, Goldman-Rakic PS. Cytoarchitectonic definition of prefrontal areas in the normal human cortex: II. Variability in locations of areas 9 and 46. Cereb Cortex. 1995;4:323-327. 83. Uylings HB, Sanz-Arigita E, de Vos K, et al. The importance of a human 3D database and atlas for studies of prefrontal and thalamic functions. Prog Brain Res. 2000;126:357-368. 84. Drevets W, Price J, Simpson JR Jr, et al. Subgenual prefrontal cortex abnormalities in mood disorders. Nature. 1997;386:824-827. 85. Miguel-Hidalgo JJ, Rajkowska G. Morphological brain changes in depression: can antidepressants reverse them? CNS Drugs. 2002;16:361-372. 86. Soares J, Mann J. The anatomy of mood disordersreview of structural neuroimaging studies. Biol Psychiatry. 1997;41:86-106. 87. Elkis H, Friedman L, Wise A, et al. Meta-analyses of studies of ventricular enlargement and cortical sulcal prominence in mood disorders. Arch Gen Psychiatry. 1995;52:735-746. 88. Drevets WC. Neuroimaging studies of mood disorders. Biol Psychiatry. 2000;48:813-829. 89. Bertolino A, Frye M, Callicott JH, et al. Neuronal pathology in the hippocampal area of patients with bipolar disorder. Biol Psychiatry. 1999;45:135S. 90. Winsberg ME, Sachs N, Tate DL, et al. Decreased dorsolateral prefrontal N-acetyl aspartate in bipolar disorder. Biol Psychiatry. 2000;47:475-481. 91. Moore GJ, Bebchuk JM, Hasanat K, et al. Lithium increases N-acetylaspartate in the human brain: in vivo evidence in support of bcl-2s neurotrophic effects? Biol Psychiatry. 2000;48:1-8. 92. Czeh B, Michaelis T, Watanabe T, et al. Stress-induced changes in cerebral metabolites, hippocampal volume, and cell proliferation are prevented by antidepressant treatment with tianeptine. Proc Natl Acad Sci U S A. 2001;98:12796-12801. 93. Evans S, Akil H, Choudary P, et al. Microarray studies in mood disorders: distinct patterns seen between major depression and bipolar disorder in two frontal cortical regions. ACNP 41st Annual Meeting. December 8-12, 2002. San Juan, Puerto. Scientific Abstract 36; 2002. 94. Tomita H, Vawter M, Evans S, et al. Gene expression profiles in postmortem brains of mood disorder patients. Society for Neuroscience, 33rd Annual Meeting, New Orleans, La. Program No. 640.19.2003 Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience, 2003. Online, 2003. Available at http://sfn.scholarone.com/itin2003/. Accessed 5 May 2004.

197

Clinical research
Neuroplasticity in mood disorders
Wayne C. Drevets, MD

he recent development of neuroimaging technologies that permit in vivo characterization of the anatomical, physiological, and receptor pharmacological correlates of mood disorders have enabled significant advances toward delineating the neurobiological correlates of mood disorders. Because these conditions were not associated with gross brain pathology or with clear animal models for spontaneous, recurrent mood episodes, the availability of tools allowing noninvasive assessment of the human brain proved critical to illuminating the pathophysiology of major depressive disorder (MDD) and bipolar disorder (BD). The results of studies applying imaging technologies and postmortem studies have Neuroimaging and neuropathological studies of major depressive disorder (MDD) and bipolar disorder (BD) have identified abnormalities of brain structure in areas of the prefrontal cortex, amygdala, striatum, hippocampus, parahippocampal gyrus, and raphe nucleus. These structural imaging abnormalities persist across illness episodes, and preliminary evidence suggests they may in some cases arise prior to the onset of depressive episodes in subjects at high familial risk for MDD. In other cases, the magnitude of abnormality is reportedly correlated with time spent depressed. Postmortem histopathological studies of these regions have shown abnormal reductions of synaptic markers and glial cells, and, in rare cases, reductions in neurons in MDD and BD. Many of the regions affected by these structural abnormalities show increased glucose metabolism during depressive episodes. Because the glucose metabolic signal is dominated by glutamatergic transmission, these data support other evidence that excitatory amino acid transmission is elevated in limbic-cortical-striatal-pallidal-thalamic circuits during depression. Some of the subject samples in which these metabolic abnormalities have been demonstrated were also shown to manifest abnormally elevated stressed plasma cortisol levels. The co-occurrence of increased glutamatergic transmission and cortisol hypersecretion raises the possibility that the gray matter volumetric reductions in these depressed subjects are partly accounted for by processes homologous to the dendritic atrophy induced by chronic stress in adult rodents, which depends upon interactions between elevated glucocorticoid secretion and N-methyl-D-aspartate (NMDA)glutamate receptor stimulation. Some mood-stabilizing and antidepressant drugs that exert neurotrophic effects in rodents appear to reverse or attenuate the gray matter volume abnormalities in humans with mood disorders. These neurotrophic effects may be integrally related to the therapeutic effects of such agents, because the regions affected by structural abnormalities in mood disorders are known to play major roles in modulating the endocrine, autonomic, behavioral, and emotional experiential responses to stressors.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:199-216.

Keywords: major depressive disorder; bipolar disorder; neuroplasticity; neuroimaging abnormalities; postmortem studies Author affiliations: Wayne C. Drevets, MD, Mood and Anxiety Disorders Program, NIH NIMH/MIB, 15K North Dr, Bethesda, Md, USA Copyright 2004 LLS SAS. All rights reserved

Address for correspondence: Wayne C. Drevets, MD, Mood and Anxiety Disorders Program, NIH NIMH/MIB, 15K North Dr, MSC 2670, Bethesda, MD 20892-2670, USA (e-mail: drevetsw@intra.nimh.nih.gov)

199

www.dialogues-cns.org

Clinical research
Selected abbreviations and acronyms
ACC BD FPDD 5-HT MDD NMDA PFC VTA anterior cingulate cortex bipolar disorder familial pure depressive disease 5-hydroxytryptamine (serotonin) major depressive disorder N-methyl-D-aspartate prefrontal cortex ventral tegmental area cation of functional imaging approaches. The regions affected by these abnormalities have been shown to play major roles in modulating emotional behavior by electrophysiological, lesion analysis, and functional neuroimaging studies in experimental animals and healthy humans. Thus, the structural abnormalities in these regions may prove relevant to the emotional dysregulation that is clinically manifest in mood disorders.

guided clinical neuroscience toward models in which both functional and structural brain pathology play roles in the pathogenesis of mood disorders. Longitudinal positron emission tomography (PET) imaging studies of MDD and BD identified abnormalities of regional cerebral glucose metabolism and cerebral blood flow (CBF), which, in some cases, persisted beyond symptom remission, and in other cases appeared mood statedependent (reviewed in reference 1; Figure 1). These reversible abnormalities presumably reflect areas where metabolic activity increases or decreases to mediate or respond to emotional and cognitive manifestations of the depressive syndrome, because local glucose metabolism and CBF (which is tightly coupled to glucose metabolism) reflect summations of the energy utilization associated with terminal field synaptic transmission during neural activity.2-4 In contrast, abnormalities that persist independently of the mood state may instead reflect neuropathological sequelae of recurrent illness or neurodevelopmental abnormalities that may confer vulnerability to MDD (eg, in cases where they are evident in otherwise healthy individuals at high familial risk for developing mood disorders). Such abnormalities in CBF and metabolism may reflect pathological changes in synaptic transmission associated with altered neurotransmitter receptor function, cerebrovascular disease, changes in neuronal arborization or synapse formation, or abnormalities in cellular viability or proliferation.5 For example, areas where CBF and metabolism appeared irreversibly decreased in depressives relative to controls in PET studies of MDD and BD were subsequently associated with focal tissue reductions in magnetic resonance imaging (MRI)based morphometric and postmortem histopathological studies of MDD and BD.6-10 Abnormalities of gray matter volume and histology have now been identified in several brain structures using volumetric MRI and postmortem neuropathological assessments, which in many cases were guided by initial appli-

Sensitivity for detecting neuroimaging abnormalities in depression

The neuroimaging abnormalities discovered to date have not had effect sizes sufficient to permit sensitive or specific classification of individual cases. Moreover, the psychiatric imaging literature is in disagreement regarding the specific location and direction of some abnormalities. Many limitations in the sensitivity in reproducing findings across studies appear to be accounted for simply by technical issues of image acquisition and/or analysis.1 In other cases, however, disagreements within the literature appear to reflect differences in subject selection criteria applied across studies, because the conditions encompassed by the diagnostic criteria for MDD appear to be heterogeneous with respect to pathophysiology and etiology. It is noteworthy that neuroimaging laboratories selecting depressed subjects according to MDD criteria alone have rarely been able to replicate their own previous findings in independent subject samples. Instead, neuroimaging abnormalities appear to be specific to subsets of MDD subjects.1 For example, requiring that subjects have familial aggregation of illness and an early age at illness onset improved sensitivity for identifying subject samples with reproducible neuroimaging abnormalities. Clinical differences related to the capacity for developing mania or psychosis or having a late age at illness onset have also been shown to influence neuroimaging data. For example, elderly MDD subjects with a late age at depression onset have an elevated prevalence of MRI signal hyperintensities (in T2-weighted MRI scans, as putative correlates of cerebrovascular disease) in the deep and periventricular white matter, which is not the case for elderly depressives with an early age at depression onset. Similarly, elderly MDD cases with a late-life onset and delusional MDD cases have been shown to have lateral ventricular enlargementa feature which is generally not present in MDD cases who are elderly but have an

200

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

early age of MDD onset, or in midlife depressives who are not delusional. In addition, enlargement of the third ventricle has been consistently reported in BD, but not in MDD. A major technical issue that influences the sensitivity for detecting neuroimaging abnormalities across studies is the low spatial resolution of imaging technology relative to the size of brain structures of primary interest. With respect to morphometric assessments of gray matter volume, the volumetric resolution of state-of-the-art image data has recently been about 1 mm3, compared with the cortex thickness of only 3 to 4 mm. MRI studies involving images of this resolution have been able to reproducibly show regionally specific reductions in mean gray matter volume across groups of clinically similar depres-

sives versus controls. However, they have lacked sensitivity to detect the relatively subtle tissue reductions extant in mood disorders in individual subjects. Moreover, studies attempting to replicate such findings using data acquired at lower spatial resolutions (ie, voxel sizes 1.5 mm3) have commonly been negative because of the substantial partial volume effects that arise when attempting to segment regions of only 3- to 4-mm cortex thickness in such low-resolution MRI images.

Volumetric MRI imaging abnormalities in mood disorders

Frontal lobe structures Volumes of the whole brain and entire frontal lobe generally have not differed between depressed and healthy control samples. In contrast, volumetric abnormalities have been identified in specific prefrontal cortical (PFC), mesiotemporal, and basal ganglia structures in mood disorders. The most prominent reductions in the cortex have been identified in the anterior cingulate gyrus ventral to the genu of the corpus callosum, where gray matter volume has been abnormally decreased 20% to 40% in depressed subjects with familial pure depressive disease (FPDD), familial BD, and psychotic depression6,11-13 relative to healthy controls or mood-disordered subjects with no first-degree relatives with mood disorders. These findings were confirmed by postmortem studies of clinically similar samples (see below). Effective treatment with selective serotonin (5-hydroxytryptamine [5-HT]) reuptake inhibitors did not alter the subgenual PFC volume in MDD,6 although the PFC appeared significantly larger in BD subjects chronically medicated with lithium or divalproex than BD subjects who were either unmedicated or medicated with other agents,1 compatible with evidence that chronic administration of these mood stabilizers increases expression of the neurotrophic factors in rodents.14 In the posterior orbital, cortex, and ventrolateral PFC, volume has also been shown to be reduced in in vivo volumetric MRI studies15,16 and in postmortem neuropathological studies of MDD.17,18 Reductions in gray matter volume were also found in the dorsomedial/dorsal anterolateral PFC in MDD subjects versus controls,19 and postmortem studies of MDD and BD reported abnormal reductions in the size of neurons and/or the density of glia.18,20,21

Dorsal anterior cingulate

Dorsal caudate Posterior cingulate

Dorsal medial/ anterolateral PFC Ventral anterior cingulate

Medial thalamus

Ventrolateral PFC Anterior insula Lateral orbital cortex Ventral striatum Amygdala

Medial cerebellum

Hippocampus

or

Areas where neuropathological/morphometric changes reported Indicates direction of metabolic abnormality relative to control

Figure 1. Summary of neuroimaging abnormalities in early-onset, primary, major depressive disorder (MDD). The regions where neurophysiological imaging abnormalities have been consistently reported in unmedicated MDD samples are listed and approximately shown on this midsagittal brain diagram in which subcortical structures are highlighted onto the medial surface. Because only the medial wall of the cortex is shown, the location of the lateral orbital/ventrolateral prefrontal cortex (PFC)/anterior insular region is better illustrated in Figure 2B. The ventral anterior cingulate region refers to both pregenual and subgenual portions (see text and Figure 2). The arrows in front of each region name indicate the direction of resting state abnormalities in glucose metabolism in unmedicated, depressed MDD samples relative to healthy control samples. In some cases, abnormalities in both directions have been reported which may depend either on the specific region involved or on the clinical state (eg, treatment responsive vs nonresponsive; see text). The red arrows have indicate histopathological and/or gray matter volumetric abnormalities in postmortem studies of primary mood disorders.

201

Clinical research
Temporal lobe structures Morphometric MRI studies of specific temporal lobe structures reported significant reductions in the hippocampal volume in MDD, with magnitudes of difference ranging from 8% to 19% with respect to healthy controls.22-28 Sheline et al23 and MacQueen et al28 reported that the hippocampal volume was negatively correlated with the total time spent depressed or with the number of depressive episodes in MDD. Other groups found no significant differences between MDD and control samples.29-35 The inconsistency in the results of MDD studies may reflect pathophysiological heterogeneity within the MDD samples studied. For example, Vythilingam et al36 reported that the hippocampal volume was abnormally decreased in depressed women who also had suffered early-life trauma, but not in women who had depression without early-life trauma. In BD, reductions in hippocampal volume were identified by Noga et al37 and Swayze et al38 relative to healthy controls, although Pearlson et al39 and Nugent et al27 found no differences between BD and control samples. In postmortem studies of BD, abnormal reductions in the mRNA concentrations of synaptic proteins40 and in apical dendritic spines of pyramidal cells41 were specifically observed in the subicular and ventral CA1 subregions of the hippocampus. A recent study using high-resolution MRI scans found that the volume of the subiculum, but not the remainder of the hippocampus, was decreased in BD relative to control samples.27 Two studies reported abnormalities of the hippocampal T1 MRI signal in MDD. Krishnan et al42 observed that the T1 relaxation time was reduced in the hippocampus, but not in the entire temporal lobe, in unipolar depressives relative to healthy controls, and Sheline et al23 observed that elderly subjects with MDD have a higher number of areas with a low MRI signal than age-matched controls in T1-weighted images. The significance of such abnormalities remains unclear. In the amygdala, the literature is in disagreement. Studies of MDD have reported that amygdala volume is decreased,43,44 increased,45 or not different26 in depressives relative to healthy controls. Similarly, in BD, amygdala volume was reported to be increased,46-48 decreased,39,49,50 or not different38 relative to healthy controls. Although the extent to which disagreements in the results across studies are accounted for by confounding factors (such as medication effects) remains unclear, it appears more likely that MRI images acquired at 1.5 tesla lack the spatial and tissue contrast resolution needed to measure amygdala volumes with sufficient validity and reliability. The amygdalas small size and proximity to other gray matter structures seriously limits the specificity (accuracy) for delimiting amygdala boundaries in images acquired using MRI scanners of 1.5-tesla field strength. High-resolution MRI images acquired at 3-tesla magnetic field strength, in contrast, permit valid and reliable volumetric measures of the human amygdala. A recent study employing this technique established that mean amygdala volumes are decreased bilaterally (P<0.001) in MDD relative to healthy control samples.51 Amygdala volumes were decreased both in currently depressed and currently remitted MDD subsamples. Although mean amygdala volumes did not differ between BD and control samples, they were smaller in BD subjects who had not been recently medicated with mood stabilizers than in BD subjects who had been taking such agents, consistent with evidence that some mood stabilizers exert neurotrophic effects.14 Basal ganglia Volumes of some basal ganglia structures have also been reported to be abnormally decreased in mood disorders. Husain et al52 reported that the putamen was smaller in depressives (mean age 55) than controls, and Krishnan et al53 found a smaller caudate nucleus volume in depressives (mean age 48) than controls. In a sample limited to elderly depressives, Krishnan et al54 also reported smaller putamen and caudate volumes relative to controls. These findings were consistent with the postmortem study of Baumann et al,55 which found that caudate and accumbens area volumes were markedly decreased in both MDD and BD samples relative to control samples. Nevertheless, Dupont et al56 and Lenze et al57 failed to find significant differences in caudate or lentiform nucleus (putamen plus globus pallidus) volumes between younger MDD subjects and controls. The factors accounting for the discrepant results across studies remain unclear. Abnormalities of corpus callosal volume in mood disorders The genual subsection of the corpus callosum was reduced in volume in both depressed women with MDD and their high-risk, female offspring (insufficient num-

202

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

bers of males were studied to determine whether the abnormality extends to males).58,59 These white matter regions contain the transcallosal fibers connecting the orbital cortex, anterior cingulate cortex (ACC), and medial PFC with their homologous cortices in the contralateral hemisphere. The volumes of the splenial subregion of the corpus callosum was also decreased in mood-disordered versus control samples, which contains transcallosal fibers from the posterior cingulate cortex. Other cerebral structures Morphometric studies of other brain structures in depression have produced less consistent results. Of MRI studies of the thalamus, Dupont et al56 reported that the thalamic volume was decreased in unipolar depressives relative to controls, but Krishnan et al42,54 found no differences between depressives and controls. Two studies of thalamic volume in BD also have reported conflicting results. Of MRI studies of the cerebellum, two reported that the vermal volume is reduced in depressives relative to controls,60,61 while a third did not.62 Consistent with evidence that the hypothalamic-pituitary-adrenal (HPA) axis function is elevated in some mood-disordered subgroups, enlargement of the pituitary and adrenal glands has been reported in MDD. Krishnan et al63 showed that MRI-based measures of cross-sectional area and volume of the pituitary were increased (by 34% and 41%, respectively) in depressives (n=19) versus controls (n=19). This observation is consistent with evidence that the adrenal gland is also abnormally enlarged in MDD,1 which would putatively result from chronically elevated stimulation of the adrenal cortex by adrenocorticotropic hormone (ACTH).

Postmortem neuropathological assessments of mood disorders

Most of the regions where MRI studies demonstrated volumetric abnormalities in mood disorders were also shown to contain histopathological changes or gray matter volumetric reductions in postmortem studies of MDD and BD. Reductions in gray matter volume, thickness, or wet weight have been reported in the subgenual ACC, posterolateral orbital cortex, and ventral striatum in MDD and/or BD subjects relative to controls.7,9,18,55 The histopathological correlates of these abnormalities included reductions in glial cells with no equivalent loss

of neurons, reductions in synapses or synaptic proteins, elevations in neuronal density, and reductions in neuronal size.9,17,18,20,40,64,65 Abnormal reductions in glial cell counts and density, and/or glia-to-neuron ratios have also been found in MDD in Brodmann area (BA) 24 cortex of the pregenual ACC,20 the dorsal anterolateral PFC (BA9),21,66 and the amygdala.1,67 Finally, the mean size of neurons was reduced in the dorsal anterolateral PFC (BA9) in MDD subjects relative to controls,18 and the density of neurons was decreased in the ACC in BD.68 In several of these studies, the decreases were largely accounted for by differences in the left hemisphere.1,7,9,17,67 In the amygdala and the dorsal anterolateral PFC (BA9), the glial type that specifically differed between MDD and control samples was the oligodendrocytes. In contrast, astrocyte and microglial cell counts did not differ significantly between MDD or BD samples and healthy control samples in the amygdala.1 Oligodendroglia are best characterized for their role in myelination, and the reduction in oligodendrocytes may conceivably arise secondary to an effect on myelin, either through demyelination, abnormal development, or atrophy in the number of myelinated axons. Notably, the myelin basic protein concentration was found to be decreased in the frontal polar cortex (BA10) in MDD subjects.69 Compatible with these data, the concentration of white matter within the vicinity of the amygdala27 and the white matter volume of the genual and splenial portions of the corpus callosum are abnormally reduced in MDD and BD.58,59 These regions of the corpus callosum were also smaller in child and adolescent offspring of women with MDD who had not yet experienced a major depressive episode, in comparison to age-matched controls, suggesting that the reduction in white matter in MDD reflects a developmental defect that exists prior to the onset of depressive episodes.58 All of these observations support the hypothesis that the glial cell loss in mood disorders is accounted for by a reduction in myelinating oligodendrocytes. Further evidence supporting this hypothesis comes from several reports that deficits in glia in the cerebral cortex depend upon laminar analysis, with the greatest effects in layers III, V, and VI.18,20,70,71 The intracortical plexuses of myelinated fibers known as bands of Baillarger are generally concentrated in layers III and V. The size of these plexuses varies across cortical areas, so if the oligodendrocytes related to these plexuses were affected, different areas would be expected to show greater or lesser deficits. Layer VI in particular has a relatively large component of

203

Clinical research
myelinated fibers running between the gray and white matter. Finally, a population of satellite oligodendrocytes exists next to neuronal cell bodies that have largely unknown functions, but do not appear to have a role in myelination under normal conditions.72 An electron microscopic study of the PFC in BD revealed decreased nuclear size, clumping of chromatin, and other types of damage to satellite oligodendrocytes, including indications of both apoptotic and necrotic degeneration.73 Fewer signs of degeneration were seen in myelin-related oligodendrocytes in white matter. Satellite oligodendrocytes may play a role in maintaining the extracellular environment for the surrounding neurons, which resembles the functions mediated by astrocytes.These oligodendrocytes are immunohistochemically reactive for glutamine synthetase, suggesting that they function like astrocytes and take up synaptically released glutamate for conversion to glutamine and cycling back into neurons.74 Many studies of glial function have not distinguished astrocytes from oligodendrocytes, and the two glial types may share several functions. In other brain regions, reductions in astroglia have been reported by postmortem studies of mood disorders. In the frontal cortex, Johnston-Wilson et al75 found that four forms of the astrocytic product glial fibrillary acidic protein (GFAP) were decreased in mood-disordered subjects relative to controls, although it remained unclear whether this decrement reflected a reduction in the astrocyte density or in GFAP expression. Using immunohistochemical staining for GFAP, Webster et al76 did not find significant differences in cortical astrocytes between controls, and MDD or BD cases. Other studies also did not find differences in GFAP between mood disorder cases and controls.66 Factors that may conceivably contribute to a loss of oligodendroglia in mood disorders include the elevated glucocorticoid secretion and glutamatergic transmission evident during depression and mania. Glucocorticoids affect glia as well as neurons,77 and elevated glucocorticoid levels decrease the proliferation of oligodendrocyte precursors.78 Moreover, oligodendrocytes express -amino-3-hydroxy5-methyl-4-isoxazolepropionate (AMPA) and kainatetype glutamate receptors, and are sensitive to excitotoxic damage from excess glutamate as well as to oxidative stress.1 These vulnerabilities putatively contribute to oligodendrocyte degeneration in ischemic brain injury and demyelinating diseases,79,80 although no data exist to establish a similar role in mood disorders. The targeted nature of the reductions in gray matter volume and glial cells to specific areas of the limbic-cortical circuits that show increased glucose metabolism during depressive episodes is noteworthy given the evidence reviewed below that the glucose metabolic signal is dominated by glutamatergic transmission. The hypothesis that glutamate transmission is elevated in these areas in depression was also supported by a postmortem study in depressed suicide victims.81 Elevations of glutamate transmission and cortisol secretion in mood disorders may also contribute to reductions in gray matter volume and synaptic markers by inducing dendritic atrophy in some brain structures. In the medial PFC and parts of the hippocampus and amygdala of adult rodents, the dendritic arbors undergo atrophy or debranching in response to specific types of repeated or chronic stress.82 The effects of stress on dendritic arborization depend both upon the type of stress applied and anatomical location. For example, chronic unpredictable stress produces dendritic atrophy in the basolateral amygdala, whereas chronic immobilization stress increased dendritic branching in pyramidal and stellate neurons within the basolateral amygdala, but did not affect dendritic arborization in the central nucleus of the amygdala.83,84 These dendritic reshaping processes depend upon interactions between N-methyl-D-aspartate (NMDA) glutamatergic receptor stimulation and glucocorticoid secretion associated with repeated stress.82 The depressives with BD and FPDD who show regional reductions in gray matter volume also show evidence of having increased cortisol secretion and glutamate transmission. Specifically, depressives with FPDD or BD are more likely to show abnormal suppression of cortisol secretion by dexamethasone and blunted hypoglycemic response to insulin8 and to release excessive amounts of cortisol during stress.8,85 Subjects with FPDD or familial BD also show elevations of glucose metabolism, which largely reflects glutamate transmission, in the medial and orbital PFC, amygdala, ventral striatum, and cingulate cortex regions that show reductions in gray matter volume and cellular elements.

Association between structural and metabolic abnormalities

The glucose metabolic signal is dominated by changes in glutamate transmission, and so the findings that gray matter reductions appear to occur specifically in regions that show hypermetabolism during depression raise the possi-

204

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

bility that excitatory amino acid transmission plays a role in the neuropathology of mood disorders.At least 85% to 90% of the glucose metabolic measure is accounted for by glutamate transmission from afferent projections originating within the same structure or from distal structures.4,86-89 In the depressed phase of familial MDD and BD, regional cerebral metabolism and CBF are abnormally increased in the amygdala, lateral orbital/ventrolateral PFC, ACC anterior to the genu of the corpus callosum (pregenual ACC), posterior cingulate cortex, ventral striatum, medial thalamus, and medial cerebellum.1 During effective antidepressant drug or electroconvulsive therapy, metabolic activity decreases in all of these regions,1,8 compatible with evidence that these treatments result in desensitization of NMDA glutamatergic receptors in the frontal cortex.90 In addition to these areas of increased metabolic activity, areas of reduced CBF and metabolism in depressives relative to controls were found in the ACC ventral to the genu of the corpus callosum (ie, subgenual ACC7) and the dorsomedial/ dorsal anterolateral PFC.19,91,92 Yet even in these regions, metabolic activity increases during the depressive relapse induced by tryptophan depletion (a dietary challenge that depletes central 5-HT transmission),93 and metabolism is increased in the subgenual ACC in the unmedicated-depressed phase relative to the unmedicated-remitted phase. In all of these regions where glucose metabolism is increased in the depressed phase relative to the remitted phase, reductions in cortex volume and/or histopathological changes have been found in in vivo MRI studies and/or postmortem studies of MDD and/or BD. The hypothesis that the elevations in glucose metabolism seen in these circuits reflect elevations in glutamatergic transmission is supported by evidence that the anatomical projections between affected areas are excitatory in nature.The abnormally increased CBF and metabolism in the ventrolateral and orbital PFC, ventral ACC, amygdala, ventral striatum, and medial thalamus evident in depression (Figure 2) implicate a limbic-thalamo-cortical circuit involving the amygdala, the mediodorsal nucleus of the thalamus and the orbital and medial PFC, and a limbicstriatal-pallidal-thalamic circuit involving related parts of the striatum and the ventral pallidum along with the components of the other circuit.95 The first of these circuits can be conceptualized as an excitatory triangular circuit, whereby the basolateral nucleus of the amygdala and the orbital and medial prefrontal regions are interconnected by excitatory (especially glutamatergic) projections with

each other and with the mediodorsal nucleus.96-100 This means that increased metabolic activity in these structures would presumably reflect increased synaptic transmission through the limbic-thalamo-cortical circuit.The limbic-striatal-pallidal-thalamic circuit constitutes a disinhibitory side loop between the amygdala or PFC and the mediodorsal nucleus. The amygdala and the PFC send excitatory projections to overlapping parts of the ventromedial striatum.101 This part of the striatum sends an inhibitory projection to the ventral pallidum,102 which in turn sends GABAergic (GABA, -aminobutyric acid), inhibitory fibers to the mediodorsal nucleus.99

Implications for the pathogenesis of emotion dysregulation

The circuits described above have also been implicated in the depressive syndromes arising secondary to lesions or degenerative illnesses. Lesions involving the PFC (eg, tumors or infarctions) and the diseases of the basal ganglia (eg, Parkinsons disease or Huntingtons disease) are associated with higher rates of depression than other similarly debilitating conditions and result in dysfunction at distinct points within these circuits and affect synaptic transmission in diverse ways.103 Consistent with this hypothesis, imaging studies of depressive syndromes arising secondary to neurological disorders have generally shown results that differ from those reported for primary mood disorders. For example, in contrast to the findings of increased CBF or metabolism in parts of the orbital cortex in primary depressives, orbital cortex flow is reportedly decreased or not significantly different in subjects with depressive syndromes arising secondary to Parkinsons disease, Huntingtons disease, or basal ganglia infarction relative to nondepressed subjects with the same illnesses.104-107 Primary and secondary depressive syndromes may thus involve the same neural network, although the direction of the physiological abnormalities within individual structures may differ across conditions. A common substrate in these cases may be dysfunction of the PFC-striatal modulation of limbic and visceral functions, because the idiopathic neuropathological changes evident in the orbital and medial PFC and ventral striatum in primary mood disorders (see above) and those found in neurodegenerative conditions all appear to be capable of inducing depressive syndromes (eg, Parkinsons disease, Huntingtons disease, and cerebrovascular disease).

205

Clinical research
PFC-amygdalar projections may also play a role in the pathogenesis of depressive and anxiety symptoms in mood disorders. Although the reciprocal PFC-amygdalar projections are excitatory in nature, these connections ultimately appear to activate inhibitory interneurons, which, in turn, lead to functional inhibition in the projected field of the amygdala (for PFC-amygdalar projections) or the medial PFC and ventrolateral PFC.96,108-110 The function of the PFC in modulating the amygdala appears to be impaired in mood disorders, according to functional MRI data showing that abnormally sustained amygdala activity in response to aversive words or sad faces in MDD is associated with blunted activation of PFC areas.108,111 Thus, the volumetric and/or histopathological changes evident in the subgenual and pregenual ACC, lateral orbital cortex, dorsomedial/dorsal anterolateral PFC, hippocampal subiculum, amygdala, and ventral striatum may interfere with the modulation of emotional behavior, as discussed below.
Figure 2. Altered metabolism in the prefrontal cortex (PFC) ventral to the genu of the corpus callosum (c.c.) (ie, subgenual PFC) in mood disorders. A. Negative voxel t values where glucose metabolism is decreased in depressives relative to controls in coronal (31 mm anterior to the anterior commissure, or y=31) and sagittal (3 mm left of midline, or x=-3) planes of a statistical parametric image comparing depressives relative to controls.7 This image localized an abnormality in the subgenual portion of the anterior cingulate cortex (subgenual ACC7), which was subsequently shown to be accounted for by a corresponding reduction in cortex volume on the left side (see text). Anterior (or left) is to the left of the image. B. Mean, normalized, glucose metabolic values for the left subgenual ACC measured using magnetic resonance imaging (MRI)based region-of-interest analysis. Metabolism is decreased in depressed subjects with either bipolar disorder (BD) or major depressive disorder (MDD) relative to healthy controls. In contrast, subjects scanned in the manic phase of BD (bipolar manic) have higher metabolism than either depressed or control subjects in this region. *P<0.025 controls versus depressed; P<0.01 depressed versus manic; P<0.05 controls versus manic. Although none of these subjects were involved in the study that generated the images shown in Figure 3, the mean glucose metabolism in this independent sample of depressives and controls also confirmed the areas of abnormally increased activity in the depressives in the amygdala, lateral orbital cortex, ventrolateral PFC, and medial thalamus (not shown in A, which only illustrates negative t values corresponding to hypometabolic areas in the depressives).
Figure 2A reproduced with permission from reference 6: Drevets WC, Price JL, Simpson JR, et al. Subgenual prefrontal cortex abnormalities in mood disorders. Nature. 1997;386:824-827. Copyright 1997, Nature Publishing Group. Figure 2B reproduced with permission from reference 94: Drevets WC. Neuroimaging and neuropathological studies of depression: Implications for the cognitive emotional manifestations of mood disorders. Curr Opin Neurobiol. 2001;11:240-249. Copyright 2001, Elsevier.

Ventral ACC The ACC ventral and anterior to the genu of the corpus callosum (subgenual and pregenual, respectively; Figure 2) shows complex relationships between CBF, metabolism, and illness state, which appear to be accounted for by a left-lateralized reduction in the corresponding cortex, initially demonstrated by MRI-based morphometric measures6,12-16,112 and later by postmortem neuropathological studies of familial BD and MDD.9 Thus, computer simulations that correct the PET data acquired from this region for the partial volume effect of the reduction in gray matter volume measured in MRI scans of the same subject conclude the actual metabolic activity in the remaining subgenual PFC tissue is increased in depressives relative to controls, and decreases to normative levels during effective treatment.113 This hypothesis appears to be compatible with the observations that effective antidepressant pharmacotherapy results in a decrease in meta-

x=-3

y=31

Subgenual PFC
c.c. -5.5 -2.8 t value 0

Mean regional/global metabolism

1.2 *

1.1

1 * 0.9

0.8 Control BD MDD Bipolar manic

206

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

bolic activity in this region in MDD,8,10,114 that during depressive episodes metabolism shows a positive relationship with depression severity,8,115,116 and that flow increases in this region in healthy, nondepressed humans during sadness induced via contemplation of sad thoughts or memories.114,117,118 The reduction in volume in this region exists early in the illness in familial MDD11 and BD.12 The gray matter deficit may nevertheless worsen or initially become apparent following illness onset based upon preliminary evidence in twins discordant for MDD that the affected twin has a smaller volume than their unaffected cotwin.119 Kimbrell et al120 reported that the subgenual ACC metabolism correlated inversely with the number of lifetime depressive episodes, compatible with the possibility that the reduction in metabolism in this region measured via PET reflects a partial volume effect of a gray matter reduction that worsens with repeated illness. In the pregenual ACC, Drevets et al95 initially found increased CBF in MDD, and subsequent studies extended this observation by demonstrating complex relationships between pregenual ACC activity and subsequent antidepressant treatment outcome. Wu et al121 reported that depressed subjects whose mood improved during sleep deprivation showed elevated metabolism in the pregenual ACC and amygdala in their pretreatment scans. Mayberg et al122 reported that, while metabolism in the pregenual ACC was abnormally increased in depressives who subsequently responded to antidepressant drugs, metabolism was decreased in depressives who later had poor treatment response. Finally, in a tomographic electroencephalographic (EEG) analysis, Pizzagalli et al123 reported that depressives who ultimately showed the best response to nortriptyline showed hyperactivity (higher theta activity) in the pregenual ACC at baseline, compared with subjects showing the poorer response. During effective antidepressant treatment, most PET studies have shown that pregenual ACC flow and metabolism decrease in posttreatment scans relative to pretreatment scans.1 The finding that this region contains histopathological changes in MDD and BD20,64,68 suggests the hypothesis that the abnormal reduction in metabolism in treatment-nonresponsive cases reflects more severe reductions in cortex. In rodents and nonhuman primates, the regions that appear homologous to human subgenual and pregenual ACC, namely the infralimbic, prelimbic, and ventral ACCs, have extensive reciprocal connections with areas implicated in the expression of behavioral, autonomic, and

endocrine responses to threat, stress, or reward/nonreward, such as the orbital cortex, lateral hypothalamus, amygdala, accumbens, subiculum, ventral tegmental area (VTA), raphe, locus ceruleus, periaqueductal grey (PAG), and nucleus tractus solitarius.7,124 Humans with lesions that include these ventromedial PFC structures show abnormal autonomic responses to emotionally provocative stimuli and an inability to experience emotion related to concepts that ordinarily evoke emotion.125 Electrical stimulation of the ACC elicits fear, panic, or a sense of foreboding in humans, and vocalization in experimental animals.126 Similarly, rats with experimental lesions of prelimbic cortex demonstrate altered autonomic, behavioral, and neuroendocrine responses to stress and fear-conditioned stimuli.The prelimbic and infralimbic cortices contain abundant concentrations of glucocorticoid receptors, which, when stimulated by corticosterone (CORT), reduce stress-related HPA activity.127 Lesions of these cortices consequently result in exaggerated plasma ACTH and CORT responses to restraint stress.127 In rats, bilateral or right-lateralized lesions of the ACC and prelimbic and infralimbic cortex attenuate sympathetic autonomic responses, stressinduced CORT secretion, and gastric stress pathology during restraint stress or exposure to fear-conditioned stimuli.128-130 In contrast, left-sided lesions of this area increase sympathetic autonomic arousal and CORT responses to restraint stress.130 These data suggest that the right subgenual PFC facilitates expression of visceral responses during emotional processing, while the left subgenual PFC inhibits or modulates such responses.130 It is thus noteworthy that the gray matter reduction in this region in MDD and BD was lateralized to the left side, suggesting that it may contribute to disinhibition of neuroendocrine and autonomic function in depression.127,131,132 The ventral ACC also appears to participate in processing of behavioral incentive and motivated behavior. These areas send efferent projections to the VTA and substantia nigra, and receive dense dopaminergic innervation from VTA.124 In rats, electrical or glutamatergic stimulation of medial PFC areas that include prelimbic cortex elicits burst firing patterns from dopamine (DA) cells in the VTA and increases DA release in the accumbens.113 These phasic, burst firing patterns of DA neurons appear to encode information regarding stimuli that predict reward and deviations between such predictions and actual occurrence of reward.133 Ventral ACC dysfunction may thus conceivably contribute to disturbances of motivated behavior and hedonic perception in mood disorders.

207

Clinical research
Dorsomedial/dorsal anterolateral PFC Metabolism and CBF are abnormally decreased in the dorsolateral and dorsomedial PFC in MDD.1 The dorsomedial PFC region includes the dorsal ACC92 and an area rostral to the dorsal ACC involving cortex on the medial and lateral surface of the superior frontal gyrus (approximately corresponding to BA9 and BA32).8,19,91 Postmortem studies of MDD and BD found abnormal reductions in the size of neurons and/or the density of glia in this portion of BA9,18,20,134 which may account for the reduction in metabolism in this region in MDD, and for the failure of antidepressant drug treatment to correct metabolism in these areas.8,19 Nevertheless, currently remitted MDD subjects who experience depressive relapse during tryptophan depletion show increased metabolic activity within these areas in the depressed versus the remitted conditions,93 similar to other structures where histopathological and gray matter volume changes exist in MDD. Flow normally increases in the vicinity of this dorsomedial/dorsal anterolateral PFC in healthy humans as they perform tasks that elicit emotional responses or require emotional evaluations.1 In healthy humans, CBF increases in this region during anxious anticipation of an electrical shock to an extent that correlates inversely with changes in anxiety ratings and heart rate, suggesting that this region functions to attenuate emotional expression. In rats, lesions of the dorsomedial PFC result in exaggerated heart rate responses to fear-conditioned stimuli, and stimulation of these sites attenuate defensive behavior and cardiovascular responses evoked by amygdala stimulation,128 although the homologue to these areas in primates has not been clearly established. In primates, the BA9 cortex sends efferent projections to the lateral PAG and the dorsal hypothalamus through which it may modulate cardiovascular responses associated with emotional behavior.124 It is thus conceivable that dysfunction of the dorsomedial/dorsal anterolateral PFC may contribute to impairments in the ability to modulate emotional responses in mood disorders. Lateral orbital/ventrolateral PFC In the lateral orbital cortex, ventrolateral PFC, and anterior insula, the resting CBF and metabolism have been abnormally increased in unmedicated subjects with primary MDD (Figure 3).1 The elevated activity in these areas in MDD appears to be mood-state dependent,95 and, during treatment with somatic antidepressant therapies, flow and metabolism decreases in these regions.1 The relationship between depression severity and physiological activity in the lateral orbital cortex/ventrolateral PFC is complex. While CBF and metabolism increase in these areas in the depressed phase relative to the remitted phase of MDD, the magnitude of these measures is inversely correlated with ratings of depressive ideation and severity.95,116,135 Moreover, while metabolic activity is abnormally increased in these areas in treatment-responsive unipolar and bipolar depressives, more severely ill or treatment-refractory samples show CBF and metabolic values lower than or not different from those of controls.81,139 This inverse relationship between orbital cortex/ventrolateral PFC activity and ratings of depression severity extends to some other emotional states as well. Posterior orbital cortex flow also increases in subjects with obsessive-compulsive disorder or simple animal phobias during exposure to phobic stimuli and in healthy subjects during induced sadness,140-142 with the change in posterior orbital CBF correlating inversely with changes in obsessive thinking, anxiety, and sadness, respectively. These data appear to be consistent with electrophysiological and lesion analysis data showing that parts of the orbital cortex participate in modulating behavioral and visceral responses associated with defensive, emotional, and reward-directed behavior as reinforcement contingencies change.124,143,144 The orbital cortex and amygdala send overlapping projections to each of these structures and to each other through which they appear to modulate each others neural transmission.124,143,145 Activation of the orbital cortex during depression may thus reflect compensatory attempts to attenuate emotional expression or interrupt unreinforced aversive thought and emotion. Consistent with this hypothesis, cerebrovascular lesions of the orbital cortex are associated with an increased risk for depression.146 These observations also suggest that the reduction of CBF and metabolism in the orbital cortex and ventrolateral PFC during antidepressant drug treatment may not be a primary mechanism through which such agents ameliorate depressive symptoms. Instead, direct inhibition of pathological limbic activity in areas such as the amygdala and ventral ACC may attenuate the mediation of depressive symptoms.8 The orbital cortex neurons may thus relax, as reflected by the return of metabolism to normal lev-

208

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

els, as antidepressant drug therapy attenuates the pathological limbic activity to which these neurons putatively respond.145 The amygdala In the amygdala, neurophysiological activity is altered both at rest and during exposure to emotionally valenced stimuli in some depressive subgroups. The basal CBF and metabolism are elevated in mood-disordered subgroups who meet criteria for FPDD (Figure 3),8,95,135,136 for MDD melancholic subtype,148 type II or nonpsychotic type I BD,136,149 or for those who are responsive to sleep deprivation.121 In contrast, metabolism has not been abnormal in unipolar depressives meeting criteria for depression spectrum disease,136,137 or in MDD samples meeting Diagnostic and Statistical Manual of Mental Health Disorders (DSM) criteria,150-152 although the interpretation of the latter results was confounded by technical problems that reduced sensitivity for measuring amygdala activity.136 During antidepressant treatment that both attenuates depressive symptoms and prevents relapse, amygdala metabolism decreases toward normative levels.8

Functional imaging data acquired as subjects view emotionally valenced stimuli that normally activate the amygdala also demonstrate altered physiological responses in MDD. In the left amygdala, the hemodynamic response to viewing fearful faces was blunted in depressed children153 and depressed adults,94 consistent with the elevation of basal CBF and metabolism in the left amygdala in such cases (physiologically activated tissue is expected to show an attenuation of further rises in the hemodynamic/metabolic signal in response to tasks that normally engage the same tissue). The duration of the amygdala response to emotionally valenced stimuli is also abnormally prolonged in response to sad stimuli in depression. Drevets et al94 observed that, although the initial amygdala CBF response to sad faces was similar in depressives and controls, this response habituated during repeated

A Amygdala

Figure 3. Areas of abnormally increased blood flow in subjects with major depressive disorder (MDD). The image sections shown are from an image of t values, produced by a voxel-by-voxel computation of the unpaired t statistic to compare regional CBF between a depressed sample selected according to criteria for familial pure depressive disease (FPDD) (n=13) and a healthy control sample (n=33).95 The positive t values shown correspond to areas where flow is increased in the depressives relative to the controls. The abnormal activity in these regions was replicated using glucose metabolism imaging in independent subject samples.8,135,136 A. Sagittal section at 17 mm left of midline illustrating areas of increased CBF in depression in the amygdala and orbital cortex. B. Area of increased flow extended through the lateral orbital cortex to the ventrolateral prefrontal cortex (VLPFC) and anterior insula.8,95 The x coordinate locates sagittal sections in millimeters to the left of midline. The PET images in A and B from which the t image was generated have been stereotaxically transformed to the coordinate system of Talairach and Tournoux,137 from which the corresponding atlas outline is shown. Anterior is left.
Figure 3A reproduced with permission from reference 126: Price JL, Carmichael ST, Drevets WC. Networks related to the orbital and medial prefrontal cortex: a substrate for emotional behavior? Prog Brain Res. 1996;107:523-536. Copyright 1996, Elsevier. Figure 3B reproduced with permission from reference 138: Drevets WC, Videen TO, Snyder AZ, MacLeod AK, Raichle ME. Regional cerebral blood flow changes during anticipatory anxiety. Soc Neurosci Abstr. 1994;20:368. Copyright 1994, Society for Neuroscience.

Orbital cortex

x=-17

2.25 t value

4.5

B Anterior insula VLPFC

x=-41

2.0 t value

4.0

209

Clinical research
exposure to the same stimuli in the controls, but not in the depressives over the imaging period. Similarly, Siegle et al44 reported that hemodynamic activity increased in the amygdala during exposure to negatively valenced words to a similar extent in depressives and controls, but, while the hemodynamic response rapidly fell to baseline in the controls, it remained elevated in the depressives. The amygdala plays major roles in organizing other behavioral, neuroendocrine, and autonomic aspects of emotional and stress responses to experiential stimuli. For example, the amygdala facilitates stress-related corticotropin-releasing hormone (CRH) release154 and electrical stimulation of the amygdala in humans increases cortisol secretion,155 suggesting a mechanism via which excessive amygdala activity may participate in inducing the CRH and cortisol hypersecretion that is evident in MDD. In PET studies of MDD and BD, CBF and metabolism in the left amygdala correlates positively with stressed plasma cortisol secretion, which may reflect the effect of either amygdala activity on CRH secretion or cortisol or CRH on amygdala function.136 If the reduction in amygdala volume is associated with reductions in synaptic contacts formed by afferent projections from regions known to modulate amygdala function, then amygdala neuronal activity may become disinhibited. The above reports that amygdala blood flow and metabolism are abnormally elevated and hemodynamic responses to emotional stimuli are abnormally persistent in MDD support this hypothesis. Notably, Siegle et al44 reported that the abnormally prolonged hemodynamic responses of the amygdala to sad words occurred particularly in the MDD subjects who had reduced amygdala volumes. If the neurotrophic effects of mood-stabilizing drugs restore and protect modulatory connections formed between the amygdala and cortex,1 then the volumetric changes observed during treatment may contribute to their therapeutic effects in mood disorders. Abnormalities in anatomically related limbic and subcortical structures In the medial thalamus and ventral striatum, CBF and metabolism are abnormally elevated in the depressed phase of MDD and BD, and decrease during antidepressant pharmacotherapy.8,95,134,136,154,156,157 Several groups also reported abnormally increased CBF in the posterior cingulate cortex in the unmedicated, depressed phase of MDD.8,112,158 Bench et al158 specifically reported that the elevation of posterior cingulate flow in depressives relative to controls correlated positively with anxiety ratings. Exposure to aversive stimuli of various types results in increased physiological activity in the posterior cingulate cortex.159 The posterior cingulate cortex sends major anatomical projections to the pregenual ACC.160

Neuroreceptor imaging abnormalities in mood disorders

Neuroreceptor imaging studies of mood disorders have demonstrated reductions in 5-HT1A receptor binding in mood disorders, which would appear to hold major implications for alterations in neuroplasticity in these conditions. Both presynaptic (in the raphe) and postsynaptic (insula, anterior, and posterior cingulate cortices, parietooccipital cortex, orbital/ventrolateral PFC) 5-HT1A binding is abnormally decreased in MDD and panic disorder (irrespective of the current presence of comorbid depression), and postsynaptic 5-HT1A receptor binding is also decreased in BD.85,116,161-165 The magnitudes of these differences have been similar to those found by postmortem studies of primary mood-disordered samples17,165 and depressed suicide victims.166 These data are also compatible with results of studies showing that MDD and panic disorder subjects show blunted thermic and adrenocorticotropin/cortisol responses to 5-HT1A receptor agonist challenge.85,162 The 5-HT1A receptor plays major roles in the neuroplasticity involving serotonergic and other neurons.167,168 In addition, during fetal development and subsequently during 5-HT neuronal injury, stimulation of astrocyte and radial glial cell-based 5-HT1A receptors results in release of the trophic factor S100, which promotes 5-HT neuronal arborization.168,169 If glial function is reduced during 5-HT system development in BD and MDD, it is conceivable that arborization of the 5-HT neurons may be attenuated, potentially reflected by the widespread reductions of 5-HT transporter and postsynaptic 5-HT1A receptor expression seen in MDD.17,85,116,163,166,170 Such a hypoplastic process may also underlie the finding that the area expressing 5-HT1A receptors in the dorsal raphe nucleus is abnormally decreased in depressed suicides.166 It is conceivable that the persistently increased anxiety behaviors and the exaggerated fear and behavioral despair responses shown by 5-HT1A receptor knockout mice at least partly reflect effects of deficient 5-HT1A receptor function on neuroplasticity during neurodevel-

210

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

opment.162 It remains unclear, however, whether the reduction in 5-HT1A receptor function and expression constitutes a neurodevelopmental or an acquired abnormality in mood disorders.165

Concluding remarks
The convergent results from studies of mood disorders conducted using neuroimaging, lesion analysis, and post-

mortem techniques support models in which the signs and symptoms of major depression can emanate from dysfunction within PFC, striatal, and brain stem systems that modulate emotional behavior. Antidepressant therapies may compensate for this dysfunction by attenuating the pathological limbic activity that mediates such symptoms,9 and by increasing genetic transmission of neurotrophic factors that exert neuroplastic effects within the pathways modulating emotional expression.14

REFERENCES
1. Drevets WC, Gadde K, Krishnan R. Neuroimaging studies of depression. In: Charney DS, Nestler EJ, Bunney BJ, eds. The Neurobiological Foundation of Mental Illness. 2nd ed. New York, NY: Oxford University Press; 2004:461-490. 2. Raichle ME. Circulatory and metabolic correlates of brain function in normal humans. In: Brookhart JM, Mountcastle VB, eds. Handbook of Physiology. The Nervous System. Baltimore, Md: American Physiology Society 1987;5:643-674. 3. DiRocco RJ, Kageyama GH, Wong-Riley MT. The relationship between CNS metabolism and cytoarchitecture: a review of 14C-deoxyglucose studies with correlation to cytochrome oxidase histochemistry. Comput Med Imaging Graph. 1989;13:81-92. 4. Magistretti PJ, Pellerin L. Cellular mechanisms of brain imaging metabolism and their relevance to functional brain imaging. Phil Trans Roy Soc London B. 1999;354:1155-1163. 5. Wooten GF, Collins RC. Metabolic effects of unilateral lesion of the substantia nigra. J Neurosci. l981;1:285-29l. 6. Drevets WC, Price JL, Simpson JR, et al. Subgenual prefrontal cortex abnormalities in mood disorders. Nature. 1997;386:824-827. 7. Drevets WC, ngr D, Price JL. Neuroimaging abnormalities in the subgenual prefrontal cortex: implications for pathophysiology of familial mood disorders. Mol Psychiatry. 1998;3:220-226. 8. Drevets WC, Bogers W, Raichle ME. Functional anatomical correlates of antidepressant drug treatment assessed using PET measures of regional glucose metabolism. Eur J Neuropharmacol. 2002;12:527-544. 9. ngr D, Drevets WC, Price JL. Glial reduction in the subgenual prefrontal cortex in mood disorders. Proc Natl Acad Sci U S A. 1998;95:13290-13295. 10. Mazziotta JC, Phelps ME, Plummer D, Kuhl DE. Quantitation in positron emission computed tomography. 5. Physical-anatomical effects. J Comput Assist Tomogr. 1981;5:734-743. 11. Botteron KN, Raichle ME, Drevets WC, Heath AC, Todd RD. Volumetric reduction in left subgenual prefrontal cortex in early onset depression. Biol Psychiatry. 2002;51:342-344. 12. Hirayasu Y, Shenton ME, Salisbury DF, et al. Subgenual cingulate cortex volume in first-episode psychosis. Am J Psychiatry. 1999;156:1091-1093. 13. Coryell W, Nopoulos P, Drevets WC, Andreasen NC. Subgenual PFC volumes in MDD and schizophrenia: diagnostic specificity and prognostic implications. Am J Psychiatry. 2004. In press. 14. Du J, Quiroz JA, Gray NA, Szabo ST, Zarate CA Jr, Manji HK. Regulation of cellular plasticity and resilience by mood stabilizers: the role of AMPA receptor trafficking. Dialogues Clin Neurosci. 2004;6:143-155. 15. Lai T, Payne ME, Byrum CE, Steffens DC, Krishnan KR. Reduction of orbital frontal cortex volume in geriatric depression. Biol Psychiatry. 2000;48:971-975. 16. Bremner JD, Vythilingham M, Vermetten E, et al. Reduced volume of orbitofrontal cortex in major depression. Biol Psychiatry. 2002;51:273-279. 17. Bowen DM, Najlerahim A, Procter AW, Francis PT, Murphy E. Circumscribed changes of the cerebral cortex in neuropsychiatric disorders of later life. Proc Natl Acad Sci U S A. 1989:86:9504-9508.

18. Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. 19. Bell KA, Kupfer DJ, Drevets WC. Decreased glucose metabolism in the dorsomedial prefrontal cortex in depression. Biol Psychiatry. 1999;45:118S. 20. Cotter DR, Mackay D, Landau S, Kerwin R, Everall I. Reduced glial cell density and neuronal size in the anterior cingulate cortex in major depressive disorder. Arch Gen Psychiatry. 2001;58:545-553. 21. Uranova NA, Vostrikov VM, Orlovskaya DD, Rachmanova VI. Oligodendroglial density in the prefrontal cortex in schizophrenia and mood disorders: a study from the Stanley Neuropathology Consortium. Schizophr Res. 2004;67:269-275. 22. Sheline YI, Sanghavi M, Mintun MA, Gado M. Depression duration but not age predicts hippocampal volume loss in medically healthy women with recurrent major depression. J Neurosci. 1999;19:5034-5043. 23. Sheline YI, Wang PW, Gado MH, Csernansky JG, Vannier MW. Hippocampal atrophy in recurrent major depression. Proc Natl Acad Sci U S A. 1996;93:3908-3913. 24. Bremner JD, Narayan M, Anderson ER, Staib LH, Miller HL, Charney DS. Hippocampal volume reduction in major depression. Am J Psychiatry. 2000;157:115-118. 25. Steffens DC, Byrum CE, McQuoid DR, et al. Hippocampal volume in geriatric depression. Biol Psychiatry. 2000;48:301-309. 26. Mervaala E, Fohr J, Kononen M, et al. Quantitative MRI of the hippocampus and amygdala in severe depression. Psychol Med. 2000;30:117-125. 27. Nugent AC, Wood S, Bain EE, et al. High resolution MRI neuromorphometric assesment of the hippocampal subiculum in mood disorders. Presented at the International Society for Magnetic Resonance in Medicine, 12th Annual Meeting, Kyoto, Japan; 2004. 28. MacQueen GM, Campbell S, McEwen BS, et al. Course of illness, hippocampal function, and hippocampal volume in major depression. Proc Natl Acad Sci U S A. 2003;100:1387-1392. 29. Ashtari M, Greenwald BS, Kramer-Ginsberg E, et al. Hippocampal/amygdala volumes in geriatric depression. Psychol Med. 1999;29:629-638. 30. Axelson D, Doraiswamy PM, McDonald WM, et al. Hypercortisolemia and amygdala hippocampal changes in depression. Psychiatry Res. 1993; 47:167-173. 31. Hauser P, Altschuler LL, Berrettini W, et al. Temporal lobe measurement in primary affective disorder by magnetic resonance imaging. J Neuropsychiatry Clin Neurosci. 1989;1:128-134. 32. Pantel J, Schroder J, Essig M, et al. Quantitative magnetic resonance imaging in geriatric depression and primary degenerative dementia. J Affect Disord. 1997;42:69-83. 33. Shah PJ, Ebmeier KP, Glabus MF, Goodwin GM. Cortical gray matter reductions associated with treatment-resistant chronic unipolar depression. Controlled magnetic resonance imaging study. Br J Psychiatry. 1998;172:527532. 34. Vakili K, Pillay SS, Lafer B, Fava M, Renshaw PF, Bonello-Cintron CM. Hippocampal volume in primary unipolar major depression: a magnetic resonance imaging study. Biol Psychiatry. 2000;47:1087-1090.

211

Clinical research
Neuroplasticidad en los trastornos afectivos
Los estudios neuropatolgicos y de neuroimgenes en la depresin mayor (DM) y en el trastorno bipolar (TB) han identificado anormalidades de la estructura cerebral en reas de la corteza prefrontal, la amgdala, el cuerpo estriado, el hipocampo, el giro parahipocmpico y el ncleo del rafe. Estas anormalidades estructurales en las neuroimgenes se mantienen ms all de los episodios de la enfermedad y las evidencias preliminares sugieren que en algunos casos ellas pueden aparecer antes del inicio de los episodios depresivos en sujetos con alto riesgo familiar de DM. En otros casos, la magnitud de la anormalidad se correlaciona con el tiempo que lleva la depresin. Estudios histopatolgicos postmortem de estas regiones han mostrado disminuciones anormales de marcadores sinpticos y de clulas gliales y, en raros casos, disminucin de neuronas en la DM y el TB. Muchas de las regiones afectadas por estas alteraciones estructurales muestran un aumento del metabolismo de la glucosa durante los episodios depresivos. Dado que la seal metablica de glucosa est comandada por la transmisin glutamatrgica, estos datos sustentan el argumento a favor del incremento de la transmisin del aminocido excitatorio en los circuitos lmbico-crtico-estriato-plido-talmicos durante la depresin. Algunos de los sujetos de las muestras en que se encontraron estas anormalidades metablicas tambin tuvieron cifras elevadas de cortisol plasmtico en respuesta al estrs. La aparicin concomitante del aumento de la transmisin glutamatrgica y de la hipersecrecin de cortisol incrementa la probabilidad que las disminuciones del volumen de sustancia gris en los sujetos con depresin se deba en parte a los procesos equivalentes a los de la atrofia dendrtica inducida por el estrs crnico en roedores adultos, lo que depende de las interacciones entre el aumento de la secrecin de glucocorticoides y la estimulacin del receptor de glutamato N-metil-D-asprtico (NMDA). Algunos antidepresivos y estabilizadores del nimo que ejercen efectos neurotrficos en roedores parece que revierten o disminuyen las anormalidades del volumen de sustancia gris en humanos con trastornos afectivos. Estos efectos neurotrficos pueden estar relacionados ntegramente con los efectos teraputicos de dichos frmacos, ya que se sabe que las regiones afectadas por alteraciones estructurales en los trastornos afectivos tienen un papel importante en la modulacin de las respuestas endocrina, autonmica, conductual y emocional que se experimenta frente al estrs.
35. von Gunten A, Fox NC, Cipolotti L, Ron MA. A volumetric study of hippocampus and amygdala in depressed patients with subjective memory problems. J Neuropsychiatry Clin Neurosci. 2000;12:493-498. 36. Vythilingam M, Heim C, Newport J, et al. Childhood trauma associated with smaller hippocampal volume in women with major depression. Am J Psychiatry. 2002;159:2072. 37. Noga JT, Vladar K, Torrey EF. A volumetric magnetic resonance imaging study of monozygotic twins discordant for bipolar disorder. Psychiatry Res. 2001;106:25-34. 38. Swayze VW 2nd, Andreasen NC, Alliger RJ, Yuh WT, Ehrhartdt JC. Subcortical and temporal structures in affective disorder and schizophrenia: a magnetic resonance imaging study. Biol Psychiatry. 1992;31:221-240. 39. Pearlson GD, Barta PE, Powers RE, et al. Medial and superior temporal gyral volumes and cerebral asymmetry in schizophrenia versus bipolar disorder. Biol Psychiatry. 1997;41:1-14. 40. Eastwood SL, Harrison PJ. Hippocampal synaptic pathology in schizophrenia, bipolar disorder, and major depression: a study of complexin mRNAs. Mol Psychiatry. 2000;5:425-432. 41. Rosoklija G, Toomayan G, Ellis SP, et al. Structural abnormalities in subicular dendrites in subjects with schizophrenia and mood disorders. Arch Gen Psychiatry. 2000;57:349-356. 42. Krishnan KRR, Doraiswamy PM, Figiel GS, et al. Hippocampal abnormalities in depression. J Neuropsychiatry Clin Neurosci. 1991;3:387-391. 43. Sheline YI, Gado MH, Price JL. Amygdala core nuclei volumes are decreased in recurrent major depression. Neuroreport. 1998;9:2023-2028. 44. Siegle G J, Konecky RO, Thase ME, Carter CS. Relationships between amygdala volume and activity during emotional information processing tasks in depressed and never-depressed individuals: an fMRI investigation. Ann N Y Acad Sci. 2003;985:481-484. 45. Frodl T, Meisenzahl E, Zetzsche T, et al. Enlargement of the amygdala in patients with a first episode of major depression. Biol Psychiatry. 2002;51:708714. 46. Altshuler LL, Bartzokis G, Thomas G, Curran J, Mintz J. Amygdala enlargement in bipolar disorder and hippocampal reduction in schizophrenia: an MRI study demonstrating neuroanatomic specificity. Arch Gen Psychiatry. 1998;55:663-664. 47. Strakowski SM, DelBello MP, Sax KW, et al. Brain magnetic resonance imaging of structural abnormalities in bipolar disorder. Arch Gen Psychiatry. 1999;56:254-260. 48. Brambilla P, Harenski K, Nicoletti M, et al. MRI investigation of temporal lobe structures in bipolar patients. J Psychiatr Res. 2003;37:287-295. 49. Blumberg HP, Kaufman J, Martin A, et al. Amygdala and hippocampal volumes in adolescents and adults with bipolar disorder. Arch Gen Psychiatry. 2003;60:1201-1208. 50. DelBello MP, Zimmerman ME, Mills NP, Getz GE, Strakowski SM. Magnetic resonance imaging analysis of amygdala and other subcortical brain regions in adolescents with bipolar disorder. Bipolar Disord. 2004;6:43-52. 51. Drevets WC, Sills R, Nugent A, et al. Volumetric assessment of the amygdala in mood disorders using high resolution, 3T MRI. Biol Psychiatry. 2004;55:182S. 52. Husain MM, McDonald WM, Doraiswamy PM, et al. A magnetic resonance imaging study of putamen nuclei in major depression. Psychiatry Res. 1991;40:95-99. 53. Krishnan KRR, McDonald WM, Escalona PR, et al. Magnetic resonance imaging of the caudate nuclei in depression: preliminary observations. Arch Gen Psychiatry. 1992;49:553-557. 54. Krishnan KRR, McDonald WM, Doraiswamy PM, et al. Neuroanatomical substrates of depression in the elderly. Eur Arch Psychiatry Neurosci. 1993;243:41-46.

212

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Neuroplasticit dans les troubles de lhumeur

Les tudes de neuropathologie et de neuro-imagerie des troubles dpressifs majeurs (TDM) et des troubles bipolaires (TB) ont identifi des anomalies de la structure crbrale dans les aires du cortex prfrontal, de lamygdale, du striatum, de lhippocampe, du gyrus parahippocampique et du noyau du raph. Ces anomalies structurelles limage persistent travers les pisodes de la maladie et des arguments antrieurs suggrent quelles peuvent se produire avant lapparition des pisodes dpressifs chez les sujets haut risque familial de TDM. Dans dautres cas, limportance des anomalies serait lie la dure de la dpression. Des tudes histopathologiques post mortem de ces rgions ont montr des rductions anormales des marqueurs synaptiques et des cellules gliales et, dans quelques rares cas, des diminutions du nombre des neurones dans les TDM et les TB. De nombreuses rgions atteintes par ces anomalies structurelles prsentent un mtabolisme du glucose augment pendant ces pisodes dpressifs. La transmission glutamatergique dominant le signal mtabolique du glucose, ces donnes confortent un autre argument savoir que la transmission de lacide amin excitateur est leve dans les circuits limbiques-corticaux-striataux-pallidaux-thalamiques pendant la dpression. Certains chantillons, chez des sujets chez qui on a trouv des anomalies mtaboliques, ont galement montr des concentrations anormalement leves de cortisol plasmatique en rponse au stress. Lapparition concomitante de laugmentation de la transmission glutamatergique et de lhyperscrtion de cortisol accrot la possibilit que les rductions de volume de la substance grise chez ces personnes dpressives soient en partie justifies par des processus identiques ceux de latrophie dendritique induite par le stress chronique chez les rongeurs adultes, qui dpend des interactions entre la scrtion leve des glucocorticodes et la stimulation du rcepteur glutamatergique N-mthyl-D-aspartate (NMDA). Certains mdicaments antidpresseurs et thymorgulateurs exerant des effets neurotrophiques chez les rongeurs semblent inverser ou attnuer les anomalies du volume de la substance grise chez les humains atteints de troubles de lhumeur. Ces effets neurotrophiques peuvent tre intgralement lis aux effets thrapeutiques de tels mdicaments, parce quil est reconnu que les rgions affectes par des anomalies structurelles au cours des troubles de lhumeur jouent un rle majeur dans la modulation des rponses aux agents stressants au niveau de lexprience endocrine, autonome, comportementale et motionnelle.

55. Baumann B, Danos P, Krell D, et al. Reduced volume of limbic systemaffiliated basal ganglia in mood disorders: preliminary data from a postmortem study. J Neuropsychol Clin Neurosci. 1999;11:71-78. 56. Dupont RM, Jernigan TL, Heindel W, et al. Magnetic resonance imaging and mood disorders. Localization of white matter and other subcortical abnormalities. Arch Gen Psychiatry. 1995;52:747-755. 57. Lenze EJ, Sheline YI. Absence of striatal volume differences between depressed subjects with no comorbid medical illness and matched comparison subjects. Am J Psychiatry. 1999;156:1989-1991. 58. Martinez P, Ronsaville D, Gold PW, Hauser P, Drevets WC. Morphometric abnormalities in adolescent offspring of depressed mothers. Soc Neurosci Abstr. 2002:32. 59. Brambilla P, Nicoletti M, Sassi RB, et al. Corpus callosum signal intensity in patients with bipolar and unipolar disorder. J Neurol Neurosurg Psychiatry. 2004;75:221-225. 60. Shah SA, Doraiswamy PM, Husain MM, et al. Posterior fossa abnormalities in major depression: a controlled magnetic resonance imaging study. Acta Psychiatr Scand. 1992;85:474-479. 61. Escalona PR, McDonald WM, Doraiswamy PM, et al. Reduction of cerebellar volume in major depression. A controlled MRI study. Depression. 1993;1:156-158. 62. Parashos IA, Tupler LA, Blitchington T, Krishnan KR. Magnetic-resonance morphometry in patients with major depression. Psychiatry Res. 1998;84:715. 63. Krishnan KRR, Doraiswamy PM, Lurie SN, et al. Pituitary size in depression. J Clin Endocrinol Metab. 1991;72:256-259. 64. Cotter D, Mackay D, Beasley C, Kerwin R, Everall I. Reduced glial density and neuronal volume in major depressive disorder and schizophrenia in the anterior cingulate cortex. Schizophr Res. 2000;41:106.

65. Eastwood SL, Harrison PJ. Synaptic pathology in the anterior cingulate cortex in schizophrenia and mood disorders. A review and a Western blot study of synaptophysin, GAP 43, and the complexins. Brain Res Bull. 2001;55:569-578. 66. Cotter DR, Pariante CM, Everall IP. Glial cell abnormalities in major psychiatric disorders: the evidence and implications. Brain Res Bull. 2001;55:585595. 67. Bowley MP, Drevets WC, ngr D, Price JL. Low glial numbers in the amygdala in mood disorders. Biol Psychiatry. 2002;52:404-412. 68. Benes FM, Vincent SL, Todtenkopf M. The density of pyramidal and nonpyramidal neurons in ACC of schizophrenic and bipolar subjects. Biol Psychiatry. 2001;50:395-406. 69. Honer WG, Falkai P, Chen C, Arango V, Mann JJ, Dworks AJ. Synaptic and plasticity-associated proteins in anterior frontal cortex in severe mental illness. Neuroscience. 1999;91:1247-1255. 70. Rajkowska G, Halaris A, Selemon LD. Reductions in neuronal and glial density characterize the dorsolateral prefrontal cortex in bipolar disorder. Biol Psychiatry. 2001;49:741-752. 71. Cotter D, Mackay D, Chana G, Beasley C, Landau S, Everall I. Reduced neuronal size and glial cell density in area 9 of the dorsolateral prefrontal cortex in subjects with major depressive disorder. Cereb Cortex. 2002;12:386394. 72. Ludwin SK. The function of perineuronal satellite oligodendrocytes: an immunohistochemical study. Neuropathol Appl Neurobiol. 1984;10:143-149. 73. Uranova N, Orlovskaya D, Vikhreva O, et al. Electron microscopy of oligodendroglia in severe mental illness. Brain Res Bull. 2001;55:597-610. 74. DAmelio F, Eng LF, Gibbs MA. Glutamine synthetase immunoreactivity present in oligodendroglia of various regions of the central nervous system. Glia. 1990;3:335-341.

213

Clinical research
75. Johnston-Wilson NL, Sims CD, Hofmann JP, et al. Disease-specific alterations in frontal cortex brain proteins in schizophrenia, bipolar disorder, and major depressive disorder. The Stanley Neuropathology Consortium. Mol Psychiatry. 2000;5:142-149. 76. Webster MJ, Knable MB, Johnston-Wilson N, Nagata K, Inagaki M, Yolken RH. Immunohistochemical localization of phosphorylated glial fibrillary acidic protein in the prefrontal cortex and hippocampus from patients with schizophrenia, bipolar disorder, and depression. Brain Behav Immunol. 2001;15:388-400. 77. Cheng JD, de Vellis J. Oligodendrocytes as glucocorticoids target cells: functional analysis of the glycerol phosphate dehydrogenase gene. J Neurosci Res. 2000;59:436-445. 78. Alonso G. Prolonged corticosterone treatment of adult rats inhibits the proliferation of oligodendrocyte progenitors present throughout white and gray matter regions of the brain. Glia. 2000;31:219-231. 79. Dewar D, Underhill SM, Goldberg MP. Oligodendrocytes and ischemic brain injury. J Cereb Blood Flow Metab. 2003;23:263-274. 80. Matute C, Sanchez-Gomez MV, Martinez-Millan L, Miledi R. Glutamate receptor-mediated toxicity in optic nerve oligodendrocytes. Proc Natl Acad Sci U S A. 1997;94:8830-8835. 81. Nowak G, Ordway GA, Paul IA. Alterations in the N-methyl-D-aspartate (NMDA) receptor complex in the frontal cortex of suicide victims. Brain Res. 1995;675:157-164. 82. McEwen BS. Structural plasticity of the adult brain: how animal models help us understand brain changes in depression and systemic disorders related to depression. Dialogues Clin Neurosci. 2004;6:119-133. 83. Vyas A, Mitra R, Shankaranarayana Rao BS, Chattarji S. Chronic stress induces contrasting patterns of dendritic remodeling in hippocampal and amygdaloid neurons. J Neurosci. 2002;22:6810-6818. 84. Vyas A, Bernal S, Chattarji S. Effects of chronic stress on dendritic arborization in the central and extended amygdala. Brain Res. 2003;965:290-294. 85. Drevets WC, Frank E, Price JC, et al. PET imaging of serotonin 1A receptor binding in depression. Biol Psychiatry. 1999;46:1375-1387 86. Rothman DL, Sibson NR, Hyder F, Shen J, Behar KL, Shulman RG. In vivo nuclear magnetic resonance spectroscopy studies of the relationship between the glutamate-glutamine neurotransmitter cycle and functional neuroenergetics. Phil Trans Roy Soc London B. 1999;354:1165-1177. 87. Sibson NR, Dhankhar A, Mason GF, Rothman DL, Behar KL, Shulman RG. Stoichiometric coupling of brain glucose metabolism and glutamatergic neuronal activity. Proc Natl Acad Sci U S A. 1998;95:316-321. 88. Shen J, Petersen KF, Behar KL, et al. Determination of the rate of the glutamate/glutamine cycle in the human brain by in vivo 13C NMR. Proc Natl Acad Sci U S A. 1999;96:8235-8240. 89. Shulman RG, Rothman DL. Interpreting functional imaging studies in terms of neurotransmitter cycling. Proc Natl Acad Sci U S A. 1998;95:11993-11998. 90. Paul IA, Nowak G, Layer RT, Popik P, Skolnick P. Adaption of the Nmethyl-D-aspartate receptor complex following chronic antidepressant treatments. J Pharmacol Exp Ther. 1994;269:95-102. 91. Baxter LR, Schwartz JM, Phelps ME, et al. Reduction of prefrontal cortex glucose metabolism common to three types of depression. Arch Gen Psychiatry. 1989;46:243-250. 92. Bench CJ, Friston KJ, Brown RG, Scott LC, Frackowiak SJ, Dolan RJ. The anatomy of melancholiafocal abnormalities of cerebral blood flow in major depression. Psychol Med. 1992;22:607-615. 93. Neumeister A, Nugent AC, Waldeck T, et al. Behavioral and neural responses to tryptophan depletion in unmedicated remitted patients with major depressive disorder and controls. Arch Gen Psychiatry. 2004. In press. 94. Drevets WC. Neuroimaging and neuropathological studies of depression: implications for the cognitive emotional manifestations of mood disorders. Curr Opin Neurobiol. 2001;11:240-249. 95. Drevets WC, Videen TO, Price JL, Preskorn SH, Carmichael ST, Raichle ME. A functional anatomical study of unipolar depression. J Neurosci. 1992;12:3628-3641. 96. Amaral DG, Insausti R. Retrograde transport of D-[3H]aspartate injected into the monkey amygdaloid complex. Exp Brain Res. 1992;88:375-388. 97. Bacon SJ, Headlam AJ, Gabbot PL, Smith AD. Amygdala input to medial prefrontal cortex (mPFC) in the rat: a light and electron microscope study. Brain Res. 1996;720:211-219. 98. Jackson ME, Moghaddam B. Amygdala regulation of nucleus accumbens dopamine output is governed by the prefrontal cortex. J Neurosci. 2001;21:676-681. 99. Kuroda M, Price JL. Synaptic organization of projections from basal forebrain structures to the mediodorsal thalamic nucleus of the rat. J Comp Neurol. 1991;303:513-533. 100. Amaral DG, Price JL. Amygdalo-cortical projections in the monkey (Macaca fascicularis). J Comp Neurol. 1984;230:465-496. 101. Russchen FT, Price JL. Amygdalostriatal projections in the rat. Topographical organization and fiber morphology shown using the lectin PHA-L as an anterograde tracer. Neurosci Lett. 1984;47:15-22. 102. Graybiel AM. The basal ganglia and the initiation of movement. Rev Neurol (Paris). 1990;146:570-574. 103. Drevets WC, Todd RD. Depression, mania and related disorders. In: Guze SB, ed. Adult Psychiatry. St Louis, Mo: Mosby Press; 2004:99-141. 104. Mayberg HS, Starkstein SE, Sadzot B, et al. Selective hypometabolism in the inferior frontal lobe in depressed patients with Parkinsons disease. Ann Neurol. 1990;28:57-64. 105. Mayberg HS, Starkstein SE, Morris PL, et al. Remote cortical hypometabolism following focal basal ganglia injury: relationship to secondary changes in mood. Neurology. 1991;41(suppl):266. 106. Mayberg HS, Starkstein SE, Peyser CE, Brandt J, Dannals RF, Folstein SE. Paralimbic frontal lobe hypometabolism in depression associated with Huntingtons disease. Neurology. 1992;42:1791-1797. 107. Ring HA, Bench CJ, Trimble MR, Brooks DJ, Frackowiak RSJ, Dolan RJ. Depression in Parkinsons disease: a positron emission study. Br J Psychiatry. 1994;165:333-339. 108. Drevets WC. Neuroimaging abnormalities in the amygdala in mood disorders. Ann N Y Acad Sci. 2003;985:420-444. 109. Garcia R, Vouimba RM, Baudry M, Thompson RF. The amygdala modulates prefrontal cortex activity relative to conditioned fear. Nature. 1999;402:294-296. 110. Rosenkranz JA, Grace AA. Cellular mechanisms of infralimbic and prelimbic prefrontal cortical inhibition and dopaminergic modulation of basolateral amygdala neurons in vivo. J Neurosci. 2002;22:324-337. 111. Siegle GJ, Steinhauer SR, Thase ME, Stenger VA, Carter CC. Cant shake that feeling: event-related fMRI assessment of sustained amygdala activity in response to emotional information in depressed individuals. Biol Psychiatry. 2002;51:693-707. 112. Buchsbaum MS, Wu J, Siegel BV, et al. Effect of sertraline on regional metabolic rate in patients with affective disorder. Biol Psychiatry. 1997;41:1522. 113. Drevets WC. Prefrontal cortical-amygdalar metabolism in major depression. Ann N Y Acad Sci. 1999;877:614-637. 114. Mayberg HS, Liotti M, Brannan SK, et al. Reciprocal limbic-cortical function and negative mood: converging PET findings in depression and normal sadness. Am J Psychiatry. 1999;156:675-682. 115. Osuch EA, Ketter TA, Kimbrell TA, et al. Regional cerebral metabolism associated with anxiety symptoms in affective disorder patients. Biol Psychiatry. 2000;48:1020-1023. 116. Drevets WC, Thase M, Bogers W, Greer P, Kupfer DJ. Glucose metabolic correlates of depression severity and antidepressant treatment response. Biol Psychiatry. 2002;51:176S. 117. George MS, Ketter TA, Parekh PI, Horwitz B, Herscovitch P, Post RM. Brain activity during transient sadness and happiness in healthy women. Am J Psychiatry. 1995;152:341-351. 118. Damasio A, Grabowski TJ, Bechara A, et al. Subcortical and cortical brain activity during the feeling of self-generated emotions. Nat Neurosci. 2000;3:1049-1056. 119. Botteron KN, Raichle ME, Heath AC, et al. An epidemiological twin study of prefrontal neuromorphometry in early onset depression. Biol Psychiatry. 1999;45:59S. 120. Kimbrell TA, Ketter TA, George MS, et al. Regional cerebral glucose utilization in patients with a range of severities of unipolar depression. Biol Psychiatry. 2002;51:237-252. 121. Wu JC, Gillin JC, Buchsbaum MS, Hershey T, Johnson JC, Bunney WE. Effect of sleep deprivation on brain metabolism of depressed patients. Am J Psychiatry. 1992;149:538-543.

214

Neuroplasticity in mood disorders - Drevets

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

122. Mayberg HS, Brannan SK, Mahurin RK, et al. Cingulate function in depression: a potential predictor of treatment response. Neuroreport. 1997;8:1057-1061. 123. Pizzagalli D, Pascual Marqui RD, Nitschke JB, et al. Anterior cingulate activity predicts degree of treatment response in major depression: evidence from brain electrical tomography analysis. Am J Psychiatry. 2001;158:405-415. 124. ngr D, Price JL. The organization of networks within the orbital and medial prefrontal cortex of rats, monkeys, and humans. Cereb Cortex. 2000;10:206-219. 125. Damasio AR. Descarte's Error: Emotion, Reason, and the Human Brain. New York, NY: Grosset/Putnam; 1995. 126. Price JL, Carmichael ST, Drevets WC. Networks related to the orbital and medial prefrontal cortex: a substrate for emotional behavior? Prog Brain Res. 1996;107:523-536. 127. Dioro D, Viau V, Meaney MJ. The role of the medial prefrontal cortex (cingulate gyrus) in the regulation of hypothalamic-pituitary-adrenal responses to stress. J Neurosci. 1993;13:3839-3847. 128. Frysztak RJ, Neafsey EJ. The effect of medial frontal cortex lesions on cardiovascular conditioned emotional responses in the rat. Brain Res. 1994;643:181-193. 129. Morgan MA, LeDoux JE. Differential contribution of dorsal and ventral medial prefrontal cortex to the acquisition and extinction of conditioned fear in rats. Behav Neurosci. 1995;109:681-688. 130. Sullivan RM, Gratton A. Lateralized effects of medial prefrontal cortex lesions on neuroendocrine and autonomic stress responses in rats. J Neurosci. 1999;19:2834-2840. 131. Veith RC, Lewis N, Linares OA, et al. Sympathetic nervous system activity in major depression. Arch Gen Psychiatry. 1994;51:411-422. 132. Carney RM, Freedland KE, Rich MW, Smith LJ, Jaffe AS. Ventricular tachycardia and psychiatric depression in patients with coronary artery disease. Am J Med. 1993;95:23-28. 133. Schultz W. Dopamine neurons and their role in reward mechanisms. Curr Opin Neurobiol. 1997;7:191-197. 134. Uranova NA, Vostrikov VM, Orlovskaya DD, Rachmanova VI. Oligodendroglial density in the prefrontal cortex in schizophrenia and mood disorders: a study from the Stanley Neuropathology Consortium. Schizophr Res. 2004;67:269-275. 135. Drevets WC, Spitznagel E, Raichle ME. Functional anatomical differences between major depressive subtypes. J Cereb Blood Flow Metab. 1995;15:S93. 136. Drevets WC, Price JL, Bardgett ME, Reich T, Todd R, Raichle ME. Glucose metabolism in the amygdala in depression: relationship to diagnostic subtype and stressed plasma cortisol levels. Pharmacol Biochem Behav. 2002;71:431-447. 137. Talairach J, Tournoux P. Co-Planar Stereotaxic Atlas of the Human Brain. Stuttgart, Germany: Thieme; 1988. 138. Drevets WC, Videen TO, Snyder AZ, MacLeod AK, Raichle ME. Regional cerebral blood flow changes during anticipatory anxiety. Soc Neurosci Abstr. 1994;20:368. 139. Mayberg HS, Lewis PJ, Reginald W, Wanger HN Jr. Paralimbic hypoperfusion in unipolar depression. J Nucl Med. 1994;35:929-934. 140. Rauch SL, Jenike MA, Alpert NM, et al. Regional cerebral blood flow measured during symptom provocation in obsessive-compulsive disorder using oxygen 15-labeled carbon dioxide and positron emission tomography. Arch Gen Psychiatry. 1994;51:62-70. 141. Drevets WC, Simpson JR, Raichle ME. Regional blood flow changes in response to phobic anxiety and habituation. J Cereb Blood Flow Metab. 1995;15:S856. 144. Schneider F, Gur RE, Alavi A, et al. Mood effects on limbic blood flow correlate with emotion self-rating: a PET study with oxygen-15 labeled water. Psychiatry Res Neuroimag. 1995;61:265-283. 143. Mogenson GJ, Brudzynski SM, Wu M, Yang CR, Yim CCY. From motivation to action: a review of dopaminergic regulation of limbic to nucleus to accumbens to ventral pallidum to pedunculopontine nucleus circitries involved in limbic-motor integration. In: Kalivas PW, Barnes CD, eds. Limbic Motor Circuits and Neuropsychiatry. London, UK: CRC Press; 1993:193-236.

144. Rolls ET. A theory of emotion and consciousness, and its application to understanding the neural basis of emotion. In: Gazzaniga M, ed. The Cognitive Neurosciences. Cambridge, Mass: MIT Press; 1995:1091-1106. 145. Timms RJ. Cortical inhibition and facilitation of the defence reaction. J Physiol Lond. 1977;266:98P-99P. 146. MacFall JR, Payne ME, Provenzale JE, Krishnan KRR. Medial orbital frontal lesions in late onset depression. Biol Psychiatry. 2001;49:803-806. 147. Oya H, Howard M, Kawasaki H, Adolphs R. Intracranial field potentials recorded from human amygdala and frontal cortex: amplitude and phase responses to emotional stimuli. Soc Neurosci Abstr. 2001;645.6. 148. Nofzinger EA, Nichols TE, Meltzer CC, et al. Changes in forebrain function from waking to REM-sleep in depression: preliminary analyses of [18F]FDG PET studies. Psychiatry Res. 1999;91:59-78. 149. Ketter TA, Kimbrell TA, George MS, et al. Effects of mood and subtype on cerebral glucose metabolism in treatment-resistant bipolar disorder. Biol Psychiatry. 2001;49:97-109. 150. Abercrombie HC, Schaefer SM, Larson CL, et al. Metabolic rate in the right amygdala predicts negative affect in depressed patients. Neuroreport. 1998:3301-3307. 151. Brody AL, Saxena S, Stoessel P, et al. Regional brain metabolic changes in patients major depressive disorder from pre- to post-treatment with paroxetine. Arch Gen Psychiatry. 2001;58:631-640. 152. Saxena S, Brody AL, Ho ML, et al. Differential cerebral metabolic changes with paroxetine treatment of obsessive-compulsive disorder vs major depression. Arch Gen Psychiatry. 2002;59:250-261. 153. Thomas KM, Drevets WC, Dahl RE, et al. Abnormal amygdala response to faces in anxious and depressed children. Arch Gen Psychiatry. 2001;58:1057-1063. 154. Herman JP, Cullinan WE. Neurocircuitry of stress: central control of the hypothalamo-pituitary-adrenocortical axis. Trends Neurosci. 1997;20:78-84. 155. Rubin RT, Mandell AJ, Crandall PH. Corticosteroid responses to limbic stimulation in man: localization of stimulus sites. Science. 1966;153:767768. 156. Videbech P, Ravnkilde B, Pedersen AR, et al. The Danish PET/depression project: PET findings in patients with major depression. Psychol Med. 2001;31:1147-1158. 157. Wilson J, Kupfer DJ, Thase M, Bogers W, Greer P, Drevets WC. Ventral striatal metabolism is increased in depression, and decreases with treatment. Biol Psychiatry. 2002;51:122S. 158. Bench CJ, Friston KJ, Brown RG, Frackowiak RS, Dolan RJ. Regional cerebral blood flow in depression measured by positron emission tomography: the relationship with clinical dimensions. Psychol Med. 1993;23:579-590. 159. Charney DS, Drevets WC. The neurobiological basis of anxiety disorders. In: Davis K, Charney DS, Coyle J, Nemeroff CB, eds. Psychopharmacology. The Fifth Generation of Progress. New York, NY: Lippencott, Williams, and Wilkins; 2002:901-930. 160. Vogt B. Structural organization of cingulate cortex. In: Vogt BA, Gabriel M, eds. Neurobiology of Cingulate Cortex and Limbic Thalamus. Boston, Mass: Birkhauser; 1993. 161. Bain EE, Nugent AC, Carson RE, et al. Decreased 5-HT1A receptor binding in bipolar depression. Biol Psychiatry. 2004;55:178S. 162. Neumeister A, Bain E, Nugent A, et al. Reduced serotonin type 1A receptor binding in panic disorder. J Neurosci. 2004;24:589-591. 163. Parsey RV, Oquendo MA, Simpson NR, et al. Altered serotonin 1A binding in major depression: a [11C]WAY100635 PET study. Biol Psychiatry. 2002;51(8S):106S. Abstract 300. 164. Sargent PA, Kjaer KH, Bench CJ, et al. Brain serotonin 1A receptor binding measured by positron emission tomography with [11C]WAY-100635: effects of depression and antidepressant treatment. Arch Gen Psychiatry. 2000;57:174-180. 165. Lopez JF, Chalmers DT, Little KY, Watson SJ. Regulation of serotonin 1A, glucocorticoid, and mineralocorticoid receptor in rat and human hippocampus: implications for the neurobiology of depression. A. E. Bennett Research Award. Biol Psychiatry. 1998;43:547-573. 166. Arango V, Underwood MD, Boldrini M, et al. Serotonin 1A receptors, serotonin transporter binding and serotonin transporter mRNA expression in the brainstem of depressed suicide victims. Neuropsychopharmacology. 2001;25:892-903.

215

Clinical research
167. Brewton LS, Haddad L, Azmitia EC. Colchicine-induced cytoskeletal collapse and apoptosis in N-18 neuroblastoma cultures is rapidly reversed by applied S-100. Brain Res. 2001;912:9-16. 168. Azmitia EC, Whitaker-Azmitia PM. Awakening the sleeping giant: anatomy and plasticity of the brain serotonergic system. J Clin Psychiatry. 1991;52(12, suppl):4-16. 169. Azmitia EC, Gannon PJ, Kheck NM, Whitaker-Azmitia PM. Cellular localization of the 5-HT1A receptor in primate brain neurons and glial cells. Neuropsychopharmacology. 1996;14:35-46. 170. Mann JJ, Huang YY, Underwood MD, et al. A serotonin transporter gene promoter polymorphism (5-HTTLPR) and prefrontal cortical binding in major depression and suicide. Arch Gen Psychiatry. 2000;57:729-738.

216

Clinical research
Cellular plasticity and resilience and the pathophysiology of severe mood disorders
Dennis S. Charney, MD; Georgette DeJesus, MD; Husseini K. Manji, MD

Recent advances in the identification of the neural circuits, neurochemicals, and signal transduction mechanisms involved in the pathophysiology and treatment of mood disorders have led to much progress toward understanding the roles of genetic factors and psychosocial stressors. The monoaminergic neurotransmitter systems have received the most attention, partly because of the observation that effective antidepressant drugs exert their primary biochemical effects by regulating intrasynaptic concentrations of serotonin and norepinephrine. Furthermore, the monoaminergic systems are extensively distributed throughout the network of limbic, striatal, and prefrontal cortical neuronal circuits thought to support the behavioral and visceral manifestations of mood disorders. Increasing numbers of neuroimaging, neuropathological, and biochemical studies indicate impairments in cellular plasticity and resilience in patients who suffer from severe, recurrent mood disorders. In this paper, we describe studies identifying possible structural, functional, and cellular abnormalities associated with depressive disorders, which are potentially the cellular underpinnings of these diseases. We suggest that drugs designed to enhance cellular plasticity and resilience, and attenuate the activity of maladaptive stress-responsive systems, may be useful for the treatment of severe mood disorders.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:217-225.

epression is a common, chronic, and often disabling psychiatric illness, which is estimated to affect 5% to 10% of the population. It frequently appears in early life, has a chronic course, and is considered a risk factor for other medical illnesses, such as coronary vascular disease, diabetes, and osteoporosis. This is not altogether surprising given the extensive bidirectional mind-body interactions mediated via the autonomic nervous system, immune system, and a host of neuroendocrine factors. According to the World Health Organization (WHO), depression is the leading global cause of years of life lived with disability and the fourth leading cause of disabilityadjusted life-years. Disability-adjusted life-years is defined as the reduction in an individuals productive life, and takes into account premature mortality.1,2 Considering the high morbidity and mortality associated with depression, it is unfortunate that the psychological and neurobiological underpinnings of depression have not been specifically defined. Although major depression is currently diagnosed by means of a diagnostic system (Diagnostic and Statistical Manual of Mental Health Disorders, Fourth Edition [DSM-IV]) based upon phenomenology, this disorder most likely embodies a heterogeneous set of disorders with multiple causes. Therefore, one of the major goals of current and future research on depression is the development of a diagnostic system based on etiology.3 This goal is becoming increasingly closer to reality due to recent progress in the identification of neural circuits, neurochemicals, and signal transduction mechanisms
Keywords: mood disorder; depression; neuroplasticity; stress; resilience; brain morphology Author affiliations: National Institute of Mental Health, Bethesda, Md, USA Address for correspondence: Prof Dennis S. Charney, National Institute of Mental Health, 15K North Drive, Room 101, Bethesda, MD 20892-2670, USA (e-mail: charneyd@nih.gov)

Copyright 2004 LLS SAS. All rights reserved

217

www.dialogues-cns.org

Clinical research
Selected abbreviations and acronyms
BDNF CREB ERK HPA LTP MAP PFC brain-derived neurotrophic factor cyclic adenosine monophosphate (cAMP) response element binding protein extracellular response kinase hypothalamo-pituitary-adrenal (axis) long-term potentiation mitogen-activated protein prefrontal cortex there should be an appreciation of the episodic and often intense mood disturbance, which can become progressive over the time. Furthermore, the phenotypic expression of the disease involves not only episodic and often profound mood disturbances, but also a constellation of cognitive, motor, autonomic, endocrine, and sleep/wake abnormalities. Additionally, while most antidepressants exert their initial effects by increasing the levels of serotonin and/or norepinephrine in the synapse, clinical antidepressant effects exclusively result after chronic administration (days to weeks). This suggests that a cascade of downstream effects is ultimately responsible for the clinical antidepressant effects of these medications. These observations have led to the recognition that, although monoaminergic neurotransmitter system dysfunction undoubtedly plays an important role in mediating some facets of the pathophysiology of depressive disorders, additional fundamental alterations in cellular plasticity cascades are most likely involved.13-15 The functional impairments during mood episodes have long been recognized; however, there is increasing evidence of significant interepisode impairment as well. The devastation of these disorders is further complicated by the fact that the medications currently used for their treatment are associated with variable rates of efficacy and not intolerable side effects.An appreciation for both the need for more efficacious treatment for mood disorders and the absence of significant advances in the development of truly innovative therapeutics has led to the investigation of intracellular signaling cascades and their role in the pathophysiology and treatment of mood disorders. Thus, while traditionally viewed exclusively as neurochemical disorders, recent evidence suggests the presence of impairments of cellular plasticity cascades, which produce not only functional, but also morphological impairments; this evidence has generated considerable excitement among the clinical neuroscience community and is reshaping views about the neurobiological underpinnings of these disorders. Thus, as we discuss in detail below, increasing neuroimaging, neuropathological, and biochemical studies suggest impairments in cellular plasticity and resilience in patients who suffer from severe, recurrent mood disorders. The term neuroplasticity encompasses diverse essential processes by which the brain perceives, adapts to, and responds to a variety of internal and external stimuli. Manifestations of neuroplasticity in the adult central nervous system (CNS) include alterations of dendritic function, synaptic remodeling, long-term potentiation (LTP),

underlying the pathophysiology and treatment of depressive illness.4,5 Advances toward specifying the contribution of genetic factors,6 psychosocial stressors,7,8 and geneenvironment interactions to susceptibility to depression are also taking place.9,10 It is anticipated that, in the next few years, combined use of genomic and proteomic strategies to refine complex psychiatric diseases into mechanism-based subcategories may ultimately facilitate the matching of specific target-based therapies to particular markers in certain patient subgroups.11 Of all brain systems, the monoaminergic neurotransmitter systems have received the greatest attention in neurobiological studies of depressive disorders. The implication of these systems in depression is based on several observations: (i) effective antidepressant drugs exert their primary biochemical effects by regulating intrasynaptic concentrations of serotonin and norepinephrine; and (ii) antihypertensives that deplete these monoamines sometimes precipitate depressive episodes in susceptible individuals. Furthermore, the monoaminergic systems are extensively distributed throughout the network of limbic, striatal, and prefrontal cortical (PFC) neuronal circuits implicated in the behavioral and visceral manifestations of mood disorders.12 Over the past 40 years, clinical studies have aimed to uncover specific flaws in these neurotransmitter systems in mood disorders by using various biochemical and neuroendocrine approaches. In fact, assessment of cerebrospinal fluid (CSF) chemistry, neuroendocrine responses to pharmacological challenge, and neuroreceptor and transporter binding have demonstrated a number of abnormalities of the serotonergic, noradrenergic, and other neurotransmitter and neuropeptide systems in mood disorders. Although such studies have been useful in the past, they have proved to be of limited value in clarifying the particular pathophysiology of depressive disorders. In order to clarify the biological underpinnings of these disorders,

218

Cellular plasticity and resilience and severe mood disorders - Charney

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

axonal sprouting, neurite extension, synaptogenesis, and neurogenesis. In this perspective paper, we describe studies identifying possible structural, functional, and cellular abnormalities associated with depressive disordersthe potential cellular underpinnings of these micro- and macromorphological brain changes.We suggest that therapeutics designed to enhance cellular plasticity and resilience, and to attenuate the activity of maladaptive stress-responsive systems may have considerable utility for the treatment of severe mood disorders.

Brain imaging studies in depressed patients

Positron emission tomography (PET) imaging studies have unveiled various abnormalities of glucose metabolism and regional cerebral blood flow (CBF) in limbic and PFC structures in patients with mood disorders.Although some disagreement exists regarding the specific locations and the direction of some of these abnormalities, unmedicated subjects with familial major depression show a consistent increase in regional CBF and metabolism in the amygdala, orbital cortex, and medial thalamus, and decreased metabolism and CBF in the dorsomedial/dorsal anterolateral PFC and anterior cingulate cortex ventral to the genu of the corpus callosum (ie, subgenual PFC) relative to healthy controls.16,17 These abnormalities suggest that limbic-thalamic-cortical and limbic-cortical-striatalpallidal-thalamic circuits, involving the amygdala, orbital and medial PFC, and anatomically related parts of the striatum and thalamus are involved in pathophysiology of depression. Additionally, these circuits have been implicated more generally in emotional behavior by electrophysiological, lesion analysis and brain mapping studies of humans and experimental animals.12,15 Some of these abnormalities reverse during symptom remission, suggesting that there are areas where neurophysiological activity may increase or decrease in order to mediate or respond to the emotional and cognitive manifestations of depression. However, CBF and metabolism do not completely normalize during effective antidepressant treatment in many of these areas. Morphometric magnetic resonance imaging (MRI) and postmortem investigations have also demonstrated alterations in cerebral structure that persist regardless of mood state and may contribute to the corresponding abnormalities of metabolic activity.16,17 Structural imaging studies have shown reduction in gray matter volumes in areas of the orbital and medial PFC, ventral striatum and

hippocampus, and enlargement of third ventricles in mood-disordered patients when compared to healthy controls. Complementary postmortem studies have demonstrated abnormal decreases in cortex volume, glial cell counts, and/or neuron size in the subgenual PFC, orbital cortex, dorsal anterolateral PFC, and amygdala.12,14,15-17 It remains unclear whether these alterations in brain structure represent developmental abnormalities that may increase an individuals susceptibility to abnormal mood episodes, compensatory changes to other pathogenic processes, or the sequelae of recurrent affective episodes per se.The clarification of these issues will in part depend on investigations that outline the onset of such abnormalities within the illness course, as well as determine whether they precede depressive episodes in individuals with a high familial risk for mood disorders. The lack of complete reproducibility among neuroimaging or postmortem studies is not altogether surprising, and the disparities likely represent variations in experimental design and in patient populations. Further research is needed in order to understand whether more specifically defined subtypes of depression or mood disorders are associated with any specific abnormality.18 The marked reduction in glial cells in these regions has been especially interesting, given the tremendous recent progress in our understanding of the critical roles of glial cells in regulating neuronal function. Thus, there is now compelling evidence that radial glial cells have the potential not only to guide newly born neurons, but also to self-renew and to generate both neurons and astrocytes. Furthermore, recent data have shown that astrocytes increase the number of mature, functional synapses on CNS neurons sevenfold, demonstrating that CNS synapse number can be profoundly regulated by glia. Glial cells are also known to play critical roles in regulating synaptic glutamate levels, CNS energy homeostasis, and the liberation of trophic factors, which in turn participate in the development and maintenance of synaptic networks formed by neuronal and glial processes.16,17,19-22 Glial function abnormalities could thus prove essential to structural plasticity impairments and overall pathophysiology of mood disorders.

Stress and brain morphology

The majority of studies of adaptive neuronal plasticity in response to stress, as well as hypothalamo-pituitaryadrenal (HPA) axis hormones, have focused on the hip-

219

Clinical research
pocampus. This is partly due to early studies on neuronal populations of the limbic brain regions, including the dentate gyrus granule cell layer, and the CA1 and CA3 pyramidal cell layers. These cell layers and their connections (mossy fiber pathway and Schaffer collateral) have long been used as cellular models of learning and memory (ie, LTP). However, it is clear that stress and glucocorticoids also influence the survival and plasticity of neurons in other brain regions (such as PFC, vide infra) that have not yet been studied in the same detail as the hippocampus. Dendritic remodeling of hippocampal neurons is one of the best-characterized effects of stress on cellular morphology.23,24 Dendritic remodeling is deeply observed in the CA3 pyramidal neurons as atrophy-decreased number and length of the apical dendritic branches.This stress-induced atrophy of CA3 neurons results after 2 to 3 weeks of exposure to restraint stress or more long-term social stress, and has been observed in rodents and tree shrews.23,24 Although the effects of chronic stress in the CA3 layer tend to be most pronounced, slight structural changes are also found in the CA1 and dentate gyrus following a 1-month multiple stress paradigm.25 Profound alterations in mossy fiber terminal morphology and significant synapse loss have also been described.The hippocampus has a very high concentration of glutamate and expresses both glucocorticoid (GR) and mineralcorticoid (MR) corticosteroid receptors, though these may be relatively scarce in the hippocampus of primates,26,27 and more prevalent in cortical regions. MR activation in the hippocampus (CA1) is associated with reduced calcium currents, while GR activation leads to increased N-methyl-D-asparate (NMDA) receptor throughput and increased calcium currents that could predispose to neurotoxicity. In fact, increasing evidence implicates glutamatergic neurotransmission in stress-induced hippocampal atrophy and death. Histopathological changes in rat PFC after corticosterone administration have recently been described although this area has not been as comprehensively studied as the hippocampus. Using a Golgi-Cox procedure, Wellman28 examined pyramidal neurons in layers II and III of the medial PFC, quantifying dendritic morphology in three dimensions. In this study, he demonstrated a significant rearrangement of apical dendrites in corticosterone-treated animals, with an increase in the dendritic material proximal to the soma and a decrease in distal dendritic material. This suggests that stress may result in a significant reorganization of the apical dendritic arbor in medial PFC in rats. It is noteworthy that glucocorticoids may exert deleterious effects on neural plasticity and morphology, since a significant percentage of mood disorder patients show some form of HPA axis activation. It has been hypothesized that the depressive subtypes most frequently associated with HPA activation are also the most likely to be associated with reductions in hippocampal volume.23 Many patients with Cushings disease, in which pituitary gland adenomas cause cortisol hypersecretion, also show marked depressive symptoms, as well as hippocampal atrophy. Moreover, some patients with Cushings disease also show reduced hippocampal volumes, correlating inversely with plasma cortisol concentrations. Corrective surgical treatment results in an enlargement of hippocampal volume in proportion to the treatment-associated decrease in urinary free cortisol concentrations.29,30 HPA axis hyperactivity in mood disorder patients has been demonstrated by a variety of techniques/measures, including increased cortisol levels in plasma (especially at the circadian nadir), urine, and CSF, increased cortisol response to adrenocorticotropic hormone (ACTH), blunted ACTH response to corticotropin-releasing hormone (CRH) challenge, enlarged pituitary and adrenal glands, and reduced CRH receptor density in the brain (presumably reflecting a compensatory downregulation to sustained CRH elevations) at postmortem examination. In both unipolar and bipolar patients, reduced corticosteroid receptor feedback has been implicated in this process by challenge studies with dexamethasone and dexamethasone plus CRF.31,32 The results of recent longitudinal studies investigating the effects of early life stress and inherited variation in monkey hippocampal volumes underscore the need for caution when interpreting the clinical neuroimaging studies described above. These longitudinal studies in monkeys randomized paternal half-siblings (monkeys raised apart from one another by different mothers in the absence of fathers) to one of three postnatal conditions that interfered with various facets of early maternal care. Paternal half-siblings with small adult hippocampal volumes showed an initial larger relative increase in cortisol level following removal of all mothers after weaning.33 However, plasma cortisol levels 3 and 7 days later did correlate with hippocampal size. These studies suggest that small hippocampal volume also reflects an inherited trait, and emphasize the need for caution in the simple attribution of causality in the cross-sectional morphometric studies of the hippocampus in humans.

220

Cellular plasticity and resilience and severe mood disorders - Charney

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Stress effects on cellular plasticity and resilience

In addition to the cellular mechanisms described above, it is now clear that stressors may exert major effects on cellular plasticity and resilience by regulating the expression and function of growth factor cascades.33,34 Neurotrophic factors (eg, nerve growth factor [NGF] and brain-derived neurotrophic factor [BDNF]), as well as cytokines, insulin-like growth factor1 (IGF-1), and glial-derived neurotrophic factor (GDNF), increase cell survival.35,36 These factors promote cell survival through the suppression of intrinsic, cellular apoptotic machinery, rather than by inducing cell survival pathways. This occurs via binding of these factors to membrane receptors and regulation of intracellular signal transduction pathways that can control apoptosis, including regulation of Bcl-2 family members. Mitogen-activated protein (MAP) kinase cascade, the phosphatidylinositol-3 kinase (PI-3K)/Akt pathway, and the PI-3K cascade are currently thought to be responsible for mediating many of the effects of neurotrophic factors.37 The family of receptors known as Trks, which contain an intrinsic tyrosine kinase domain, mediates neurotrophic factor signaling. Nerve growth factor binds to the TrkA receptor, while BDNF binds to TrkB. The resulting receptor activation results in phosphorylation and activation of effectors, including PI-3K, as well as protein coupling leading to of the MAP kinase cascade activation. Recent studies have shown that MAP kinase cascade activation can inhibit apoptosis by inducing the phosphorylation of Bad (a major proapoptotic protein) and increasing the expression of Bcl-2 (a major antiapoptotic protein). This increased Bcl-2 expression likely involves a protein known as the cyclic adenosine monophosphate (cAMP) response element binding protein (CREB).38,39 Phosphorylation of Bad takes place via activation of a downstream target of the MAP kinase cascade, ribosomal S-6 kinase (Rsk). This phosphorylation by Rsk promotes the inactivation of Bad. Additionally, Rsk activation mediates the actions of the MAP kinase cascade and neurotrophic factors on the expression of Bc1-2. Rsk can phosphorylate CREB, leading to induction of Bcl-2 gene expression. A growing body of evidence indicates that not only is Bcl-2 neuroprotective, but also that it exerts neurotrophic effects and promotes neurite sprouting, neurite outgrowth, and axonal regeneration.40-43

Recently, it has been demonstrated that chronic stress (21 days foot-shock) induces a marked and persistent hyperphosphorylation of an extracellular response kinase (ERK) in higher PFC layer dendrites, while phosphoCREB was reduced in the frontal cortex and other cortical regions.44 Since CREB is phosphorylated and activated by phospho-ERK1/2 directly, this reduction indicates that chronic stress could downregulate CREB phosphorylation indirectly, and subsequently downregulate the transcription of some genes such as Bcl-2 and BDNF. In this context, it is worth mentioning that a recent study revealed that severe stress exacerbates stroke outcome by suppressing Bcl-2 expression.45 In this study, stressed mice expressed approximately 70% less Bcl-2 mRNA than unstressed mice following stroke. In addition, stress greatly exacerbated stroke in control mice, but not in transgenic mice that express increased neuronal Bcl-2. High corticosterone concentrations were significantly correlated with a greater stroke size in wild-type mice, but not in transgenic mice overexpressing Bcl-2. Therefore, enhanced Bcl-2 expression seems to offset the potentially harmful consequences of stress-induced neuronal endangerment, and suggests that pharmacologically induced upregulation of Bcl-2 may be useful in the treatment of a various disorders that have been linked to endogenous or acquired impairments of cellular resilience. It is now clear that the neurotrophic factorERK1/2MAPKBcl-2 signaling cascade has a critical role in cell survival in the CNS and that a fine balance exists between the levels and activities of cell survival and cell death factors. BDNFERK1/2CREBBcl-2 cascade dysregulation may be a key mechanism via which prolonged stress induces atrophy of select vulnerable neuronal subpopulations, distal dendrites, or both. Although dysregulation of this cascade most likely results in decreased neuronal survival, the differential survival is likely modulated not only by region-specific expression of protective factors, but also by the network properties of vulnerable structures. Therefore, it is likely that the dynamics of the impairments of cellular plasticity and resilience are determined by intrinsic properties of the affected regions. There is emerging evidencemainly from postmortem studiessupporting a role for abnormalities in neurotrophic signaling pathways in depression. Decreased levels of CREB, BDNF, and the TrkB receptor have been described in suicide victims.46-48 Depressed individuals may also have genetic abnormalities in CREB and BDNF.

221

Clinical research
Sequence variations in the CREB1 gene have been observed in depressed women.6 A coding variant of BDNF may be associated with the personality trait of neuroticism, which is a risk factor for depression.49 Furthermore, two recent studies50,51 suggest that a polymorphism in the proBDNF molecule is associated with bipolar disorder (a condition in which depressive episodes are accompanied by manic episodes). This polymorphism is associated with alterations in BDNF trafficking and secretion in vitro, as well as with alterations in hippocampal working memory in humans.52 Therefore, an opportunity exists to study the interaction of life stress, signal transductionrelated genes, neuroimaging abnormalities consistent with deficient structural plasticity, and susceptibility to depression.15 been reported to occur with conditions that stimulate neuronal activity (eg, enriched environment, learning, exercise). This suggests that neurogenesis is positively regulated by, and might be reliant on, neuronal plasticity. The enhancement of hippocampal neurogenesis following chronic antidepressant use highlights the level to which these efficacious treatments can regulate long-term neuroplastic processes in the brain. Since stress and antidepressants have opposite effects on hippocampal neurogenesis, it is likely that the clinical symptoms of depression are related to changes in hippocampal neurogenesis. In order to assess whether antidepressant-induced hippocampal neurogenesis is functionally relevant, Santarelli and associates64 utilized both genetic and radiological methods to show that disruption of antidepressantinduced neurogenesis blocked behavioral responses to antidepressants. In this study, serotonin 1A receptor null mice were insensitive to the neurogenic and behavioral effects of fluoxetine, a serotonin selective reuptake inhibitor. In mice, X-irradiation of the hippocampus prevented the neurogenic and behavioral effects of two classes of antidepressants. Together, the above findings suggest that some of the behavioral effects observed with chronic antidepressant use may be mediated by the stimulation of neurogenesis in the hippocampus. However, as Kempermann65 clearly articulated, much more research is required in order to adequately link changes in adult hippocampal neurogenesis to the pathophysiology and treatment of depression. Agents capable of reversing the hypothesized impairments of cellular resilience, reductions in brain volume, and cell death or atrophy in depression have the potential of becoming new therapeutic classes of antidepressant drugs. New molecular targets might include phosphodiesterase inhibitors that increase CREB phosphorylation, MAP kinase phosphatase inhibitors that increase expression of the antiapoptotic protein bcl-2, presynaptic glutamate receptor subtypes that attenuate glutamate release, amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) potentiators that increase BDNF expression, and NMDA antagonists that enhance plasticity and cell survival.14

Antidepressant mechanisms and neurotrophic signaling cascades

An increasing amount of evidence suggests that antidepressants regulate neurotrophic signaling cascades. Antidepressant treatment increases CREB phosphorylation and CREB-mediated gene expression in mice limbic brain regions.53 Various classes of chronic antidepressant treatments, as well as electroconvulsive treatment (ECT), upregulate CREB and BDNF expression, suggesting that the CREB cascade and BDNF are common post-receptor targets of antidepressants.54,55 This increase is exclusively seen after chronic use, thus corresponding to the onset clinical antidepressant effects with these therapies. Additional evidence that relates upregulation of these pathways and antidepressant treatment comes from antidepressant-like performance in behavioral models.56 In rats, CREB overexpression in the dentate gyrus or BDNF injection leads to an antidepressant-like effect in the learned-helplessness paradigm and the forced swim test model of antidepressant efficacy.57-59 Chronic antidepressant treatment also increases the neurogenesis of dentate gyrus granule cells.60-62 This effect has not been observed with acute antidepressant treatment. These studies show that chronic administration of different classes of antidepressants and ECT lead to an increase in the proliferation and survival of new neurons. Lithium, an effective antidepressant potentiating agent, also increases neurogenesis in the dentate gyrus.63 It is noteworthy that in contrast to the findings seen with chronic antidepressant use, increases in neurogenesis do not occur with chronic administration of nonantidepressant psychotropic medications. Increases in neurogenesis have

Concluding comments
A substantial body of evidence suggests that impairments in neuroplasticity and cellular resilience play a central role in the underlying biology of mood disorders.Additionally, there is a growing appreciation that new medications that

222

Cellular plasticity and resilience and severe mood disorders - Charney

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

simply imitate traditional drugs, those aiming to directly or indirectly alter monoaminergic throughput, may be of limited benefit to those patients with refractory depression. Those strategies assume that the target circuits are functionally intact and that changes in synaptic activity will alter the postsynaptic throughput of the system. The evidence discussed here indicates that, in addition to neurochemical changes, many patients suffering from mood disorders also have marked structural alterations in crucial neuronal circuits. Therefore, in order to obtain an optimal treatment response, it will most likely be crucial to provide both trophic and neurochemical support. The aim of the trophic support would be to enhance and maintain normal synaptic connectivity, therefore permitting the chemical signal to restore maximum functioning of vital circuits essential for normal affective functioning. In fact, preliminary studies suggest that regional structural changes in the brains of patients with mood disorders may be related with not only severity and duration of the illness, but also with altered treatment response to pharmacotherapy and ECT. The evidence also suggests that, somewhat similar to the treatment of other chronic medical conditions, such as hypertension and diabetes, prompt and sustained treatment may be necessary to prevent many of the injurious long-term sequelae associated with mood disorders. Although the evidence hints at an association between hippocampal atrophy and illness duration in depressed patients, it remains unclear whether the volumetric and cellular changes observed in other brain areas are related

to affective episodes. In fact, some studies have described reduced gray matter volumes and increased ventricle size in patients with mood disorders at the time of their first episode and in early onset of the disease.12,15 In conclusion, relevant genotypes for mood disorders are being identified, and clinical research techniques are now capable of defining neurobiological phenotypes. Similarly, results from transcriptomic and proteomic studies which identified neurotrophic signaling as targets for the longterm actions of antidepressants and mood stabilizers have played a role (along with neuroimaging and postmortem brain studies) in a reconceptualization about the pathophysiology, course, and optimal long-term treatment of severe mood disorders. These data suggest that, while mood disorders are clearly not classical neurodegenerative diseases, they are in fact associated with impairments of cellular plasticity and resilience.As a consequence, there is a growing appreciation that optimal long-term treatment will most likely be achieved by attempting to prevent the underlying disease progression and its attendant cellular dysfunction, rather than exclusively focusing on the treatment of signs and symptoms.We are optimistic that a new generation of research will clarify the relation among environmental and genetic risk factors to quantify the risk for the development of depression more precisely. These advances will result in a dramatically different diagnostic system based upon etiology, and ultimately in the discovery of new approaches to the prevention and treatment of some of mankinds most devastating and least understood illnesses.

REFERENCES
1. World Health Organization, Mental Health: New Understanding, New Hope. Geneva, Switzerland: World Health Organization; 2001. 2. Evans DS, Charney DS. Mood disorders and medical illness: a major public health problem. Biol Psychiatry. 2003;54:177-180. Editorial. 3. Charney DS, Barlow DH, Botteron K, et al. In: Kupfer DJ, First MB, Reiger DA, eds. A Research Agenda for DSM-V. Washington, DC: American Psychiatric Association; 2002:31-83. 4. Manji HK, Duman RS. Impairments of neuroplasticity and cellular resilience in severe mood disorders: implications for the development of novel therapeutics. Psycopharmacol Bull. 2001;35:5-49. 5. Nestler EJ, Barrot M, DiLeone RJ, Eisch AJ, Gold SJ, Monteggia LM. Neurobiology of depression. Neuron. 2002;34:13-25. 6. Zubenko GS, Stiffler HBH, III, Brechbiel A, Zubenko WN, Maher BS, Marazita ML. Sequence variations in CREB1 cosegregate with depressive disorders in women. Mol Psychiatry. 2003;8:622-618.

7. Kendler KS, Karkowski LM, Prescott CA. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry. 1999;156:837-841. 8. Kendler KS, Thornton LM, Prescott, Prescott CA. Gender differences in the rates of exposure to stressful life events and sensitivity to their depressogenic effects. Am J Psychiatry 2001;158:587-593. 9. Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science. 2003;301:386389. 10. Kendler KS, Karkowski-Shuman L. Stressful life events and genetic liability to major depression: genetic control of exposure to the environment. Psychol Med. 1997;7:539-547. 11. Gould TG, Manji HK. The molecular medicine revolution and psychiatry: bridging the gap between basic neuroscience research and clinical psychiatry. J Clin Psychiatry. 2004. In press. 12. Drevets WC. Neuroimaging and neuropathological studies of depression: implications for the cognitive-emotional features of mood disorders Curr Opin Neurobiol. 2001;11:240-249.

223

Clinical research
Plasticidad celular, resiliencia y fisiopatologa de los trastornos afectivos graves
Los avances recientes en la identificacin de los circuitos neurales, los mecanismos neuroqumicos y de transduccin de seales que participan en la fisiopatologa y el tratamiento de los trastornos afectivos han conducido hacia un significativo progreso en la comprensin de los papeles de los factores genticos y de los estresores psicosociales. Los sistemas de neurotransmisin monoaminrgica han concitado la mayor atencin, en parte, debido a la observacin que los antidepresivos eficaces ejercen sus efectos bioqumicos primarios a travs de la regulacin de las concentraciones intrasinpticas de serotonina y noradrenalina. Adems, los sistemas monoaminrgicos se distribuyen extensamente a travs de la red de circuitos neuronales lmbicos, estriatales y corticales prefrontales, que se piensa son los responables de las manifestaciones conductuales y viscerales de los trastornos afectivos. Un nmero creciente de estudios de neuroimgenes, neuropatolgicos y bioqumicos revelan deterioros en la plasticidad celular y la resiliencia en pacientes que padecen trastornos afectivos graves y recurrentes. En este artculo se describen estudios que identifican posibles anormalidades estructurales, funcionales y celulares que se asocian con los trastornos depresivos, los que constituyen potencialmente los fundamentos celulares de estas enfermedades. Se sugiere que los frmacos diseados para incrementar la plasticidad celular y la resiliencia, y atenuar la actividad de los sistemas que determinan una mala adaptacin al estrs, pueden ser tiles para el tratamiento de los trastornos afectivos graves.

Plasticit cellulaire, rsilience et physiopathologie des troubles de lhumeur svres

Les progrs rcents concernant lidentification des substances chimiques et circuits neuronaux et des mcanismes de transduction du signal impliqus dans la physiopathologie et le traitement des troubles de lhumeur ont amlior la comprhension des rles des facteurs gntiques et des facteurs psychosociaux de stress. Les systmes de neurotransmetteurs monoaminergiques ont retenu le plus dattention, en partie parce que lon a remarqu que les principaux effets biochimiques des antidpresseurs efficaces sexercent en rgulant les concentrations intrasynaptiques de srotonine et noradrnaline. De plus, les systmes monoaminergiques sont largement distribus dans le rseau des circuits neuronaux limbiques, striataux et du cortex prfrontal supposs dterminer les manifestations comportementales et organiques des troubles de lhumeur. De plus en plus dtudes de neuro-imagerie, neuropathologie et biochimie soulignent laltration de la plasticit cellulaire et de la rsilience chez les patients souffrant de troubles de lhumeur svres et rcurrents. Dans cet article, nous dcrivons des tudes identifiant de possibles anomalies structurelles, fonctionnelles et cellulaires associes aux troubles dpressifs, qui constituent les bases cellulaires potentielles de ces pathologies. Nous suggrons que les mdicaments conus pour augmenter la plasticit cellulaire et la rsilience et attnuer lactivit des systmes inadapts de rponse aux stress pourraient tre utiles au traitement des troubles de lhumeur svres.

13. Manji HK, Drevets WC, Charney DS. The cellular neurobiology of depression. Nat Med. 2001;7:541-547. 14. Payne J, Quiroz J, Gould T, Zarate C, Manji H. The cellular neurobiology of bipolar disorder. In: Charney DS, Nestler EJ, eds. Neurobiology of Mental Illness. New York, NY: Oxford University Press; 2004:397-420. 15. Charney DS, Manji HK. Life stress, genes, and depression: multiple pathways lead to increased risk and new opportunities for intervention. Science STKE. 2004;225:1-11. 16. Ongur D, Drevets WC, Price JL. Glial reduction in the subgenual prefrontal cortex in mood disorders. Proc Natl Acad Sci U S A. 1998;95:1329013295. 17. Rajkowska G. Postmortem studies in mood disorders indicate altered numbers of neurons and glial cells. Biol Psychiatry. 2000;48:766-777.

18. Lenox RH, Gould TD, Manji HK. Endophenotypes in bipolar disorder. Am J Med Genet. 2002;114:391-406. 19. Coyle JT, Schwarcz R. Mind glue: implications of glial cell biology for psychiatry. Arch Gen Psychiatry. 2000;57:90-93. 20. Haydon PG. Glia: listening and talking to the synapse. Nat Rev Neurosci. 2001;2:185-193. 21. Rajkowska G, Miguel-Hidalgo JJ, Wei J, et al. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol Psychiatry. 1999;45:1085-1098. 22. Ullian EM, Sapperstein SK, Christopherson KS, Barres BA. Control of synapse number by glia. Science. 2001;291:657-661. 23. Sapolsky RM. Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders. Arch Gen Psychiatry. 2000;57:925-935.

224

Cellular plasticity and resilience and severe mood disorders - Charney

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

24. McEwen BS. Stress and hippocampal plasticity. Annu Rev Neurosci. 1999;22:105-122. 25. Sousa N, Lukoyanov NV, Madeira MD, Almeida OF, Paula-Barbosa MM. Reorganization of the morphology of hippocampal neuritis and synapses after stress-induced damage correlates with behavioral improvement. Neuroscience. 2000;97:253-266. 26. Sanchez MM, Young LJ, Plotsky PM, Insel TR. Distribution of corticosteroid receptors in the rhesus brain: relative absence of glucocorticoid receptors in the hippocampal formation. J Neurosci. 2000;20:4657-4668. 27. Patel PD, Lopez JF, Lyons DM, Burke S, Wallace M, Schatzberg AF. Glucocorticoid and mineralocorticoid receptor mRNA expression in squirrel monkey brain. J Psychiatr Res. 2000;34:383-392. 28. Wellman CL. Dendritic reorganization in pyramidal neurons in medial prefrontal cortex after chronic corticosterone administration. J Neurobiol. 2001;49:245-253. 29. Starkman MN, Giordani B, Gebarski SS, Berent S, Schork MA, Schteingart DE. Decrease in cortisol reverses human hippocampal atrophy following treatment of Cushings disease. Biol Psychiatry. 1999;46:1595-1602. 30. Simmons NE, Alden TD, Thorner MO, Laws ER. Serum cortisol response to transsphenoidal surgery for Cushing disease. J Neurosci. 2001;95:1-8. 31. McQuade R, Young AH. Future therapeutic targets in mood disorders: the glucocorticoid receptor. Br J Psychiatry. 2000;177:390-395. 32. Reul JM, Holsboer F. Corticotropin-releasing factors receptors 1 and 2 in anxiety and depression. Curr Opin Pharmacol. 2002;2:23-33. 33. Lyons DM, Yasng C, Sawyer-Glover AM, Moseley ME, Schatzberg AF. Early life stress and inherited variation in monkey hippocampal volumes. Arch Gen Psychiatry. 2001;58:1145-1151. 34. Nibuya M, Takahashi M, Russell DS, Duman RS. Repeated stress increases catalytic TrkB mRNA in rat hippocampus. Neurosci Lett. 1999;267:81-84. 35. Mamounas LA, Blue ME, Siuciak JA, ltar CA. Brain-derived neurotrophic factor promotes the survival and sprouting of serotonergic axons in rat brain. J Neurosci. 1995;15:7929-7939. 36. Henderson CE, Pettman B. Neuronal cell death. Neuron.1998;20:633-647. 37. Theonen H. Neurotrophins and neuronal plasticity. Science. 1995;270:593598. 38. Segal RA, Greenberg ME. Intracellular signaling pathways activated by neurotrophic factors. Annu Rev Neurosci. 1996;19:463-489. 39. Riccio A, Ahn S, Davenport CM, Blendy JA, Ginty DD. Mediation by a CREB family transcription factor of NGF-dependent survival of sympathetic neurons. Science. 1999;286:2358-2361. 40. Bonni A, Brunet A, West AE, Datta SR, Takasu MA, Grenberg ME. Cell survival promoted by the ras-MAPK signaling pathway by transcriptiondependent and -independent mechanisms. Science. 1999;286:1358-1362. 41. Chen DF, Schneider GE, Martinou JC, Tonegawa S. Bcl-2 promotes regeneration of severed axons in mammalian CNS. Nature. 1997;385:434-439. 42. Chen DF, Tonegawa S. Why do mature CNS neurons of mammals fail to re-establish connections following injuryfuctions of bcl-2. Cell Death Differentiation. 1998;5:816-822. 43. Holm KH, Cicchetti F, Bjorklund L, et al. Enhanced axonal growth from fetal human bcl-2 transgenic mouse dopamine neurons transplanted to the adult rat striatum. Neuroscience. 2001;104:397-405. 44. Trentani A, Kuipers SD, Ter Horst GJ, Den Boer JA. Selective chronic stress-induced in vivo ERK1/2 hyperphosphorylation in medial prefrontocortical dendrites: implications for stress-related cortical pathology? Eur J Neurosci. 2002;15:1681-1691. 45. DeVries AC, Joh HD, Bernard O, Hattori K, Hurn PD, Traystman RJ, Alkayed NJ. Social stress exacerbates stroke outcome by suppressing Bcl-2 expression. Proc Natl Acad Sci U S A. 2001;98:11824-11828.

46. Dwivedi Y, Rao JS, Rizavi HS, et al. Abnormal expression and functional characteristics of cyclic adenosine monophosphate response element binding protein in postmortem brain of suicide subjects. Arch Gen Psychiatry. 2003;60:273-282. 47. Dwivedi Y, Rizavi HS, Conley RR, Roberts RC, Tamminga CA, Pandey GN. Altered gene expression of brain-derived neurotrophic factor and receptor tyrosine kinase B in postmortem brain of suicide subjects. Arch Gen Psychiatry. 2003;60:804-814. 48. Yamada Y, Yamamoto M, Ozawa H, Riederer P, Sito T. Reduced phosphorylation of cyclic AMP-responsive element binding protein in the postmortem orbitofrontal cortex of patients with major depressive disorder. J Neural Transm. 2003;110:671-680. 49. Sen S, Neese RM, Stoltenberg SF, et al. A BNDF coding variant is associated with the NEO personality inventory domain neuroticism, a risk factor for depression. Neuropsychopharmacology. 2003;28:397-401. 50. Sklar P, Gabriel SB, McInnis MG, et al. Family-based association study of 76 candidate genes in bipolar disorder; BDNF is a potential risk locus. Mol Psychiatry. 2002;7:579-593. 51. Neves-Pereira M, Mundo E, Muglia P, King N, Macciardi F, Kennedy JL. The brain-derived neurotrophic gene confers susceptibility to bipolar disorder: evidence from a family-based association study. Am J Hum Genet. 2002;71:651-655. 52. Egan MF, Kojima M, Callicott JH, et al. The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell. 2003;112:257-269. 53. Thome J, Sakai N, Shin K, et al. cAMP response element-mediated gene transcription is upregulated by chronic antidepressant treatment. J Neurosci. 2000;20:4030-4036. 54. Nibuya M, Morinobu S, Duman RS. Regulation of BDNF and TrkB mRNA in rat brain by chronic electroconvulsive seizure and antidepressant drug treatments. J Neurosci. 1995;15:7539-7547. 55. Nibuya M, Nestler EJ, Duman RS. Chronic antidepressant administration increased the expression of cAMP response element binding protein (CREB) in rat hippocampus. J Neurosci. 1996;16:2365-2372. 56. Duman RS, Malberg J, Thome J. Neural plasticity to stress and antidepressant treatment. Biol Psychiatry. 1999;46:1181-1191. 57. Siuciak JA, Lewis DR, Wiegand SJ, Lindsay RM. Antidepressant-like effect of brain-derived neurotrophic factor (BDNF). Pharmacol Biochem Behav. 1997;56:131-137. 58. Shirayama Y, Chen AC, Nakagawa S, Russell DS, Duman RS. Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. J Neurosci. 2002;22:3251-3261. 59. Chen AC, Shirayama Y, Shin KH, Neve RL, Duman RS. Expression of the cAMP response element binding protein (CREB) in hippocampus produces an antidepressant effect. Biol Psychiatry. 2001;49:753-762. 60. Jacobs BL, Praag H, Gage FH. Adult brain neurogenesis and psychiatry: a novel theory of depression. Mol Psychiatry. 2000;5:262-269. 61. DSa C, Duman R. Antidepressants and neuroplasticity. Bipolar Disord. 2002;4:183. 62. Manev H, Uz T, Smalheiser NR, Manev R. Antidepressants alter cell proliferation in the adult brain I vivo and in neural cultures in vitro. Eur J Pharmacol. 2001;411:67-70. 63. Chen G, Rajkowska G, Du F, Seraji-Bozorgzaed N, Manji HK. Enhancement of hippocampal neurogenesis by lithium. J Neurochem. 2002;75:1729-1734. 64. Santarelli L, Saxe M, Gross C, et al. Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science. 2003;301:805-809. 65. Kempermann G. Regulation of adult hippocampal neurogenesisimplications for novel theories of major depression. Bipolar Disord. 2002;4:17-33.

225

Free paper
Texture analysis of the brain: from animal models to human applications
Jean-Franois J. Nedelec, PhD; Olivier Yu, MD, PhD; Jacques Chambron, MD, PhD; Jean-Paul Macher, MD
Animal model

Magnetic resonance imaging (MRI) is widely used to image brain in vivo both in studies in animal models and for human diagnosis. A large part of the value of MRI is due to the fact that soft tissue contrast is enhanced by the substantial variation in the T1 and T2 relaxation times between tissues. It may be possible to use an alternative approach, which does not rely on the absolute measurement of relaxation times. Generally speaking, textures are complex visual patterns composed of entities, or subpatterns, that have characteristic brightness, color, slope, size, etc. Thus, texture can be regarded as a similarity grouping in an image. The properties of the local subpattern give rise to the perceived lightness, uniformity, density, roughness, regularity, linearity, frequency, phase, directionality, coarseness, randomness, fineness, smoothness, and granulation. The purpose here is to illustrate how texture analysis can be used in animal models and in human clinical applications, as well as in the search for further pharmacological applications in humans. Thus, this article summarzes three different MRI studies in (i) rats, using the lipocarpine epileptic rat model as an animal model; (ii) patients with Alzheimers disease; and (iii) patients with schizophrenia.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:227-233.

he epilepsy induced in the rat by lithium pilocarpine (Li-Pilo) constitutes an animal model of human mesial temporal lobe epilepsy.1 Neuronal damage is mainly detected in hippocampus, thalamus, piriform cortex, entorhinal cortex, and neocortex. At present, magnetic resonance imaging (MRI) is the most sensitive imaging method for the study of mesial temporal lobe epilepsy, but the examination is often restricted to the detection of hyperintensities. In previous studies, we used MRI to explore the morphological changes resulting from an injection of Li-Pilo that leads to epilepsy.2,3 In order to improve the predictive value of MRI images, we performed a texture analysis4 of MRI images combined with a discriminant analysis. The results presented here indicate that this procedure can detect defects that cannot be visualized by classic examination and permits a more correct classification of the images. Materials and methods MRI protocol MRI images were recording using an MRI scanner operating at 4.7 tesla (SMIS, UK).The rats were anaesthetized for MRI by an intramuscular injection of 37 mg/kg ketaAuthor affiliations: FORENAP, Rouffach, France (Jean-Franois J. Nedelec, PhD; Jean-Paul Macher, MD); Institut de Physique Biologique, Facult de Mdecine, Strasbourg, France (Olivier Yu; Jacques Chambron) Address for correspondence: Jean-Franois J. Nedelec, PhD, FORENAP, Institute for Research in Neuroscience and Neuropsychiatry, BP29, 68250 Rouffach, France (e-mail: jeanfrancois.nedelec@forenap.asso.fr) This article is published following the 14th Biological Interface Conference held in Rouffach, France, between October 1 and 5, 2002, on the theme of Drug Development. Other articles from this meeting can be found in Dialogues in Clinical Neuroscience (2002, Vol 4, No 4). www.dialogues-cns.org

Keywords: texture analysis; brain; rat; epilepsy; MRI; gray-level dependence histogram; Alzheimers disease; schizophrenia Copyright 2004 LLS SAS. All rights reserved

227

Free paper
Selected abbreviations and acronyms
AD GLDH Li-Pilo MMSE ROI SE Alzheimers disease gray-level dependence histogram lithium-pilocarpine Mini-Mental State Examination region of interest status epilepticus ity of the texture). The software MaZda was used to analyze the texture of the digitized images within all regions of interest (ROI) and yielded 300 parameters.5 Statistical analysis The statistical analysis was carried out using software from Statistica, Statsoft Inc. Discriminant analysis was used for multigroup classification. Using stepwise analysis, we checked the ability of each texture parameter to discriminate between two groups of ROIs, ie, presence or absence of lesions in piriform or entorhinal cortices. As a preliminary step, we determined the most important parameters that best discriminated the lesion ROIs from the safe ROIs observed before the Li-Pilo protocol. The question to be answered here is whether the two groups are well distinguished on the basis of the set of texture parameters. If the discrimination is successful on the basis of the set of selected parameters, it makes sense to classify particular piriform or entorhinal cortices in terms of group membership, ie, in terms of into which group they are most likely to be classified. The search for hidden defects could then be undertaken in the nonmodified images, obtained after the Li-Pilo protocol, in order to discriminate between lesion and safe ROIs. Results In all 21-day-old rats (n=20), pilocarpine injections led to SE within about 50 min. However, only 80% of rats were still epileptic after a mean delay of 70.224.6 days (meanSD). MRI images obtained before Li-Pilo treatment were considered as control group images (Figure 1) (64 ROIs were used for the texture analysis). Among the 20 rats followed for 4 months, 16 exhibited seizures, whereas 4 did not. Retrospectively, three groups of rats could be characterized according to type of images and the possibility of late epilepsy: Group A: 6 rats with obvious lesions characterized by a hypersignal on T2-weighted images in the piriform or entorhinal cortices 24 h after the SE (Figure 1; 44 ROIs were used for texture analysis); all these rats exhibited late epileptic seizures. Group B: 4 rats with control-like images (without any hypersignals), as shown in Figure 1, which did not present late epilepsy (34 ROIs were used for texture analysis). Group C: 10 rats with control-like images (without any hypersignals), as shown Figure 1, but which subse-

mine and 5.5 mg/kg xylazine.A T2-weighted spin-echo fastimaging method sequence (repetition time [TR]=3800 ms and echo time [TE]=80 ms) was used with 4-cm field of view, a 256256 pixel matrix, 1-mm thickness, upon coronal slices of the whole brain. Animals and Li-Pilo protocol Eleven 21-day-old, male, Sprague-Dawley rats were used for the experiments. The images of the 11 rats obtained before the injection of Li-Pilo served as control. All the rats first received lithium chloride (3 mEq/kg) intraperitoneally.After 18 h, the rats received a subcutaneous injection of pilocarpine (30 mg/kg) and 30 min later 1 mg/kg methylscopolamine intraperitoneally, in order to reduce the peripheral consequences of pilocarpine administration. Two hours after onset of status epilepticus (SE), the rats received 2 mg/kg diazepam by deep intramuscular injection in order to improve their survival. Images of all the rats were performed 24 h after onset of SE. Texture analysis Conventional texture analysis was performed using statistical methods, mostly based on first-order and secondorder histograms derived from the co-occurrence matrix, which describes the spatial gray level dependencies. Another possibility is the run-length matrix, which is the matrix of the run-length frequency occurring in the image for a certain angle of sight (lines of the same pixel level). This method has been fully described by Haralick.4 The co-occurrence matrix is based on the probability that pairs of pixels with a given level will appear. For each orientation (0, 45, 90, and 135) and for each distance between two pixels forming a pair, a number of co-occurrence matrix parameters may be calculated: contrast (an uneven texture provides large/high contrast values); correlation (relationship between two pixels); homogeneity (uniformity of the gray levels); and entropy (coarse-grained qual-

228

Texture analysis of the brain - Nedelec et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Control

Group A

ROI was then classified into one group or the other, according to the highest classification score. The above classification process was then used as a basis for prediction for the 114 apparently normal ROIs from the 57 brain slices of the rats in groups B and C. The resulting classification gave 84 control ROIs and 30 lesion ROIs. Indeed, only 2 rats had control ROIs and were safe (group B). About 50% of the lesion ROIs of the other 12 rats were distributed bilaterally (10 rats in group C and 2 rats in group B). During the 4 months clinical follow-up, 10 rats became epileptic and 4 rats remained nonepileptic, among which 2 had been incorrectly classified as epileptic (Table I). Discussion and conclusion The MRI study that was based only on the presence of hyperintensity signals in the piriform and entorhinal cortices predicted 6 late chronic epilepsy and 5 safe rats.This missed latent disease in 3 rats. The combined texture and discriminant analyses that were based on pixel pattern abnormalities selected 3 texture parameters that characterized structural abnormalities relevant to the hypersignal, both in the modified images of 6 rats and in the images of 4 rats with apparently nonmodified images, predicting the late chronic epilepsy in 10 rats.The classification based on early texture abnormalities in the piriform and entorhinal cortices improved the results of the regular MRI study.6

Group B

Group C

Figure 1. Magnetic resonance imaging (MRI) scans in rats before treatment (Control) and after treatments with lithium pilocarpine (Lipilo). Group A exhibited late epileptic seizures. Group B presented no late epilepsy. Group C had control-like images, but subsequently became epileptic.

quently became spontaneously epileptic (80 ROIs were used for texture analysis). Therefore, the conventional MRI study could not predict the fate of the 10 rats in group C, which did not display visible lesions in their brain images 24 h after SE, but subsequently became epileptic. The results of the texture analysis yielded 200 texture parameters in each ROI. Preliminary discriminant analysis yielded a classification function corresponding to the control group or group A. Each function was a linear combination of the features (or texture parameters) that yielded the best discrimination. For a given ROI, described by the texture parameters, a classification score was calculated from the classification functions. Each
Group Rats (n) Number of ROIs used in texture analysis Hyperintensity of piriform and entorhinal cortices Texture and discriminant analysis classification of ROIs Prediction of disease Effective chronic epilepsy

Human applications in AD
The method of gray-level dependence histograms (GLDH) as defined by Chetverikov for 2D7 and generalized to 3D by Kovalev and Petrou8-10 leads to derived features of texture anisotropy from MRI data. The aim
A 6 44 Yes 42 lesions 2 control 6 epilepsy Yes B 4 34 No 8 lesions 26 control 2 safe 2 epilepsy No C 10 80 No 22 lesions 58 control 10 epilepsy Yes

Table I. Groups and classification of rats. Group A exhibited late epileptic seizures. Group B presented no late epilepsy. Group C had control-like images, but subsequently became epileptic. ROI, region of interest.

229

Free paper
was to evaluate Alzheimers disease (AD) patients for a correlation between the anisotropic features and their score on the Mini-Mental State Examination (MMSE), which is routinely used to help diagnose AD.11 Methods Two groups of subjects were investigated and analyzed in this study: 12 control volunteers and 13 AD patients. The control group was matched with the AD group in terms of age and gender.The mean age (range) at time of investigation was 56.77 (39-72) years for the AD patients and 58.33 (47-72) years for the control volunteers. MRI T1-weighted images with coronal orientation were recorded for each subject. Each data set had 180180124 pixels and the voxel size was 0.93750.93751.5 mm. The scans were segmented to isolate the brain from external structures (eyes, ventricles, bones, etc).12 The brains were further segmented to isolate the white and gray matter, as well as the border between the two types of tissues. Because the texture analysis technique effectively counts the number of pairs of voxels that appear in the same relative position and have certain fixed gray values, the relative gray values of the voxels are extremely important. Thus, a normalization set is used in order to have the same relative gray level values for different scans: the smallest gray-level value is assigned to 1 and the highest to 255 for the segmented scan; 0 is assigned to the voxels that do not belong to the ROI. 3D texture representation: isotropy or anisotropy A coordinate system is defined as a cube of data cube in which the x and y axes form the plane of each slice, and the z axis is perpendicular to each slice. The azimuthal angle is measured on the x,y plane away from the direction of the x axis. The pair of values ,z defines a unique orientation in 3D space. We can then calculate the quantity h (,z;d). One component of h is the number of pairs of voxels that are at distance d from each other, along the direction ,z with one member of the pair having a gray value k and the other l. If the data are isotropic, the function h must be independent of direction and therefore a 3D representation is a sphere. Any deviation from this shape indicates anisotropy in the data. The 3D representation for a fixed distance is a closed digital surface, which is called an indicatrix. Projections of the orientation histograms can be obtained as illustrated in Figure 2 for a control subject and an AD patient. Feature extraction Three features are used to analyze the shape of the 3D indicatrix9: the anisotropy coefficient, the integral anisotropy measure or standard deviation, and the local mean curvature. Another set of features can be extracted by expanding the indicatrix in terms of spherical harmonics. The coefficients of such an expansion can characterize any 3D closed surface: coefficient A0,0 is the mean radius of the indicatrix; any other nonzero coefficient represents different types of anisotropy. Anisotropic features were extracted from four brain regions: the whole brain, white matter, gray matter, and the border between gray and white matter. In every single region, five different distances d were used: 0.9375, 1.5, 2, 2.5, and 3 mm. MMSE score and correlation with the isotropy coefficient The MMSE score is used to detect dementia. The maximum score is 30 (typically above 29 for healthy volunteers). Scores between 10 and 24 are considered to indicate mildto-moderate dementia cases, and scores below 10 indicate severe dementia. The scores obtained in the AD patients (named AD1 to AD13) and the control volunteers (named CO1 to CO12) are displayed in Table II.Two of the scores do not match the clinical diagnosis: AD3 and CO2. While many features correlate well with the MMSE scores, Figure 3 illustrates the best correlation (-0.876) with the MMSE score, which was obtained for the feature A1,1 in gray matter for a distance of 0.9375 mm. Subject AD3 is interesting because this patient was imaged before the onset of the first clinical symptoms, at a time when there may have been ongoing structural brain changes.

Figure 2. Projections of the orientation histogram on the z=0 plane obtained from MRI T1 images: from an Alzheimers disease patient (left) and a healthy volunteer (right). The isotropic features of the histogram are related to brain pathology.

230

Texture analysis of the brain - Nedelec et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Discussion and conclusion The GLDH method can be used to produce many features that strongly correlated with the MMSE scores when applied to the gray matter components of the MRI T1 scans. The features computed reflected the microtextural properties of the brain, ie, the texture anisotropy at scales of the order of 1 mm, rather than the shape of each ROI. Moreover, they correlate better with the condition of the subject rather than with age. In general, the AD brains presented greater anisotropy in their gray matter texture than the control brains. Both 2D features and 3D features correlate with the MMSE score, indicating that the information is already available in each individual layer.13

Methods Two groups of subjects were investigated and analyzed in this study: 19 control subjects and 21 patients with schizophrenia. The controls were matched for age, gender, and social class. 3D MRI T2-weighted images were collected for each subject. Each data set consisted of slices with a 0.856-mm spatial resolution with interslice distance of 3 mm. The images were segmented such that only the brain component was extracted for further analysis. The anisotropic sampling of the data along the z axis (interslice direction) is handled by an appropriate scaling factor of 3.5 (3/0.856) for all data sets. For the texture analysis, we use generalized co-occurrence matrices.14 3D texture analysis

Human applications in schizophrenia

Texture analysis can also provide feature parameters for different classes, which can then be used for classification. Generalized 4D co-occurrence matrices have been used to analyze the 3D MRI T2-weighted brain images from controls and patients with schizophrenia. The ROI approach has a number of potential problems: inter- and intraoperator reproducibility; difficulties detecting neuroanatomic boundaries; and the requirements that are of interest have to be specified from the outset. This approach does not need prior hypotheses and, because it is automated, reproducibility and comparability are ensured. The co-occurrence matrix is a generalized histogram, which records the frequency with which a certain combination of characteristics appear in the relevant position. Usually the main characteristic used is the gray value of the image, but other features can be used, such as gradient magnitude or relative orientation of the gradient vectors. Because they are independent of rotation and translation of the data, co-occurrence matrices offer

AD13 10

Alzheimers disease Patient MMSE score AD1 25 AD2 8 AD3 30 AD4 25 AD5 23 AD6 25 AD7 28 AD8 22 AD9 19 AD10 AD11 14 AD12 24 AD13 12

Controls Subject MMSE score CO1 30 CO2 28 CO3 30 CO4 29 CO5 30 CO6 30 CO7 29 CO8 30 CO9 30 CO10 30 CO11 30 CO12 30

AD2 8

6 A 1,1

AD11 AD9 AD8 AD4 AD1 AD12 CO5 CO4 CO10 CO6 CO11 CO1 CO12 CO3 CO6 CO9 30

2 AD6 0 10 15 20 MMSE AD5 25

AD3 CO7 AD7 CO2

Figure 3. Feature |A1, 1| in gray matter for d=0.9375 mm versus the score on the Mini-Mental State Examination (MMSE). G Alzheimers disease patient (AD1, AD2, AD3, etc); G control volunteers (CO1, CO2, CO3, etc).
Reproduced from reference 13: Segovia-Martinez M, Petrou M, Crum W. Texture features that correlate with the Mini Mental State Examination (MMSE) Score. EMBS Proc. 2001:910-913. Copyright 2001 IEEE.

Table II. Mini-Mental State Examination (MMSE) score for subjects with Alzheimers disease (AD1, AD2, AD3, etc) and controls (CO1, CO2, CO3, etc).

231

Free paper
descriptors that include these properties. The calculated co-occurrence matrix was w[g(i), g(j), a(i,j), d(i,j)]], where w is the frequency of the occurrence of a voxel pair i,j, with gradient magnitude g(i) and g(j) respectively, an angle a(i,j) between their gradient vectors, and a distance d(i,j) from each other. Results Three series of experiments were conducted: one considering the whole brain, one considering the bottom half of the brain, and one considering the bottom quarter of the brain.The division of the brain was performed by identifying the slices that are anatomically most similar to slices 12 and 24 of the anatomical atlas of Talairach and Tournoux.15 In each experiment, each element of the co-occurrence matrix was tested as a class discriminator according to the t test. The feature were selected by thresholding the t values using various limits. The classification was then performed using Statistica software. For the whole brain, the best results were obtained by retaining the features with t>4.5, and 14 subjects was misclassified. Using the bottom half of the brain, there were too many features with t>4.5 and so only the features with t>5.5 were used, and only 7 subjects were misclassified. When the bottom quarter of the brain was used, the features were so good that only features with t>7.5 were retained and only 2 subjects were misclassified, as illustrated in Figure 4. Conclusions The most significant conclusion is that the brains of patients with schizophrenia show structural differences from the brains of the control subject. Moreover, from the three series of analysis performed, it appears that these differences are located in the bottom quarter of the brain. Finally, it was demonstrated that the co-occurrence matrices could characterize the two classes of subjects with 90% accuracy using 3D T2-weighted MRI.

Perspectives
8 6 4 2 0 -2 -4 -6 -20

Control subject Schizophrenic patient -15 -10 -5 0 5 10 15 20

Texture analysis is a new approach for image analysis. Once the pharmacological aspect in the rat model is clearly demonstrated, extension to potential applications for humans can be considered. In fact, brain plasticity could be assessed with such a technique in brain diseases such as epilepsy, dementia, and schizophrenia. Drug effects could also be investigated in order to evaluate whether brain anisotropy or asymmetry varies during drug therapy. Finally, such an analysis could be correlated with neurocognitive tests to measure improvements in subjects performance.
The authors thank Dr I. J. Namer, Dr M. Petrou, and V. A. Kovalev, who allowed their original data to be included in this overview. This was handled inside the EEC COST-B 11 Action entitled Quantitative magnetic resonance imaging texture, led by Dr R. Lerki.

Figure 4. 2D scattergram produced by projecting the high-dimensional feature space onto a 2D feature space and preserving the euclidean distances between the points.

REFERENCES
1. Turski L, Ikonomidou C, Turski WA, Bortolotto ZA, Cavalheiro EA. Cholinergic mechanisms and epileptogenesis. The seizures induced by pilocarpine: a novel experimental model of intractable epilepsy. Synapse. 1989;3:154-171. 2. Roch C, Leroy C, Nehlig A, Namer IJ. Magnetic resonance imaging in the study of the lithium-pilocarpine model of temporal lobe epilepsy in adult rats. Epilepsia. 2002;43:325-335.

3. Roch C, Leroy C, Nehlig A, Namer IJ. Predictive value of cortical injury on the development temporal lobe epilepsy in immature rats: a magnetic resonance imaging (MRI) approach using the lithium-pilocarpine model. Epilepsia. 2001;42(suppl 7):233. 4. Haralick RM. Statistical and structural approaches to texture. IEEE Proc. 1979;67:786-804. 5. Szczypiski P, Kocioek M, Materka A, Strzelecki M. Computer program for image texture analysis. Proceedings of the International Conference on Signals and Electronic Systems. 18-21 September 2001. Lodz, Poland. 2001:255-261.

232

Texture analysis of the brain - Nedelec et al

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

6. Yu O, Mauss Y, Roch C, Namer IJ, Chambron J. Detection of late epilepsy by texture analysis of MR brain images in the lithium pilocarpine model. Magn Reson Imaging. 2002;20:771-775. 7. Chetverikov D. GLDH-based analysis of texture anisotoropy and symmetry: an experimental study. In: Proceedings 12th IPCR, Jerusalem, Israel. 1994:1071-1073. 8. Kovalev VA, Kruggel F, Gertz HJ, von Cramon DY. Three-dimensional texture analysis of MRI brain datasets. IEEE Trans Med Imaging. 2001;20:424-433. 9. Kovalev VA, Petrou M. Texture anisotropy in 3D images. IEEE Trans Image Processing. 1999;8:346-360. 10. Kovalev V, Petrou M. Multidimensional co-occurrence matrices for object recognition and matching. Graph Models Image Processing. 1996;58:187-197.

11. Fosltein M, Folstein S, McHugh P. Mini-Mental State: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189-198. 12. Freeborough P, Fox N. MR image texture analysis applied to the diagnosis and tracking of Alzheimers disease. IEEE Trans Med Imaging. 1998;17:475-479. 13. Segovia-Martinez M, Petrou M, Crum W. Texture features that correlate with the Mini-Mental state Examination (MMSE) Score. EMBS Proc. 2001:910-913. 14. Kovalev VA, Petrou M, Bondar YS. Texture anisotropy in 3D images. IEEE Trans Image Processing. 1999;l8:346-360. 15. Talairach J, Tournoux P. Coplanar Stereotaxic Atlas of the Human Brain. New York, NY: Thieme; 1988.

Anlisis de la textura del cerebro: desde los modelos animales a las aplicaciones en el hombre
Las imgenes de resonancia magntica (IRM) se utilizan ampliamente para estudiar cerebros in vivo, tanto a nivel experimental en modelos animales como para orientaciones diagnsticas en el hombre. Gran parte del valor de las IRM se relaciona con el hecho que el contraste del tejido blando se acenta por la variacin importante de los tiempos de relajacin T1 y T2 entre los tejidos. Puede ser posible el empleo de una aproximacin alternativa que no se basa en la medicin absoluta de los tiempos de relajacin. En general, las texturas son patrones visuales complejos compuestos de entidades o subconfiguraciones que tienen brillo, color, ngulo, tamao, etc. caractersticos. De esta forma, una textura puede considerarse una agrupacin de unidades semejantes en una imagen. Las propiedades de la subconfiguracin local determinan la percepcin de luminosidad, uniformidad, densidad, aspereza, regularidad, alineacin, frecuencia, fase, orientacin, grosor, aspecto aleatorio, fineza, tersura y granulaciones. El presente artculo se propone ilustrar cmo se puede aplicar el anlisis de la textura en modelos animales y en clnica humana, al igual que en la investigacin de futuras aplicaciones farmacolgicas en el hombre. Se presentan tres estudios diferentes de IRM: (1) en animales, utilizando el modelo de rata epilptica por lipocarpina, (2) en pacientes con enfermedad de Alzheimer y (3) en pacientes con esquizofrenia.

Analyse de la texture du cerveau : des modles animaux aux applications humaines

Limagerie par rsonance magntique (IRM) est largement utilise dans les tudes du cerveau in vivo, tant sur le plan exprimental dans les modles animaux que dans une vise diagnostique chez lhomme. Une grande partie de la valeur de lIRM est lie au fait que le contraste du tissu mou est accentu par la variation importante des temps de relaxation T1 et T2 dans les divers tissus. Il est peut-tre possible dutiliser une autre approche, qui ne sappuie pas sur la mesure absolue des temps de relaxation. De faon gnrale, les textures rsultent de motifs visuels complexes composs dentits et de sousmotifs ayant une brillance, une couleur, un angle, une taille, etc., caractristiques. Une texture donne peut donc tre considre comme un groupement de similitudes au sein dune image. Les proprits dun sous-motif local sont lorigine de ce qui est peru en termes de luminosit, duniformit, de densit, dingalit, de rgularit, de linarit, de frquence, de phase, de direction, de grossiret, daspect alatoire, de finesse, daspect lisse et de granulation. Le prsent article se propose dillustrer comment lanalyse de la texture peut tre utilise dans des modles animaux et dans des applications cliniques humaines, ainsi que dans la recherche dapplications pharmacologiques futures chez lhomme. Trois tudes IRM y sont prsentes, lune mene dans un modle animal de rat rendu pileptique par la pilocarpine, la seconde chez des patients atteints de maladie dAlzheimer et la troisime chez des schizophrnes.

233

Free paper
Problems in texture analysis with magnetic resonance imaging
Lothar R. Schad, PhD

Since its introduction in the 1980s, magnetic resonance imaging (MRI) has become recognized as a powerful in vivo diagnostic tool. The objective of this article is to discuss developments in quantitative MRIand particularly texture analysisthat maximize diagnostic information. A fundamental part of the work involves careful study of the optimal MRI data collection strategies for texture analysis. This is critical, because different centers may vary their measuring sequences and acquisition protocols for clinical reasons, and may be reluctant to vary these for texture investigation. Different measuring techniques, such as spin echo, gradient echo, and echo planar, and different measuring parameters produce totally different patterns in texture. Careful investigation of the dependence of all these variables using texture phantoms (test objects) will help understand how MRI image texture is formed from tissue structures. Therefore, it is essential to design and test reliable and accurate test objects for a detailed assessment of texture analysis methods in MRI. The main feature of these test objects is their ability to simulate tissue-like textures with tissue-like MR relaxation properties. Long-term stability is also vital, as is uniformity of the overall texture. Another aspect is to examine the test objects under a whole range of MRI measuring sequences and imaging conditions using different scanners to determine their stability and utility.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:235-242.

agnetic resonance imaging (MRI) is one of the most exciting imaging technologies for texture analysis: it offers the best soft tissue contrast, which can be dramatically varied during imaging. Careful study of the dependence of texture parameters on MRI data collection strategy is essential for texture analysis in order to avoid artificial texture from the scanner.This is critical, since different centers may vary their measuring sequences and acquisition protocols for their clinical investigations. The basic problem in quantitative MRI texture analysis is the large number of different measuring techniques and imaging parameters, which can be easily changed during a clinical examination. Thus, different techniques and imaging parameters produce totally different patterns in the texture parameters of the same tissues in clinical examinations with different sensitivity to artificial texture overlaid by the scanner. The main problem in texture analysis with MRI is to avoid this artificial texture and minimize its influence. The presented work was performed in the framework of a European research project COST (Cooperation in the Field of Scientific and Technical Research) B11 between 1998 and 2002 by institutions from 13 European countries, aimed at the development of quantitative methods for MRI texture analysis.1 For further detail of texture analysis, parameters, and software, see the article by Materka in this volume2 or references 3 to 7.
Keywords: texture analysis; magnetic resonance imaging; brain; trabecular bone Author affiliations: Department of Biophysics and Medical Radiation Physics, German Cancer Research Centre, Heidelberg, Germany Address for correspondence: Department of Biophysics and Medical Radiation Physics, German Cancer Research Centre, Im Neuenheimer Feld 280, D-69009 Heidelberg, Germany (e-mail: L.Schad@DKFZ-Heidelberg.de) This article is published following the 14th Biological Interface Conference held in Rouffach, France, between October 1 and 5, 2002, on the theme of Drug Development. Other articles from this meeting can be found in Dialogues in Clinical Neuroscience (2002, Vol 4, No 4).

Copyright 2004 LLS SAS. All rights reserved

235

www.dialogues-cns.org

Free paper
Selected abbreviations and acronyms
EPI FLASH FOV MA RF ROI SNR TE TH TR echo planar imaging fast low angle shot field of view matrix size radio frequency region of interest signal-to-noise ratio spin echo time slice thickness repetition time (T2), which are most responsible for tissue contrast and texture.According to the theoretical equation for the spin echo signal9: S (1e-TR/T1) e-TE/T2 [1]

Material and methods

The complexity of this problem can be demonstrated by considering a typical measuring spin echo sequence as measured by a commercial whole-body imager. Various parameters can be easily changed during clinical investigation: image contrast is mainly defined by repetition time (TR) and spin echo time (TE); image resolution is defined by slice thickness (TH), field of view (FOV), and matrix size (MA), which also influence texture analysis.The parameters of k-space acquisition and reconstruction are very important: k-space is the artificial space in which the raw MRI data are collected, and the image contrast and texture is very sensitive to k-space strategies. Other parameters like coil setup and number of active coil segments are also responsible for signal and flip angle () variations in the image. Careful investigation of the dependence of all these variables will help understand how MRI image texture is formed in tissue structures. In our studies, MRI acquisition was performed in the standard head coil of a 1.5-T scanner (Siemens Vision, Erlangen, Germany). Spin echo technique One of the most important measuring techniques in clinical diagnosis is the spin echo sequence, in which 90 and 180 radio frequency (RF) pulses produce the spin echo signal. In addition, gradients are used in x, y, and z directions to localize the signal.8 The advantages of this technique are reduced artifacts, clearly defined contrast, and common availability. The disadvantages are the contrast dependency on RF pulse quality, and slice cross-talking, which is typical of a two-dimensional (2D) technique. This imaging technique allows measurement of the three relevant MRI tissue parameters: spin density (), spin-lattice relaxation time (T1), and spin-spin relaxation time

in which S is the spin echo signal, the contrast can be created by a long TR and short TE, resulting in a flat image contrast and texture at high signal intensity (Figure 1a).T1 contrast can be created by short TR and short TE in spin echo imaging (Figure 1b). On the other hand, T2 contrast is created by long TR and long TE, mainly reflecting the water content of the tissue (Figure 1c). These three physical tissue parameters are described in reference 1.The real physical properties of tissues may be obscured by artificial contrast and texture from the scanner. Slice profile Slice profile is defined by the slice gradient and the shape of the RF pulse. Ideally, we would like to measure a rectangular slice, but due to technical reasons the real slice profile is Gaussian shaped. The consequence is that we have signal contributions from neighboring slices that influence the tissue texture. To minimize this effect, an interleaved slicing scheme is used in multislice 2D imaging. k-space Another aspect of artificial texture is connected to the kspace, which describes the strategy for raw data collection. The k-space contains the measured signal frequencies kx and ky, the so-called hologram from which the real MRI image can be calculated by a Fourier transform

Figure 1. Spin echo images of a patient with meningioma. A. -image (TR/TE = 2000 ms/10 ms). B. T1 image (TR/TE = 600 ms/10 ms). C. T2 image (TR/TE 2000 ms/100 ms). TR, repetition time; TE, spin echo time.

236

Problems in texture analysis with MRI - Schad

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

(Figure 2). Some imaging techniques measure only every second line in the k-space to speed up the imaging sequence, which results in a reduction in the signal-tonoise ratio (SNR) by 1/2 and aliasing artifacts, with consequences for image texture. Restriction to the center of the k-space with zero filling of the outer part results in the same SNR effect without aliasing. RF excitation Another important variable is the RF characteristic and sensitivity of the transmitting and/or receiving coil, which can produce a lot of artificial texture from the scanner. This is demonstrated in Figure 3 using hard image scaling, which shows a clear signal inhomogeneity due to nonideal RF pulses at the outer range of the phantom (ie, coil). Another coil effect on image texture is produced with coil arrays, where the summarized image is a result of the combination of single coils, each of which contributes its own coil characteristics (eg, SNR, sensitivity, and RF excitation profile) to the summarized image. This means that image texture could slightly differ between the object center and the object boundary, where protons are close or far away from the center of the coil. Gradient echo techniques Significant effects on image contrast and texture are introduced by the imaging sequence itself, since the imaging signal can have a very complex dependence on the physical properties of the underlying tissue. One example is the so-

called gradient echo technique like FLASH (fast low angle shot),10 where the 90 and 180 RF pulses are replaced by a low-angle RF pulse with a bipolar gradient scheme resulting in a gradient echo signal. This measuring technique can be used as fast imaging 2D technique or as a real 3D imaging technique because of the compact timing of the sequence. On the other hand, the FLASH signal has a complex dependence on T1, the local spin-spin relaxation time (T2*), and the flip angle , according to: S A e-TE/T2* with A = sin 1 e-TR/T1 1 cos e-TR/T1 [2]

B
y

Fourier transformation

x ky

k-space kx

Figure 2. The effect of k-space filling on image contrast and texture. The example demonstrates the strong dependence of image texture on the k-space filling factor as used in so-called keyhole techniques.

Figure 3. Effect of radio frequency (RF) profile of the head coil on image contrast and texture. A. Phantom measurement demonstrating the signal inhomogeneity due to nonideal RF pulses at the outer range of the image (ie, coil). B to E. Result of a measurement using coil arrays, where the summarized image is a result of the combination of the single coils, which contribute with their own coil characteristic (signal-to-noise ratio, sensitivity, and RF excitation profile) to the contrast and texture of the image.

237

Free paper
Thus, the different flip angle distributions produced by the coil characteristics result in different signals, and as a consequence in different image texture patterns as demonstrated in Figure 4. A very complex signal and texture situation is present in so-called single shot imaging techniques like echo planar imaging (EPI),11 where k-space is filled in one shot with multiple gradient echoes. This is achieved by a gradient scheme in which the upper corner of the k-space is reached by a single gradient pulse followed by a series of blips resulting in a rectangular movement through the kspace. This technique is very sensitive to local susceptibility artifacts, resulting in image distortions and strong T2* contrast dependence. Some special imaging techniques like spiral imaging can produce a very complex pattern in the image texture, since this single shot technique moves on a spiral through the k-space, which can be achieved by oscillating gradients with a phase shift of 90 in the x and y directions. This technique requires data interpolation in k-space to bring the measured data onto a Cartesian coordinate system before Fourier transform. This interpolation can produce spurious artifacts with the consequence that the image texture is dependent on k-space interpolation and image reconstruction. In addition, the problem of texture dependence on measuring technique is more complicated due to the large number of imaging sequences available on modern scanners, as illustrated in Figure 5.

Results and discussion

SNR dependence Figures 6a and 6b show the results of a FLASH experiment in a normal volunteer for SNR dependence measurement of texture parameters. The measuring parameters of the FLASH experiment were: TR/TE/ = 2 ms/ 9 ms/30; bandwidth (BW) = 195 Hz/pixel; MA = 512512; FOV = 280 mm; TH = 2 mm; and acquisitions (AQ) = 1 to 324 resulting in an SNR = 1 to 18. Texture parameters (SNR, entropy 55, correlation 55) of white matter, gray matter, and noise are shown as a function of the number of acquisitions (=SNR2). Figure 6c demonstrates that no texture can be measured in white matter using standard image resolution (0.50.52 mm3) as described above, since the SNR of white matter has the same characteristics as noise. In contrast, the SNR of gray matter reaches a nearly constant value at about 16 acquisitions and no further improvement can be reached due to the true underlying texture of the tissue. The same observation holds for a typical parameter of microtexture, like entropy 55 (Figure 6d), while no dependence

Figure 4. Example of a 3D FLASH (fast low angle shot) dataset of a normal volunteer measured by the conventional head coil. A. Excellent T1 contrast in the original transversal images. B. A strong signal inhomogeneity is obvious in the reconstructed sagittal images as seen in the homogeneous phantom and corresponding anatomical structures, like upper part of the brain and cerebellum. This artifact is due to the radio frequency excitation profile of the head coil, which produces different T1 weightings and signal amplitudes as a function of the flip angle according to the FLASH formula (Equation 2).

238

Problems in texture analysis with MRI - Schad

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

on SNR can be detected for a typical parameter of macrotexture, like correlation 55 (Figure 6e). Based on this observation, a sufficient SNR>4 is necessary to measure the real textural behavior of the human brain.12,13 Normalization A texture test object (PSAG) was developed on the basis of polystyrene (PS) and agar solution (AG) to mimic texture properties artificially. PS spheres are available from the technological process of PS production. Two types of spheres were used for the phantom construction: randomly distributed spheres of diameter 0.2 to 3.15 mm; or mechanically separated spheres of diameter 0.8 to 1.25 mm, 1.25 to 2 mm, or 2 to 3.15 mm. Polyethylene tubes of diameter 1.5 and 2.8 cm were filled with spheres and by
EPI TGSE

a hot solution of 4% agar (free and doped with DyCl3). One milliliter of 0.1% NaN3 was added per liter of agar for microbiological stability.14 A second texture test object containing foam at different densities in Gd-DTPA solution was used to describe microtexture properties. Phantom tubes containing foams with coarse, middle, and fine density were constructed and filled with a Magnevist (Schering, Berlin) solution at a concentration of 1:4000. Problems with the foam phantoms are air bubbles, which create susceptibility artifacts in the images, and so a careful preparation of the foam phantoms is necessary. Both types of phantoms were placed next to the head of a volunteer and a position for the imaging slab was chosen such that all vials and part of the volunteers brain were contained in the 3D slab. With this setup several 3D data sets with different imaging parameters were

MPR

EPI TGSE GSE GRASE TIR TIR, TIRM, TFLAIR HASTE MPRAGE

turboFLASH snapshotFLASH FSPGR PSLF trueFISP CISS DESS PSIF CE-FAST

FISP GRASS FAST T2-FFE FLASH SPGR T1-FFE spoiled FAST FLASH

TSE, FSE RARE TSE

IR CSE SE Fourier-coding K-space

GRE

Figure 5. Sketch of the family of imaging techniques available on modern scanners. There are many strategies of mixing spin echoes with gradient echoes to speed up imaging time with the consequence of very complex image contrast and texture.CISS, constructive interference in the steady state; CSE, conventional SE; DESS, dual echo steady state; EPI, echo planar imaging; FAST, Fourier-acquired steady state; FISP, fast imaging with steady precession; FLASH, fast low angle shot; FSE, fast spin echo; FSPGR, fast spoiled gradient recalled acquisition into steady state; GRASE, gradient and spin echo; GRASS, gradient recalled acquisition into steady state; GSE, gradient SE; HASTE, half-Fourier acquisition single-shot turbo SE; IR, inversion recovery; MPRAGE, magnetization prepared gradient echo imaging; RARE, rapid acquisition with relaxation enhancement; SE, spin echo; SPGR, spoiled gradient recalled acquisition into steady state; T1-FFE, T1-weighted fast field echo; T2-FFE, T2weighted fast field echo; TFLAIR, turbo fluid attenuated IR; TGSE, turbo GSE; TIR, turbo IR; TIRM, turbo IR magnitidue; TSE, turbo SE.
Copyright Siemens.

239

Free paper
acquired to demonstrate the influence of resolution and SNR, as well as the dependence of the texture parameters on different imaging parameters (eg, ,TR,TE). In a pilot study, texture parameters such as mean gradient show the same behavior in phantoms as in white matter for different patients, indicating that a normalization of texture parameters using test objects is possible (Figure 7). However, texture normalization is necessary, but it is not possible to mimic all texture features by phantoms.15
White matter

A C
60 50

SNR (mean/SD)

40 30 20 10 0 1 4 9 16 25 36 49 64

Noise

A
Gray matter

81 100 121 144 169 196 225 256 289 324

Number of acquisitions

D
1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 1

Entropy 5x5

Gray matter

White matter Noise

4 9 16 25 36 49 64 81 100 121 144 169 196 225 256 289 324

Number of acquisitions

B
0.25

0.5

Correlation 5x5

Mean gradient

0.3

0.20

0.1 -0.1

Phantom spheres
0.15 0.10

-0.3

White matter
0.05

-0.5 1 4 9 16 25 36 49 64 81 100 121 144 169 196 225 256 289 324

733

735

737

807

Number of acquisitions

Patient number

Figure 6. FLASH (fast low angle shot) images of a normal volunteer for measuring signal-to-noise (SNR) dependence of texture parameters at (A) SNR =1 (1 acquisition) and (B) SNR =18 (324 acquisitions). C to E. Texture parameters (SNR, entropy 55, correlation 55) of white matter, gray matter, and noise are shown as a function of the number of acquisitions (=SNR2). SD, standard deviation.

Figure 7. A. Three-dimensional FLASH (fast low angle shot) image of a patient with glioblastoma with texture test objects located beside the head for testing texture normalization. B. Texture parameters such as mean gradient show the same behavior in the phantom as in white matter for different patients, indicating that normalization of texture parameters using test objects is possible.

240

Problems in texture analysis with MRI - Schad

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

Clinical application The aim of this pilot study was to assess the possibility of quantitative description of texture directivity in trabecular bone with an attempt to quantitative description of trabecular bone structural anisotropy using texture analysis of 3D FLASH MRI. A series of 3D FLASH images, all of 256256 pixels, with the voxel size of 0.40.40.4 mm3, were measured on a standard 1.5-T scanner (Siemens Vision, Erlangen, Germany) using a small flex coil. The images in Figure 8 represent trabecular bone cross-sections in the sagittal and reconstructed transversal direction. For bone image texture analysis, circular regions of interest (ROI) were marked on corresponding bone cross-sections and effort has been made

to maintain a large-size ROI for better statistical significance of texture parameters. The texture of the bone image shows apparent directivity, which reflects anisotropy of its physical structure according to the direction of gravity (Figure 8c). Quantitative analysis of this directivity is important to medical diagnosis, eg, in early detection of osteoporosis, as the directivity may vary according to the development of the disease.
The author like to thank Michael Bock (DKFZ Heidelberg, Germany), Milan Hajek (University Prague, Czech Republic), Richard Lerski (University Dundee, Scotland), Arvid Lundervold (University Bergen, Norway), Andrzej Materka (University Lodz, Poland), Lubomir Pousek (Technical University Prague, Czech Republic), Yan Rolland (University Rennes, France), and Ivan Zuna (DKFZ Heidelberg, Germany) for their help during the COST B11 action in many aspects of texture measurements and analysis.

REFERENCES
1. COST B11. Quantitation of Magnetic Resonance Image Texture. Department of Physiology, University of Bergen, Norway. Available at http://www.uib.no/costb11/. Accessed 3 December 2003. 2. Materka A. Texture analysis methodologies for magnetic resonance imaging. Dialogues Clin Neurosci. 2004;6:243-250. 3. Haralick R, Shanmugam K, Dinstein I. Textural features for image classification. IEEE Trans Syst Man Cybern. 1973;3:610-621. 4. Haralick R. Statistical and structural approaches to texture. Proc IEEE. 1979;67:786-804. 5. Lerski RA, Straughan K, Schad LR, Boyce D, Blml S, Zuna I. MR image texture analysisan approach to tissue characterization. Magn Reson Imaging. 1993;11:873-887. 6. Materka A. MaZda Users Manual. Available at http://www.eletel.p.lodz.pl/ cost/ Accessed 3 December 2003. 7. Materka A, Strzelecki M, Lerski R, Schad L. Feature evaluation of texture test objects for magnetic resonance imaging. In: Pietikainen M, ed. Texture Analysis in Machine Vision, Series in Machine Perception and Artificial Intelligence. Vol 40. Singapore: World Scientific; 2000:197-206. 8. Hahn EL. Spin echoes. Phys Rev. 1950;80:580-594. 9. Schad LR, Brix G, Zuna I, Hrle W, Lorenz WJ, Semmler W. Multiexponential proton spin-spin relaxation in MR imaging of human brain tumors. J Comput Assist Tomogr. 1989;13:577-587. 10. Haase A, Frahm J, Matthaei D, Hnicke W, Merboldt KD. FLASH imaging. Rapid NMR imaging using low flip-angle pulses. J Magn Reson. 1986;67:258-266. 11. Mansfield P. Multi-planar image formation using NMR spin echoes. J Phys Chem Solid State Phys. 1977;10:L55-L58. 12. Lysaker M, Lundervold A, Tai XC, Bock M, Schad LR. Noise removal with tissue boundary preservation using fourth-order partial differential equations. In: Proceedings of the International Society for Magnetic Resonance in Medicine (ISMRM). Glasgow, UK, 2000. Abstracts 2001 (ISSN 1524-6965). 2001;9:126. 13. Lysaker M, Lundervold A, Tai XC, Bock M, Schad L. Noise removal with tissue boundary preservation. Collection of Abstracts First SIAM-EMS Conference. Applied Mathematics in our Changing World. Berlin, Germany, 2 to 6 September 2001. 14. Jirk D, Hjek M, Herynek V. New type of phantom for texture analysis based on polystyrene and agar gel. European Society for Magnetic Resonance in Medicine and Biology, 17th Annual Meeting, Paris, France, 14 to 17 September 2000. MAGMA. 2000;11:343. 15. Schad LR, Blml S, Zuna I. MR tissue characterization of intracranial tumors by means of texture analysis. Mag Reson Imaging. 1993;11:889-896.

A C
1780 Run length nonuniformity

1740

6%

1700

1660 Direction Horizontal Vertical 45 135

1620

Figure 8. Example of a trabecular bone 3D FLASH (fast low angle shot) measurement of a normal volunteer in sagittal (A) and reconstructed transversal direction (B) for testing texture directivity. (C) Texture parameters like run length nonuniformity show a clear dependence on direction of gravity. Quantitative analysis of this directivity is important to medical diagnosis, eg, in early detection of osteoporosis, as the directivity may vary according to the development of the disease.

241

Free paper
Problemas en el anlisis de la textura con imgenes de resonancia magntica
Las imgenes de resonancia magntica (IRM) se han reconocido como una poderosa herramienta para el diagnstico in vivo desde su introduccin en la dcada de 1980. El objetivo de este artculo es discutir el desarrollo de las IRM de tipo cuantitativo y en particular el anlisis de la textura, lo que permite maximizar la informacin para el diagnstico. Una parte fundamental de este trabajo incluye un cuidadoso estudio de las mejores estrategias de recoleccin de datos de las IRM para el anlisis de la textura. Este aspecto es central ya que diferentes centros, por razones clnicas, pueden variar la secuencia de medicin y los protocolos de adquisicin de datos y por lo tanto pueden ser reacios a modificar stos para la investigacin de la textura. Diversas tcnicas de medicin como el eco de spin, el eco de gradiente y el eco planar y distintos parmetros de medicin producen patrones de textura totalmente diferentes. Una investigacin cuidadosa de la influencia de todas estas variables mediante el empleo de texturas fantasma (objetos de prueba) ayudar a comprender cmo se forman las texturas de las IRM a partir de las estructuras de los tejidos. Por lo tanto, resulta esencial disear y ensayar objetos de prueba exactos y confiables que permitan una evaluacin detallada de los mtodos de anlisis de la textura de las IRM. La caracterstica principal de estos objetos de prueba es su capacidad para simular texturas parecidas a los tejidos y con propiedades semejantes a estos ltimos. La estabilidad a largo plazo de estos modelos es de gran importancia, como tambin la uniformidad global de la textura. Otro aspecto consiste en examinar la estabilidad y utilidad de los objetos de prueba mediante diversas secuencias de medicin y condiciones en que se realicen las IRM empleando distintos resonadores.

Problmes poss par les analyses de texture au cours de limagerie par rsonance magntique
Depuis son introduction au milieu des annes quatre-vingt, limagerie par rsonance magntique (IRM) est devenue un puissant outil de diagnostic in vivo. Lobjectif de cet article est dvaluer les avances de lIRM quantitative, notamment en ce qui concerne les analyses de texture, qui optimisent linformation diagnostique. Cet article sera en grande partie consacr ltude minutieuse des meilleures stratgies de recueil de donnes dIRM pour les analyses de texture. Ceci est essentiel, dans la mesure o les squences de mesures et les protocoles de saisie cliniques peuvent varier selon les centres et que ces derniers peuvent tre rticents les modifier pour les analyses de texture. Des diffrences aussi bien dans les techniques de mesure utilises, quil sagisse de lcho de spin, lcho de gradient, lcho planaire, etc., que dans les paramtres mesurs retentissent sur la texture en entranant lapparition de motifs totalement diffrents. Lvaluation prcise de linfluence de ces variables en utilisant des images talon ( fantmes ) de texture (ou objets de test) facilitera une meilleure comprhension du mode de constitution des textures des structures tissulaires en IRM. Par consquent, il est essentiel de dvelopper et dexprimenter des objets de test exacts et fiables pour une valuation dtaille des mthodes utilises en IRM pour les analyses de texture. La principale caractristique de ces modles rside dans leur capacit imiter des textures semblables celles produites par les tissus et ayant des proprits de relaxation RM identiques. La stabilit long terme de ces modles revt galement une grande importance, tout comme luniformit globale des textures qui en rsultent. Un autre objectif est dexaminer la stabilit et lutilit de ces objets de test en les soumettant toute une gamme de squences de mesures et de conditions dimagerie par IRM et de les tudier avec divers types dappareils dIRM.

242

Free paper
Texture analysis methodologies for magnetic resonance imaging
Andrzej Materka, MSEE, PhD, DSc

Methods for the analysis of digital-image texture are reviewed. The functions of MaZda, a computer program for quantitative texture analysis developed within the framework of the European COST (Cooperation in the Field of Scientific and Technical Research) B11 program, are introduced. Examples of texture analysis in magnetic resonance images are discussed.
2004, LLS SAS Dialogues Clin Neurosci. 2004;6:243-250.

Keywords: quantitative texture analysis; magnetic resonance imaging Author affiliations: Institute of Electronics, Technical University of Lodz, Lodz, Poland Address for correspondence: Technical University of Lodz, Institute of Electronics, ul Wolczanska 223, 90-924 Lodz, Poland (e-mail: materka@p.lodz.pl) This article is published following the 14th Biological Interface Conference held in Rouffach, France, between October 1 and 5, 2002, on the theme of Drug Development. Other articles from this meeting can be found in Dialogues in Clinical Neuroscience (2002, Vol 4, No 4). Copyright 2004 LLS SAS. All rights reserved

agnetic resonance imaging (MRI) is a unique and powerful tool for medical diagnosis, in that it is a noninvasive technique that allows visualization of soft tissues. There is an increasingly growing interest in using MRI for early detection of many diseases, such as brain tumors, multiple sclerosis, and others. The diagnostic information is often included in the image texture.1,2 In such cases, it is not sufficient to analyze image properties on the basis of point-wise brightness only; higher-order statistics of the image must be taken into account. Texture quantitation, ie, its description by precisely defined parameters (features) is then needed to extract information about tissue properties. Numerical values of texture parameters can be used for classification of different regions in the image, eg, representing either tissues of different origin or normal and abnormal tissues of a given kind. Changes of properly selected texture parameters in time can quantitatively reflect changes in tissue physical structure, eg, to monitor progress in healing. A European research project COST (Cooperation in the Field of Scientific and Technical Research) B11 was performed between 1998 and 2002 by institutions from 13 European countries, aimed at development of quantitative methods of MRI texture analysis.3 It gathered experts of complementary fields (physics, medicine, and computer science) to seek MRI acquisition and processing techniques that would make medical diagnoses more precise and repetitive. One of the unique outcomes of this project is MaZda,4 a package of computer programs that allows interactive definition of regions of interest (ROIs) in images, computation of a variety of texture parameters for each ROI, selection of most informative parameters, exploratory analysis of the texture data obtained, and automatic classification of ROIs on the basis of their texture. The MaZda software has been designed and implemented as a package of two MS Windows, PC applications: MaZda.exe and B11.exe.4 Its functionality extends beyond the needs of analysis of MRI, and applies to the investigawww.dialogues-cns.org

243

Free paper
Selected abbreviations and acronyms
ANN LDA NDA PCA ROI artificial neural network linear discriminant analysis nonlinear discriminant analysis principal component analysis region of interest Texture classification: To determine to which of a finite number of physically defined classes a homogeneous texture region belongs (eg, normal or abnormal tissue). Shape from texture: To reconstruct 3D surface geometry from texture information. Feature extraction is the first stage of image texture analysis. The results obtained from this stage are used for texture discrimination, texture classification, or object shape determination. Consider an MRI cross-section of human skull (Figure 1a) in which an elongated, kidney-shape bright object in the upper-right quadrant of the image is needed to be numerically characterized. This object is our ROI in this example, and is marked in red in Figure 1b. For a population of images, the subimage covered by ROI is a random variable. Assuming that texture is homogeneous within the ROI and that the area of the ROI is sufficiently large, one can compute a number, say N, of statistical parameters based on image points contained in the ROI. Depending on definition of these statistics, different properties of the ROI texture can be highlighted; these parameters are called texture features. In the example illustrated, the calculated parameters can be arranged to form a feature vector [p1, p2, , pN]. Such a vector is a compact description of the image texture. Comparison of vectors computed for images measured for different patients indicates whether the texture covered by ROI represents normal or abnormal tissue. Feature vectors can be applied to the input of a device called a classifier. On the basis of its input, the classifier takes the decision as to which predefined texture classes its input represents. Consider a population of K images, each showing a different instance of texture A. A feature vector is computed for each image, and applied to the input of the classifier. In an ideal case, seeing a vector drawn from texture of class A, the classifier responds with the information class A at its output. Similarly, for a population of K images, K feature vectors can be computed. Any of these could be applied to the input of the classifier. In an ideal case, the response of the classifier to a feature vector computed for texture class B is class B. (Sometimes a classifier cannot make a correct decision; in such cases, it wrongly recognizes a texture class different from the one represented at the input, or it is unable to make a choice between assumed texture classes.) The concept of textured image segmentation is illustrated in Figure 2.The left and right halves of the image in Figure 2a have different textures. In the process of image seg-

tion of digital images of any kind, where information is carried in texture. The essential properties of the MaZda package is described in this report, illustrated by examples of its application to selected MRI texture analysis.

Texture analysis methods

Although there is no strict definition of the image texture, it is easily perceived by humans and is believed to be a rich source of visual information about the internal structure and three-dimensional (3D) shape of physical objects. Generally speaking, textures are complex visual patterns composed of entities or subpatterns that have characteristic brightness, color, slope, size, etc. Thus, texture can be regarded as a similarity grouping in an image.5 The local subpattern properties give rise to the perceived lightness, uniformity, density, roughness, regularity, linearity, frequency, phase, directionality, coarseness, randomness, fineness, smoothness, granulation, etc, of the texture as a whole.6 A large collection of examples of natural textures is contained in the album by Brodatz.7 There are four major issues in texture analysis: Feature extraction: To compute a characteristic of a digital image that can numerically describe its texture properties. Texture discrimination: To partition a textured image into regions, each corresponding to a perceptually homogeneous texture (leading to image segmentation).
A B

Figure 1. A cross-section of human skull (A), with the region of interest (ROI) marked in red (B).

244

Texture analysis methodologies for MRI - Materka

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

mentation, the two regions are automatically identified and marked in different colors, eg, orange and blue in Figure 2b. (Some parts of the image are wrongly recognized as regions of yet other texture types, though.) There are two main techniques of image segmentation: supervised, where texture classes are known in advance; and unsupervised, where they are unknown, and so the segmenting device has to identify not only the texture classes, but also their number. There exist a variety of different texture segmentation methods, such as region growing, maximum likelihood, split-and-merge algorithms, Bayesian classification, probabilistic relaxation, clustering, and neural networks.2 All of these are based on feature extraction, which is the initial step and is necessary to describe (measure and analyze) the texture properties. This paper is confined mainly to feature extraction and texture classification techniques, which are typically the basic steps performed to support medical diagnosis. Approaches to texture analysis are usually categorized into structural, statistical, model-based, and transform methods. Structural approaches Structural approaches6,8 represent texture by well-defined primitives (microtexture) and a hierarchy of spatial
A

arrangements (macrotexture) of those primitives. To describe the texture, one must define the primitives and the placement rules. The choice of a primitive (from a set of primitives) and the probability of the chosen primitive to be placed at a particular location can be a function of location or the primitives near the location. The advantage of the structural approach is that it provides a good symbolic description of the image; however, this property is more useful for texture synthesis than analysis tasks. The abstract descriptions can be ill defined for natural textures because of the variability of both micro- and macrostructure and no clear distinction between them. A powerful tool for structural texture analysis is provided by mathematical morphology.9,10 This may prove to be useful for bone image analysis, eg, for the detection of changes in bone microstructure. Statistical approaches In contrast to structural methods, statistical approaches do not attempt to explicitly understand the hierarchical structure of the texture. Instead, they represent the texture indirectly by the nondeterministic properties that govern the distributions and relationships between the gray levels of an image. Methods based on second-order statistics (ie, statistics given by pairs of pixels) have been shown to achieve higher discrimination rates than the power spectrum (transform-based) and structural methods.11 Human texture discrimination in terms of the statistical properties of texture is investigated in reference 12.Accordingly, the textures in gray-level images are discriminated spontaneously only if they differ in second-order moments. Equal secondorder moments, but different third-order moments, require deliberate cognitive effort.This may be an indication that, for automatic processing, statistics up to the second order may be the most important.13 The most popular secondorder statistical features for texture analysis are derived from the so-called co-occurrence matrix.8 These have been demonstrated to feature a potential for effective texture discrimination in biomedical images.1,14 Model-based approaches Model-based texture analysis15-20 using fractal and stochastic models attempts to interpret an image texture by use of a generative image model and a stochastic model, respectively. The parameters of the model are estimated and then used for image analysis. In practice, the com-

Figure 2. Textured image segmentation.

245

Free paper
putational complexity arising in the estimation of stochastic model parameters is the primary problem. The fractal model has been shown to be useful for modeling some natural textures. It can be used also for texture analysis and discrimination16,21-23; however, it lacks orientation selectivity and is not suitable for describing local image structures. Transform methods Transform methods of texture analysis, such as Fourier24-26 and wavelet27-29 transforms, produce an image in a space whose coordinate system has an interpretation that is closely related to the characteristics of a texture (such as frequency or size). Methods based on the Fourier transform perform poorly in practice, due to lack of spatial localization. Gabor filters provide means for better spatial localization; however, their usefulness is limited in practice because there is usually no single filter resolution at which one can localize a spatial structure in natural textures. Compared with the Gabor transform, the wavelet transform have several advantages: Varying the spatial resolution allows it to represent textures at the most appropriate scale. There is a wide range of choices for the wavelet function, and so the best-suited wavelets for texture analysis can be chosen a specific application. Wavelet transform is thus attractive for texture segmentation. The problem with wavelet transform is that it is not translation-invariant.30,31 Regardless of their definition and underlying approach to texture analysis, texture features should allow good discrimination between texture classes, show weak mutual correlation, preferably allow linear class separability, and demonstrate good correlation with physical structure properties. For a more detailed review of basic techniques of quantitative texture analysis, the reader is referred to reference 2. In this paper, we will discuss practical implementation of these techniques, in the form of MaZda computer program. gram. (The name MaZda is an acronym derived from Macierz Zdarzen, which is Polish for co-occurrence matrix.Thus, MaZda has no connection with the Japanese car manufacturer.) Up to 16 ROIs can be defined for an image; they may overlap each other. Once ROIs are established, MaZda allows calculation of texture parameters available from a list of about 300 different definitions that cover most of the features proposed in the known literature. The parameters can be stored in text files. One can demonstrate using properly designed test images that some of the higher-order texture parameters, especially those derived from the co-occurrence matrix, show correlation to first-order parameters, such as the image mean or variance. To avoid this unwanted phenomenon, prior to feature extraction, image normalization is preferably performed. Typically, the features computed by MaZda are mutually correlated. Morover, not all of them are equally useful for classification of given texture classes or to measure properties of the underlying physical object structure. Thus,

Image acquisition

MR scanner

ROI selection

Image normalization MaZda program Feature extraction

Feature selection

MaZda: a software package for quantitative texture analysis

The main steps of the intended image texture analysis are illustrated in Figure 3. First, the image is acquired by means of a suitable scanner. The ROIs are defined using the interactive graphics user interface of the MaZda pro-

Data preprocessing B11 program Texture classification

Figure 3. Main steps of digital image texture analysis.

246

Texture analysis methodologies for MRI - Materka

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

there is a need for feature selection.To do that, a subset of the computed parameters can be selected manually. MaZda also offers the possibility of selecting the best 10 features automatically following two different selection criteria. One criterion maximizes the ratio of between-class to within-class variance, which resembles the Fisher coefficient.32,33 The second minimizes a combined measure of probability of classification error and correlation between features.33 The selected best 10 parameters can be used for texture classification. Still, it is too difficult for a human operator to imagine and understand relationships between parameter vectors in the 10-dimensional data space. The B11 program of the MaZda package allows further data processing to transform them to a new, lower-dimension data space. The data preprocessing employs linear transforms, such as principal component analysis (PCA)34 and linear discriminant analysis (LDA),32 as well as nonlinear operations leading to artificial neural network (ANN)based nonlinear discriminant analysis (NDA).35 The B11 program displays both input and transformed data in a form of a scatter-plot graph. B11 allows also raw and transformed data vectors classification, and evaluation of the usefulness of texture features calculated using MaZda to classification of different texture classes present in image regions. For data classification, the nearest-neighbor classifier (k-NN)36 and ANN classifiers33,37,38 are implemented. Neural networks of the architecture defined during training can be tested using data sets composed of feature vectors calculated for images not used for the training. At the time of writing this paper, MaZda version 3.20 was available. It implements procedures which allow4: Loading image files in most popular MRI scanner standards (such as Siemens, Picker, Brucker, and others). MaZda can also load images in the form of Windows Bitmap files, DICOM files, and unformatted gray-scale image files (raw images) with pixel intensity encoded with 8 or 16 bits. Additionally, details of image acquisition protocol can be extracted from the image information header. Image normalization. There are three options: default (analysis is made for original image); 3 (image mean m value and standard deviation is computed, then analysis is performed for gray scale range between m-3 and m+3); or 1%-99% (gray-scale range between 1% and 99% of cumulated image histogram is taken into consideration during analysis).

Definition of ROIs. The analysis is performed within these regions. Up to 16 regions of any shape can be defined; they can be also edited, loaded, and saved as disk files. A histogram of defined ROIs may be visualized and stored. Image analysis, which is computation of texture feature values within defined ROIs. The feature set (almost 300 parameters) is divided into following groups: histogram, co-occurrence matrix, run-length matrix, gradient, autoregressive model, and Haar waveletderived features. Detailed description of feature definitions can be found in reference 33. Computation of feature maps that represent distribution of a given feature over an analyzed image. It is possible to save or load feature maps into a special floating-point file format or into Windows Bitmap file. Display of image analysis reports, saving, and loading reports into disk files. Feature reduction and selection, in order to find a small subset of features that allows minimum error classification of analyzed image textures. This is performed by means of two criteria, as explained above. Selected features can be transferred to B11 program for further processing and/or classification. Image analysis automation by means of text scripts containing MaZda language commands. Scripts allow loading analyzed images and their ROI files, running the analysis, and saving report files on disk. The number of features computed by MaZda is still increasing. New feature groups are added according to suggestions of the project members and other MaZda users.Also, new procedures for data processing are being appended.

MaZda module
MaZda generates windows and selected window elements when the program main functions are invoked. The report window contains list of numerical values of all parameters computed by MaZda for defined ROIs. The selection of texture parameters that are actually computed is made using appropriate options window. Once computed, the reports can be stored as text files. There is a possibility of manual and automatic feature selection that gives lists of 10 parameters maximizing selection criteria. These lists can be stored as text files. Texture parameters can also be computed by MaZda in a rectangular ROI moved automatically around the analyzed image. Parameter values calculated for each ROI

247

Free paper
position can be arranged to form a new image, a so-called feature map. Options of feature maps calculations (ROI size and step size of ROI movement) can be selected using appropriate window tools. B11 module This program can either be called from the MaZda windows or run as a separate MS Windows application. Input to this program is a file containing data vectors corresponding to selected texture parameters. The content of this file is displayed in the left panel of B11 window. The results of data transformation and classification are shown in the right panel of the window. The input data preprocessing and classification options can be selected in the options window. Nonlinear transforms and classifiers implemented in B11 employ feedforward ANNs. Training techniques of these networks are described in the Users Manual of MaZda.33 The clusters formed in the transform data spaces can be visualized in the form of 2D or 3D scatter plots, which are generated by B11 program and discussed in the next section. on some selected texture parameters computed using the MaZda software. There were 22 images showing different cross-sections of the test objects, leading to 22 examples of texture of each class. Numerical values of about 300 texture statistical parameters were computed using MaZda module. This step produced eighty-eight 300-dimensional data vectors. A list of 10 best features was then automatically generated based on Fisher coefficient criterion (maximization of the ratio F of between-class to within-class variance).The best parameters were then passed to the B11 module.Thus, the input to B11 was made of eighty-eight 10-dimensional data vectors, with 22 vectors for each texture class. A scatter plot of the input data in the 3D data space was made of first three best texture features. The raw data were transformed to lower-dimensional spaces, using the PCA, LDA, and NDA projections. In each case, the Fisher coefficient F was calculated for the obtained data vectors. They were also classified using a 1-NN classifier, and tested using a leave-one-out technique.36 The PCA projection to a lower-dimensionality data space does not improve the classification accuracy. This can be explained by the fact that PCA is optimized for representation of data variability, which is not the same as data suitability for class discrimination (which is the case of LDA). Although the LDA gives lower value of the Fisher coefficient F, it eliminates the classification errors. Thus,
noise

Application example 1
Figure 4a shows an MRI image that contains regions of different texture, and its respective ROIs are illustrated in Figure 4b. Each circular region in Figure 4a represents a cross-section through a tube filled with polystyrene spheres of different diameters, which were test objects manufactured in the Institute of Clinical and Experimental Medicine in Prague. The example experiment described here was made to verify whether texture classes represented in the image in Figure 4a could be classified based
A B

ps3

foam 1 foam 2

ps2

ps1 foam 3

Figure 4. A. Magnetic resonance image cross-section of four test objects of different texture. B. Four regions of interest (four texture classes) defined for the image in A.

Figure 5. Magnetic resonance image with eight regions of interest (ROIs) marked with different colors. ps1, ps2, ps3, foam 1, foam 2, and foam 3 are test objects.

248

Texture analysis methodologies for MRI - Materka

Dialogues in Clinical Neuroscience - Vol 6 . No. 2 . 2004

the lower F coefficient does not necessarily indicate worse classification. Extremely large F can be obtained using NDA; however, one should verify (using a separate test dataset) whether the ANN does not suffer from the overtraining problem.38 An overtrained network does not generalize the training data well and, consequently, it may wrongly classify unseen data points.

Application example 2
Figure 5 shows an MRI image that contains cross-section of human scull, along with cross-section of six artificial test objects (phantoms designed and manufactured to generate standard texture patterns), three on each side of the scull. There are altogether eight ROIs defined for this image, each marked with a different color. The numerical experiment carried out on seven images of the described category aimed to verify whether one can find features able to separate mixed, artificial, and natural textures. The experiment was performed in three parts. First, higherorder features were considered only.Those were co-occurrence matrix, run-length matrix, gradient, and autoregressive modelderived parameters. The best of these were automatically selected by MaZda. Using the B11 program, the two sets of best features were transformed (PCA and LDA) and the transform data were used as new features for classification (by means of a 1-NN classifier tested using the leave-one-out technique). The results are
As computed data Standardized data Raw PCA LDA Raw PCA LDA Best higher-order features (histogram and wavelet-based features excluded) Fisher 22 31 19 22 31 19 POE+ACC 26 28 3 6 4 4 Best higher-order features (wavelet-based features excluded) Fisher 1 1 1 9 9 1 POE+ACC 24 24 0 4 3 2 Best higher-order features (wavelet-based features only) Fisher 3 4 0 0 0 0 POE+ACC 3 6 0 0 0 0 Table I. Number of classification errors (out of 56 samples) for best higherorder features (histogram and wavelet-based features excluded; wavelet-based features excluded; and wavelet-based features only). POE, probability of classification error; ACC, average correlation coefficient; PCA, principal component analysis; LDA, linear discriminant analysis.

shown in Table I, which indicates, that lowest error figure (3/56) was obtained for the LDA data, with no possibility of perfect classification. In the second part of the experiment, histogram-based features were added to the higher-order ones used in the first part. Table I shows significance of these parameters in region discrimination. Perfect classification was achieved for LDA-transformed data. One can notice that even if histogram data do not represent texture, they are significant to ROI classification. In the third part, wavelet-based features only were used. Table I shows that perfect ROI discrimination is possible even in the raw data space.This family of features seems to describe texture for classification purposes extremely well. The results collected (Table I) indicate that one cannot specify in advance which particular texture features will be useful for discrimination of texture classes, and that raw-data texture features usually do not allow perfect discriminationsome pre-processing is necessary, eg, by means of linear or nonlinear discriminant transforms.

Summary
Texture analysis applied to MRI (and other modalities) is one of the methods that provide quantitative information about internal structure of physical objects (eg, human body tissue) visualized in images.This information can be used to enhance medical diagnosis by making it more accurate and objective. Within the framework of a European COST B11 action, a unique package of computer programs has been developed for texture quantitative analysis in digital images.The package consists of two modules: MaZda.exe and B11.exe.The modules are seamlessly integrated, and each of the modules can be run as a separate application. Using the package, one can compute a large variety of different texture features and use them for classification of regions in the image. Moreover, MaZda allows generation of feature map images that can be used for visual analysis of image content in a new feature space, highlighting some image properties.The package has already been used for quantitative analysis of MRI images of different kinds,39 such as images of human liver40 and computed tomography images for early detection of osteoporosis.23 The program appears to be a useful research tool in a PhD student laboratory.41 The MaZda package is available on the Internet.33 So far more than 300 researchers from all over the world have downloaded it onto their computers.

249

Free paper
Delivery of MRI images by Dr Richard Lerski of Dundee University and Hospital (Figure 2a), Prof Milan Hajek of the Institute of Clinical and Experimental Medicine in Prague (Figure 4), Prof Lothar Schad of German Cancer Research Centre in Heidelberg (Figures 1 and 5), and Dr Michal Strzelecki (Figure 2) is very much appreciated.

Metodologas de anlisis de la textura para imgenes de resonancia magntica

Se revisan los mtodos de anlisis de la textura de imgenes digitales. Se presentan las funciones del MaZda, un programa computacional para el anlisis cuantitativo de la textura, desarrollado dentro del marco del Programa Europeo de Cooperacin en el Campo de la Investigacin Cientfica y Tcnica (COST) B11. Se discuten ejemplos de anlisis de la textura en imgenes de resonancia magntica. REFERENCES
1. Lerski R, Straughan K, Shad L, et al. MR image texture analysisan approach to tissue characterisation. Magn Reson Imaging. 1993;11:873-887. 2. Materka A, Strzelecki M. Texture analysis methodsA review. COST B11 Technical Report, Lodz-Brussels: Technical University of Lodz; 1998. Available at http://www.eletel.p.lodz.pl/cost/publications.html. Accessed 3 December 2003. 3. COST B11. Quantitation of Magnetic Resonance Image Texture. Department of Physiology, University of Bergen, Norway. Available at http://www.uib.no/costb11/. Accessed 3 December 2003. 4. MaZda Software. Medical Electronic Division, Technical University of Lodz, Lodz, Poland. Available at http://www.eletel.p.lodz.pl/cost/software.html. Accessed 3 December 2003. 5. Rosenfeld A, Kak A. Digital Picture Processing. New York, NY: Academic Press; 1982. 6. Levine M. Vision in Man and Machine. New York, NY: McGraw-Hill; 1985. 7. Brodatz P. TexturesA Photographic Album for Artists and Designers. New York, NY: Dover; 1966. 8. Haralick R. Statistical and structural approaches to texture. Proc IEEE. 1979;67:786-804. 9. Serra J. Image Analysis and Mathematical Morphology. London, UK: Academic Press; 1982. 10. Chen Y, Dougherty E. Grey-scale morphological granulometric texture classification. Opt Eng. 1994;33:2713-2722. 11. Weszka J, Deya C, Rosenfeld A. A comparative study of texture measures for terrain classification. IEEE Trans Syst Man Cybern. 1976;6:269-285. 12. Julesz B. Experiments in the visual perception of texture. Sci Am. 1975;232:34-43. 13. Niemann H. Pattern Analysis. Berlin, Germany: Springer-Verlag; 1981. 14. Strzelecki M. Segmentation of Textured Biomedical Images Using Neural Networks [in Polish]. PhD Thesis, Technical University of Lodz, Poland; 1995. 15. Cross G, Jain A. Markov random field texture models. IEEE Trans Patt Anal Mach Int. 1983;5:25-39. 16. Pentland A. Fractal-based description of natural scenes. IEEE Trans Patt Anal Mach Int. 1984;6:661-674. 17. Chellappa R, Chatterjee S. Classification of textures using Gaussian Markov random fields. IEEE Trans Acous Speech Sig Proc. 1985;33:959-963. 18. Derin H, Elliot H. Modeling and segmentation of noisy and textured images using Gibbs random fields. IEEE Trans Patt Anal Mach Int. 1987;9:39-55. 19. Manjunath B, Chellappa R. Unsupervised texture segmentation using Markov random fields. IEEE Trans Patt Anal Mach Int. 1991;13:478-482. 20. Strzelecki M, Materka A. Markov random fields as models of textured biomedical images. Proceedings of the 20th National Conference Circuit Theory Electronic Networks 1997. Kolobrzeg, Poland; 1997:493-498. 21. Chaudhuri B, Sarkar N. Texture segmentation using fractal dimension. IEEE Trans Patt Anal Mach Int. 1995;17:72-77.

Mthodologies danalyse de la texture pour lIRM

Les mthodes danalyse de la texture de limage numrise sont ici passes en revue. Les fonctions de MaZda, un logiciel dvelopp dans le cadre du programme europen COST (Cooperation in the Field of Scientific and Technical Research) B11 pour lanalyse quantitative de la texture, sont prsentes. Des exemples danalyse de la texture sur des images par rsonance magntique sont discuts.
22. Kaplan L Kuo C. Texture roughness analysis and synthesis via extended self-similar (ESS) model. IEEE Trans Patt Anal Mach Int. 1995;17:1043-1056. 23. Materka A, Cichy P, Tuliszkiewicz J. Texture analysis of X-ray images for detection of changes in bone mass and structure. In: Pietikainen M, ed. Texture Analysis in Machine Vision, Series in Machine Perception and Artificial Intelligence. Vol 40. Singapore: World Scientific; 2000:189-195. 24. Rosenfeld A, Weszka J. Picture recognition. In: Fu K, ed. Digital Pattern Recognition. Berlin, Germany: Springer-Verlag; 1980:135-166. 25. Daugman J. Uncertainty relation for resolution in space, spatial frequency and orientation optimised by two-dimensional visual cortical filters. J Opt Soc Am. 1985;2:1160-1169. 26. Bovik A, Clark M, Giesler W. Multichannel texture analysis using localised spatial filters. IEEE Trans Patt Anal Mach Int. 1990;12:55-73. 27. Mallat S. Multifrequency channel decomposition of images and wavelet models. IEEE Trans Acous Speech Sig Proc. 1989;37:2091-2110. 28. Laine A, Fan J. Texture classification by wavelet packet signatures. IEEE Trans Patt Anal Mach Int. 1993;15:1186-1191. 29. Lu C, Chung P, Chen C. Unsupervised texture segmentation via wavelet transform. Patt Rec. 1997;30:729-742. 30. Brady M, Xie Z. Feature selection for texture segmentation. In: Bowyer K, Ahuja N, eds. Advances in Image Understanding. Los Alamitos, Calif: IEEE Computer Society Press; 1996. 31. Lu C, Chung P, Chen C. Unsupervised texture segmentation via wavelet transform. Patt Recog. 1997;30:729-742. 32. Fukunaga K. Introduction to Statistical Pattern Recognition. San Diego, Calif: Academic Press; 1991. 33. Materka A. MaZda Users Manual. Available at http://www.eletel.p.lodz.pl/ cost/ Accessed 3 December 2003. 34. Krzanowski W. Principles of Multivariable Data Analysis, Oxford, UK: Oxford University Press; 1988. 35. Mao J, Jain A. Artificial neural networks for feature extraction and multivariate data projection. IEEE Trans Neural Networks. 1995;6:296-316. 36. Duda R, Hart P. Pattern Classification and Scene Analysis. New York, NY: Wiley; 1973. 37. Freeman J, Skapura D. Neural NetworksAlgorithms, Applications and Programming Techniques. Redwood City, Calif: Addison-Wesley; 1991. 38. Hecht-Nielsen R. Neurocomputing. Reading, Mass: Addison-Wesley; 1989. 39. Materka A, Strzelecki, Lerski R, et al. Feature evaluation of texture test objects for magnetic resonance imaging. In: Pietikainen M, ed. Texture Analysis in Machine Vision, Series in Machine Perception and Artificial Intelligence. Vol 40. Singapore: World Scientific; 2000:197-206. 40. Jirak D, Dezortova M, Taimy P, et al. Texture analysis of human liver. J Magn Reson Imaging. 2002;15:68-74. 41. Szczypinski P, Kociolek M, Materka A, et al. Computer program for image texture analysis in PhD student laboratory. ICSES 2001. Lodz. 2001:255-261.

250

NEW

Dialogues
in

is now online

www.dialogues-cns.org

Dialogues in Clinical Neuroscience e-mail alert service offers the opportunity to receive the Contents of each issue.
Supported by an educational grant from Servier

clinical neuroscience

w w w. s e r v i e r. c o m

Instructions for authors

AIM AND SCOPE
Dialogues in Clinical Neuroscience is a quarterly publication that aims to serve as an interface between clinical neuropsychiatry and the neurosciences by providing state-of-the-art information and original insights into relevant clinical, biological, and therapeutic aspects. Each issue addresses a specific topic, and also publishes free contributions in the field of neuroscience as well as other nontopic-related material.

FIGURES AND TABLES

Submit figures as two sets of clear glossy prints. Figures should be of good quality or professionally prepared, numbered according to their order, with proper orientation indicated (eg, top or left), and fully explicit legends provided, not repetitive of text. All abbreviations used should be expanded in the legends. As figures and graphs may need to be reduced or enlarged, all absolute values and statistics should be provided. Illustrations will be reproduced in full color only when clearly necessary, eg, images from nuclear medicine or histology. Provide each table on a separate sheet, with title above and description below. All figures and tables should be cited in the text, with distinct numbering for figures and tables.

GENERAL INSTRUCTIONS
Manuscripts should be provided on word-processor disks (3.5-in, for IBM, IBM-compatible, or Apple computers) with three hard copies (text and figures) printed on one side of standard-sized white bond paper, double-spaced, with 1-inch/2.5-cm margins. Pages must be numbered. Standard typed page = 25 lines; 250 words per page. All texts should be submitted in English. In the case of translations, the text in the original language should be included. Provide black and white photograph(s) of author(s). On the title page, provide: a title (concise and informative); full names of authors (first name, middle name initial, and last name); highest academic degrees (in country-of-origin language); affiliations (names of department[s] and institution[s]) at the time the work was done; a short running title (no more than 50 letters and spaces); keywords (5 to 10); corresponding authors complete mailing address and telephone, fax, and e-mail; acknowledgments. Provide an abstract in English (125-150 words) (except for Editorials, Posters, and Letters to the Editor). Abstracts that are too short or long will be edited by the editorial department or returned to the author. Abstracts will be translated into French and Spanish by the publishers editorial department. Abbreviations should be used sparingly and expanded at first mention. A list of selected abbreviations and acronyms should be provided (or will be prepared by the editorial department) where necessary. In the text include headings using a consistent style for the various levels of headings, to highlight key points and facilitate comprehension of the text. The editorial department reserves the right to add, modify, or delete headings when necessary. Use SI units. Use generic names of drugs.

SPECIFIC FORMATS
Editorial: no more than 400 words. Pharmacological aspects; Clinical research; Basic research: 30005000 words (exclusive of abstract, references, tables, and figure legends). State of the art: maximum 7000 words (exclusive of abstract, references, tables, and figure legends). Free papers: no more than 2500 words (exclusive of abstract, references, tables, and figure legends). Posters / images in neuroscience: take up 2 facing printed pages of the journal. Posters may need to be edited to meet formatting requirements. Legends must be provided with all illustrations. Letters to the editor: (of no more than 500 words, and including no more than 10 references) are welcomed. Letters will be selected by the Editor, and published according to available space. Unpublished letters will not be returned to the authors.

EDITORIAL ASSESSMENT AND PROCESSING

All contributions to Dialogues in Clinical Neuroscience will be reviewed by members of the Editorial Board and submitted to expert consultants for peer review. All contributions to Dialogues in Clinical Neuroscience will be styled by the publishers editorial department according to the specifications of the current edition of the American Medical Association Manual of Style (Williams & Wilkins). Page proofs and editorial queries will be sent to the corresponding author for approval and should be returned within 48 hours by fax and express mail. If this deadline is not met, the changes made by the editorial department will be assumed to be accepted by the author. Authors are responsible for all statements made in their work, including changes made by the editorial department and authorized by the author. Articles and abstracts will be edited to required length or returned to the author if specific requirements are not complied with.

REFERENCES
All references should be cited in the text and numbered consecutively using superscript arabic numerals. Presentation of the references should be based on the Uniform Requirements for Manuscripts Submitted to Biomedical Journals. Ann Intern Med. 1997;126:36-47 (Vancouver style). The author-date system of citation is not acceptable. In press references should be avoided. In the bibliography, titles of journals should be abbreviated according to the Index Medicus. All authors should be listed up to six; if there are more, only the first three should be listed, followed by et al. Where necessary, references will be styled by the editorial department to Dialogues in Clinical Neuroscience copyediting requirements. Authors bear total responsibility for the accuracy and completeness of all references and for correct text citation. Example of style for references:
1. Schneider LS, Olin JT, Lyness SA, Chui HC. Eligibility of Alzheimers disease clinic patients for clinical trials. J Am Geriatr Soc. 1997;45:1-6. 2. Dubovsky J, Weber JL, Orr HT, et al. A second gene for familial febrile convulsions maps on chromosome 19p. Am J Hum Genet. 1996;59(suppl 1):A223. 3. Carpenter WT Jr, Buchanan RW. Domains of psychopathology relevant to the study of etiology and treatment in schizophrenia. In: Schulz SC, Tamminga CA, eds. Schizophrenia: Scientific Progress. New York, NY: Oxford University Press; 1989:13-22.

COPYRIGHT
Copyright of articles will be transferred to the publisher of Dialogues in Clinical Neuroscience. The Copyright Transfer Agreement must be signed by the main author and all coauthors and returned to the publisher. For reproduction of copyrighted work, the author is required to obtain authorizations from the author(s) (including self) and the publisher(s) and provide copies of these authorizations with the manuscript.

252

Dialogues

e
2004 Volume 6 No. 1

Contents of latest issues

Editor in Chief: Jean-Paul MACHER

2004 Volume 6 No. 2

Predictors of Response to Treatment in Neuropsychiatry

Editorial
Jean-Paul Macher, Marc-Antoine Crocq ________________________ 1

Neuroplasticity
Editorial
Jean-Paul Macher, Marc-Antoine Crocq ______________________ 113

State of the art

Genetic variation and pharmacogenomics: concepts, facts, and challenges Margret R. Hoehe, Thomas Kroslak ____________________________ 5

State of the art

Structural plasticity of the adult brain: how animal models help us understand brain changes in depression and systemic disorders related to depression Bruce S. McEwen ______________________________________________ 119

Basic research
The future of genetic testing for drug response Deborah J. Morris-Rosendahl, Bernd L. Fiebich______________ 27

Basic research
Structural plasticity of the adult brain Fred H. Gage __________________________________________________ 135 Regulation of cellular plasticity and resilience by mood stabilizers: the role of AMPA receptor trafficking Jing Du, Jorge A. Quiroz, Neil A. Gray, Steve T. Szabo, Carlos A. Zarate Jr, Husseini K. Manji ______________________ 143

Pharmacological aspects
Sex-dependent modulation of treatment response David R. Rubinow, Molly Moore ______________________________ 39 Current perspectives in the management of treatment-resistant depression Rajesh M. Parikh, Barry D. Lebowitz __________________________ 53 Treatment-refractory schizophrenia Asaf Caspi, Michael Davidson, Carol A. Tamminga __________ 61 Differing response to antipsychotic therapy in schizophrenia: pharmacogenomic aspects Manfred Ackenheil, Klaus Weber ______________________________ 71

Pharmacological aspects
Neural plasticity: consequences of stress and actions of antidepressant treatment Ronald S. Duman ______________________________________________ 157 Cellular consequences of stress and depression Eberhard Fuchs, Gabriele Flgge ____________________________ 171

Poster
Spectral EEG sleep profiles as a tool for prediction of clinical response to antidepressant treatment Jean-Paul Macher, Rmy Luthringer, Luc Staner ______________ 78

Clinical research
Cellular abnormalities in depression: evidence from postmortem brain tissue Craig A. Stockmeier, Grazyna Rajkowska ____________________ 185 Neuroplasticity in mood disorders Wayne C. Drevets ______________________________________________ 199 Cellular plasticity and resilience and the pathophysiology of severe mood disorders Dennis S. Charney, George DeJesus, Husseini K. Manji ____ 217

Clinical research
Treatment goals: response and nonresponse Jean-Paul Macher, Marc-Antoine Crocq ______________________ 83 Poor response to treatment: beyond medication Csar Carvajal __________________________________________________ 93 Clinicians predictions of patient response to psychotropic medications Pierre Schulz, Patricia Berney ________________________________ 105

Free papers
Texture analysis of the brain: from animal models to human applications Jean-Franois J. Nedelec, Olivier Yu, Jacques Chambron, Jean-Paul Macher________________________ 227 Problems in texture analysis with magnetic resonance imaging Lothar R. Schad ______________________________________________ 235 Texture analysis methodologies for magnetic resonance imaging Andrzej Materka ______________________________________________ 243

If you wish to receive future issues of Dialogues in Clinical Neuroscience, please contact mail.dialneuro@fr.netgrs.com This publication is supported by an educational grant from Servier www.servier.com www.dialogues-cns.org

05 ST 001 BA

For information on this journal and subscriptions, contact: Dr Marc-Antoine CROCQ FORENAP - Institute for Research in Neuroscience and Neuropsychiatry - BP29 - 68250 Rouffach - France Tel: +33 3 89 78 71 20 / 70 18 - Fax: +33 3 89 78 51 24 - E-mail: macrocq@forenap.asso.fr