Stress-Induced Depression and Comorbidities: …...Stress-Induced Depression and Comorbidities: From...

Stress-Induced Depression and

Comorbidities: From Bench to Bedside

Stress-Induced Depression and Comorbidities: From Bench to Bedside.

This collection of articles represents the output of a group of international research institutions (informally referred to as EUMOOD) who collaborated around the causal link between stress exposure and depression vulnerability.

Within the collection, preclinical and clinical research papers present an integrated experimental effort, employing a variety of methods and concepts from different disciplines such as biological psychiatry, neuroscience, and neuroendocrinology.

Editorial oversight, and coordination of the peer-review, was provided by Bernhard Baune, PLoS ONE Section Editor for Neuroscience and Psychiatry.

Overview

Stress and Depression: Preclinical Research and Clinical Implications

Alessandro Bartolomucci, Rosario Leopardi

Conceptual Challenges of a Tentative Model of Stress-Induced Depression

Bernhard Baune

Research Article

Novel Biochemical Markers of Psychosocial Stress in Women

Marie Åsberg, Åke Nygren, Rosario Leopardi, Gunnar Rylander, Ulla Peterson, Lukas Wilczek, Håkan Källmén, Mirjam Ekstedt, Torbjörn Åkerstedt, Mats Lekander, Rolf Ekman

The Combined Dexamethasone/CRH Test (DEX/CRH Test) and Prediction of Acute Treatment Response in Major Depression

Cornelius Schüle, Thomas C. Baghai, Daniela Eser, Sibylle Häfner, Christoph Born, Sascha Herrmann, Rainer Rupprecht

Regulation of Kainate Receptor Subunit mRNA by Stress and Corticosteroids in the Rat Hippocampus

Richard G. Hunter, Rudy Bellani, Erik Bloss, Ana Costa, Katharine McCarthy, Bruce S. McEwen

Enriched Environment Experience Overcomes Learning Deficits and Depressive-Like Behavior Induced by Juvenile Stress

Yana Ilin, Gal Richter-Levin

Exercise Improves Cognitive Responses to Psychological Stress through Enhancement of Epigenetic Mechanisms and Gene Expression in the Dentate Gyrus

Andrew Collins, Louise E. Hill, Yalini Chandramohan, Daniel Whitcomb, Susanne K. Droste, Johannes M. H. M. Reul

Metabolic Consequences and Vulnerability to Diet-Induced Obesity in Male Mice under Chronic Social Stress

Alessandro Bartolomucci, Aderville Cabassi, Paolo Govoni, Graziano Ceresini, Cheryl Cero, Daniela Berra, Harold Dadomo, Paolo Franceschini, Giacomo Dell'Omo, Stefano Parmigiani, Paola Palanza

Expression of the Axonal Membrane Glycoprotein M6a Is Regulated by Chronic Stress

Ben Cooper, Eberhard Fuchs, Gabriele Flügge

Opposite Effects of Early Maternal Deprivation on Neurogenesis in Male versus Female Rats

Charlotte A. Oomen, Carlos E. N. Girardi, Rudy Cahyadi, Eva C. Verbeek, Harm Krugers, Marian Joëls, Paul J. Lucassen

Rhythmicity in Mice Selected for Extremes in Stress Reactivity: Behavioural, Endocrine and Sleep Changes Resembling Endophenotypes of Major Depression

Chadi Touma, Thomas Fenzl, Jörg Ruschel, Rupert Palme, Florian Holsboer, Mayumi Kimura, Rainer Landgraf

Chronic Mild Stress (CMS) in Mice: Of Anhedonia, ‘Anomalous Anxiolysis’ and Activity

Martin C. Schweizer, Markus S. H. Henniger, Inge Sillaber

Timing Is Critical for Effective Glucocorticoid Receptor Mediated Repression of the cAMP-Induced CRH Gene

Siem van der Laan, E. Ronald de Kloet, Onno C. Meijer

Overview

Stress and Depression: Preclinical Research and ClinicalImplicationsAlessandro Bartolomucci1*, Rosario Leopardi2*

1 Department of Evolutionary and Functional Biology, University of Parma, Parma, Italy, 2 Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden

Major depression (MD) is a severe, life-

threatening, and highly prevalent psychi-

atric disorder, predicted to soon become

one of the major causes of death world-

wide. Despite extensive investigations, the

exact mechanisms responsible for MD

have not been identified. This Overview

focuses on the role of stress in depression.

It discusses current advancements in

biological psychiatry, neuroscience, and

neuroendocrinology, highlighting key find-

ings presented in the research papers

included in the Special Collection ‘‘Stress-

Induced Depression and Comorbidities:

From Bench to Bedside,’’ published in this

issue of PLoS ONE. The Overview encom-

passes the problematic diagnosis of MD, as

well as preclinical evidences linking genet-

ic predisposition, early life events and

social factors, imbalanced HPA axis,

molecular pathways within the central

nervous system, and metabolic comorbid-

ities with depression-like disorders. It is

emphasized how the link between stress

and depression can be deeper than

previously recognized, following the de-

scription of a potentially common depres-

sion subtype, named tentatively ‘‘stress-

induced depression’’ (STRID). Due to the

inherent biological perspective underlying

the STRID concept, both preclinical and

clinical research will be pivotal in clarify-

ing the validity of this new subtype of MD

and in improving predictors for treatment

response, and will provide a better basis

for genetic studies as well as stimulating

new drug discovery programs.

Introduction

Major depression (MD) is a severe, life-

threatening, and widespread psychiatric

disorder having an incidence of about 340

million cases worldwide. MD ranks fifth

among leading causes of global disease

burden including developing countries, and

by year 2030 it is predicted to represent one

of the three leading causes of burden of

disease worldwide [1,2]. MD is also a risk

factor for cardiovascular and metabolic

diseases, and a major risk factor for suicide

[3]. Despite extensive investigations, the

exact mechanisms responsible for MD have

not been identified, and current therapeu-

tics are based on serendipitous discoveries

rather than on bench-to-bedside, targeted

drug discovery [4]. In addition, although

clinically efficient antidepressant drugs do

exist, the situation is in many cases far from

ideal. Shortcomings such as low remission

and/or high treatment-resistance rates,

slow onset of action, side effects, and

drug–drug interactions merit the explora-

tion of all plausible agents that are effective,

tolerable, and safe, and that improve

maintenance of wellness [5–8]. According-

ly, there is an enormous need for joint

experimental efforts between preclinical

and clinical scientists.

Understanding MD in its etiology and

biological phenomenological characteristics

could improve its recognition and treat-

ment [8–11]. The present Overview high-

lights current trends in modern biological

psychiatry, neuroscience, and neuroendo-

crinology by discussing key aspects present-

ed in research papers included in the Special

Collection ‘‘Stress-Induced Depression and

Comorbidities: From Bench to Bedside,’’

published in this issue of PLoS ONE.

The Problematic Diagnosis ofMajor Depression

Presently accepted diagnostic criteria

for MD [12] are five (or more) specific

symptoms having been present during the

same two-week period and representing a

change from previous functioning; at least

one of the symptoms should be either

depressed mood or loss of interest or

pleasure. Although their definition fol-

lowed rigorous statistical validating criteria

and years-long investigations, they are

often criticized as being subjective–quali-

tative rather than objective–quantitative

[13,14]. The current Diagnostic and Statisti-

cal Manual (DSM-IV) classification ignores

etiology, and distinguishes between bipolar

and unipolar conditions, and within the

unipolar group between cases with and

without melancholia, or with and without

psychotic symptoms, as well as atypical

depression. In addition, the current diag-

nostic criteria represent clusters of symp-

toms and characteristics of clinical courses

that do not necessarily describe homoge-

nous disorders and may rather reflect

common final pathways of different path-

ological processes [15]. MD is also a highly

heterogeneous disease. Subtypes of de-

pression may differ not only in etiology

and clinical picture, but also in clinical

response to medical treatments [15].

However, no past or present classification

includes biological criteria except for

changes in body weight or sleep parame-

ters [16]. Therefore, there is an urgent

need for neurological, biological, and

genetic data in future DSM classifications

[17]. Inclusion of biological diagnostic

criteria requires extensive investigation

on the biological correlates of MD as well

as on the implementation of mechanistic-

based investigations. In this respect, As-

berg and coworkers [18] and Schule and

coworkers [19] add to the current discus-

sion important biological correlates in MD

patients. Increased plasma monocyte che-

moattractant protein-1 (MCP-1), epider-

mal growth factor (EGF), and vascular

endothelial growth factor (VEGF) are

increased in a population of women under

prolonged psychosocial stress and can thus

be considered potential biomarkers for

screening and early interventions [18].

These data are particularly intriguing

because they extend a growing body of

evidence linking increased plasma concen-

tration of signaling molecules such as

Citation: Bartolomucci A, Leopardi R (2009) Stress and Depression: Preclinical Research and ClinicalImplications. PLoS ONE 4(1): e4265. doi:10.1371/journal.pone.0004265

Editor: Bernhard Baune, James Cook University, Australia

Received December 5, 2008; Accepted January 7, 2009; Published January 30, 2009

Copyright: � 2009 Bartolomucci, Leopardi.

Funding: The authors have no support or funding to report.

Competing Interests: The authors have declared that no competing interests exist.

* E-mail: [email protected] (AB); [email protected] (RL)

PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4265

cytokines and growth factors with MD

[13,20,21]. On the other hand, hypotha-

lamo–pituitary–adrenocortical (HPA)-axis

dysregulation is confirmed to be an

important parameter for treatment out-

come in MD by Schule and coworkers

[19], although dysfunction of the HPA

system as determined by the classical

dexamethasone (DEX)/corticotrophin re-

leasing hormone (CRH) test as well as with

other neuroendocrine assays, seems to be

neither a necessary nor a sufficient deter-

minant for acute treatment response. Of

interest from a biological perspective,

Czeh and coworkers [22] recently ob-

served that chronic tianeptine treatment

may reverse neurobiological alterations

associated with chronic psychosocial stress

in male tree shrews without an improve-

ment in HPA functions (see also [23]).

Genetic Predisposition

Although no single gene could be

responsible for a complex and multifacto-

rial disorder like MD, association and

pharmacogenetic studies identified a num-

ber of loci associated with vulnerability to

MD or antidepressant efficacy [8,10,24–

26]. In parallel, increasing evidences

report gene6environment (G6E) effect

in MD, with stress often representing the

key environmental trigger of MD onset in

vulnerable individuals [27]. Preclinical

studies are ongoing to elucidate specific

genes and environmental context that

could precipitate psychopathologies [28].

An alternative approach by Touma and

coworkers [29] is described here: They

selected for seven generations of mice

based on their corticosterone reactivity to

an environmental challenge. They dem-

onstrate that individuals selected to be

high responders show blunted circadian

rhythmicity of HPA-axis hormones, be-

havioral hyperactivity, and changes in

rapid eye movement (REM) and non-

REM sleep, as well as slow wave activity,

indicative of reduced sleep efficacy. Con-

sidering the incidence of disturbed HPA

axis and sleep disorders in MD patients,

these selected mouse lines may offer a new

important experimental tool.

G6E effects are easily accessible for

preclinical investigations, and the elevated

number of inbred strains of mice (i.e.,

animals showing almost null heterozygosis)

available offers a powerful experimental

tool. In this respect, Schweizer and

coworkers [30] further clarify how differ-

ent inbred strains of mice can have very

different vulnerabilities to the chronic mild

stress model of depression in both behav-

ioral and physiological parameters, thus

offering an invaluable tool to understand-

ing G6E interaction in stress-induced

disturbances.

A Biological Pathogenesis: TheStress Model and HPA AxisActivity

Stress is usually defined as a state of

disturbed homeostasis inducing somatic

and mental adaptive reactions, globally

defined as ‘‘stress response,’’ aiming to

reconstitute the initial homeostasis or a

new level of homeostasis after successful

adaptation, i.e., allostasis [31–34]. There

is wide consensus and support from

preclinical and clinical data that stress

exposure conceivably plays a causal role in

the etiology of MD and depression-like

disorders [11,27,31,34]. However, no spe-

cific mechanism linking stress exposure

and stress response to the occurrence of

MD has yet been fully elucidated. Grow-

ing evidence indicates several classical

candidates, including neurotransmitters

and neuropeptides, as well as conceptually

novel immune and inflammatory media-

tors, as likely intermediate links between

stress exposure, depressive symptoms, and

MD [9,21,34–38]. Related to the latter,

Asberg and coworkers [18] discuss in their

paper in this Collection a potential role for

some inflammatory mediators in a cohort

of patients under prolonged psychosocial

stress, providing further epidemiological

support (results discussed above in this

Overview).

One of the hallmarks of the stress

response has long been considered the

activation of the HPA axis. Hypothalamic

CRH activation is a pivotal signaling

molecule in the regulation of the HPA

axis in particular and of the stress response

in general. Therefore, comprehension of

the mechanism responsible for the nega-

tive feedback regulation of CRH is of

paramount importance. In the present

Collection, van der Laan and coworkers

[39] demonstrate that the timing of

glucocorticoid receptors (GR) activation

determines the effective repression of the

cAMP-induced transcription of the CRH

gene, thus clarifying that in vivo a critical

time window may exist for effective

repression of the CRH gene and HPA

axis by glucocorticoids.

Knowledge on the functioning of the

HPA axis under acute or chronic chal-

lenge is also a key to understanding the

intimate link between stress response and

the pathogenesis of depression [40]. In-

deed, in all MD syndromes, a certain

degree of HPA-axis disturbance is often

present, visible either at the baseline or

with functional tests. Despite the fact that

observed changes of HPA regulation are

so far not specific for the diagnosis of

depression or for any of its clinical

syndromes [8], altered HPA-axis parame-

ters are considered important biomarkers,

particularly in preclinical studies. In-

creased circulating hormones such as

adrenocorticotropic hormone (ACTH)

and cortisol/corticosterone or increased

adrenal gland weight are considered

biomarkers of stress response in preclinical

models [41], including in several papers in

this Collection [19,29,42–46]. Despite the

bulk of data available, surprisingly current

knowledge has not yet been developed to a

point where HPA-axis reactivity can be

rationally exploited for targeted drug

treatment, as opposed to the major

achievements of drugs targeting the

CRH receptors [47]. Present data offer

reliable experimental tools to stimulate

future drug discovery programs [48].

Behavioral Neuroscience

The DSM-IV identifies specific behav-

ioral and cognitive diagnostic criteria for

MD patients. Among these, depressed

mood, anhedonia, locomotor disturbanc-

es, and anxiety are accessible for preclin-

ical investigation, while others such as

feelings of worthlessness and thoughts of

death or suicide cannot be reliably mim-

icked in animal models. A number of

animal models have been developed and

validated [11,38,49–51]. In particular,

models involving a chronic (i.e., continu-

ous exposure to a threatening stimulus for

a significant amount of time, usually

weeks) or intermittent (i.e., daily short

exposure to a threat for subsequent days)

exposure to negative stressful events can

be considered the most effective in mod-

eling MD-associated behavioral and phys-

iological disturbances (but see [52]). In line

with this conclusion, the preclinical papers

included in the Special Collection of PLoS

ONE make use of chronic or intermittent

models of stress [30,42–45]. Furthermore,

papers in the Collection also describe animal

models that are increasingly being regard-

ed as the most promising to model

etiological factors and key features of

MD patients, i.e., i) models in which the

threatening stimulus is social in nature

[42,46], ii) models in which exposure to

stressful stimuli occurs in the early post-

natal or juvenile age [44,46].

From a nosological point of view,

original research presented in the Collection

further clarifies that at least some behav-

ioral disturbances present in MD diagnos-

tic criteria can be reliably induced and

Stress & Depression


experimentally determined in animals

models, including for anhedonia [30],

locomotor disturbances [29,42,44,53],

anxiety [44,53], sleep pattern [29], and

learning/memory [44,53]. Of main inter-

est is the experimental evidence [30] that a

shift in the circadian rhythm induced by

overnight illumination is the single most

important experimental factor influencing

the intake of a sweetened solution, which is

the currently accepted animal equivalent

of evidence of anhedonia [54]. Further-

more, Ilin and Richter-Levine [44] show

that daily exposure to a stressful stimulus

for three consecutive days at a juvenile age

(named juvenile stress, JS), determines

long-lasting behavioral and motivational

effects in rats, i.e., increased anxiety, lower

exploratory drive, and increased learned

helplessness. On the contrary, housing JS

rats in an enriched environment complete-

ly abolished JS-induced behavioral effects.

A similar ameliorating effect on stress

coping has also been determined by

Collins and coworkers [53], by exposing

rats to a different form of environmental

enrichment, i.e., home cage presence of a

running wheel which induces in rodents

spontaneous physical exercise. Exercised

rats show increased behavioral coping and

reduced anxiety and depression-like be-

haviors in the open field test and in the

forced swimming test when compared

with sedentary rats.

Molecular Neuroscience

In the postgenomic era, high-throughput

techniques allow the identification of genes

overexpressed or downregulated in selected

brain regions after chronic stress exposure

or in MD. Following these observations,

researchers now aim at translating omics

evidences into experimentally based results

(based on on purpose experimental designs).

For example, the ‘‘sequenced treatment

alternatives to relieve depression’’

(STAR*D) clinical trial identified several

loci associated with response to anti-

depressants and suicidal ideation in MD

patients [55,56]. Among these is the gene

encoding for a class of ionotropic glutamate

receptors known as kainate receptors (KA).

Hunter and coworkers [45] now report that

KA1 subunit mRNA is selectively modu-

lated by stress- and HPA-axis activity in the

dentate gyrus and CA3 region of the

hippocampal formation. Another fruitful

approach follows the identification of

potential biomarkers in postmortem brain

tissue of MD patients. A recent study found

in the prefrontal and parieto-occipital

cortex of MD patients an altered level of

the L1-cell adhesion molecule (L1-CAM)

[57]. Ilin and Richter-Levin [44] firstly

demonstrate that juvenile stress in rats is

able to upregulate L1-CAM expression in

the basolateral amygdala and the thalamus

(in parallel with behavioral disturbances

described above), and secondly they prove

that environmental enrichment is able to

reverse stress-induced alterations. A final

example concerns the role of membrane

glycoprotein M6a in stress and neuroplas-

ticity. M6a mRNA was found to be

upregulated in the hippocampus of both

mice and tree shrews under chronic stress

[58,59]. Cooper and coworkers [43] now

establish for the first time that only a splice

variant, M6a-Ib, is modulated in a region-

ally dependent manner, i.e., downregulated

in the dentate gyrus granule neurons and in

CA3 pyramidal neurons while upregulated

in the medial prefrontal cortex.

According to the examples above, it can

be concluded that changes in gene expres-

sion and their association with behavioral

traits or psychopathologies remain among

the more powerful experimental tools to

uncover the mechanisms leading to a

brain disorders. In addition, recent find-

ings demonstrate that complex ‘‘epigenet-

ic’’ mechanisms, which regulate gene

activity without altering the DNA code,

have long-lasting effects within mature

neurons [60]. An example of the former

is presented by Collins and coworkers

[53], who establish that histone (H3)

phospho-acetylation and c-Fos immunore-

activity increase in the dentate gyrus upon

exposure to a novel environment or to

forced swimming and that their expression

is further augmented in exercised rats.

Another fruitful research area in bio-

logical psychiatry is the link between

neural plasticity, MD, and antidepressants

[9,13,32,61]. In particular, neurogenesis in

the granular layer of the dentate gyrus is

impaired by stress exposure and increased

by other environmental factors including

environmental enrichment or exercise

[9,61]. Of main interest was the demon-

stration that hippocampal neurogenesis is

required for the beneficial effect of some

but not all antidepressant classes [62–64].

Although impaired neurogenesis could not

be confirmed in a human cohort of MD

patients [65], the study of adult hippo-

campal neurogenesis in MD has benefited

tremendously from the attention it has

received, and results will ultimately dem-

onstrate its role in the etiology and/or

treatment of MD [60]. Oomen and

coworkers [46] now demonstrate that

maternal deprivation in rats at postnatal

day 3, which induces a transient increase

of maternal care, also determines impaired

neurogenesis in the dentate gyrus (DG) of

female rats, while increasing neurogenesis

in male rats. Therefore, early environment

may have a critical influence on establish-

ing long-held sex differences in neural

plasticity. This finding is particularly

interesting because MD incidence in

women is about twice that for men [66].

Metabolic Functions

In addition to neuroanatomical chang-

es, MD is also associated with severe

vegetative and biological disturbances,

including sleep and eating disorders, body

weight changes, and neuroendocrine ab-

normalities. The DSM-IV indicates as

diagnostic criteria increased or diminished

appetite/body weight which should repre-

sent a change from pre-MD onset. Most

animal models including several discussed

in the present Collection [30,43–45] de-

scribe stress-associated weight loss, which

has long being considered a face-validity

criterion for a valid animal model of MD

[67]. Until recently there was a paucity of

animal models of chronic stress-induced

weight gain. Bartolomucci and coworkers

[42] now report a mouse model of social

subordination stress with behavioral de-

pression-like responses and neuroendo-

crine disturbances, which also determine

hyperphagia, weight gain, and increased

vulnerability to obesity. In addition, an-

other study in the present Collection [46]

reports that maternal deprivation deter-

mined increased weight gain in juvenile

rats when compared with undisturbed

controls. These data offer new experimen-

tal tools to investigate the link between

mood disorders and metabolic functions.

In this respect it is remarkable that obesity

is often found in comorbidity with MD

and particularly so with the atypical

depression subtype [68], while clinical

efficacy of antidepressants is reduced in

obese individuals [69]. Accordingly, there

is a great need to rule out the mechanism

responsible for stress-induced positive or

negative energy balance in different ani-

mal models as well as in MD patients.

‘‘Stress-Induced Depression’’(STRID): A New DepressionSubtype?

The notion that stress may cause

depression has been an underlying con-

cept in the choice of papers included in the

PLoS ONE Collection discussed here. The

link between stress and depression is not

novel, and several authors have aimed at

identifying new subtypes of depression

based on their functional link with stress

exposure (e.g., [70–72]). Of special interest

Stress & Depression


for this Collection is to highlight a poten-

tially common depression subtype, named

tentatively ‘‘stress-induced depression’’

(STRID), recently described in Sweden

by Asberg and coworkers [71]. A dramatic

increase in the number of workers on long-

term sick leave was observed between the

years 1997 and 2003 (Statistics Sweden,

2004; http://www.scb.se). Studies of con-

secutive cases with psychiatric diagnoses

culled from the databases of two large

Swedish insurance companies showed that

about 80% of patients met DSM-IV

criteria for MD (Asberg et al., unpublished

data). The depression episodes were mild

to moderate (MADRS ,20), and accom-

panied by significant working memory

impairment [71]. Follow up showed that

STRID tended to have a prolonged

course, and that the patients often re-

mained in a state of exhaustion after the

depressive symptoms had remitted. Typi-

cally, the remaining clinical picture was

one of deep mental and physical fatigue,

disturbed and non-restorative sleep, irrita-

bility, perceptual hypersensitivity, emo-

tional liability, and pronounced cognitive

disturbances (mainly memory and concen-

tration problems). A closer examination of

the case histories revealed that a majority

was clearly induced by psychosocial stress,

either at the workplace or often in

combination with stress factors in the

family. This was confirmed by data

obtained in a cohort of almost 5,000

Swedish workers on long-term sick leave

with a psychiatric diagnosis [73]. Findings

are consistent with the life event stress

literature showing that specific, enduring

work-related stressful experiences contrib-

ute to depression [74]. From an endocrine

standpoint, disturbances of the HPA axis

may be distinctive pathophysiological fea-

tures of this depression subtype. HPA-axis

hyper-reactivity has long been known and

considered a classical feature of depres-

sion, particularly with the severe, melan-

cholic type. An opposite situation, i.e.,

HPA-axis hypo-reactivity was found in-

stead in STRID patients [71]. In addition

to the HPA-axis disturbance, the STRID

subtype of MD is expected to be linked to

different neurobiological, immunological

[18], and metabolic features, thus requir-

ing joint forces between preclinical and

clinical research.

Overall, the studies presented in this

Special Collection of PLoS ONE propose an

integrated effort on how to move in the

direction of joint studies. Both preclinical

and clinical research will be pivotal in

clarifying the validity of this new subtype

of MD, in improving predictors for

treatment response, and in providing a

better basis for genetic studies, as well as in

stimulating new drug discovery processes.

Acknowledgments

We thank Professor Marie Asberg (Karolinska

Institute, Sweden) for sharing unpublished data,

and Professor Eberhard Fuchs (German Pri-

mate Center, Germany) and Professor Paola

Palanza (University of Parma, Italy) for helpful

comments and suggestions.

References

1. World Health Organization (2005) WHO global

report: Preventing chronic diseases: A vital

investment. Geneva: World Health Organization.

Available: http://www.who.int/chp/chronic_

disease_report/en/. Accessed 8 January 2009.

2. Mathers CD, Loncar D (2007) Projections of

global mortality and burden of disease from 2002

to 2030. PLoS Medicine 3: e442. doi: 10.1371/

journal.pmed.0030442.

3. Malhi GS, Moore J, McGuffin P (2000) The

genetics of major depressive disorder. Curr

Psychiatry Rep 2: 165–169.

4. Binder EB, Holsboer F (2006) Pharmacogenomics

and antidepressant drugs. Ann Med 38: 82–94.

5. Bosker FJ, Westerink BH, Cremers TI, Gerrits M,

van der Hart MG, et al. (2004) Future antide-

pressants: what is in the pipeline and what is

missing? CNS Drugs 18: 705–732.

6. Greden JF (2002) Unmet need: what justifies the

search for a new antidepressant? J Clin Psychiatry

63(Suppl 2): 3–7.

7. Lucas G, Rymar VV, Du J, Mnie-Filali O,

Bisgaard C, et al. (2007) Serotonin(4) (5-HT(4))

receptor agonists are putative antidepressants

with a rapid onset of action. Neuron 55: 712–725.

8. Holsboer F (2008) How can we realize the

promise of personalized antidepressant medi-

cines? Nat Rev Neurosci 9: 638–646.

9. Duman RS, Monteggia LM (2006) A neuro-

trophic model for stress-related mood disorders.

Biol Psychiatry 59: 1116–1127.

10. Lesch KP (2004) Gene-environment interaction

and the genetics of depression. J Psychiatry

Neurosci 29: 174–184.

11. Nestler EJ, Gould E, Manji H, Buncan M,

Duman RS, et al. (2002) Preclinical models:

status of basic research in depression. Biol

Psychiatry 52: 503–528.

12. American Psychiatric Association (2000) Diagnos-

tic and Statistical Manual of Mental Disorders,

(4th ed.), Text Revised (DSM-IV-TR). Washing-

ton, DC: American Psychiatric Association.

13. Krishnan V, Nestler EJ (2008) The molecular

neurobiology of depression. Nature 455: 894–

902.

14. Parker G (2006) Beyond major depression.

Psychol Med 35: 467–474.

15. Antonijevic IA (2006) Depressive disorders—is it

time to endorse different pathophysiologies?

Psychoneuroendocrinology 31: 1–15.

16. Rush AJ (2007) The varied clinical presentations

of major depressive disorder. J Clin Psychiatry

68(Suppl 8): 4–10.

17. Hyman SE (2007) Can neuroscience be integrat-

ed into the DSM-V? Nat Rev Neurosci 8:

725–732.

18. Asberg M, Nygren A, Leopardi R, Rylander G,

Peterson U, et al. (2009) Novel biochemical

markers of psychosocial stress in women. PLoS

ONE 3: e?.

19. Schule C, Baghai TC, Eser D, Hafner S, Born C,

et al. (2009) The combined dexamethasone/CRH

test (DEX/CRH test) and prediction of acute

treatment response in major depression. PLoS

ONE 3: e.

20. Warner-Schmidt JL, Duman RS (2008) VEGF as

a potential target for therapeutic intervention in

depression. Curr Opin Pharmacol 8: 14–19.

21. Dantzer R, O’Connor JC, Freund GG,

Johnson RW, Kelley KW (2008) From inflam-

mation to sickness and depression: when the

immune system subjugates the brain. Nat Rev

Neurosci 9: 46–56.

22. Czeh B, Michaelis T, Watanabe T, Frahm J, de

Biurrun G, et al. (2001) 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 98: 12796–12801.

23. Sapolsky RM (2001) Depression, antidepressants,

and the shrinking hippocampus. Proc Natl Acad

Sci U S A 98: 12320–12322.

24. Sullivan P, de Geus E, Willeen G, James MR,

Smit JH, et al. (2009) Genomwide Association for

Major Depressive Disorder: possible role for the

presynaptic protein piccolo. Mol Psych, In press.

25. Baune BT, Hohoff C, Berger K, Neumann A,

Mortensen LS, et al. (2008) Association of the

COMT val158met variant with antidepressant

treatment response in Major Depression. Neuro-

psychopharmacology 33: 924–932.

26. Gizatullin R, Zaboli G, Jonsson EG, Asberg M,

Leopardi R (2006) Haplotype analysis reveals

tryptophan hydroxylase (TPH) 1 gene variants

associated with major depression. Biol Psychiatry

59: 295–300.

27. Caspi A, Sugden K, Moffitt TE, Taylor A,Craig IW, et al. (2003) Influence of life stress on

depression: moderation by a polymorphism in the5-HTT gene. Science 301: 386–389.

28. Urani A, Chourbaji S, Gass P (2005) Mutant

mouse models of depression: candidate genes andcurrent mouse lines. Neurosci Biobehav Rev 29:

805–828.

29. Touma C, Fenzl T, Ruschel J, Palme R,

Holsboer F, et al. (2009) Rhythmicity in Mice

Selected for Extremes in Stress Reactivity:Behavioural, Endocrine and Sleep Changes

Resembling Endophenotypes of Major Depres-sion. PLoS ONE 3: e.

30. Schweizer MC, Henniger MSH, Sillaber I (2009)

Chronic Mild Stress (CMS) in Mice: of Anhedo-nia, ‘Anomalous Anxiolysis’ and Activity. PLoS

ONE 3: e.

31. McEwen BS (2000) Allostasis and allostatic load:

implications for neuropsychopharmacology. Neu-

ropsychopharmacology 22: 108–124.

32. McEwen BS (2007) Physiology and neurobiology

of stress and adaptation: central role of the brain.Physiol Rev 87: 873–904.

33. Sapolsky RM, Romero LM, Munck AU (2000)

How do glucocorticoids influence stress respons-es? Integrating permissive, suppressive, stimula-

tory, and preparative actions. Endocr Rev 21:55–89.

34. de Kloet ER, Joels M, Holsboer F (2005) Stress

and the brain: from adaptation to disease. NatRev Neurosci 6: 463–475.

35. Anisman H, Merali Z, Hayley S (2008) Neuro-transmitter, peptide and cytokine processes in

relation to depressive disorder: comorbidity

between depression and neurodegenerative disor-ders. Prog Neurobiol 85: 1–74.

36. Moreau M, Andre C, O’Connor JC, Dumich SA,Woods JA, et al. (2008) Inoculation of Bacillus

Calmette-Guerin to mice induces an acute

episode of sickness behavior followed by chronicdepressive-like behavior. Brain Behav Immun 22:

1087–1095.

37. Bierhaus A, Wolf J, Andrassy M, Rohleder N,

Humpert PM, et al. (2003) A mechanism

converting psychosocial stress into mononuclear

Stress & Depression


cell activation. Proc Natl Acad Sci U S A 100:

1920–1925.

38. Bartolomucci A (2007) Social stress, immune

functions and disease in rodents. Front Neuroen-

docrinol 28: 28–49.

39. van der Laan S, de Kloet ER, Meijer OC (2009)

Timing is critical for effective glucocorticoid

receptor mediated repression of the cAMP-

induced CRH gene. PLoS ONE 4(1): e4327.

40. Charney DS, Manji HK (2004) Life stress, genes,

and depression: multiple pathways lead to in-

creased risk and new opportunities for interven-

tion. Sci STKE 205: re5.

41. Cryan JF, Mombereau C (2004) In search of a

depressed mouse: utility of models for studying

depression-related behavior in genetically modi-

fied mice. Mol Psychiatry 9: 326–357.

42. Bartolomucci A, Cabassi A, Covoni P, Ceresini G,

Cero C, et al. (2009) Metabolic consequences and

vulnerability to diet-induced obesity in male mice

under chronic social stress. PLoS ONE 4(1):

e4331.

43. Cooper B, Fuchs E, Flugge G (2009) Expression

of the axonal membrane glycoprotein M6a is

regulated by chronic stress. PLoS ONE 4(1):

e4659.

44. Ilin Y, Richter-Levin G (2009) Enriched environ-

ment experience overcomes learning deficits and

depressive-like behavior induced by juvenile

(early life) stress. PLoS ONE 4(1): e4329.

45. Hunter RG, Bellani R, Bloss E, Costa A,

McCarthy K, et al. (2009) Regulation of Kainate

Receptor Subunit mRNA by Stress and Cortico-

steroids in the Rat Hippocampus. PLoS ONE

4(1): e4328.

46. Oomen CA, Girardi CEN, Cahyadi R,

Verbeek EC, Krugers H, et al. (2009) Opposite

effects of early maternal deprivation on neuro-

genesis in male versus female rats. PLoS ONE

4(1): e3675.

47. Holsboer F, Ising M (2008) Central CRH system

in depression and anxiety—evidence from clinical

studies with CRH1 receptor antagonists.

Eur J Pharmacol 583: 350–357.

48. Berton O, Nestler EJ (2006) New approaches to

antidepressant drug discovery: beyond mono-

amines. Nat Rev Neurosci 7: 137–151.

49. Miczek KA, Yap JJ, Covington HE III (2008)

Social stress, therapeutics and drug abuse:

Preclinical models of escalated and depressed

intake. Pharmacol Ther 120: 102–128.

50. Willner P, Mitchell PJ (2002) The validity of

animal models of predisposition to depression.Behav Pharmacol 13: 169–188.

51. Koolhaas JM, Meerlo P, De Boer SF, Strubbe JH,

Bolhus B (1995) Social stress in rats: an animalmodel of depression? Acta Neuropsychiatrica 7:

27–29.52. Koolhaas JM, Meerlo P, De Boer SF, Strubbe JH,

Bohus B (1997) The temporal dynamics of the

stress response. Neurosci Biobehav Rev 21:775–782.

53. Collins A, Hill LE, Chandramohan Y,Whitcomb D, Droste SK, et al. (2009) Exercise

improves cognitive responses to psychologicalstress through enhancement of epigenetic mech-

anisms and gene expression in the dentate gyrus.

PLoS ONE 4(1): e4330.54. Anisman H, Matheson K (2005) Stress, depres-

sion, and anhedonia: caveats concerning animalmodels. Neurosci Biobehav Rev 29: 525–546.

55. McMahon FJ, Buervenich S, Charney D,

Lipsky R, Rush AJ, et al. (2006) Variation inthe gene encoding the serotonin 2A receptor is

associated with outcome of antidepressant treat-ment. Am J Hum Genet 78: 804–814.

56. Paddock S, Laje G, Charney D, Rush AJ,Wilson AF, et al. (2007) Association of GRIK4

with outcome of antidepressant treatment in the

STAR*D cohort. Am J Psychiatry 164:1181–1188.

57. Laifenfeld D, Karry R, Klein E, Ben-Shachar D(2005) Alterations in cell adhesion molecule L1

and functionally related genes in major depres-

sion: a postmortem study. Biol Psychiatry 57:716–725.

58. Alfonso J, Frick LR, Si lberman DM,Palumbo ML, Genaro AM, et al. (2006)

Regulation of hippocampal gene expression isconserved in two species subjected to different

stressors and antidepressant treatments. Biol

Psychiatry 59: 244–251.59. Alfonso J, Pollevick GD, Van Der Hart MG,

Flugge G, Fuchs E, et al. (2004) Identification ofgenes regulated by chronic psychosocial stress and

antidepressant treatment in the hippocampus.

Eur J Neurosci 19: 659–666.60. Tsankova N, Renthal W, Kumar A, Nestler EJ

(2007) Epigenetic regulation in psychiatric disor-ders. Nat Rev Neurosci 8: 355–367.

61. Sahay A, Hen R (2007) Adult hippocampalneurogenesis in depression. Nat Neurosci 10:

1110–1115.

62. Santarelli L, Saxe M, Gross C, Surget A,

Battaglia F, et al. (2003) Requirement of hippo-campal neurogenesis for the behavioral effects of

antidepressants. Science 301: 805–809.

63. Surget A, Saxe M, Leman S, Ibarguen-Vargas Y,Chalon S, et al. (2008) Drug-dependent require-

ment of hippocampal neurogenesis in a model ofdepression and of antidepressant reversal. Biol


64. Li Y, Luikart BW, Birnbaum S, Chen J,Kwon CH, et al. (2008) TrkB regulates hippo-

campal neurogenesis and governs sensitivity toantidepressive treatment. Neuron 59: 399–412.

65. Reif A, Fritzen S, Finger M, Strobel A, Lauer M,et al. (2006) Neural stem cell proliferation is

decreased in schizophrenia, but not in depression.

Mol Psychiatry 11: 514–522.66. Kornstein SG (1997) Gender differences in

depression: implication for treatment. J ClinPsychiatry 58: 12–18.

67. Willner P (1997) Validity, reliability and utility of

the chronic mild stress model of depression: a 10-year review and evaluation. Psychopharmacology

134: 319–29.68. Stunkard AJ, Faith MS, Allison KC (2003)

Depression and obesity. Biol Psychiatry 54:330–337.

69. Kloiber S, Ising M, Reppermund S, Horstmann S,

Dose T, et al. (2007) Overweight and obesityaffect treatment response in major depression.

Biol Psychiatry 62: 321–326.70. van Praag HM (2004) Can stress cause depres-

sion? Prog Neuropsychopharmacol Biol Psychia-

try 28: 891–907.71. Rydmark I, Wahlberg K, Ghatan PH, Modell S,

Nygren A, et al. (2006) Neuroendocrine, cognitiveand structural imaging characteristics of women

on longterm sickleave with job stress-induceddepression. Biol Psychiatry 60: 867–873.

72. Kendler KS, Karkowski LM, Prescott CA (1999)

Causal relationship between stressful life eventsand the onset of major depression. Am J Psychiatry

156: 837–841.73. Vaez M, Rylander G, Nygren A, Asberg M,

Alexanderson K (2007) Sickness absence and

disability pension in a cohort of employeesinitially on long-term sick leave due to psychiatric

disorders in Sweden. Soc Psychiatry PsychiatrEpidemiol 42: 381–388.

74. Tennant C (2001) Work-related stress anddepressive disorders. J Psychosom Res 51:

697–704.

Stress & Depression


Overview

Conceptual Challenges of a Tentative Model of Stress-Induced DepressionBernhard Baune*

Department of Psychiatry and Psychiatric Neuroscience, James Cook University, Queensland, Australia

The concept of a stress-induced depres-

sion has recently been proposed in light of

new findings from animal and human

studies. Depression associated with stress

involves a number of body systems such as

the neuroendocrine and neurotransmitter

system and the immune system including

cytokines and the dysregulation of the HPA

axis interacting in complex pathways.

However, numerous research challenges

present when addressing a tentative con-

cept of stress-induced depression. One of

them is the requirement to establish a

causative relationship between stressful

environmental factors and stress-related

neurochemical and genetic pathways in a

complex model of interaction using valid

and etiological relevant animal models.

Another challenge is the establishment of

animal models compatible with the concept

of stress-induced depression; however,

chronic mild and social stress models are

promising models for the study of stressfully

perceived environmental events assembling

stressors relevant in depression. Moreover,

the consideration of individual psycholog-

ical ‘‘neurotic’’ factors presents another

major challenge in animal and human

models of stress-induced depression. In

addition, the study of translational implica-

tions is needed to enhance research into the

validity and relevance of a tentative concept

of stress-induced depression.

Stress and Depression

An enduring clinical literature suggests

that individual vulnerability to stress and

subsequent predisposition to develop cer-

tain disease states, notably depression, are

related at least in part to a history of early

environmental adversity. Exposure to ear-

ly trauma, for example sexual and physical

abuse or other types of early disadvantage,

can increase several-fold the risk of being

diagnosed with a depressive illness in

adulthood [1,2].

Similarly, the onset and recurrence of

adult depression can reliably be predicted

by the presence of environmental stressors,

often labeled ‘‘life events.’’ Some individ-

uals may have a genetic propensity to

select themselves into high-risk environ-

ments, but epidemiological studies using

identical and non-identical twins have

shown that there is still a substantial causal

relationship between stressful life events

and depression [3].

Since the mechanisms by which stress is

mediated in the central nervous system are

multiple and include the autonomic ner-

vous system, the neuroendocrine and

neurotransmitter systems, and the immune

system, it appears challenging to identify a

single ‘‘stress’’-pathway leading to or caus-

ing depression. However, it is obvious that

stress may have an impact on a number of

other systems relevant to depression, in-

cluding the autonomic nervous system, the

neuroendocrine system, and the immune

system. In addition, stress is related to

symptom clusters such as sleep disturbance,

impaired learning, and impaired memory,

which have been suggested to form en-

dophenotypes of depression [4].

Specific Stress Models Relevantto Depression

The study of the relationship between

stress and depression depends on the

concepts and models used for defining

stress and depression. Clinical studies have

consistently implicated abnormalities in the

regulation of key neuroendocrine responses

to stress in a proportion of patients with

depression, with a hyperactivity of the HPA

axis that is probably driven by hypersecre-

tion of the hypothalamic peptide cortico-

tropine releasing hormone (CRH) [5,6].

Certain areas of the brain, including parts

of the hippocampal formation, are more

sensitive to damage from high levels of

glucocorticoids [7].

Inflammation and cytokines appear to

play an important role in mediating the

relationship between stress and the devel-

opment of depression and indicate the

complex relationship between stress and

the immune and neuroendocrine systems.

In humans, psychological stress significant-

ly increases pro-inflammatory (but inhibits

anti-inflammatory) cytokine production in

patients responding to stress and anxiety.

In depressed patients, increases in macro-

phage activity and the production of pro-

inflammatory cytokines complement, and

some acute-phase proteins have been

consistently reported [8].

Furthermore, animal experiments have

demonstrated that pro-inflammatory cyto-

kines, such as interleukin (IL)-1beta, IL-6,

and TNF-alpha can stimulate the hypo-

thalamus to release corticotrophin releas-

ing hormone (CRH), which, via adreno-

corticotropic hormone (ACTH), induces

glucocorticoid (GC) secretion. Excessive

secretion of GC may downregulate GC

receptors in the hippocampus, which

impairs the GC feedback system. Similar

neuroendocrine changes also occur in

depressed patients. From the neurotrans-

mitter perspective, pro-inflammatory cy-

tokines have been found to reduce both

serotonin and norepinephrine availability

to the brain to levels similar to those

observed in depression [9].

There are a number of animal stress

models of depression, including learned

helplessness, which is perhaps the best-

known stress model of depression; other

models are the inescapable foot shock and

intracranial-self stimulation model, the

behavioural despair model, and the chron-

ic unpredictable mild stress and the social

stress models of depression. All models

have presented with significant validity

problems relevant to a hypothesized etio-

logical stress model of depression. A

promising group of stress models of

Citation: Baune B (2009) Conceptual Challenges of a Tentative Model of Stress-Induced Depression. PLoSONE 4(1): e4266. doi:10.1371/journal.pone.0004266

Editor: Peter Binfield, Public Library of Science, United States of America

Received December 16, 2008; Accepted January 5, 2009; Published January 30, 2009

Copyright: � 2009 Bernhard Baune.

Funding: The author has no support or funding to report.


* E-mail: [email protected]


depression is called ‘‘chronic social stress

models’’ [10], which is considered a model

of social defeat or subordination [11], and

therefore may mimic situations occurring

in humans and may be an appropriate

model for depressive disorders [12]. It is

suggested that the rat chronic social stress

model may be useful to describe depressive

disorders; however, further research into

the context of the immune system [13] and

the neuroendocrine and neurotransmitter

systems is required to explore the validity

of the chronic social stress model in the

context of stress-induced depression.

Conceptual Challenges

Although a tentative new subtype of

depression has been proposed and called

stress-induced depression [14], the scientific

question of whether stress can cause

depression consistent with existing diag-

nostic criteria such as in DSMIV (Diagnostic

and Statistical Manual of Mental Disorders, 4th

edition) or ICD10 (International Statistical

Classification of Disease and Related Health

Problems, 10th revision) is unresolved. Is it

possible that stress can be a causative

factor in depression, or should the re-

search approach be more precise to

unravel the stress-induced specific molec-

ular mechanisms eventually inducing

symptoms that are defined as depression?

In the latter case, the term stress-associated

depression might be better-suited to describe

the complex interactions between environ-

mental stress and molecular mechanisms

in a complex phenotype of depression. In

line with this is the observation that

individuals developing ‘‘stress-associated

depression’’ are characterized by a genet-

ically and socially determined higher

susceptibility to stress. Diathesis-stress

theories of depression predict that genes

influence individuals’ sensitivity to stressful

events, consistent with a potentially im-

portant role of gene-by-environment in-

teractions played in the etiology of depres-

sive psychopathology [15].

Although the concept of a clinically

relevant stress-induced depression, which

is characterized by a heterogeneous phe-

notype, intuitively may have clinical ap-

plication and a relatively low threshold of

acceptance from a clinical point of view as

well as from a basic science perspective

into the molecular mechanisms of stress,

the definition of a circumscribed and

specific phenotype of stress-induced de-

pression is lacking. Given the lack of

specificity between stressors and psycho-

pathological outcomes [16], one may

hypothesize that gene–stressor interactions

account for a better outcome specificity

than stress alone. Therefore, psychopath-

ological constructs reflecting gene-by-en-

vironment interactions might be among

the most specific and most useful endo-

phenotypes for major depression. As an

example, Caspi et al. [17] have shown in a

representative prospective study that 5-

HTT genotypes moderate the influence of

stressful life events on major depression.

The establishment of a causative rela-

tionship between stress and a phenotype of

depression is most challenging as it

requires a chain of evidence linking a

number of crucial factors built into

complex systems: (1) the environmental

factor of stress, (2) the individual percep-

tion of and vulnerability to stress for which

the diagnostic construct of neuroticism

defined as general vulnerability to anxiety

and depressive symptoms under stress

might be useful, (3) the genetic level and

corresponding (4) neurochemical/neuro-

anatomical characteristics of stress-in-

duced changes, and (5) the psychopathol-

ogy phenotype consistent with symptoms

of depression.

In light of such complexity, the estab-

lishment of a causative relationship in a

concept of stress-induced depression is still

facing conceptually and methodologically

unresolved problems, some of which have

been discussed in this Overview.

Conclusions

Since stressfully perceived environmen-

tal events activate a number of neuro-

chemical systems including the immune

system in a complex interaction of path-

ways in the individual, it appears difficult

at this stage to define a homogenous

psychopathological and neurochemical

endophenotype required for a model of

stress-induced depression. Research in this

area is required to establish a causative

relationship between stressful environmen-

tal factors, individual psychological ‘‘neu-

rotic’’ factors, and stress-related neuro-

chemical and genetic pathways. The

development of adequate human and

animal models and the study of their

translational implications will enhance the

research into the validity and relevance of

a concept of stress-induced depression.

In the PLoS ONE Special Collection

‘‘Stress-Induced Depression and Comor-

bidities: From Bench to Bedside,’’ some of

the conceptual challenges relevant for the

scientific discussion of a concept of stress-

induced depression will be addressed by

presenting empirical data from animal and

human studies.

References

1. Nemeroff CB (2004) Neurobiological consequenc-es of childhood trauma. J Clin Psychiatry

65(Suppl 1): 18–28.2. Gilmer WS, McKinney WT (2003) Early expe-

rience and depressive disorders: human and non-human primate studies. J Affect Disord 75:

97–113.

3. Tennant C (2002) Life events, stress and depres-sion: a review of recent findings. Aust N Z J

Psychiatry 36: 173–182.4. Hasler G, Drevets WC, Manji HK, Charney DS

(2004) Discovering endophenotypes for major

depression. Neuropsychopharmacology 29:1765–1781.

5. Binder EB, Kunzel HE, Nickel T, Kern N,Pfennig A, et al. (2008) HPA-axis regulation at in-

patient admission is associated with antidepres-sant therapy outcome in male but not in female

depressed patients. Psychoneuroendocrinology

34: 99–109.6. Aihara M, Ida I, Yuuki N, Oshima A, Kumano H,

et al. (2007) HPA axis dysfunction in unmedicatedmajor depressive disorder and its normalization

by pharmacotherapy correlates with alteration of

neural activity in prefrontal cortex and limbic/paralimbic regions. Psychiatry Res 155: 245–256.

7. McEwen BS (2005) Glucocorticoids, depression,and mood disorders: structural remodeling in the

brain. Metabolism 54: 20–23.8. Garcia-Bueno B, Caso JR, Leza JC (2008) Stress

as a neuroinflammatory condition in brain:

damaging and protective mechanisms. NeurosciBiobehav Rev 32: 1136–1151.

9. Anisman H, Merali Z, Hayley S (2008) Neuro-transmitter, peptide and cytokine processes in

relation to depressive disorder: comorbidity

between depression and neurodegenerative disor-ders. Prog Neurobiol 85: 1–74.

10. Tamashiro KL, Nguyen MM, Sakai RR (2005)Social stress: from rodents to primates. Front

Neuroendocrinol 26: 27–40.11. Huhman KL (2006) Social conflict models: can

they inform us about human psychopathology?

Horm Behav 50: 640–646.12. Rygula R, Abumaria N, Flugge G, Fuchs E,

Ruther E, et al. (2005) Anhedonia and motiva-tional deficits in rats: impact of chronic social

stress. Behav Brain Res 162: 127–134.

13. Bartolomucci A (2007) Social stress, immune

functions and disease in rodents. Front Neuroen-

docrinol 28: 28–49.

14. Rydmark I, Wahlberg K, Ghatan PH, Modell S,

Nygren A, et al. (2006) Neuroendocrine, cognitive

and structural imaging characteristics of women

on longterm sickleave with job stress-induced

depression. Biol Psychiatry 60: 867–873.

15. Anisman H, Merali Z, Stead JD (2008) Experi-

ential and genetic contributions to depressive-

and anxiety-like disorders: clinical and experi-

mental studies. Neurosci Biobehav Rev 32:

1185–1206.

16. McMahon SD, Grant KE, Compas BE,

Thurm AE, Ey S (2003) Stress and psychopa-

thology in children and adolescents: is there

evidence of specificity? J Child Psychol Psychiatry

44: 107–133.

17. Caspi A, Sugden K, Moffitt TE, Taylor A,

Craig IW, et al. (2003) Influence of life stress on

depression: moderation by a polymorphism in the

5-HTT gene. Science 301: 386–389.

Depression Overview


Novel Biochemical Markers of Psychosocial Stress inWomenMarie Asberg1, Ake Nygren1, Rosario Leopardi2*, Gunnar Rylander2, Ulla Peterson1, Lukas Wilczek1,

Hakan Kallmen1, Mirjam Ekstedt3, Torbjorn Akerstedt3, Mats Lekander2, Rolf Ekman4

1 Department of Clinical Sciences, Karolinska Institute, Stockholm, Sweden, 2 Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden, 3 Institute for

Psychosocial Medicine and Department of Public Health, Karolinska Institute, Stockholm, Sweden, 4 Institute of Clinical Neuroscience and Physiology, Goteborg University,

Goteborg, Sweden

Abstract

Background: Prolonged psychosocial stress is a condition assessed through self-reports. Here we aimed to identifybiochemical markers for screening and early intervention in women.

Methods: Plasma concentrations of interleukin (IL) 1-a, IL1-b, IL-2, IL-4, IL-6, IL-8, IL-10, interferon-c (INF-c), tumor necrosisfactor-a (TNF-a), monocyte chemotactic protein-1 (MCP-1), epidermal growth factor (EGF), vascular endothelial growthfactor (VEGF), thyroid stimulating hormone (TSH), total tri-iodothyronine (TT3), total thyroxine (TT4), prolactin, andtestosterone were measured in: 195 women on long-term sick-leave for a stress-related affective disorder, 45 women at riskfor professional burnout, and 84 healthy women.

Results: We found significantly increased levels of MCP-1, VEGF and EGF in women exposed to prolonged psychosocialstress. Statistical analysis indicates that they independently associate with a significant risk for being classified as ill.

Conclusions: MCP-1, EGF, and VEGF are potential markers for screening and early intervention in women under prolongedpsychosocial stress.

Citation: Asberg M, Nygren A, Leopardi R, Rylander G, Peterson U, et al. (2009) Novel Biochemical Markers of Psychosocial Stress in Women. PLoS ONE 4(1):e3590. doi:10.1371/journal.pone.0003590


Received August 21, 2008; Accepted October 10, 2008; Published January 29, 2009

Copyright: � 2009 Asberg et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The study was supported by the Swedish Science Council (5454 My, 07517 RE), the Sylvan1s foundation, the State of Sweden under the LUAagreement, and the insurance companies AFA and Alecta. The Randox Laboratories Ltd (Crumlin, UK) supplied the assay kits and performed the measurements.The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.



Introduction

Sustained psychosocial stress, often related to work, is an

increasingly important factor in the development of illness,

physical as well as mental [1–3]. In Sweden, public expenditure

for sick-leave more than doubled in a few years, and in 2003 the

number of workers on long-term sick-leave (more than 30 days)

increased to all time high levels [4]. Women represent about 70–

80% of this patient group. We have studied more than 400

patients on long-term sick leave because of an affective disorder,

and found that about 80% met DSM-IV criteria for major

depression at some time during their illness (our unpublished data).

However, mental and physical exhaustion were the most

prominent symptoms, which tended to persist after the depressive

symptoms had cleared. Many patients remained incapacitated for

a very long period with a pronounced tendency to recurrence.

The sequence of pathophysiological events set forward by

prolonged exposure to mental stress in humans is not completely

characterized, but it involves the hypothalamus-pituitary-adrenal

(HPA) axis, as well as the endocrine and the immune systems [5–

7]. The cortisol response to corticotropin-releasing hormone

(CRH) was recently analyzed in our patients [8]. The results

showed an attenuated dexamethasone-CRH response, a feature

opposite to that seen in other patients with major depression [9].

Such a finding suggests an imbalance in other physiological

systems. Metabolic processes occurring under chronic stress might

be reflected in altered hormones, cytokines and cellular growth

factors levels [10–11].

Here, we have taken advantage of the microarray technology,

which uses miniaturized microspot ligand-binding assays to enable

the simultaneous measurement of a large number of biochemical

markers from a minute plasma sample within a single assay [12].

Using this technique we assessed a panel of 17 biological mediators

(hormones, cytokines, and cellular growth factors), aiming to

investigate whether plasma concentrations of any of these

molecules could be related to prolonged psychosocial stress, and

if so, whether they could be used, individually or in combination,

as markers for screening and diagnostic purposes.

Materials and Methods

Patients on long term sick leaveThe studies were cleared by the Research Ethics Committees of

the Karolinska Institute, and the Medical Faculty of Linkoping

University, respectively. One hundred ninety-five women (mean


age 44, SD = 9.1 years, age range 21–60 years) on long-term

sickleave (more than three months) for any affective or stress-

related mental disorder (depression, anxiety disorder, stress

disorder, burnout syndrome, exhaustion disorder) were recruited

from one of the major Swedish insurance companies, the Alecta

Company, which insures about 600,000 employees in private

enterprises, mainly white collar workers. The company keeps a

database which includes diagnostic information for all cases of sick

leave exceeding 3 months. Letters were sent to all patients on sick

leave with any of the above mentioned diagnoses on their sickness

certificate. Patients were then approached by telephone and

invited to participate in a clinical study involving interviews and

questionnaires, as well as an offer of further treatment. Written

informed consent was obtained from all participants. All patients

were ambulatory and none had received inpatient care for their

current illness. They were diagnosed according to the Diagnostic

and Statistical Manual of the American Psychiatric Association,

4th Edition [DSM-IV; 13] by specially trained physicians, using

the Structured Clinical Interview for DSM-IV [14]. Eighty-two

per cent met DSM-IV criteria for Major Depressive Disorder at

some time during their current illness episode. Likely eliciting

factors were: work-related stress (39%), stressful family relation-

ships (9.3%), a combination of work and family stressors (49.3%),

or not identified (2.1%).

Health care personnel with occupational stressA group of women experiencing work stress was selected from

the results of a questionnaire sent to all 6118 health care

employees of a Swedish county council (Kalmar). Of the 3976

employees who replied to the questionnaire, those who scored

above the 75th percentile on the Oldenburg Burnout Inventory

[OLBI; 15], which measures degree of professional burnout, were

invited to participate in a randomized controlled study of the

possible beneficial effect of a series of structured group discussions

with colleagues. Those who were randomised to active treatment

were asked to leave blood samples prior to treatment. The

resulting group consisted of 45 women, ranging in age from 39 to

62 years, mean age 52.8, SD = 5.2 years.

Healthy control workersThe control group included 84 women (mean age 36.1,

SD = 8.4, range 23–62 years), recruited among the employees of

a Swedish IT-company. Out of 560 employees, both women and

men, the above 84 individuals had agreed to undergo a

physiological examination as part of a health screening in a stress

prevention program at the company. The subjects were all full

time workers, of whom 34% were managers and 54% project

leaders. Two women who were pregnant at the time of the

physiological examination were excluded.

Analytical methodsVenous blood was drawn into tubes containing EDTA, and

immediately centrifuged. Plasma was separated and stored in

aliquots at 220uC or below until analyzed. The following

cytokines and growth factors were analyzed: interleukin 1-a(IL1-a), interleukin 1-b (IL1-b), interleukin 2 (IL-2), interleukin 4

(IL-4), interleukin 6 (IL-6), interleukin 8 (IL-8), interleukin 10 (IL-

10), interferon-c (INF-c), tumor necrosis factor-a (TNF-a),

monocyte chemotactic protein-1 (MCP-1), epidermal growth

factor (EGF), and vascular endothelial growth factor (VEGF).

Furthermore, thyroid stimulating hormone (TSH), total tri-

iodothyronine (TT3), total thyroxine (TT4), prolactin, and

testosterone were included in the panel analyzed on a high

throughput automated biochip immunoassay system, EvidenceH,

Randox Laboratories Ltd [Crumlin, UK; 16]

Statistical analysisAll analyses were performed with the SPSS 11.5 for Windows.

One-way ANOVA’s were calculated by using the group variable

as independent variable with 3 levels, and each biochemical

marker as a dependent variable. The significance levels were

adjusted according to Bonferroni’s method. Pair-wise post hoc

comparisons were made according to Scheffe’s method. To

determine the optimal cut-off point maximizing the sensitivity

(true positive rate) and specificity (true negative rate), receiver

operating curve (ROC) analyses were performed [17]. An area

under the ROC curve of 0.9 or above is needed for a reliable

differentiation of the groups. To calculate the odds of being

classified as ill or not, when having a value above, compared with

below the established cut-off, logistic regression analyses were

used.

Results

Means and standard deviations for the 17 markers in the three

groups are shown in Tables 1 and 2. Neither the interleukins nor

IFN-c or TNF-a differed between the groups. Differences were

observed, however, between the three subject groups for MCP-1,

EGF, and VEGF. MCP-1 levels were more than twice as high in

the sick leave group compared to the healthy controls, with the

occupational stress group in between. VEGF levels were three

times as high in the sick leave group, and EGF levels were more

than twice as high, compared to the healthy group, once again

with the occupational stress group in between. The sick leave

group also had significantly lower levels of prolactin and TSH

(Table 1).

Since the correlations between some of the markers and age

were significant, and the mean age differed significantly between

the groups (one-way ANOVA F = 59.09 df = 2,338, P = 0.000), we

controlled for age in an analysis of covariance. This resulted in one

additional significant difference, namely in testosterone, which was

higher in the sick leave group. In order to examine the usefulness

of these markers for screening and diagnostic purposes, a receiver

operating characteristic (ROC) curve analysis was performed

(Table 2). As seen from the table, the best sensitivity and specificity

was obtained for MCP-1, VEGF, and EGF.

The relative value of MCP-1, EGF, and VEGF as risk factors

for classification as ill or healthy was also tested. The results are

shown in Table 3, and indicate that each of these markers

independently associates with a significantly increased risk for

being classified as ill.

Discussion

Our results show a direct correlation between plasma

concentrations of MCP-1, EGF, and VEGF, and psychosocial

stress-related illness. Plasma levels were elevated in subjects with

occupational stress, and more so in a group of subjects on long

term sick leave for an affective disorder following exposure to

chronic stress. Statistically, each of these three markers associated

independently with a significantly increased risk for being classified

as ill.

However, our results should be seen as preliminary, and need to

be replicated because of a number of intrinsic limitations:

(i) The study design has involved only one single time point

observation during the course of a prolonged stress

condition. This design may in principle suffer from possible

Stress Biomarkers in Women


short time-course variations. However, we have data from

an almost complete 2-year follow up study on our recovering

patient cohort, showing that MCP-1, EGF, and VEGF in

plasma decrease very slowly within months to years (Asberg

et al, unpublished).

(ii) The daytime point for blood withdrawal varied among

individuals approximately between 9 AM and 3 PM, making

circadian rhythm a possible confounding factor. Circadian

rhythms have been described for EGF and VEGF, and

MCP-1 [18,19,20]. However, both EGF and VEGF plasma

levels are relatively stable during daytime [21,22]. Thus, it is

unlikely that collection time points could significantly

influence results on these two mediators. On the other

hand, the studies available on MCP-1 were done in mice

peritoneal macrophages [20], and to our knowledge similar

studies have not been confirmed in any human tissue. Thus,

it remains entirely possible that collection time points may

have played a role for MCP-1, as well as for other markers

for which we have not found a significant association with

stress and disease. Indeed, several of these markers are

known to have circadian rhythms [23], and for this reason

our negative results should be taken with caution.

(iii) We have not controlled for nicotine consumption or for sex

hormone variables such as menstrual phase or oral

contraceptive use. Also, data on nutrition and physical

activity have not been consistently collected across groups.

While our study is mainly concerned with their potential use as

markers of disease, plasma MCP-1, EGF, and VEGF may also be

related to pathophysiological outcomes that are worth to consider.

For instance, MCP-1 mediates inflammatory-like disorders and

oxidative stress [24], and it also contributes to macrophage

infiltration into adipose tissue and insulin resistance [25].

Moreover, at least for acute stress there are examples of an

MCP-1 effect on chemotaxis and immune cell redistribution.

Table 1. Biochemical markers in women experiencing different levels of stress.

Marker

Sick leave(group 1)

Occupational stress(group 2)

Healthy Subjects(group 3) ANOVA

Significant group pairwise comparisons

M SD M SD M SD F Df P

IL1-a 5.1 14.1 5.6 6.1 5.4 5.1 0.92 2,338 .402

IL1-b 2.9 6.4 2.5 4.0 2.1 2.7 0.59 2,338 .554

IL-2 17.6 39.4 18.6 24.0 14.6 24.0 0.31 2,338 .734

IL-4 3.8 3.5 3.2 0.8 3.3 1.7 1.82 2,338 .163

IL-6 6.6 16.4 9.4 26.2 13.7 33.4 2.87 2,333 .058

IL-8 5.4 10.8 3.3 2.9 3.1 1.3 2.98 2,338 .052

IL-10 3.4 6.6 3.0 5.6 6.5 20.3 2.37 2,338 .095

IFN-c 0.7 0.7 0.7 0.7 0.7 0.7 0.39 2,338 .678

TNF-a 5.1 16.2 3.3 2.5 5.3 19.6 0.26 2,338 .773

MCP-1 348.4 126.7 217.8 92.6 160.2 85.7 97.82 2,338 .000* 1–2, 1–3, 2–3

EGF 117.0 77.2 70.6 53.0 29.4 47.5 56.16 2,338 .000* 1–2, 1–3, 2–3

VEGF 30.9 22.7 18.4 15.4 10.3 7.1 41.24 2,338 .000* 1–2, 1–3

Prolactin 388.6 216.4 534.0 246.3 684.1 748.0 15.47 2,346 .000* 1–3

TT3 0.5 0.3 0.6 0.2 0.5 0.3 6.46 2,329 .002* 1–2

TT4 6.9 1.6 7.1 1.6 7.2 1.8 0.82 2,441 .442

TSH 1.8 1.0 2.4 1.6 2.4 1.2 11.11 2,342 .000* 1–3

Testosterone 4.1 1.5 4.0 1.8 3.5 1.4 5.39 2,330 .005

Concentration is given in pg/ml except for prolactin (mIU/L), TT3 (ng/ml), TT4 (g/L), and TSH (mIU/ml), and testosterone (nmol/L). Mean (M) and standard deviation (SD)values are indicated for the three test groups. Significant comparisons between groups (P,.05 level) are indicated in the last column.doi:10.1371/journal.pone.0003590.t001

Table 2. Optimal cut-off, area under the ROC-curve (Area),and diagnostic sensitivity and specificity of statisticallysignificant biochemical markers.

Marker Area Cut-off Sensitivity Specificity

MCP-1 0.886 243.00 0.85 0.92

VEGF 0.805 7.80 0.78 0.85

EGF 0.798 68.00 0.69 1.00

Prolactin 0.699 380.00 0.60 0.79

TSH 0.624 1.80 0.58 0.69

Testosterone 0.618 3.73 0.63 0.65

doi:10.1371/journal.pone.0003590.t002

Table 3. Relative risks of being classified as ill, using theestablished cut-off points of MCP-1, EGF and VEGF (Table 3).

Marker Beta Wald P OR 95% CI

MCP-1 3.55 52.17 0.000 34.85 13.30–91.35

EGF 2.12 18.02 0.000 8.35 3.13–22.25

VEGF 2.08 18.99 0.000 8.02 3.14–20.44

Results from multiple logistic regression analyses are shown. OR, Observed Risk,(95% CI).doi:10.1371/journal.pone.0003590.t003



Similarly, there is a documented relationship between MCP-1 and

atherosclerosis. [26]. On the other hand, EGF is a mediator of

stress-related events influencing the cell cycle [27]. This might be

particularly important in heart diseases, where adverse effects of

angiotensin II could be partly mediated by EGF [28]. Similarly,

VEGF has been reported to be upregulated by angiotensin II, and

thereby activate vascular inflammation [29]. However, VEGF has

also been reported to act as a neuroprotective factor, which makes

our results somewhat intriguing [30]. Also, VEGF has recently

been shown to selectively recruit stem and progenitor cells to

specific organs [31].

While psychosocial stress in humans is a complex phenomenon

not readily captured by bodily biochemical modifications, animal

models have began to shed light on translational mediating events.

In mice, psychosocial stress has been shown to be converted into

cell cycle signalling, and the mechanism is mediated by the

transcription factor nuclear factor kappaB [NF-kappaB; 32].

Interestingly, NF-kappaB expression in humans has been reported

to be influenced by EGF and VEGF during various pathological

conditions [33,34]. In turn, MCP-1 expression has been shown to

be controlled by NF-kappaB [35]. Taken together, these data

suggest that psychosocial stress may broadly influence pathophys-

iological changes acting on a cellular level also in humans.

Previous studies have shown that patients on sick leave because

of occupational burnout resulting from psychosocial stress, have a

disrupted sleep, with more arousals and sleep fragmentation, more

wake time, and lower sleep efficiency [36]. Sleep fragmentation is

associated with elevated levels of metabolic and cardiovascular risk

indicators of stress-related disorders, such as morning cortisol,

heart rate, systolic and diastolic blood pressure, total cholesterol,

high-density lipoprotein (HDL)- and low-density lipoprotein

(LDL)-cholesterol, and LDL/HDL-ratio [37]. It is conceivable

that some of the pathophysiological changes developing during

exposure to psychosocial chronic stress reflect sleep disturbances.

Some studies suggest that MCP1, EGF en VEGF levels may

indeed be related with sleep quality, and be altered as an effect of

disrupted sleep [38–40]. It would be therefore important to test a

correlation of MCP1, EGF and VEGF levels in relationship to

qualitative and quantitative sleep alterations.

Taken together with the recently reported dexamethasone-

CRH data [8], our results indicate that women under prolonged

psychosocial stress develop so-far unique neuro-endocrine-im-

mune alterations. If confirmed, our results may be developed into

novel work hypotheses to construct models for further investiga-

tions both in the preclinical and in the clinical settings. Also, our

data may bring to the clinician a potential tool for diagnosis of a

condition that is poorly understood, not diagnosable through

laboratory tests, yet progressively more common in industrialized

areas of the word.

Author Contributions

Conceived and designed the experiments: M GR TA ML RE. Performed

the experiments: UP ME. Analyzed the data: M N RL GR LW HK ME

TA ML RE. Contributed reagents/materials/analysis tools: N LW HK TA

RE. Wrote the paper: M RL RE.

References

1. Tennant C (2001) Work-related stress and depressive disorders. J PsychosomaticRes 51: 697–704.

2. Vahtera J, Kivimaki M, Pentti J (1997) Effect of organisational downsizing on

health of employees. Lancet 350: 1124–1128.

3. Levi L (2000) Guidance on work-related stress: Spice of life or kiss of death?Luxembourg: Office for Official Publications of the European Communities.

4. Statistics Sweden (2004) (Accessed Nov 11, 2004, at www.rfv.se.).

5. Eskandari F, Webster J, Sternberg E (2003) Neural immune pathways and their

connection to inflammatory diseases. Arthritis Res Therapy 5: 251–265.

6. Haddad J, Saade N, Safieh-Garabedian B (2002) Cytokines and neuro-immuno-endocrine interactions: A role for the hypothalamic-pituitary-adrenal revolving

axis. J Neuroimmunol 133: 1–19.

7. Lundberg U (2005) Stress hormones in health and illness: the roles of work and

gender. Psychoneuroendocrinology 30: 1017–1021.

8. Rydmark I, Wahlberg K, Ghatan PH, Modell S, Nygren A, et al. (2006)

Neuroendocrine, cognitive and structural imaging characteristics of women on

longterm sickleave with job stress-induced depression. Biol Psychiatry 60: 867–873.

9. Heuser I, Yassouridis A, Holsboer F (1994) The combined dexamethasone/CRH test: a refined laboratory test for psychiatric disorders. J Psychiatr Res 28:

341–356.

10. Anisman H, Merali Z (2002) Cytokines, stress, and depressive illness. BrainBehavior Immunity 16: 513–524.

11. Padgett D, Glaser R (2003) How stress influences the immune response. Trends

Immunol 24: 444–448.

12. Molloy RM, McConnell RI, Lamont JV, FitzGerald SP (2005) Automation ofbiochip array technology for quality results. Clin Chem Lab Med 43:

1303–1313.

13. American Psychiatric Association (1994) Diagnostic and statistical manual of

mental disorders (4th ed.). Washington DC: American Psychiatric Association.

14. First M, Spitzer R, Gibbon M, Williams J (1997) Structured Clinical Interview

for DSM-IVTM Axis I Disorders (SCID-I), Clinician Version. Arlington, VA:

American Psychiatric Press, Inc.

15. Demerouti E, Bakker AB, Nachreiner F, Schaufeli WB (2001) The job demands-resources model of burnout. J Appl Psychol 86: 499–512.

16. FitzGerald SP, Lamont JV, McConnell RI, Benchiki EO (2005) Development of

a high-throughput automated analyzer using biochip array technology. ClinChem 51: 1165–1176.

17. Metz C (1978) Basic principles of ROC analysis. Semin Nucl Med 8: 283–298.

18. Krieger DT, Hauser H, Liotta A, Zelenets (1976) A circadian periodicity of

epidermal growth factor and its abolition by superior cervical ganglionectomy.Endocrinology 99: 1589–1596.

19. Robertson D, Selleck K, Suikkari AM, Hurley V, Moohan J, et al. (1995)

Urinary vascular endothelial growth factor concentrations in women withundergoing gonadotropin treatment. Hum Reprod 10: 2478–2482.

20. Hayashi M, Shimba S, Tezuka M (2007) Characterization of the molecular clock

in mouse peritoneal macrophages. Biol Pharm Bull 30: 621–6.

21. Haus E, Dumitriu L, Nicolau GY, Bologa S, Sackett-Lundeen L (2001)

Circadian rhythms of basic fibroblast growth factor (bFGF), epidermal growth

factor (EGF), insulin-like growth factor-1 (IGF-1), insulin-like growth factor

binding protein-3 (IGFBP-3), cortisol, and melatonin in women with breast

cancer. Chronobiol Int 18: 709–727.

22. Hetland ML, Christensen IJ, Lottenburger T, Johansen JS, Svendsen MN, et al.

(2008) Circulating VEGF as a biological marker in patients with rheumatoid

arthritis? Preanalytical and biological variability in healthy persons and in

patients. Dis Markers 24: 1–10.

23. Haus E, Smolensky MH (1999) Biologic rhythms in the immune system.

Chronobiol Int 16: 581–622.

24. Egashira K (2003) Molecular mechanisms mediating inflammation in vascular

disease: special reference to Monocyte Chemoattractant Protein-1. Hypertension

41: 834–841.

25. Kanda H, Tateya S, Tamori Y, Kotani K, Hiasa K, et al. (2006) MCP-1

contributes to macrophage infiltration into adipose tissue, insulin resistance, and

hepatic steatosis in obesity. J Clin Invest 116: 1494–1505.

26. Amasyali B, Kose S, Kursaklioglu H, Barcin C, Kilic A (2008) Monocyte

chemoattractant protein-1 in acute coronary syndromes: Complex vicious

interaction. Int J Cardiol, In press (Epub ahead of print).

27. Zhang Y-W, Vande Woude GF (2007) Mig-6, Signal Transduction, Stress

Response and Cancer. Cell Cycle 6: 507–513.

28. Smith NJ, Chan HW, Osborne JE, Thomas WG, Hannan RD (2004) Hijacking

epidermal growth factor receptors by angiotensin II: new possibilities for

understanding and treating cardiac hypertrophy. Cell Mol Life Sci 61:

2695–703.

29. Granger DN, Vowinkel T, Petnehazy T (2004) Modulation of the inflammatory

response in cardiovascular disease. Hypertension 43: 924–931.

30. Vogel G (2005) Developmental biology. The unexpected brains behind blood

vessel growth. Science 307: 665–667.

31. Rabbany SY, Heissig B, Hattori K, Rafii S (2003) Molecular pathways

regulating mobilization of marrow-derived stem cells for tissue revascularization.

Trends Mol Med 9: 109–117.

32. Bierhaus A, Wolf J, Andrassy M, Rohleder N, Humpert PM, et al. (2003) A

mechanism converting psychosocial stress into mononuclear cell activation. Proc

Natl Acad Sci U S A 100: 1920–1925.

33. Graham B, Gibson SB (2005) The two faces of NF-KappaB in cell survival

responses. Cell Cycle 4: 1342–1345.

34. Wilkinson-Berka JL, Wraight C, Werther G (2006) The role of growth hormone,

insulin-like growth factor and somatostatin in diabetic retinopathy. Curr Med

Chem 13: 3307–3317.



35. Nam NH (2006) Naturally occurring NF-kappaB inhibitors. Mini Rev Med

Chem 6: 945–51.

36. Ekstedt M, Soderstrom M, Akerstedt T, Nilsson J, Søndergaard HP, et al. (2006)

Disturbed sleep and fatigue in occupational burnout. Scand J Work Environ

Health 32: 121–131.

37. Ekstedt M, Akerstedt T, Soderstrom M (2004) Microarousals during sleep are

associated with increased levels of lipids, cortisol, and blood pressure. Psychosom

Med 66: 925–931.

38. Hu J, Chen Z, Gorczynski CP, Gorczynski LY, Kai Y, et al. (2003) Sleep-

deprived mice show altered cytokine production manifest by perturbations inserum IL-1ra, TNFa, and IL-6 levels. Brain Behav Immun 17: 498–504.

39. Kramer A, Yang F-C, Snodgrass P, Li X, Scammell TE, et al. (2001) Regulation

of daily locomotor activity and sleep by hypothalamic EGF receptor signaling.Science 21: 2511–2515.

40. Teramoto S, Kume H, Yamamoto H, Ishii T, Miyashita A (2003) Effects of oxygenadministration on the circulating vascular endothelial growth factor (VEGF) levels

in patients with obstructive sleep apnea syndrome. Intern Med 42: 681–685.



The Combined Dexamethasone/CRH Test (DEX/CRH Test)and Prediction of Acute Treatment Response in MajorDepressionCornelius Schule*, Thomas C. Baghai, Daniela Eser, Sibylle Hafner, Christoph Born, Sascha Herrmann,

Rainer Rupprecht

Department of Psychiatry and Psychotherapy, Ludwig-Maximilian-University of Munich, Munich, Germany

Abstract

Background: In this study the predictive value of the combined dexamethasone/CRH test (DEX/CRH test) for acuteantidepressant response was investigated.

Methodology/Principal Findings: In 114 depressed inpatients suffering from unipolar or bipolar depression (sample 1) theDEX/CRH test was performed at admission and shortly before discharge. During their stay in the hospital patients receiveddifferent antidepressant treatment regimens. At admission, the rate of nonsuppression (basal cortisol levels .75.3 nmol/l)was 24.6% and was not related to the later therapeutic response. Moreover, 45 out of 114 (39.5%) patients showed anenhancement of HPA axis function at discharge in spite of clinical improvement. In a second sample, 40 depressed patientswere treated either with reboxetine or mirtazapine for 5 weeks. The DEX/CRH test was performed before, after 1 week, andafter 5 weeks of pharmacotherapy. Attenuation of HPA axis activity after 1 week was associated with a more pronouncedalleviation of depressive symptoms after 5-week mirtazapine treatment, whereas downregulation of HPA system activityafter 5 weeks was related to clinical response to reboxetine. However, early improvement of HPA axis dysregulation was notnecessarily followed by a beneficial treatment outcome.

Conclusions/Significance: Taken together, performance of a single DEX/CRH test does not predict the therapeuticresponse. The best predictor for response seems to be an early attenuation of HPA axis activity within 1 or 2 weeks.However, early improvement of HPA system dysfunction is not a sufficient condition for a favourable response. Since asubstantial part of depressive patients display a persistence of HPA axis hyperactivity at discharge, downregulation of HPAsystem function is not a necessary condition for acute clinical improvement either. Our data underline the importance ofHPA axis dysregulation for treatment outcome in major depression, although restoration of HPA system dysfunction seemsto be neither a necessary nor a sufficient determinant for acute treatment response.

Citation: Schule C, Baghai TC, Eser D, Hafner S, Born C, et al. (2009) The Combined Dexamethasone/CRH Test (DEX/CRH Test) and Prediction of Acute TreatmentResponse in Major Depression. PLoS ONE 4(1): e4324. doi:10.1371/journal.pone.0004324


Received August 20, 2008; Accepted October 28, 2008; Published January 29, 2009

Copyright: � 2009 Schule et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The authors have no support or funding to report.



Introduction

According to the corticosteroid-receptor hypothesis of depres-

sion hypothalamic-pituitary-adrenocortical (HPA) system dysreg-

ulation plays an important role in the pathophysiology of

depression [1,2]. In depressed patients, elevated cortisol (COR)

and adrenocorticotrophic hormone (ACTH) concentrations in the

plasma [3–6] or in the cerebrospinal fluid (CSF) [7] have been

found. Additionally, HPA axis hyperactivity is obviously reflected

by elevated urinary free cortisol (UFC) levels, which appear to be

approximately twofold higher in depressed patients as compared

to healthy controls [8]. Further investigations using neuroendo-

crine challenge tests confirmed the hypothesis of a profound HPA

axis dysregulation in depression: Several studies using the

corticotropin-releasing hormone (CRH)-stimulation test reported

a blunted ACTH response whereas the COR stimulation was

indistinguishable from normal controls [9,10]. In contrast to a

reduced ACTH response to CRH, depressive patients show both

an enlargement of the adrenal gland [11,12] and elevated COR

stimulation patterns indicating an enhanced adrenal sensitivity

after challenge with ACTH in most [13–15] but not all [16]

studies. Findings in depressed patients of increased CRH levels in

the CSF [17] and elevated numbers of CRH [18] and arginine-

vasopressin (AVP) [19] expressing neurons in the paraventricular

nucleus of the hypothalamus as well as the observation of reduced

CRH binding sites in the frontal cortex of suicide victims [20] gave

reason to the assumption that depression is characterized by a

hypothalamic overdrive of CRH and/or AVP which in conse-

quence leads to receptor down-regulation in the corticotrophs of

the pituitary gland.

Moreover, it has been suggested that an impaired signalling

pathway via corticosteroid-activated mineralocorticoid and gluco-

corticoid receptors, leading to an impaired negative feedback

regulation of the HPA system, causes this hyperactivity [21]. With


regard to the glucocorticoid receptor, a disturbed negative feedback

control in depressed patients is reflected by COR escape from

dexamethasone suppression [22,23] as well as an increased ACTH

and COR release in the combined dexamethasone-suppression/

CRH-stimulation test (DEX/CRH test) [24,25]. The DEX/CRH

test at present is considered to be the most sensitive tool to

demonstrate a disturbed regulation of the HPA axis in depressed

patients and has been shown to have a sensitivity of more than 80% if

subjects are clustered into different age ranges [25].

A gradual down-regulation of HPA axis hyperactivity in

depressed patients as measured by serial DEX/CRH tests has

been demonstrated for tricyclic antidepressants such as amitrip-

tyline [26], doxepin [27], trimipramine [28–31], for the selective

serotonin-reuptake inhibitor paroxetine [32], for tianeptine which

enhances the presynaptic reuptake of serotonin [32], and for the

selective norepinephrine reuptake inhibitor reboxetine [33].

Proponents of the corticosteroid receptor hypothesis of depression

emphasize that a gradual normalization of HPA system dysreg-

ulation as measured by the DEX/CRH test precedes or coincides

with the response to antidepressant treatment and is a necessary

prerequisite for clinical remission to become manifest, whereas

persisting COR hypersecretion during the DEX/CRH test at

discharge in spite of clinical improvement may be an indicator for

an enhanced risk for relapse within the following six months

[34,35]. In addition, in outpatients with clinically remitted major

depression, higher cortisol levels in the DEX/CRH test are

apparently associated with relapse of major depression [36,37].

Interestingly, persisting nonsuppression in the single dexametha-

sone suppression test (DST) also indicates a higher risk for relapse

within the following months [38–43]. It has further been

postulated that antidepressants may exert their therapeutic effects

at least partly through their actions on the HPA system and that all

antidepressants developed so far may have a uniform dampening

impact on HPA axis function irrespective of their type of action

within monoaminergic systems [1,2,44–46].

The present study aims to answer the following questions:

– What is the nonsuppression rate in the DEX/CRH test in

acutely depressed inpatients within the first week after

admission to a psychiatric hospital before starting antidepres-

sant therapy?

– Is the normalization of HPA axis hyperactivity during

antidepressant therapy (as measured by serial DEX/CRH

tests) a necessary condition for clinical response? Are there

depressed patients who respond to antidepressant therapy

although their HPA axis hyperactivity is not attenuated or even

further increased?

– Is the normalization of HPA axis hyperactivity during

antidepressant treatment a sufficient condition for clinical

response? Are there depressed patients who do not respond to

subsequent antidepressant therapy in spite of early improve-

ment of HPA axis dysregulation?

– Is the DEX/CRH test at admission or its change during

antidepressant treatment suitable for prediction of acute

therapeutic response?


Sample 1114 depressed inpatients (53 men, 61 women) aged between 18

and 74 years (mean age 46.48613.81 years) entered the study after

the procedures had been fully explained and written informed

consent had been obtained. The patients were diagnosed by

experienced and trained psychiatrists according to DSM-IV

criteria [47] using the Structured Clinical Interview for DSM-

IV, German version [48]. Inclusion criteria for the depressed

patients were a) a major depressive episode with melancholic

features, according to DSM-IV criteria (DSM-IV: 296.2, 296.3) or

bipolar depression (DSM-IV: 296.5) b) a sum score of at least 18

on the 17-item version of the Hamilton Depression Rating Scale

(17-HAMD) [49] c) exclusion of major medical disorders;

availability of normal laboratory parameters; normal blood

pressure; normal electrocardiogram; and normal encephalogram

d) exclusion of addiction or other comorbid psychiatric diagnoses

e) no psychotropic drugs for at least 3 days before the first DEX/

CRH test with the exception of zopiclone (up to 7.5 mg per day) in

case of sleep difficulties and lorazepam (up to 2 mg per day) in case

of inner tension and anxiety f) exclusion of pregnancy or use of

oral contraceptives g) no use of oral steroid hormones or hormonal

replacement therapy which may influence the results of the DEX/

CRH test. Further clinical characteristics are given in Table 1.

With regard to the DEX/CRH test procedure, participants

received an oral dose of 1.5 mg dexamethasone at 11:00 PM the

day before stimulation. On the following day, patients had to rest

supine on a bed at 02:00 PM. An intravenous catheter was

inserted into the antecubital vein before 02:15 PM and kept open

with physiological saline solution. Blood samples were collected at

03:00, 03:15, 03:30, 03:45, 04:00, and 04:15 PM. Each sample

was immediately centrifuged and stored at 280u C for COR

measurements. At 03:02 PM 100 mg hCRH (Clinalfa AG,

Laufelfingen, Switzerland) reconstituted in 1 ml 0.02% HCL in

0.9% saline were injected within 30 sec. For determination of

COR serum concentrations, a commercial radioimmunoassay kit

was employed (Cortisol-RIA, DPC Biermann, Germany) with a

sensitivity of 8.27 nmol/l. Our intra- and interassay coefficients of

variation were below 5%. We abstained from reporting the ACTH

levels, since COR has been demonstrated to be the best parameter

for analyzing DEX/CRH test results and since most established

cut-off criterions are related to COR levels but not ACTH

concentrations. For the DEX/CRH test the total COR AUC

values (total area under the concentration time curve), determined

by the trapezoid rule according to Simpson [50], were used for

determination of the COR response to the hCRH challenge in the

dexamethasone pretreated patients representing the combined

effects of altered glucocorticoid receptor (GR) function and

hyperdrive of endogenous CRH and vasopressin.

The first DEX/CRH test was administered within the first week

after admission to the Department of Psychiatry and Psychother-

apy, Ludwig-Maximilian-University of Munich. A wash-out

period of 3 days before neuroendocrine testing was mandatory.

After the first DEX/CRH test, the patients were treated according

to clinical judgement with pharmacological and non-pharmaco-

logical antidepressant treatment options at the discretion of the

doctor in attendance. Within the following 4 weeks after the first

DEX/CRH test, 6 patients were treated with SSRIs, 22 patients

with reboxetine, 12 with mirtazapine, 3 with venlafaxine, 3 with

tricyclic antidepressants, 6 with antidepressant and lithium

augmentation, 2 with antidepressant and anticonvulsant augmen-

tation (not carbamazepine), 29 with pharmacological combination

therapies, 8 received electroconvulsive therapy (ECT), and 23

were treated with 2-week transcranial magnetic stimulation (TMS)

followed by antidepressant pharmacotherapy. Clinical response

was defined by a reduction of at least 50% in the 17-HAMD sum

score after 4 weeks of antidepressant treatment. Remission after

week 4 was assumed if the 17-HAMD sum score was lower than

9 points. In case of nonresponse, either an augmentation/

combination strategy or use of another antidepressant with a

DEX/CRH Test and Prediction


Ta

ble

1.

De

mo

gra

ph

ican

dcl

inic

alp

aram

ete

rsin

11

4in

pat

ien

tssu

ffe

rin

gfr

om

un

ipo

lar

or

bip

ola

rd

ep

ress

ion

wh

ou

nd

erw

en

tth

eD

EX/C

RH

test

atad

mis

sio

nan

dat

dis

char

ge

(sam

ple

1).

all

pa

tie

nts

sup

pre

sso

rsn

on

sup

pre

sso

rsS

tati

stic

sC

OR

pe

ak

ImC

OR

pe

ak

NIm

Sta

tist

ics

(n=

11

4)

test

1(n

=8

6)

test

1(n

=2

8)

sup

pre

ssio

nte

st2

(n=

69

)te

st2

(n=

45

)C

OR

pe

ak

imp

rov

em

en

t

dia

gn

ose

s

MD

,fi

rst

ep

iso

de

35

25

10

26

9

MD

,re

curr

en

t7

25

61

6x

2=

0.5

77

;p

=0

.74

93

93

3x

2=

5.0

26

;p

=0

.13

4

bip

ola

rd

ep

ress

ion

75

24

3

ge

nd

er

(M/F

)5

3/6

14

4/4

29

/19

x2

=3

.07

2;

p=

0.0

80

34

/35

19

/26

x2

=0

.54

5;

p=

0.4

61

age

46

.56

13

.84

6.4

61

3.5

46

.86

15

.0F

=0

.01

4;

p=

0.9

07

47

.56

12

.84

4.9

61

5.3

F=

0.9

90

;p

=0

.32

2

he

igh

t1

71

.06

13

.01

70

.76

14

.31

71

.96

8.0

F=

0.1

78

;p

=0

.67

41

71

.36

8.3

17

0.6

61

8.1

F=

0.0

79

;p

=0

.77

9

BM

I2

8.7

64

.23

0.1

64

.82

4.3

63

.8F

=0

.40

7;

p=

0.5

25

25

.06

4.5

34

.46

66

.6F

=1

.37

8;

p=

0.2

43

age

of

on

set

38

.06

12

.83

8.2

61

3.3

37

.16

11

.2F

=0

.16

6;

p=

0.6

85

40

.16

11

.63

4.7

61

4.0

F=

5.1

21

;p

=0

.02

6

nu

mb

er

of

de

pre

ssiv

ee

pis

od

es

3.4

76

3.7

73

.426

3.8

93

.646

3.4

5F

=0

.07

4;

p=

0.7

86

3.1

96

3.3

33

.916

4.3

7F

=0

.99

9;

p=

0.3

20

du

rati

on

of

inp

atie

nt

stat

us

64

.66

37

.26

6.6

64

0.7

58

.56

22

.8F

=1

.00

8;

p=

0.3

18

58

.46

27

.07

4.2

64

7.8

F=

5.0

46

;p

=0

.02

7

17

-HA

MD

sum

sco

reat

bas

elin

e2

4.4

65

.42

3.8

65

.12

6.2

65

.8F

=4

.41

5;

p=

0.0

38

24

.76

5.5

23

.96

5.2

F=

0.5

20

;p

=0

.47

2

resp

on

se(N

Rs/

Rs)

58

/56

44

/42

14

/14

x2

=0

.01

1;

p=

0.9

15

32

/37

26

/19

x2

=1

.41

7;

p=

0.2

34

rem

issi

on

(NR

m/R

m)

86

/28

64

/22

22

/6x

2=

0.1

97

;p

=0

.65

74

9/2

03

7/8

x2

=1

.84

6;

p=

0.1

74

Me

an6

stan

dar

dd

evi

atio

n(S

D)

isin

dic

ate

d.

Pat

ien

tsar

esu

bd

ivid

ed

into

sup

pre

sso

rs/n

on

sup

pre

sso

rsat

adm

issi

on

and

into

CO

Rp

eak

Im(i

mp

rove

rs)/

CO

Rp

eak

NIm

(no

nim

pro

vers

)at

dis

char

ge

.Su

pp

ress

or:

CO

R1

50

0in

the

DEX

/C

RH

test

1at

adm

issi

on

#7

5.3

nm

ol/

l.C

OR

pe

akIm

(=C

OR

pe

akim

pro

ver)

:Re

du

ctio

no

fth

eC

OR

pe

akva

lue

inth

eD

EX/C

RH

test

be

twe

en

adm

issi

on

and

dis

char

ge

.MD

=M

ajo

rD

ep

ress

ion

.M=

mal

e,F

=fe

mal

e.B

MI=

bo

dy

mas

sin

de

x.1

7-H

AM

D=

Ham

ilto

nD

ep

ress

ion

Rat

ing

Scal

e,

17

-ite

mve

rsio

n.

NR

s=

no

nre

spo

nd

ers

;R

s=

resp

on

de

rs.

NR

m=

no

nre

mit

ters

;R

m=

rem

itte

rs.

Re

spo

nse

and

rem

issi

on

rate

sat

we

ek

4ar

esh

ow

n.

Stat

isti

cs:

resu

lts

ofx

2-t

est

or

on

ew

ayA

NO

VA

are

pro

vid

ed

.Si

gn

ific

ant

resu

lts

are

giv

en

inb

old

lett

ers

.d

oi:1

0.1

37

1/j

ou

rnal

.po

ne

.00

04

32

4.t

00

1



different pharmacological profile was initiated. Patients were

discharged after they had recovered from the depressive episode.

Within the last week before discharge, a second DEX/CRH test

was performed.

The dexamethasone suppression status (suppression versus

nonsuppression) within the DEX/CRH test at admission was

defined by a cut-off criterion of 27.5 ng/ml (,75.3 nmol/l)

applied to the baseline COR level (COR concentration at 03:00

PM, i.e. after administration of 1.5 mg dexamethasone, but

immediately before CRH-challenge) which was derived from a

normative database from the Max-Planck-Institute of Psychiatry in

Munich after correction of a linear bias (Heuser criterion) [25,51].

Moreover, a further criterion was employed which has been

proposed by a Japanese research group [52,53] defining subjects as

nonsuppressors (baseline COR$50 ng/ml [,137.95 nmol/l]),

intermediate suppressors (baseline COR,50 ng/ml and peak

COR$50 ng/ml), and suppressors (peak COR,50 ng/ml) (Ku-

nugi criterion). HPA axis activity at the time of the DEX/CRH

test at discharge was categorized in improvers and nonimprovers

according to the change in the peak COR level after CRH

challenge between DEX/CRH test 1 (admission) and test 2

(discharge). A COR peak improver was defined by a lower COR

peak concentration during test 2; otherwise, a COR peak

nonimprover was presumed. The peak COR level was used for

the categorization into HPA system improvers and nonimprovers

instead of the COR AUC value to be in line with previous

definitions of HPA system improvement in remitted depression

[34,35,54].

Sample 2Data of the second patient sample have been already reported

in a previous publication of our research group [33]. Clinical and

demographic characteristics of sample 2 are provided in Table 2.

This sample was now re-analyzed with respect to the predictive

value of COR peak improvement during serial DEX/CRH tests

for the therapeutic response. 40 drug-free patients suffering from a

major depressive episode (DSM-IV criteria) were treated with

either reboxetine (8 mg/day; n = 20) or mirtazapine (45 mg/day;

n = 20) monotherapy for 5 weeks. Before, after 1 week and after

5 weeks of therapy, the dexamethasone/CRH-test was performed

as described above and COR concentrations were measured.

COR peak week 1 improvement was defined as lowering of the

COR peak value between DEX/CRH test 1 (week 0 before

treatment) and test 2 (after 1 week of treatment with either

reboxetine or mirtazapine). COR peak week 5 improvement was

established as a reduction of the COR peak level between test 1

(week 0) and test 3 (week 5). Likewise, COR basal week 1 or week

5 improvement was defined as lowering of the basal COR value at

03:00 PM (after administration of 1.5 mg DEX, but immediately

before hCRH injection) between test 1 (week 0) and test 2 (week 1)

or between test 1 (week 0) and test3 (week 5), respectively. In this

sample, response was defined by a reduction of at least 50% in the

21-HAMD sum score after five weeks of treatment with either

reboxetine or mirtazapine.

StatisticsDemographic and clinical parameters were compared between

suppressors and nonsuppressors (Heuser criterion) or between

COR peak improvers and nonimprovers by the Pearson Chi-

Square test for contingency tables or by Fisher’s exact test with

respect to qualitative variables and by one-way ANOVA for

independent samples with regard to quantitative variables.

Correlations between quantitative variables and endocrinological

parameters were calculated using the rank order coefficient

(Spearman’s rho) since hormonal data were not normally

distributed. Moreover, in sample 2 the baseline-corrected decrease

in 21-HAMD sum scores during 5-week treatment was compared

between COR peak week 1 improvers and nonimprovers and

between COR peak week 5 improvers and nonimprovers using

ANOVAs for repeated measurements. Thereby ‘‘time’’ (week 0–5)

and ‘‘group’’ (improvers vs nonimprovers) were considered as

within-subjects and between-subjects factors with six (‘‘time’’) and

two (‘‘group’’) levels, respectively. Post-hoc tests with contrasts

were additionally performed when ‘‘group’’ was among the

significant influential factors. For the ANOVA procedures, a

correction was applied to the F-value by means of adjusting the

Table 2. Demographic and clinical parameters in 40 inpatients suffering from unipolar depression treated with either reboxetine(n = 20; 8 mg/day) or mirtazapine (n = 20; 45 mg/day) for 5 weeks (sample 1) [33].

all patients COR peak week 1 Statistics COR peak week 5 Statistics

(n = 40) Im (n = 30) NIm (n = 10) COR peak week 1 Im (n = 24) NIm (n = 16) COR peak week 5

diagnoses

MD, first episode 10 9 1 x2 = 1.600; p = 0.206 7 8 x2 = 0.556; p = 0.456

MD, recurrent 30 21 9 17 13

gender (M/F) 17/23 14/16 3/7 x2 = 0.853; p = 0.356 9/15 8/8 x2 = 0.614; p = 0.433

age 47.9614.6 49.7614.3 42.5615.1 F = 1.876; p = 0.179 49.6614.2 45.4615.3 F = 0.767; p = 0.387

height [cm] 170.169.2 169.668.3 171.8611.9 F = 0.432; p = 0.515 169.268.7 171.4610.2 F = 0.531; p = 0.470

BMI 25.064.1 24.964.2 25.364.2 F = 0.063; p = 0.803 25.164.2 24.864.2 F = 0.076; p = 0.784

age of onset 40.0615.0 40.5614.9 38.3615.9 F = 0.158; p = 0.693 39.5614.3 40.7616.4 F = 0.063; p = 0.803

number of depressive episodes 2.5861.74 2.6361.92 2.4061.08 F = 0.132; p = 0.718 2.5061.93 2.6961.45 F = 0.109; p = 0.743

duration of inpatient status 67.5638.7 64.0635.3 78.2648.1 F = 1.013; p = 0.321 61.8632.1 76.1646.8 F = 1.306; p = 0.260

21-HAMD sum score at baseline 24.363.9 24.763.9 22.963.8 F = 1.594; p = 0.214 25.263.1 22.864.7 F = 3.818; p = 0.058

Mean6standard deviation (SD) is indicated. Patients are subdivided into COR week 1 Im (improvers)/COR peak NIm (nonimprovers) and into COR week 5 Im/NIm. CORpeak week 1/week 5 Im ( = COR peak week 1/week 5 improver): patient with reduction of COR peak value in the DEX/CRH test after 1 week/5 weeks of treatment, ascompared to baseline (week 0). MD = Major Depression. M = male, F = female. BMI = body mass index. 21-HAMD = Hamilton Depression Rating Scale, 21-item version.Statistics: results of x2-test or oneway ANOVA are provided.doi:10.1371/journal.pone.0004324.t002



degrees of freedom by a factor Epsilon, if the sphericity test

(Mauchly W test) was significant indicating a heterogeneity of

covariances (Huyn-Feldt correction). In addition, Cramer’s Phi

was calculated in sample 2 for all patients and also separately in

the reboxetine and the mirtazapine group in order to investigate

putative associations between COR week 1 improvement/COR

week 5 improvement and the clinical outcome after 5 weeks of

treatment (response, remission).

As a nominal level of significance, alpha = 0.05 was accepted.

The software program SPSS version 15.0 (SPSS Inc., Chicago,

Illinois, USA) was used for data analysis.

The study was carried out according to the Declaration of

Helsinki (http://www.wma.net) and had been approved by a local

ethics committee (intramural review panel of the Ludwig-

Maximilian-University of Munich, Faculty of Medicine).

Results

Using a cut-off criterion of 75.3 nmol/l (COR level at 03:00

PM) for the definition of nonsuppression in the DEX/CRH test 1

at admission (Heuser criterion), in sample 1 only 28 out of 114

(,24.6%) depressed inpatients were nonsuppressors whereas 86

(,75.4%) acutely depressed patients were suppressors already

before the beginning of antidepressant therapy (Figure 1A).

Moreover, when also using the Kunugi criterion the suppressors

(n = 74, i.e. 64.9%) were predominant as compared to nonsup-

pressors (n = 19 [,16.7%]) or intermediate suppressors (n = 21

[,18.4%]) (Figure 1B). With regard to the DEX/CRH test 1

suppressors and nonsuppressors (categorized by the Heuser

criterion) did not differ in qualitative variables such as diagnoses,

gender distribution, response or remission rates (p.0.05 in x2-

tests, respectively) (Table 1). Suppressors and nonsuppressors

were also comparable in quantitative parameters such as age,

height, BMI, age of onset, number of depressive episodes, and

duration of inpatient status (p.0.05 in one-way ANOVA,

respectively). However, nonsuppressors were prone to have higher

17-HAMD sum scores at baseline which was statistically

significant (p,0.05) (Table 1). In addition, with respect to test 1

at admission severity of depressive symptoms (17-HAMD sum

scores) was positively correlated with COR AUC values (Spear-

man’s Rho = 0.238, p = 0.011).

The overall group of depressed inpatients (n = 114; sample 1)

showed a significant decrease in COR AUC values during the

DEX/CRH tests between admission and discharge (Figure 2).

However, when the sample was subdivided in COR peak

improvers (COR peak value test 1.COR peak value test 2;

n = 69) and in COR peak nonimprovers (COR peak value test

1#COR peak value test 2; n = 45), COR peak improvers

displayed a marked reduction in COR AUC values during

inpatient treatment whereas COR peak nonimprovers were

characterized by a pronounced increase in COR AUC values in

spite of clinical recovery and discharge (Figure 2). The same

finding was observed if the patients were classified in patients

receiving psychopharmacological drugs (n = 83) and patients

treated with non-pharmacological treatment strategies such as

TMS or ECT (n = 31). In the psychopharmacotherapy group,

there were 52 COR peak improvers (mean COR AUC at

admission: 9444.2568606.81 nmol/l6min; mean COR AUC at

discharge: 4599.7564734.71 nmol/l6min) and 31 nonimprovers

(mean COR AUC at admission: 4510.4366065.80 nmol/l6min;

mean COR AUC at discharge: 7790.3967825.23 nmol/l6min).

The non-pharmacological treatment group consisted of 17 COR

peak improvers (mean COR AUC at admission:

12,531.01610,626.87 nmol/l6min; mean COR AUC at dis-

charge: 6442.8268027.38 nmol/l6min) and 14 nonimprovers

(mean COR AUC at admission: 6286.0967256.49 nmol/l6min;

mean COR AUC at discharge: 10,669.75610,608.55 nmol/

l6min). Considering the total sample (n = 114), COR peak

improvers and nonimprovers were comparable with respect to

diagnoses, gender distribution, response and remission rates (x2-

tests: p.0.05, respectively) (Table 1). Moreover, there were no

significant differences between COR peak improvers and nonim-

provers regarding age, height, BMI, number of depressive episodes

or 17-HAMD sum score at baseline (oneway ANOVA: p.0.05,

respectively). However, in COR peak nonimprovers a significantly

earlier age of onset of the depressive illness and a significantly

longer duration of time between admission and discharge

(inpatient status) were found (p,0.05, respectively) (Table 1).

Demographic and clinical characteristics of sample 2 (40

depressed inpatients, treated with either reboxetine or mirtazapine

Figure 1. DEX/CRH test at admission in 114 depressedinpatients. (A) Subdivision into suppressors (n = 86) and nonsup-pressors (n = 28) according to the Heuser criterion (nonsuppression:baseline COR$75.3 nmol/l) as indicated by the cross bar. (B)Subdivision into nonsuppressors (NS; baseline COR$50 ng/ml[,137.95 nmol/l]; n = 19), intermediate suppressors (IS; baselineCOR,50 ng/ml and peak COR$50 ng/ml; n = 21), and suppressors (S;peak COR,50 ng/ml; n = 74) according to the Kunugi criterion.Baseline COR = COR at 03:00 PM. Mean+/2standard error of mean (SEM)is given.doi:10.1371/journal.pone.0004324.g001



for 5 weeks) are given in Table 2. There were no significant

differences between COR week 1/week 5 improvers and

nonimmprovers with regard to diagnoses, gender distribution,

age, height, body mass index, age of onset, number of episodes,

duration of total inpatient status, or severity of depression at

baseline. COR peak week 1 improvement (reduction of the COR

peak value in the DEX/CRH test after 1 week of treatment) was

associated with alleviation of depressive symptoms. Regarding

COR peak week 1 improvement in all patients (n = 40), repeated-

measures ANOVA revealed a highly significant ‘‘time’’ effect, i.e.

decrease in 21-HAMD sum scores (F = 50.173; d.f. = 2.642,

100.401; p,0.001). Moreover, a significant ‘‘group’’ effect was

observed (F = 4.638; d.f. = 1, 38; p = 0.038) indicating a more

pronounced amelioration of depressive symptoms in COR peak

week 1 improvers than in nonimprovers (Figure 3). Post-hoc tests

demonstrated significant differences between COR peak week 1

improvers and nonimprovers at week 1 and week 3 (p,0.05,

respectively). No associations were found between COR peak

week 1 improvement and response or remission rates when

regarding all patients (Cramer’s Phi: p.0.05, respectively;

Table 3). It is also worth to be mentioned that 8 COR peak

week 1 improvers were nonresponders at week 5, i.e. improvement

of COR peak values after 1 week was not a guarantee (sufficient

condition) for clinical response after 5 weeks. When analyzing

separately depressed patients treated with reboxetine (n = 20) or

mirtazapine (n = 20), significant ‘‘group’’ effects in the repeated-

measures ANOVAs were obtained in the mirtazapine group

(F = 5.738; d.f. = 1, 18; p = 0.028), but not in the reboxetine group

(F = 1.410; d.f. = 1,18; p = 0.250) indicating better alleviation of

depressive symptomatology in COR peak week 1 improvers

treated with mirtazapine than in nonimprovers receiving this

antidepressant. Moreover, in the mirtazapine group (Phi = 0.572;

p = 0.010), but not in the reboxetine group (Phi = 0.121; p = 0.589)

a significant association between COR peak week 1 improvement

and response rate after 5 weeks of treatment was demonstrated

(Table 3). Similar results were obtained if COR basal

improvement after 1 week of treatment was used instead.

Figure 3. Analysis of COR week 1 improvers and nonimprovers.Mean value graphs of the decrease in 21-HAMD sum scores in 40depressed patients treated with either reboxetine (n = 20) or mirtaza-pine (n = 20) for 5 weeks, subdivided into COR week 1 improvers andnonimprovers. COR week 1 improver = patient with reduction of CORpeak value in the DEX/CRH test after 1 week of treatment, as comparedto baseline (week 0). Mean+/2standard error of mean (SEM) is given.Significant group effects in the ANOVA for repeated measurementsindicated. * = significant group differences in post-hoc test (p,0.05).** = highly significant group differences in post-hoc test (p,0.01).doi:10.1371/journal.pone.0004324.g003

Figure 2. COR AUC values of DEX/CRH tests in 114 depressedinpatients at admission and at discharge. All patients (n = 114),and subgroups of COR improvers (n = 69) and of nonimprovers (n = 45)are shown. Mean+standard error of mean (SEM) is given. COR improver:reduction of COR peak value in the DEX/CRH test between admissionand discharge. ** = highly significant differences in COR AUC valuesbetween admission and discharge (p,0.01).doi:10.1371/journal.pone.0004324.g002



Reduction of the basal COR value between test 1 and test 2 was

significantly associated with better clinical response when

regarding all patients, but also in separate analyses of the

reboxetine and the mirtazapine groups (Table 4).

Analyzing putative associations between the attenuation of

COR peak values at week 5 and antidepressant efficacy in the

whole sample (n = 40), there was no significant ‘‘group’’ effect in

the repeated-measures ANOVA between COR peak week 5

improvers and nonimprovers (F = 0.358; d.f. = 1, 38; p = 0.553)

(Figure 4). In addition, when analyzing separately for reboxetine

and mirtazapine, no significant group effects were seen in the

repeated-measures ANOVA between COR peak week 5 improv-

ers and nonimprovers either (reboxetine: F = 2.314; d.f. = 1, 18;

p = 0.146; mirtazapine: F = 0.692; d.f. = 1, 18; p = 0.416). How-

ever, when using Cramer’s Phi as effect size parameter, relevant

associations between COR week 5 improvement and clinical

outcome (response rate, remission rate) could be shown for the

reboxetine group which were nearly significant (response rate:

Phi = 0.435; p = 0.052; remission rate: Phi = 0.546; p = 0.051),

whereas no such association were seen in the mirtazapine group

(Table 3).

Using COR basal improvement after 5 weeks of treatment as

parameter for changes in HPA axis activity, the analysis revealed a

significant association between COR basal week 5 improvement

and response/remission in the reboxetine group, but not in the

mirtazapine group (Table 4).

Discussion

One of our main results is the finding that in acutely depressed

inpatients the nonsuppression rate in the DEX/CRH test at

admission was 24.6% (28 out of 114) according to the Heuser

criterion which focuses on the 1.5 mg dexamethasone suppression

status and does not consider the CRH-stimulated COR concen-

trations [25]. When using the Kunugi criterion which also involves

the COR levels after CRH challenge [52,53], the rates of

nonsuppression (16.7%) or intermediate suppression (18.4%) were

somewhat higher if added together (35.1%). However, in any case

the proportion of acutely and severely depressed inpatients who

were identified by nonsuppression in the DEX/CRH test was

much lower in our study than originally expected [25]. In a large

meta-analysis [55] including more than 5,000 depressed patients, a

sensitivity of the single dexamethasone suppression test (DST) of

44% was found using the ‘‘Carroll criterion’’ [56] (nonsuppression

in the DST: COR level .5 mg/dl the day after oral administration

of 1 mg dexamethasone). In the original study of the Max-Planck-

Institute of Psychiatry in Munich, introducing the combined

DEX/CRH test in the literature, it was reported that the

sensitivity of the DEX/CRH test in depression is about 80 to

90% if the control subjects are matched for age and gender and

thus greatly exceeds the sensitivity of the standard DST.

Moreover, a dichotomous cut-off criterion of 40 ng/ml

(110 nmol/l) for the baseline COR concentration in the DEX/

Table 3. COR peak week 1 improvement and COR peak week 5 improvement.

response statistics remission statistics

(NRs/Rs) Cramer’s phi p-value (NRm/Rm) Cramer’s phi p-value

all patients (n = 40) 14/26 22/18

COR peak week 1 Im (n = 30) 8/22 15/15

Phi = 0.303 p = 0.056 Phi = 0.174 p = 0.271

COR peak week 1 NIm (n = 10) 6/4 7/3

COR peak week 5 Im (n = 24) 7/17 11/13

Phi = 0.150 p = 0.343 Phi = 0.226 p = 0.154


reboxetine group (n = 20) 7/13 10/10

COR peak week 1 Im (n = 13) 4/9 6/7

Phi = 0.121 p = 0.589 Phi = 0.105 p = 0.639


COR peak week 5 Im (n = 14) 3/11 5/9

Phi = 0.435 p = 0.052 Phi = 0.436 p = 0.051


mirtazapine group (n = 20) 7/13 12/8

COR peak week 1 Im (n = 17) 4/13 9/8

Phi = 0.572 p = 0.010 Phi = 0.343 p = 0.125


COR peak week 5 Im (n = 10) 4/6 6/4

Phi = 0.105 p = 0.639 Phi = 0.000 p = 1.000


Response and remission rates after 5 weeks of treatment in 40 depressed patients (sample 2) treated with either reboxetine (n = 20) or mirtazapine (n = 20) [33]. Patientsare subdivided into COR peak week 1 Im (improvers)/NIm (nonimprovers) and COR peak week 5 Im (improvers)/NIm (nonimprovers). COR peak week 1/week 5 Im( = COR peak week 1/week 5 improver): patient with reduction of COR peak value in the DEX/CRH test after 1 week/5 weeks of treatment, as compared to baseline(week 0). NRs = nonresponders; Rs = responders. NRm = nonremitters; Rm = remitters. Statistics: Cramer’s phi as measure of association for the chi-square test isprovided. Significant results (p,0.05) or trends for significance (p,0.10) are given in bold letters.doi:10.1371/journal.pone.0004324.t003



CRH test (1.5 mg dexamethasone suppression status) was

proposed to differentiate between suppressors and nonsuppressors

[25,57]. However, in a cross-laboratory validation of the Max-

Planck-Institute it was discovered that the RIA analysis of plasma

COR which had been performed produced concentration values

with a linear bias [35]. Since the biased COR levels were elevated

by a factor of 1.46 in comparison with those at other laboratories,

the cut-off criterion was corrected accordingly and was now

proposed as 27.5 ng/ml (75.3 nmol/l) [32,51].

Even if the corrected cut-off criterion (75.3 nmol/l) is used, in

more recent studies the sensitivity (nonsuppression rate) in acutely

depressed inpatients using the DEX/CRH test is surprisingly low,

possibly lower than the sensitivity of the standard DST (44%), and

amounts to 31.6% (12 out of 38 [32]), 16.6% (35 out of 211 [51]),

or 24.6% (28 out of 114; present study). Moreover, when using the

Kunugi criterion (a) nonsuppression: baseline COR$50 ng/ml

[,137.95 nmol/l]) b) intermediate suppression: baseline

COR,50 ng/ml and peak COR$50 ng/ml c) suppression: peak

COR,50 ng/ml) which considers also the CRH effects within the

combined DEX/CRH test, the rates of nonsuppression or

intermediate suppression in this test (35.1% in our sample) is

lower than expected. In the original investigation of Heuser and

coworkers [25] a sensitivity of 80 to 90% in the DEX/CRH test

was only reached if depressed patients and control subjects were

clustered into different age ranges and highly sophisticated

statistical methods such as multivariate analysis of variance or

discriminant analysis were used. However, these analyses are not

practicable in the clinical situation which requires clear dichoto-

mous variables to differentiate between suppression and non-

suppression. No study has been performed so far confirming the

originally reported high sensitivity of the DEX/CRH test of more

than 80% by using a criterion which is applicable under clinical

conditions. Apparently an ideal cut-off criterion has not been

established yet for the DEX/CRH test.

Nevertheless, a considerable part of acutely depressed patients

shows normally regulated HPA axis activity in the DEX/CRH test

already before antidepressant treatment and may though benefit

from this therapy. In fact the severity of depression was

significantly higher in baseline nonsuppressors than in suppressors

in our study and a significant positive correlation between baseline

21-HAMD sum score and COR AUC values (test 1) could be

demonstrated in our investigation as it has been reported in

previous studies [35,53,58,59]. However, the response and

remission rates in nonsuppressors and suppressors on test 1 were

comparable. Therefore, a single DEX/CRH test performed

within the first week after admission is obviously not suitable for

prediction of the acute treatment response.

The best predictor for acute antidepressant efficacy seems to be

the responsiveness of the HPA system and the change of DEX/

CRH test results within the first one or two weeks of

antidepressant treatment. In our investigation, attenuation of

HPA axis activity (reduction of COR basal value, reduction of

COR peak value) in the whole patient sample after one week of

pharmacotherapy was significantly associated with the subsequent

Table 4. COR basal week 1 improvement and COR basal week 5 improvement.

response statistics remission statistics

(NRs/Rs) Cramer’s phi p-value (NRm/Rm) Cramer’s phi p-value

all patients (n = 40) 14/26 22/18

COR basal week 1 Im (n = 28) 5/23 12/16

Phi = 0.549 p = 0.001 Phi = 0.373 p = 0.018

COR basal week 1 NIm (n = 12) 9/3 10/2


Phi = 0.257 p = 0.104 Phi = 0.226 p = 0.154


reboxetine group (n = 20) 7/13 10/10


Phi = 0.560 p = 0.012 Phi = 0.314 p = 0.160



Phi = 0.663 p = 0.003 Phi = 0.436 p = 0.051


mirtazapine group (n = 20) 7/13 12/8


Phi = 0.545 p = 0.015 Phi = 0.471 p = 0.035



Phi = 0.105 p = 0.639 Phi = 0.000 p = 1.000


Response and remission rates after 5 weeks of treatment in 40 depressed patients (sample 2) treated with either reboxetine (n = 20) or mirtazapine (n = 20) [33]. Patientsare subdivided into COR basal week 1 Im (improvers)/NIm (nonimprovers) and COR basal week 5 Im (improvers)/NIm (nonimprovers). COR basal week 1/week 5 Im( = COR basal week 1/week 5 improver): patient with reduction of COR basal value in the DEX/CRH test after 1 week/5 weeks of treatment, as compared to baseline(week 0). NRs = nonresponders; Rs = responders. NRm = nonremitters; Rm = remitters. Statistics: Cramer’s phi as measure of association for the chi-square test isprovided. Significant results (p,0.05) or trends for significance (p,0.10) are given in bold letters.doi:10.1371/journal.pone.0004324.t004



alleviation of depressive symptoms. This is in line with studies

reported by Ising and colleagues who found improved HPA system

regulation in a second DEX/CRH test (performed 2 or 3 weeks

after the first test at the beginning of the study) to be associated

with beneficial treatment response after 5 weeks [51,54]. Thus,

performance of two DEX/CRH tests in acutely depressed patients

with an interval of 1 up to 3 weeks seems to be a potential

biomarker with certain significance for the subsequent therapeutic

response.

Since improvement of basal COR, representing a single

dexamethasone suppression test (DST), was at least as powerful

in prediction of therapeutic response as improvement of the COR

peak value (as part of the combined DEX suppression/CRH

stimulation test), one may assume that performance of 2

subsequent dexamethasone suppression tests (DST) may be a

feasible and appropriate predictor of clinical outcome. One single

DST prior to treatment does not reliably predict response to

antidepressant therapy [60]. However, reduction of COR in 2

subsequent DST is of predictive value (our data). As it has been

shown by Carroll and colleagues 4 PM and 11 PM samples are

much better to detect COR nonsuppression (as defined by the

5 mg/dl-criterion in the standard DST) than 8 AM samples [61].

Furthermore, the same research group could demonstrate that

with the 1-mg DEX dose the sensitivity of the DST greatly exceeds

that of the 2-mg DST [61]. Thus, performance of 2 subsequent 1-

mg DST using 4 PM or 11 PM samples [61] or two subsequent

1.5 mg DST using 3 PM samples (our study) may be an easy and

appropriate way to predict therapeutic response. Our results are

confirmed by former studies performing serial DST and suggesting

that downregulation of the HPA system activity as measured by

the DST precedes or coincides with the amelioration of depressive

symptoms [42,62–65].

However, there seem to be differences between antidepressant

drugs which are related to their distinct biochemical properties.

Reboxetine is a norepinephrine reuptake inhibitor which acutely

stimulates COR secretion [66] whereas mirtazapine does not

cause reuptake inhibition but is an antagonist at a2-, 5-HT2-, 5-

HT3-, and histamine H1 receptors and acutely inhibits COR

secretion [67,68]. Apparently, early change of HPA axis activity

(week 1) induced by mirtazapine is related to clinical outcome after

5 weeks whereas 5-week response to reboxetine is associated with

late change in DEX/CRH test results at week 5 (Table 2,

Figure 3, Figure 4). Reuptake inhibiting antidepressants such as

reboxetine, selective serotonin reuptake inhibitors (SSRIs) or

tricyclic antidepressants acutely stimulate cortisol and ACTH

secretion both in healthy subjects [69] and in depressed patients

[70,71] after single administration and may gradually normalize

HPA axis hyperactivity in depressed patients if they are given daily

for several weeks via up-regulation of mineralocorticoid receptor

and glucocorticoid receptor mRNA levels [44,46,72,73], down-

regulation of pro-opiomelanocortin mRNA expression in the

pituitary gland [74], and decrease of CRH gene expression and

CRH mRNA synthesis in the paraventricular nucleus [75,76]

thereby enhancing mineralocorticoid receptor and glucocorticoid

receptor function and restoring the disturbed feedback control. On

the contrary, mirtazapine rapidly reduces HPA axis hyperactivity

in depressed patients within one week which can be explained

most likely by direct pharmacoendocrinological effects of mirta-

zapine such as antagonism at central 5-HT2- and H1-receptors

thereby inhibiting the hypothalamic CRH release. After 5 weeks

of mirtazapine therapy in depressed patients, there is a partial

‘‘rebound’’ phenomenon which can probably be explained by a

compensatory up-regulation of CRH receptors at the pituitary

gland during several weeks of mirtazapine treatment leading to a

Figure 4. Analysis of COR week 5 improvers and nonimprovers.Mean value graphs of the decrease in 21-HAMD sum scores in 40depressed patients treated with either reboxetine (n = 20) or mirtaza-pine (n = 20) for 5 weeks, subdivided into COR week 5 improvers andnonimprovers. COR week 5 improver = patient with reduction of CORpeak value in the DEX/CRH test after 5 weeks of treatment, ascompared to baseline (week 0). Mean+/2standard error of mean(SEM) is given.doi:10.1371/journal.pone.0004324.g004



partial re-enhancement of cortisol and ACTH output after

exogenous administration of 100 mg hCRH during the DEX/

CRH test [33]. Obviously, these different effects on the time

course of HPA axis activity in depressed patients during reboxetine

or mirtazapine treatment are also reflected by diverse associations

with clinical outcome (response related to early changes in HPA

axis activity during mirtazapine treatment and to late changes in

HPA system during reboxetine therapy).

However, two limiting issues have to be pointed out in this

context: First, an early improvement of HPA axis hyperactivity

(e.g. within 1 week of treatment) is not necessarily followed by a

favourable response and therefore is not a sufficient condition for a

beneficial treatment outcome. In the present study, 8 out of 40

depressed patients were classified as COR peak week 1 improvers

but were nonresponders after 5-week treatment with either

reboxetine or mirtazapine. Moreover, in a former study of our

research group, mirtazapine effectively reduced the overshoot of

COR during the DEX/CRH test within 1 week of treatment in 40

depressed inpatients, but this attenuation of HPA axis activity

occurred both in 5-week responders and nonresponders and was

not related to clinical improvement [77]. Therefore, the

importance of an early improvement of HPA axis dysregulation

for the prediction of the acute antidepressant response is limited.

Second, the association between clinical response to the norepi-

nephrine reuptake inhibitor reboxetine and late changes in HPA

system activity (week 5) in our investigation is not confirmed by

other clinical trials investigating the impact of reuptake inhibiting

antidepressants on HPA axis function in depression, since the

decrease in COR levels during serial DEX/CRH tests after 4 to

6 weeks of pharmacotherapy has been found to be comparable in

responders and nonresponders in these studies [26–33].

Moreover, it is remarkable in our study that a considerable

proportion of depressed inpatients (39.5%, i.e. 45 out of 114)

showed a pronounced enhancement of HPA axis activity shortly

before discharge in spite of clinical recovery. Our finding is

supported by other researchers who also found an enhanced HPA

system activity at discharge in a notable part of depressed patients

[34,35], e.g. in 21 out of 74 (28.4%) investigated patients [35]. It is

important to note that in our study HPA system nonimprovers at

discharge were prone to have an earlier age of onset of the

depressive illness and a longer duration of the inpatient stay as

compared to improvers. Furthermore it is known that HPA axis

activity at discharge in spite of clinical improvement is associated

with a higher risk for relapse of depression with regard to medium-

term or long-term outcome [34–37], which has not been

investigated in the present study. Nevertheless attenuation of

HPA axis activity during antidepressant therapy is obviously not a

necessary condition for acute clinical recovery.

Taken together, it can be concluded from our data that the

sensitivity (rate of nonsuppression) of the combined DEX/CRH

test in acutely depressed patients is much lower than originally

reported. Moreover, the performance of a single DEX/CRH test

shortly after admission does not predict the therapeutic response.

The best predictor for response seems to be the early

responsiveness and downregulation of HPA axis activity within

the first 1 or 2 weeks of antidepressant treatment as measured by 2

subsequent DEX/CRH tests. Possibly, the performance of 2

subsequent standard DST may be of comparable predictive value

and can be offered to depressed patients more easily in the clinical

situation. However, the significance of these potential biomarkers

is limited since early improvement of HPA axis dysregulation is

not necessarily followed by a favourable therapeutic response and

is therefore not a sufficient condition for a beneficial treatment

outcome. After 4–6 weeks of antidepressant treatment, the

attenuation of HPA axis activity is comparable in responders

and nonresponders in most studies although an association

between COR week 5 improvement and clinical response to

reboxetine could be demonstrated in the present investigation. At

discharge, a substantial part of depressive patients show even an

enhancement of HPA axis activity in spite of clinical recovery.

Thus, downregulation of HPA system function is not a necessary

condition for clinical improvement. However, patients with

persistence of HPA axis hyperactivity at discharge are known to

have a higher risk for relapse during the following 6 months. Our

data underline the importance of HPA axis dysregulation for

treatment outcome in major depression, although restoration of

HPA system dysfunction seems to be neither a necessary nor a

sufficient determinant for acute treatment response.

Acknowledgments

The study was done in the framework of the doctoral thesis of Mr. Sascha

Herrmann which will be submitted to the Faculty of Medicine, University

of Munich.


Conceived and designed the experiments: CS TCB DE RR. Performed the

experiments: CS TCB DE SH CB SH. Analyzed the data: CS RR.

Contributed reagents/materials/analysis tools: CS. Wrote the paper: CS

TCB DE SH CB SH RR.

References

1. Holsboer F (2000) The corticosteroid receptor hypothesis of depression.

Neuropsychopharmacology 23: 477–501.2. Holsboer F (2001) Stress, hypercortisolism and corticosteroid receptors in

depression: implications for therapy. J Affect Disord 62: 77–91.

3. Halbreich U, Asnis GM, Shindledecker R, Zumoff B, Nathan RS (1985) Cortisolsecretion in endogenous depression. I. Basal plasma levels. Arch Gen Psychiatry

42: 904–908.4. Linkowski P, Mendlewicz J, Leclercq R, Brasseur M, Hubain P, et al. (1985) The

24-hour profile of adrenocorticotropin and cortisol in major depressive illness.

J Clin Endocrinol Metab 61: 429–438.5. Board F, Wadeson R, Persky H (1957) Depressive affect and endocrine

functions: Blood levels of adrenal cortex and thyroid hormones in patientssuffering from depressive reactions. Arch Neurol Psychiatry 78: 612–620.

6. Gibbons JL, McHugh PR (1963) Plasma cortisol in depressive illness. PsychiatryRes 1: 162–171.

7. Traskman L, Tybring M, Asberg L, Bertilsson O, Lantto O, et al. (1980) Cortisol

in the CSF of depressed and suicidal patients. Arch Gen Psychiatry 37: 761–767.8. Carroll BJ, Curtis GC, Davies BM, Mendels J, Sugerman AA (1976) Urinary

free cortisol excretion in depression. Psychol Med 6: 43–50.9. Holsboer F, Von Bardeleben U, Gerken A, Stalla GK, Muller OA (1984)

Blunted corticotropin and normal cortisol response to human corticotropin-

releasing factor in depression. N Engl J Med 311: 1127.

10. Gold PW, Chrousos G, Kellner C, Post R, Roy A, et al. (1984) Psychiatric

implications of basic and clinical studies with corticotropin-releasing factor.Am J Psychiatry 141: 619–627.

11. Nemeroff CB, Krishnan KR, Reed D, Leder R, Beam C, et al. (1992) Adrenal

gland enlargement in major depression. A computed tomographic study. ArchGen Psychiatry 49: 384–387.

12. Rubin RT, Phillips JJ, Sadow TF, McCracken JT (1995) Adrenal gland volumein major depression. Increase during the depressive episode and decrease with

successful treatment. Arch Gen Psychiatry 52: 213–218.

13. Amsterdam JD, Winokur A, Lucki I, Snyder P (1983) Neuroendocrineregulation in depressed postmenopausal women and healthy subjects. Acta

Psychiatr Scand 67: 43–49.14. Gerken A, Holsboer F (1986) Cortisol and corticosterone response after syn-

corticotropin in relationship to dexamethasone suppressibility of cortisol.Psychoneuroendocrinology 11: 185–194.

15. Jaeckle RS, Kathol RG, Lopez JF, Meller WH, Krummel SJ (1987) Enhanced

adrenal sensitivity to exogenous cosyntropin (ACTH alpha 1–24) stimulation inmajor depression. Relationship to dexamethasone suppression test results. Arch

Gen Psychiatry 44: 233–240.16. Rubin RT, Miller TH, Rhodes ME, Czambel RK (2006) Adrenal cortical

responses to low- and high-dose ACTH(1–24) administration in major

depressives vs. matched controls. Psychiatry Res 143: 43–50.



17. Nemeroff CB, Widerlov E, Bissette G, Walleus H, Karlsson I, et al. (1984)

Elevated concentrations of CSF corticotropin-releasing factor-like immunoreac-

tivity in depressed patients. Science 226: 1342–1344.

18. Raadsheer FC, Hoogendijk WJ, Stam FC, Tilders FJ, Swaab DF (1994)

Increased numbers of corticotropin-releasing hormone expressing neurons in the

hypothalamic paraventricular nucleus of depressed patients. Neuroendocrinol-

ogy 60: 436–444.

19. Purba JS, Hoogendijk WJ, Hofman MA, Swaab DF (1996) Increased number of

vasopressin- and oxytocin-expressing neurons in the paraventricular nucleus of

the hypothalamus in depression. Arch Gen Psychiatry 53: 137–143.

20. Nemeroff CB, Owens MJ, Bissette G, Andorn AC, Stanley M (1988) Reduced

corticotropin releasing factor binding sites in the frontal cortex of suicide victims.

Arch Gen Psychiatry 45: 577–579.

21. Pariante CM, Miller AH (2001) Glucocorticoid receptors in major depression:

relevance to pathophysiology and treatment. Biol Psychiatry 49: 391–404.

22. Carroll BJ, Curtis GC, Mendels J (1976) Neuroendocrine regulation in

depression. II. Discrimination of depressed from nondepressed patients. Arch

Gen Psychiatry 33: 1051–1058.

23. Carroll BJ (1984) Dexamethasone suppression test for depression. Adv Biochem

Psychopharmacol 39: 179–188.

24. Von Bardeleben U, Holsboer F (1989) Cortisol response to a combined

dexamethasone-hCRH challenge in patients with depression. J Neuroendocrinol

1: 485–488.

25. Heuser I, Yassouridis A, Holsboer F (1994) The combined dexamethasone/

CRH test: a refined laboratory test for psychiatric disorders. J Psychiatr Res 28:

341–356.

26. Heuser IJ, Schweiger U, Gotthardt U, Schmider J, Lammers CH, et al. (1996)

Pituitary-adrenal-system regulation and psychopathology during amitriptyline

treatment in elderly depressed patients and normal comparison subjects.

Am J Psychiatry 153: 93–99.

27. Deuschle M, Schmider J, Weber B, Standhardt H, Korner A, et al. (1997) Pulse-

dosing and conventional application of doxepin: effects on psychopathology and

hypothalamus-pituitary-adrenal (HPA) system. J Clin Psychopharmacol 17:

156–160.

28. Holsboer-Trachsler E, Stohler R, Hatzinger M (1991) Repeated administration

of the combined dexamethasone-human corticotropin releasing hormone

stimulation test during treatment of depression. Psychiatry Res 38: 163–171.

29. Holsboer-Trachsler E, Hemmeter U, Hatzinger M, Seifritz E, Gerhard U, et al.

(1994) Sleep deprivation and bright light as potential augmenters of

antidepressant drug treatment–neurobiological and psychometric assessment of

course. J Psychiatr Res 28: 381–399.

30. Hatzinger M, Hemmeter UM, Baumann K, Brand S, Holsboer-Trachsler E

(2002) The combined DEX-CRH test in treatment course and long-term

outcome of major depression. J Psychiatr Res 36: 287–297.

31. Frieboes RM, Sonntag A, Yassouridis A, Eap CB, Baumann P, et al. (2003)

Clinical outcome after trimipramine in patients with delusional depression - a

pilot study. Pharmacopsychiatry 36: 12–17.

32. Nickel T, Sonntag A, Schill J, Zobel AW, Ackl N, et al. (2003) Clinical and

neurobiological effects of tianeptine and paroxetine in major depression. J Clin

Psychopharmacol 23: 155–168.

33. Schule C, Baghai TC, Eser D, Zwanzger P, Jordan M, et al. (2006) Time course

of hypothalamic-pituitary-adrenocortical axis activity during treatment with

reboxetine and mirtazapine in depressed patients. Psychopharmacology (Berl)

186: 601–611.

34. Zobel AW, Yassouridis A, Frieboes RM, Holsboer F (1999) Prediction of

medium-term outcome by cortisol response to the combined dexamethasone-

CRH test in patients with remitted depression. Am J Psychiatry 156: 949–951.

35. Zobel AW, Nickel T, Sonntag A, Uhr M, Holsboer F, et al. (2001) Cortisol

response in the combined dexamethasone/CRH test as predictor of relapse in

patients with remitted depression. a prospective study. J Psychiatr Res 35: 83–94.

36. Appelhof BC, Huyser J, Verweij M, Brouwer JP, van Dyck R, et al. (2006)

Glucocorticoids and relapse of major depression (dexamethasone/corticotropin-

releasing hormone test in relation to relapse of major depression). Biol Psychiatry

59: 696–701.

37. Aubry JM, Gervasoni N, Osiek C, Perret G, Rossier MF, et al. (2006) The

DEX/CRH neuroendocrine test and the prediction of depressive relapse in

remitted depressed outpatients. J Psychiatr Res.

38. Coryell W, Zimmerman M (1983) The dexamethasone suppression test and

ECT outcome: a six-month follow-up. Biol Psychiatry 18: 21–27.

39. Yerevanian BI, Olafsdottir H, Milanese E, Russotto J, Mallon P, et al. (1983)

Normalization of the dexamethasone suppression test at discharge from hospital.

Its prognostic value. J Affect Disord 5: 191–197.

40. Katona CL, Aldridge CR, Roth M, Hyde J (1987) The dexamethasone

suppression test and prediction of outcome in patients receiving ECT.

Br J Psychiatry 150: 315–318.

41. Peselow ED, Baxter N, Fieve RR, Barouche F (1987) The dexamethasone

suppression test as a monitor of clinical recovery. Am J Psychiatry 144: 30–35.

42. Schweitzer I, Maguire KP, Gee AH, Tiller JW, Biddle N, et al. (1987) Prediction

of outcome in depressed patients by weekly monitoring with the dexamethasone

suppression test. Br J Psychiatry 151: 780–784.

43. Charles GA, Schittecatte M, Rush AJ, Panzer M, Wilmotte J (1989) Persistent

cortisol non-suppression after clinical recovery predicts symptomatic relapse in

unipolar depression. J Affect Disord 17: 271–278.

44. Reul JM, Stec I, Soder M, Holsboer F (1993) Chronic treatment of rats with the

antidepressant amitriptyline attenuates the activity of the hypothalamic-

pituitary-adrenocortical system. Endocrinology 133: 312–320.

45. Reul JM, Labeur MS, Grigoriadis DE, De Souza EB, Holsboer F (1994)

Hypothalamic-pituitary-adrenocortical axis changes in the rat after long-term

treatment with the reversible monoamine oxidase-A inhibitor moclobemide.

Neuroendocrinology 60: 509–519.

46. Barden N, Reul JM, Holsboer F (1995) Do antidepressants stabilize mood

through actions on the hypothalamic-pituitary-adrenocortical system? Trends

Neurosci 18: 6–11.

47. American Psychiatric Association (1994) Diagnostic and Statistical Manual of

Mental Disorders. Washington, DC: American Psychiatric Press.

48. Wittchen HU, Wunderlich U, Zaudig M, Fydrich T (1997) Strukturiertes

Klinisches Interview fur DSM-IV. Achse I: Psychische Storungen. Gottingen:

Hogrefe.

49. Hamilton M (1960) A rating scale for depression. J Neurol Neurosurg Psychiatry

23: 56–62.

50. Forsythe AI, Keenan TA, Organick EI, Stenberg W (1969) Computer Science:

A First Course. New York: John Wiley and Sons Inc.

51. Ising M, Kunzel HE, Binder EB, Nickel T, Modell S, et al. (2005) The combined

dexamethasone/CRH test as a potential surrogate marker in depression. Prog

Neuropsychopharmacol Biol Psychiatry 29: 1085–1093.

52. Kunugi H, Urushibara T, Nanko S (2004) Combined DEX/CRH test among

Japanese patients with major depression. J Psychiatr Res 38: 123–128.

53. Kunugi H, Ida I, Owashi T, Kimura M, Inoue Y, et al. (2006) Assessment of the

dexamethasone/CRH test as a state-dependent marker for hypothalamic-

pituitary-adrenal (HPA) axis abnormalities in major depressive episode: a

Multicenter Study. Neuropsychopharmacology 31: 212–220.

54. Ising M, Horstmann S, Kloiber S, Lucae S, Binder EB, et al. (2007) Combined

dexamethasone/corticotropin releasing hormone test predicts treatment re-

sponse in major depression - a potential biomarker? Biol Psychiatry 62: 47–54.

55. Arana GW, Baldessarini RJ, Ornsteen M (1985) The dexamethasone

suppression test for diagnosis and prognosis in psychiatry. Commentary and

review. Arch Gen Psychiatry 42: 1193–1204.

56. Carroll BJ (1980) Dexamethasone suppression test in depression. Lancet 2: 1249.

57. Modell S, Lauer CJ, Schreiber W, Huber J, Krieg JC, et al. (1998) Hormonal

response pattern in the combined DEX-CRH test is stable over time in subjects

at high familial risk for affective disorders. Neuropsychopharmacology 18:

253–262.

58. Rybakowski JK, Twardowska K (1999) The dexamethasone/corticotropin-

releasing hormone test in depression in bipolar and unipolar affective illness.

J Psychiatr Res 33: 363–370.

59. Von Bardeleben U, Holsboer F (1991) Effect of age on the cortisol response to

human corticotropin-releasing hormone in depressed patients pretreated with

dexamethasone. Biol Psychiatry 29: 1042–1050.

60. Ribeiro SC, Tandon R, Grunhaus L, Greden JF (1993) The DST as a predictor

of outcome in depression: a meta-analysis. Am J Psychiatry 150: 1618–1629.

61. Carroll BJ, Feinberg M, Greden JF, Tarika J, Albala AA, et al. (1981) A specific

laboratory test for the diagnosis of melancholia. Standardization, validation, and

clinical utility. Arch Gen Psychiatry 38: 15–22.

62. Holsboer F, Liebl R, Hofschuster E (1982) Repeated dexamethasone

suppression test during depressive illness. Normalisation of test result compared

with clinical improvement. J Affect Disord 4: 93–101.

63. Greden JF, Gardner R, King D, Grunhaus L, Carroll BJ, et al. (1983)

Dexamethasone suppression tests in antidepressant treatment of melancholia.

The process of normalization and test-retest reproducibility. Arch Gen


64. Gerken A, Maier W, Holsboer F (1985) Weekly monitoring of dexamethasone

suppression response in depression: its relationship to change of body weight and

psychopathology. Psychoneuroendocrinology 10: 261–271.

65. Bowie PC, Beaini AY (1985) Normalisation of the dexamethasone suppression

test: a correlate of clinical improvement in primary depressives. Br J Psychiatry

147: 30–35.

66. Schule C, Baghai T, Schmidbauer S, Bidlingmaier M, Strasburger CJ, et al.

(2004) Reboxetine acutely stimulates cortisol, ACTH, growth hormone and

prolactin secretion in healthy male subjects. Psychoneuroendocrinology 29:

185–200.

67. Schule C, Baghai T, Goy J, Bidlingmaier M, Strasburger C, et al. (2002) The

influence of mirtazapine on anterior pituitary hormone secretion in healthy male

subjects. Psychopharmacology (Berl) 163: 95–101.

68. Schule C, Baghai T, Rackwitz C, Laakmann G (2003) Influence of mirtazapine

on urinary free cortisol excretion in depressed patients. Psychiatry Res 120:

257–264.

69. Schule C (2007) Neuroendocrinological mechanisms of actions of antidepressant

drugs. J Neuroendocrinol 19: 213–226.

70. Asnis GM, Halbreich U, Rabinovich H, Ryan ND, Sachar EJ, et al. (1985) The

cortisol response to desipramine in endogenous depressives and normal controls:

preliminary findings. Psychiatry Res 14: 225–233.

71. Asnis GM, Sanderson WC, van Praag HM (1992) Cortisol response to

intramuscular desipramine in patients with major depression and normal control

subjects: a replication study. Psychiatry Res 44: 237–250.

72. Brady LS, Whitfield HJ Jr, Fox RJ, Gold PW, Herkenham M (1991) Long-term

antidepressant administration alters corticotropin-releasing hormone, tyrosine



hydroxylase, and mineralocorticoid receptor gene expression in rat brain.

Therapeutic implications. J Clin Invest 87: 831–837.

73. Seckl JR, Fink G (1992) Antidepressants increase glucocorticoid and mineral-

ocorticoid receptor mRNA expression in rat hippocampus in vivo. Neuroendo-

crinology 55: 621–626.

74. Jensen JB, Mork A, Mikkelsen JD (2001) Chronic antidepressant treatments

decrease pro-opiomelanocortin mRNA expression in the pituitary gland: effects

of acute stress and 5-HT(1A) receptor activation. J Neuroendocrinol 13:

887–893.

75. Mori S, Zanardi R, Popoli M, Garbini S, Brunello N, et al. (1998) cAMP-

dependent phosphorylation system after short and long-term administration ofmoclobemide. J Psychiatr Res 32: 111–115.

76. Stout SC, Owens MJ, Nemeroff CB (2002) Regulation of corticotropin-releasing

factor neuronal systems and hypothalamic-pituitary-adrenal axis activity bystress and chronic antidepressant treatment. J Pharmacol Exp Ther 300:

1085–1092.77. Schule C, Baghai T, Zwanzger P, Ella R, Eser D, et al. (2003) Attenuation of

hypothalamic-pituitary-adrenocortical hyperactivity in depressed patients by

mirtazapine. Psychopharmacology (Berl) 166: 271–275.



Regulation of Kainate Receptor Subunit mRNA by Stressand Corticosteroids in the Rat HippocampusRichard G. Hunter*, Rudy Bellani, Erik Bloss, Ana Costa, Katharine McCarthy, Bruce S. McEwen

Laboratory of Neuroendocrinology, The Rockefeller University, New York, New York, United States of America

Abstract

Kainate receptors are a class of ionotropic glutamate receptors that have a role in the modulation of glutamate release andsynaptic plasticity in the hippocampal formation. Previous studies have implicated corticosteroids in the regulation of thesereceptors and recent clinical work has shown that polymorphisms in kainate receptor subunit genes are associated withsusceptibility to major depression and response to anti-depressant treatment. In the present study we sought to examinethe effects of chronic stress and corticosteroid treatments upon the expression of the mRNA of kainate receptor subunitsGluR5-7 and KA1-2. Our results show that, after 7 days, adrenalectomy results in increased expression of hippocampal KA1,GluR6 and GluR7 mRNAs, an effect which is reversed by treatment with corticosterone in the case of KA1 and GluR7 and byaldosterone treatment in the case of GluR6. 21 days of chronic restraint stress (CRS) elevated the expression of the KA1subunit, but had no effect on the expression of the other subunits. Similarly, 21 days of treatment with a moderate dose ofcorticosterone also increased KA1 mRNA in the dentate gyrus, whereas a high corticosterone dose has no effect. Our resultssuggest an interaction between hippocampal kainate receptor composition and the hypothalamic-pituitary-adrenal (HPA)axis and show a selective chronic stress induced modulation of the KA1 subunit in the dentate gyrus and CA3 that hasimplications for stress-induced adaptive structural plasticity.

Citation: Hunter RG, Bellani R, Bloss E, Costa A, McCarthy K, et al. (2009) Regulation of Kainate Receptor Subunit mRNA by Stress and Corticosteroids in the RatHippocampus. PLoS ONE 4(1): e4328. doi:10.1371/journal.pone.0004328


Received September 11, 2008; Accepted October 31, 2008; Published January 29, 2009

Copyright: � 2009 Hunter et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was supported by NIH MH 15125, MH41256 and MH065749. RGH was supported by the Gary R. Helman Foundation. The funders had no rolein study design, data collection and analysis, decision to publish, or preparation of the manuscript.



Introduction

The hippocampal formation, due to its high levels of expression of

receptors for corticosteroid stress hormones, is particularly suscep-

tible to weathering and structural changes as a result of chronic stress

and stress related diseases such as depression. The interplay of

corticosteroids and ionotropic excitatory amino acid receptors in

producing structural and physiologic changes in the hippocampal

formation has been the subject of a significant amount of research,

but most of this research has focused upon NMDA and AMPA

receptors while relatively little has sought to describe the effects of

corticosteroids upon the expression of kainate receptors (KAR).

There are five members to the KAR gene family: GluR5, 6 and

7 and KA 1 and 2 [1] and kainate receptors are comprised of

various admixtures of the five subunit proteins produced by these

genes. The KARs contribute to both excitatory neurotransmission

and the presynaptic modulation of neurotransmitter release [2–4].

Notably, KAR activation contributes to LTP in the hippocampus,

particularly at the mossy fiber synapse of the CA3 [5].

A number of recent clinical studies have shown KARs have

potentially important roles in a number of major mental disorders,

particularly depression. Polymorphisms in the KA1 receptor have

been associated with response to the anti-depressant citalopram

and GluR6 has been associated with suicidal ideation during

treatment with the same drug [6,7]. The GluR7 gene has also

been connected to recurrent major depression [8]. KA1, GluR5

and GluR6 have also shown association with schizophrenia and

bipolar disorder [9,10].

The first study to examine the effects of corticosteroids upon

hippocampal KAR was performed by Clark and Cotman [11],

who tested the effects of adrenalectomy and corticosterone

(CORT) replacement on binding at AMPAR, KAR and

NMDAR and found no replicable effect of corticosterone or

adrenalectomy on 3H kainate binding. Watanabe [12], did

however, observe a decrease in 3H kainate binding after

adrenalectomy, an effect which was blocked by replacement

with the selective mineralocorticoid receptor (MR) agonist

aldosterone but not the glucocorticoid receptor (GR) selective

agonist RU28362. A study examining the expression of mRNA

for KAR subunits 3 days after adrenalectomy and in response to

acute high and low dose CORT showed that low doses of

CORT, which presumptively occupy only MR, increased

expression of all subunits, while ADX or high dose CORT

(occupying both MR and GR) failed to significantly alter

expression of any subunit [13]. Finally, chronic peripheral

administration of the GR agonist dexamethasone increases

expression of GluR6 protein in the dentate gyrus and CA3

[14]. To date these studies constitute most of what is known

about the interactions of corticosteroids and the KAR. The

present study aims to add to our understanding of these

interactions by examining the extent to which different kainate

receptor subunit mRNA’s are regulated differentially by stress

and adrenal steroids, using adrenalectomy, hormone replace-

ment, chronic restraint stress and chronic corticosterone

treatment of adrenally intact animals.


cbaker

Cross-Out

cbaker

Inserted Text

30

Results

Effects of adrenalectomy and adrenal steroidreplacement

In order to examine the effect of corticosteroids on the

expression of KAR subunits, we performed In situ hybridization

(ISH) after adrenalectomy and subacute treatment with cortico-

sterone, aldosterone and RU28362. ISH revealed KAR subunit

specific patterns of expression in the subfields of the hippocampal

formation, consistent with previous reports [15]. There was a

significant main effect of treatment (F (4,28) = 3.824, see Figure 1.)

upon KA1 mRNA expression in the CA3 and dentate gyrus

regions of the dorsal hippocampus, but no main effects were seen

in the CA1 or CA2 regions. Adrenalectomy (ADX) increased KA1

message by 68611% in the CA3 and by 54613% in the dentate

(p,0.05 versus sham, n = 8), while neither the selective MR

agonist aldosterone nor the selective GR agonist RU28362, given

alone, reversed this effect. However, treatment with corticosterone

significantly reduced KA1 mRNA from ADX levels (p,0.05),

suggesting that MR/GR heterodimers may regulate expression

specifically in the dentate gyrus and downstream in the CA3.

The GluR6 results, shown in Figure 2, suggest that this receptor

subtype is predominately regulated via MR in the DG, CA1 and

CA3. There was a main effect of treatment on GluR6 mRNA

expression in the CA1 (F (4,28) = 3.778), CA3 (F (4,26) = 2.991)

and dentate gyrus (F (4,26) = 4.338) after corticosteroid manipu-

lations. In the DG, CA1 and CA3, ADX treatment showed a

modest but non-significant trend toward increased GluR6

expression, and ADX+Aldosterone treatment significantly reduced

mRNA compared to ADX treatment (p,0.05, n = 8), with no

effect compared to sham. These data suggest that GluR6 mRNA is

predominantly regulated through MR in the dentate gyrus, CA1

and CA3.

For GluR7, there was a main effect of treatment on mRNA

expression in the dentate gyrus (F (4,32) = 7.789, see Figure 3.).

Relative to Sham, ADX increased expression by 3965% (p,0.05)

and ADX+RU362 increased expression by 37.966% (p,0.05).

ADX+CORT replacement significantly reduced GluR7 message

levels from both ADX+Vehicle, ADX+Aldo, and ADX+RU362.

No main effects of treatment were observed with either GluR5

or KA2 after chronic corticosteroid manipulations (see Figure 4.

for representative autoradiograms of KA2 and GluR5 expression

in the hippocampal formation) .

Effects of chronic restraint stressTo examine the effects of a stress paradigm known to cause

structural and functional changes in the hippocampus on KAR

expression, KAR subunit levels were measured in the hippocam-

pus in response to 21-day chronic restraint stress (CRS).

Spironolactone, an MR antagonist, was used concurrently to

assess the possible contribution of MR to any stress effect. As can

be seen in Figure 5, a main effect of treatment was observed on

KA1 subunit expression in both the CA3 (F (2,21) = 7.817) and

DG (F(2,21) = 4.285). In the CA3, CRS and CRS+Spironolactone

significantly elevated expression compared to control. In the DG,

a similar increase was seen with CRS, but not with CRS+Spir-

onolactone. No other effects of CRS or CRS+Spironolactone were

seen with KA2, GluR5, GluR6, or GluR7. Final body weights of

both stressed groups (393.467.3g for CRS alone and 372.468.1g

for CRS and spironolactone) were significantly lower than controls

(430.266.6g, p,0.05), confirming that CRS was effective

systemically.

Effects of chronic corticosterone in drinking waterTo confirm that the effects of CRS were corticosteroid

dependent, we treated rats for 21 days with vehicle, 25 mg/ml

Figure 1. ROD of KA1 mRNA in the dentate gyrus (A) and CA3 (B) after adrenalectomy and treatment with vehicle (ADX),aldosterone (ADX+Aldo) RU28,362 (ADX+RU362) or corticosterone (ADX+CORT). (C) A representative autoradiogram of KA1 mRNA.*-significantly different from sham and ADX+Cort (p,0.05, n = 8). **-significantly different from ADX (p,0.05, n = 8).doi:10.1371/journal.pone.0004328.g001

KAR, Stress & Steroids


or 400 mg/ml corticosterone. As shown in Figure 6, treatment

with a moderate dose of 25 mg/ml, but not a high dose of 400 mg/

ml of corticosterone significantly (p,0.05) increased KA1 mRNA

levels in the dentate gyrus (F (2,16) = 6.504) but did not reach

significance in the CA3.

Apoptosis in the dentate gyrusAdrenalectomy produced a 46% (p,0.00001) increase (from

3.2 to 4.8% of total cell profiles) in the number of pyknotic cells in

the dentate gyrus relative to sham adrenalectomized animals (data

not shown).

Figure 2. ROD of GluR6 mRNA in the dentate gyrus (A) CA3 (B) and CA1 (C) after adrenalectomy and treatment with vehicle (ADX),aldosterone (ADX+Aldo) RU28,362 (ADX+RU362) or corticosterone (ADX+CORT). (D) A representative autoradiogram of GluR6 mRNA.*-significantly different from sham and ADX+CORT (p,0.05, n = 8).doi:10.1371/journal.pone.0004328.g002

Figure 3. ROD of GluR7 mRNA in the dentate gyrus (A) after adrenalectomy and treatment with vehicle (ADX), aldosterone(ADX+Aldo) RU28,362 (ADX+RU362) or corticosterone (ADX+CORT). (B) A representative autoradiogram of GluR6 mRNA. *-significantlydifferent from sham and ADX+CORT (p,0.05, n = 8).doi:10.1371/journal.pone.0004328.g003



Figure 4. Representative photomicrographs showing KA2 mRNA signal on the right and GluR5 mRNA on the left. We did not observechanges in expression of either of these transcripts.doi:10.1371/journal.pone.0004328.g004

Figure 5. ROD of KA1 mRNA in the dentate gyrus (A) and CA3 (B) after CRS. *-significantly different from unstressed controls (p,0.05,n = 8).doi:10.1371/journal.pone.0004328.g005

Figure 6. ROD of KA1 mRNA in the dentate gyrus (A) and CA3 (B) after 21 day treatment with either vehicle, 25 mg/ml or 400 mg/mlcorticosterone in drinking water. *-significantly different from vehicle treated animals (p,0.05, n = 8).doi:10.1371/journal.pone.0004328.g006



Discussion

Our studies reveal a complex pattern of changes in kainate

receptor subunit expression induced by adrenalectomy and cortico-

steroid replacement and a significant and somewhat paradoxical

effect of CRS and chronic corticosterone on KA1 mRNA levels, but

no effect of CRS upon either KA2 mRNA levels or levels of GluR5-7

mRNA. This pattern, which was found in the dentate gyrus (DG) and

CA3 region of the hippocampal formation, demonstrates that CRS

involves more than adrenal steroid mediation and that increased

KA1 mRNA levels may help explain morphological changes caused

by CRS in the DG and CA3. Moreover, the results for KA1 mRNA

levels highlight the potential of adrenal steroids to oppose certain

actions of stress, which is analogous to their ability to inhibit

inflammatory cytokine production.

Effects of adrenalectomy and steroid replacementThe use of chronic adrenalectomy might have potentially

confounded the interpretation of our results as adrenalectomy can

produce apoptosis in dentate granule cells [16–18]. In our

experiment, adrenalectomy did increase the number of pyknotic

cells observed in the dentate gyrus, though the total percentage of

pyknotic cells was never higher than 5%. While we cannot exclude

dentate apoptosis as the reason for the change in KAR mRNA levels

we observe in that region, it seems an improbable explanation for a

number of reasons. First, the changes in mRNA levels we observed

after adrenalectomy were generally increases. Further, the changes

we saw in the dentate were mirrored in the CA3 (GluR6 and KA1)

and CA1 (KA1), suggesting that in these cases at least, the change is

more likely due to a direct effect of our manipulations of steroid levels,

rather than an indirect one due to cell death in the dentate gyrus.

GluR6 mRNA levels increased after adrenalectomy in all

regions of the hippocampus examined. This effect was reversed by

aldosterone treatment, but not by the specific glucocorticoid

receptor agonist, RU28362. This implicates the MR in the control

of GluR6 mRNA levels in the hippocampal formation. Joels [13],

also observed a non-significant increase in GluR6 in the DG after

3 days of adrenalectomy. Collectively, these observations suggest

that MR activation inhibits GluR6 expression within the

hippocampal formation.

KA1, but not KA2, mRNA expression also increased after

7 days of adrenalectomy, but the effect was reversed by high dose

corticosterone rather than either the MR or GR selective agonists.

GluR7 mRNA expression showed a similar pattern to KA1. We

observed no changes in KA2 or GluR5 though expression of the

latter was very low, which may have limited our ability to detect

subtle changes. That KA1 and GluR7 were regulated by

corticosterone but not by selective GR or MR agonists suggests

they may be regulated by MR/GR heterodimers, a permutation of

classical steroid receptor signaling recently described in cell culture

[19], but as yet undescribed in vivo.

Effects of chronic restraint stressKA1 expression also increased after CRS; in fact, it was the only

KAR subunit to do so. This is interesting because KA1, in contrast to

KA2, appears to have a largely pre-synaptic localization at the mossy

fiber synapse [20]. Presynaptic KARs have been shown to act as

facilitating autoreceptors at the mossy fiber synapse [2,21–23].

These findings, therefore, suggest a potential mechanism for the

increase in hippocampal glutamate levels observed after stress

[24,25], namely, that they mediate a feed-forward enhancement of

glutamate release from mossy fiber terminals. Mossy fiber

activation by glutamate has been identified as a key factor in the

damaging effects of kainic acid on CA3 neurons [26–28].

Effects of chronic corticosterone treatmentSimilarly to the effects of CRS, chronic treatment with a

moderate dose of corticosterone produced an elevation of KA1

mRNA in the dentate, similar to that produced by chronic

restraint stress. In the CA3, which has comparatively little GR

[29,30], this effect was not present, suggesting that the changes

observed in the CA3 with CRS are the result of other mediators of

the response to chronic stress, such as increased activity of the

glutamate system in the hippocampus [25] . Interestingly, the

response to chronic corticosterone showed an inverted-U shaped

dose response, an effect often seen with regard to the effects of

glucocorticoids on brain [31]. Chronic restraint, which produces a

moderate elevation of corticosterone levels similar to that

produced by our low dose treatment, but not as high as those

produced by the 400 mg/ml dose [32,33] fits with this interpre-

tation, as do the findings of Joels[13], who also found that KA1

mRNA expression was enhanced more by a lower dose of cort

than by a high dose. our results suggest that KA1 is also subject to

regulation by corticosteroids in an inverted U shaped fashion.

Adrenal steroids oppose effects of CRS in CA3 anddentate gyrus

The role of adrenal steroids, at least based on the effects of

adrenalectomy and hormone replacement reported in this study, is

somewhat paradoxical and not unlike their anti-inflammatory

effects [34]. Moreover, the observation of increased KA1

expression in both adrenalectomy and CRS, however, is similar

to what has been observed for the glutamate transporter, GLT-1,

namely, an increased expression of GLT-1 after CRS but also an

increase after ADX that is reversed by adrenal steroid replacement

[35,36]. It is possible, for both GLT-1 and KA1, that two different

processes are operating in the two different treatment schemes.

One may speculate that, under basal conditions, adrenal

steroids may help to maintain the basal level of kainate receptors,

as well as GLT-1, so as to homeostatically regulate the level of

glutamate release and glutamatergic activity. According to the

present study, this type of regulation also applies to GluR6 and

GluR7, but not to GluR5 or KA2 mRNA expression. Yet, at the

same time, acute restraint stress elevates extracellular glutamate

levels, measured by microdialysis, and these elevations are blocked

by adrenalectomy [24]. Moreover, we show in the present study

that CRS produces a feed forward, allostatic up-regulation of the

KA1 subunit that may contribute to the dendritic retraction

caused by CRS, which is mediated in part by excitatory amino

acids [37]. Finally, our finding that moderate doses of CORT in

the drinking water mimic the CRS induced increase of KA1

mRNA levels whereas high oral doses of CORT fail to elevate

KA1mRNA indicates that a hormetic inverted U shaped dose

response is operating [38]. Moreover, this hormetic dose response

relationship may help explain the paradoxical finding that, while

both CRS and chronic CORT each separately cause shrinkage of

dendrites of CA3 neurons via a process dependent on glutamate

release, the combination of CRS plus chronic CORT treatment,

which presumptively elevates CORT levels beyond those pro-

duced by either treatment alone, prevented the dendritic

remodeling [39].

Future work, when specific antibodies become available, needs

to determine whether this up-regulation at the mRNA level is

reflected in increased KA1 protein expression (as subunit specific

radioligands are as yet unavailable), as well as determine the extent

to which stress-induced glucocorticoid secretion may be involved

in these changes. Examination of the behavior of KARs after

chronic stress using electrophysiology might also provide us with a

window on the functional role of these receptors in the adaptation



of the hippocampus to stress, although this approach may also be

impaired by the lack of selective drugs. Another important

question to answer will be the extent to which chronic stress or

corticosteroid treatment alters the response of KARs to and acute

stressor or corticosteroid treatment, as this will allow us to begin to

assess the extent to which KARs are involved in resilience to stress

versus stress induced pathophysiology.

These findings are made more interesting by recent findings

associating KA1 and GluR6 and 7 with major depression and

other major mental disorders [6–10], all the more so because the

subunit which definitively did not change expression levels in our

experiments, KA2, has thus far shown no association with affective

disorders either. Further understanding of these changes could

permit an improved understanding of both stress induced

pathologies and the reasons why these pathologies can take a

substantial amount of time to reverse, as is the case with major

depression.

Methods

AnimalsAdult male Sprague-Dawley rats were obtained from Charles

River Laboratories (Kingston, NY) at 70 days of age. Animals

were housed 2–3 per cage (same age cage mates) in clear

polycarbonate cages with wood chip bedding. All animals were

maintained on a 12 h light-dark schedule (lights on at 0800 h) and

the temperature was kept at 2162uC. All animals had ad libitum

access to food and water. All procedures were carried out in

accordance with the guidelines established by the NIH Guide for

the Care and Use of Laboratory Animals.

Chronic Restraint StressAnimals were left undisturbed after arrival for one week after

delivery. Stressed animals were restrained in wire mesh restrainers,

secured at the head and tail ends with large binder clips. Chronic

stress was administered for 6 hours daily for 21 days from 10:00 to

16:00. Animals were returned to their home cages immediately

after termination of the stressor. These animals were sacrificed by

decapitation roughly 24 hours after the last stress (i.e. between

1300 and 1700 h). Brains were removed and flash frozen on dry

ice and then stored at 280uC until processing.

Steroid TreatmentsThese treatments follow those administered in [12] with some

modification. We chose to follow the one week time period used by

Watanabe for two reasons: first, he observed changes in KAR

levels after one week of steroid replacement. Secondly, after

adrenalectomy there is a progressive apoptosis of dentate gyrus

granule cells [16] and while we have successfully detected changes

in mRNA at the seven day time point in the past [12,40], we were

concerned that at later time points the potential for confounds

would be much greater. Animals were anesthetized using ketamine

and xylazine and the adrenal glands removed, save for one group

which received a sham surgery. During the same surgery, osmotic

mini-pumps (Alzet, Cupertino, CA) were implanted subcutane-

ously between the scapulae. These pumps delivered vehicle (50%

polyethylene glycol), the mineralocorticoid receptor agonist

aldosterone at 10 mg/hour or the glucocorticoid receptor agonist

RU28,362 at 10 mg/hour. Animals who underwent ADX received

0.9% saline in their drinking water and one group received

400 mg/ml corticosterone in addition to the saline. Seven days

after the completion of the surgeries, the animals were sacrificed

by decapitation and their brains removed and frozen as described

above.

Chronic Corticosterone TreatmentAnimals were provided with either 2.5% ETOH (vehicle),

25 mg/ml corticosterone or 400 mg/ml corticosterone in their

home cage drinking water for a period of 21 days.

In Situ HybridizationBrain sections were cut at 20 mm on a cryostat and placed on

Fisher Biotech ProbeOn Plus slides (Fisher, Pittsburgh, PA). In situ

hybridization began with a tailing reaction to radioactively label

the oligonucleotide probes with 35S. The probe sequences follow

those described by [41], two probe sequences were used in a

cocktail in order to improve sensitivity: KA1 59-TCC AGA GAG

GAG AAA TAG CCC GGT CTG CGT CCC ATA TGA ACT

CTG -39, 59-CTT GTA GTT GAA CCG TAG GAT CTC AGC

GAA CTC CTT GAG CAT GTC-39; KA2 59-TTC CAC TCG

GGC CTT GGC TGG GAC CTC GAT GAT CCC ATT GAT

CTG-39, 59-GTT CTC CAG GAT ATG GGG ACG CGC CCG

AAG ACA CGG GTG AGG GTT-39; GluR5 59-AAA TCC

CTC CGA TCC TGA GCA CT TGA GGG GAG GTC TGA

GGG AGG-39, 59-CCC GGG TTG GTT CCA TTG GGC TTC

CGC GTA AAG GAT GCT AAT GCC-39; GluR6 59-GGT

TCC TTG CGA ATA TCC GAT CCA CAA TAA GCA GAG

CAG G, 59- GGT TCC TTG CGA ATA TCC GAT CCA CAA

TAA GCA GAG CAG G-39, 59-ACT AAA CCT GGC TAT

GAC AAA GAG CAC ACA ACT GAC ACC CAA GTA-39;

GluR7 59-CTC AGC GTT CAT GAC CTG GGC GTT GGG

GCC GTC CGC GTA CTC AAA-39, 59-ATT CTC CAC CAC

CTC AGA GCC GGG GTT GCA GGG GTG GGC ATC

ATA-39. Processing of the slides followed methods as previously

described in [42]. Anatomical locations were determined with the

assistance of the atlas of Paxinos and Watson [43]. Optical density

was determined using MCID 5.0 (Imaging Research, St.

Catharine’s, OT, Canada).

Pyknotic Cell CountsNumbers of pyknotic cells were assessed following the method

of Frye and McCormick [44]. Sections were serial to those used

for autoradiography and in situ. Slides containing these sections

were processed to reveal Nissl substance beginning with a brief

fixation in 4% paraformaldehyde in 0.1M PB for 15 minutes

after which they were washed in distilled water three times for

2 minutes per wash. Sections were then dipped in 0.1% Cresyl

Violet for 2 minutes and then dehydrated in ascending

concentrations of ethanol prior to clearing in xylenes for

4 minutes. After drying, the slides were coverslipped with

permount. Pyknotic cells in the granule cell layer and subgranule

zone of the dentate gyrus were identified in a 1006visual field as

those having a small volume, membrane blebbing, and dark

condensed nucleus and chromatin.

StatisticsOptical density measurements were analyzed by a one way

ANOVA for the chronic steroid study and the chronic stress study.

Significant main effects and interactions in ANOVA were further

analyzed using Fisher’s protected least significant difference test

and Tukey’s test, respectively. Differences are considered signif-

icant at p,0.05. All data are presented as mean6SEM.


Conceived and designed the experiments: RGH BM. Performed the

experiments: RGH RB EB AC KMM. Analyzed the data: RGH EB AC

KMM. Wrote the paper: RGH BM.



References

1. Hollmann M, Heinemann S (1994) Cloned glutamate receptors. Annu Rev

Neurosci 17: 31–108.2. Contractor A, Swanson G, Heinemann SF (2001) Kainate receptors are involved

in short- and long-term plasticity at mossy fiber synapses in the hippocampus.Neuron 29: 209–216.

3. Chittajallu R, Vignes M, Dev KK, Barnes JM, Collingridge GL, et al. (1996)

Regulation of glutamate release by presynaptic kainate receptors in thehippocampus. Nature 379: 78–81.

4. Mulle C, Sailer A, Swanson GT, Brana C, O’Gorman S, et al. (2000) Subunitcomposition of kainate receptors in hippocampal interneurons. Neuron 28:

475–484.

5. Bortolotto ZA, Nistico R, More JC, Jane DE, Collingridge GL (2005) Kainatereceptors and mossy fiber LTP. Neurotoxicology 26: 769–777.

6. Laje G, Paddock S, Manji H, Rush AJ, Wilson AF, et al. (2007) Genetic markersof suicidal ideation emerging during citalopram treatment of major depression.

Am J Psychiatry 164: 1530–1538.7. Paddock S, Laje G, Charney D, Rush AJ, Wilson AF, et al. (2007) Association of

GRIK4 with outcome of antidepressant treatment in the STAR*D cohort.

Am J Psychiatry 164: 1181–1188.8. Schiffer HH, Heinemann SF (2007) Association of the human kainate receptor

GluR7 gene (GRIK3) with recurrent major depressive disorder. Am J MedGenet B Neuropsychiatr Genet 144: 20–26.

9. Beneyto M, Kristiansen LV, Oni-Orisan A, McCullumsmith RE, Meador-

Woodruff JH (2007) Abnormal glutamate receptor expression in the medialtemporal lobe in schizophrenia and mood disorders. Neuropsychopharmacology

32: 1888–1902.10. Blackwood DH, Pickard BJ, Thomson PA, Evans KL, Porteous DJ, et al. (2007)

Are some genetic risk factors common to schizophrenia, bipolar disorder anddepression? Evidence from DISC1, GRIK4 and NRG1. Neurotox Res 11:

73–83.

11. Clark AS, Cotman CW (1992) Adrenal hormone effects on hippocampalexcitatory amino acid binding. Brain Res 585: 161–168.

12. Watanabe Y, Weiland NG, McEwen BS (1995) Effects of adrenal steroidmanipulations and repeated restraint stress on dynorphin mRNA levels and

excitatory amino acid receptor binding in hippocampus. Brain Res 680:

217–225.13. Joels M, Bosma A, Hendriksen H, Diegenbach P, Kamphuis W (1996)

Corticosteroid actions on the expression of kainate receptor subunit mRNAs inrat hippocampus. Brain Res Mol Brain Res 37: 15–20.

14. Strutz-Seebohm N, Seebohm G, Shumilina E, Mack AF, Wagner HJ, et al.(2005) Glucocorticoid adrenal steroids and glucocorticoid-inducible kinase

isoforms in the regulation of GluR6 expression. J Physiol 565: 391–401.

15. Wisden W, Seeburg PH (1993) A complex mosaic of high-affinity kainatereceptors in rat brain. J Neurosci 13: 3582–3598.

16. Sloviter RS, Valiquette G, Abrams GM, Ronk EC, Sollas AL, et al. (1989)Selective loss of hippocampal granule cells in the mature rat brain after

adrenalectomy. Science 243: 535–538.

17. Woolley CS, Gould E, Sakai RR, Spencer RL, McEwen BS (1991) Effects ofaldosterone or RU28362 treatment on adrenalectomy-induced cell death in the

dentate gyrus of the adult rat. Brain Res 554: 312–315.18. Gould E, McEwen BS (1993) Neuronal birth and death. Curr Opin Neurobiol 3:

676–682.19. Nishi M, Tanaka M, Matsuda K, Sunaguchi M, Kawata M (2004) Visualization

of glucocorticoid receptor and mineralocorticoid receptor interactions in living

cells with GFP-based fluorescence resonance energy transfer. J Neurosci 24:4918–4927.

20. Darstein M, Petralia RS, Swanson GT, Wenthold RJ, Heinemann SF (2003)Distribution of kainate receptor subunits at hippocampal mossy fiber synapses.

J Neurosci 23: 8013–8019.

21. Schmitz D, Mellor J, Nicoll RA (2001) Presynaptic kainate receptor mediation offrequency facilitation at hippocampal mossy fiber synapses. Science 291:

1972–1976.

22. Lauri SE, Delany C, VR JC, Bortolotto ZA, Ornstein PL, et al. (2001) Synaptic

activation of a presynaptic kainate receptor facilitates AMPA receptor-mediatedsynaptic transmission at hippocampal mossy fibre synapses. Neuropharmacology

41: 907–915.23. Kullmann DM (2001) Presynaptic kainate receptors in the hippocampus: slowly

emerging from obscurity. Neuron 32: 561–564.

24. Lowy MT, Gault L, Yamamoto BK (1993) Adrenalectomy attenuates stress-induced elevations in extracellular glutamate concentrations in the hippocam-

pus. J Neurochem 61: 1957–1960.25. Raudensky J, Yamamoto BK (2007) Effects of chronic unpredictable stress and

methamphetamine on hippocampal glutamate function. Brain Res 1135:

129–135.26. Nadler JV, Perry BW, Cotman CW (1978) Intraventricular kainic acid

preferentially destroys hippocampal pyramidal cells. Nature 271: 676–677.27. Nadler JV, Cuthbertson GJ (1980) Kainic acid neurotoxicity toward hippocam-

pal formation: dependence on specific excitatory pathways. Brain Res 195:47–56.

28. de Montigny C, Weiss M, Ouellette J (1987) Reduced excitatory effect of kainic

acid on rat CA3 hippocampal pyramidal neurons following destruction of themossy projection with colchicine. Exp Brain Res 65: 605–613.

29. Han F, Ozawa H, Matsuda K, Nishi M, Kawata M (2005) Colocalization ofmineralocorticoid receptor and glucocorticoid receptor in the hippocampus and

hypothalamus. Neurosci Res 51: 371–381.

30. Van Eekelen JA, Jiang W, De Kloet ER, Bohn MC (1988) Distribution of themineralocorticoid and the glucocorticoid receptor mRNAs in the rat

hippocampus. J Neurosci Res 21: 88–94.31. Lupien SJ, McEwen BS (1997) The acute effects of corticosteroids on cognition:

integration of animal and human model studies. Brain Res Brain Res Rev 24:1–27.

32. Luine VN, Spencer RL, McEwen BS (1993) Effects of chronic corticosterone

ingestion on spatial memory performance and hippocampal serotonergicfunction. Brain Res 616: 65–70.

33. Watanabe Y, Gould E, McEwen BS (1992) Stress induces atrophy of apicaldendrites of hippocampal CA3 pyramidal neurons. Brain Res 588: 341–345.

34. Munck A, Naray-Fejes-Toth A (1994) Glucocorticoids and stress: permissive and

suppressive actions. Ann N Y Acad Sci 746: 115–130; discussion 131–113.35. Autry AE, Grillo CA, Piroli GG, Rothstein JD, McEwen BS, et al. (2006)

Glucocorticoid regulation of GLT-1 glutamate transporter isoform expression inthe rat hippocampus. Neuroendocrinology 83: 371–379.

36. Reagan LP, Gorovits N, Hoskin EK, Alves SE, Katz EB, et al. (2001)Localization and regulation of GLUTx1 glucose transporter in the hippocampus

of streptozotocin diabetic rats. Proc Natl Acad Sci U S A 98: 2820–2825.

37. McEwen BS (1999) Stress and hippocampal plasticity. Annu Rev Neurosci 22:105–122.

38. Calabrese EJ (2008) Neuroscience and hormesis: overview and general findings.Crit Rev Toxicol 38: 249–252.

39. Magarinos AM, Orchinik M, McEwen BS (1998) Morphological changes in the

hippocampal CA3 region induced by non-invasive glucocorticoid administra-tion: a paradox. Brain Res 809: 314–318.

40. Hunter RG, Bellani R, Bloss E, Costa A, Romeo RD, et al. (2007) Regulation ofCART mRNA by stress and corticosteroids in the hippocampus and amygdala.

Brain Res 1152: 234–240.41. Wullner U, Standaert DG, Testa CM, Penney JB, Young AB (1997) Differential

expression of kainate receptors in the basal ganglia of the developing and adult

rat brain. Brain Res 768: 215–223.42. Hunter RG, Jones D, Vicentic A, Hue G, Rye D, et al. (2006) Regulation of

CART mRNA in the rat nucleus accumbens via D3 dopamine receptors.Neuropharmacology.

43. Paxinos G, Watson C (1986) The Rat Brain in Stereotaxic Coordinates. New

York: Academic Press.44. Frye CA, McCormick CM (2000) Androgens are neuroprotective in the dentate

gyrus of adrenalectomized female rats. Stress 3: 185–194.



Enriched Environment Experience Overcomes LearningDeficits and Depressive-Like Behavior Induced byJuvenile StressYana Ilin, Gal Richter-Levin*

Department of Psychology, The Institute for the Study of Affective Neuroscience (ISAN), University of Haifa, Mount Carmel, Haifa, Israel

Abstract

Mood disorders affect the lives and functioning of millions each year. Epidemiological studies indicate that childhoodtrauma is predominantly associated with higher rates of both mood and anxiety disorders. Exposure of rats to stress duringjuvenility (JS) (27–29 days of age) has comparable effects and was suggested as a model of induced predisposition for thesedisorders. The importance of the environment in the regulation of brain, behavior and physiology has long been recognizedin biological, social and medical sciences. Here, we studied the effects of JS on emotional and cognitive aspects ofdepressive-like behavior in adulthood, on Hypothalamic-Pituitary-Adrenal (HPA) axis reactivity and on the expression of celladhesion molecule L1 (L1-CAM). Furthermore, we combined it with the examination of potential reversibility by enrichedenvironment (EE) of JS – induced disturbances of emotional and cognitive aspects of behavior in adulthood. Three groupswere tested: Juvenile Stress –subjected to Juvenile stress; Enriched Environment – subjected to Juvenile stress and then, fromday 30 on to EE; and Naıves. In adulthood, coping and stress responses were examined using the elevated plus-maze, openfield, novel setting exploration and two way shuttle avoidance learning. We found that, JS rats showed anxiety- anddepressive-like behaviors in adulthood, altered HPA axis activity and altered L1-CAM expression. Increased expression of L1-CAM was evident among JS rats in the basolateral amygdala (BLA) and Thalamus (TL). Furthermore, we found that EE couldreverse most of the effects of Juvenile stress, both at the behavioral, endocrine and at the biochemical levels. Theinteraction between JS and EE resulted in an increased expression of L1-CAM in dorsal cornu ammonis (CA) area 1 (dCA1).

Citation: Ilin Y, Richter-Levin G (2009) Enriched Environment Experience Overcomes Learning Deficits and Depressive-Like Behavior Induced by JuvenileStress. PLoS ONE 4(1): e4329. doi:10.1371/journal.pone.0004329



Copyright: � 2009 Ilin et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricteduse, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This study was supported by the EU’s PROMEMORIA grant #512012 to GRL, and by a grant from The Institute for the Study of Affective Neuroscience(ISAN) (2007) to GRL. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.



Introduction

Mood disorders affect the lives and functioning of millions each

year. A greater understanding of the neural circuits underlying

mood in both normal and abnormal affective states has been

identified as one of the critical needs in the field of mood disorders

research [1].

Stress, particularly when uncontrollable, excessive and/or

prolonged, can produce a myriad of emotional and cognitive

alterations [2–4]. In some individuals, stress can eventually trigger

or exacerbate mood disorders, among which depression and

bipolar disorders appear to be particularly linked to aversive life

experiences [5]. Chronic stress procedures are currently widely

used in experimental animals (mainly rodents) to model depression

[6–9].

Many of the hormones secreted during stress have been shown

to affect learning and memory processes [3,10,11]. Thus, stress has

been shown to affect synaptic plasticity [3,12], particularly

hippocampal plasticity, dendrite morphology, neurotoxicity and

neurogenesis within the dentate gyrus [13,14]. Stress diminishes

hippocampal synaptic plasticity, producing morphological changes

in dendritic development, and decreasing neurogenesis in the

dentate granule cells. Stress effects on the hippocampal formation

and on memory involve other neural structures (e.g., hypothala-

mus) and neuromodulators (norepinephrine and c-aminobutric

acid (GABA)) [12].

Also, numerous studies have demonstrated that early-life

stressful experiences affect both acute and long-term development

of neuroendocrine, cognitive and behavioral systems. Exposure to

stress or trauma during early childhood may disturb the formation

of functional brain pathways, in particular, of the limbic circuits

[15–18].

Previous findings from our group indicate that an exposure of

rats to a relatively brief stressful experience during juvenility (27–

29 days of age) has profound and long-lasting behavioral effects

[19,20]. In addition, a short-term juvenile exposure to variable

stressors produced two types of impaired avoidance learning

reminiscent of symptoms of both mood and anxiety disorders

[20,21].

The importance of the environment in the regulation of brain,

behavior and physiology has long been recognized in biological,

social and medical sciences [22]. Animals maintained under

enriched conditions (EE) have clearly been shown to have reduced

aggression [23], reduction of anxiety, fear and excitability [24–27],


reduction of stress [28–30], brain function [31,32] and better

learning abilities [33] than those maintained under standard

conditions. However, most of these studies have been carried out

in animals with no history of early insults.

Several studies have demonstrated that cell adhesion molecules

(CAMs) are involved in corticosterone (CORT) actions in memory

and neuroplasticity. Relevant to the current study, the expression

levels of a member of the immunoglobulin superfamily, L1-CAM,

are considered to be regulated through glucocorticoid-mediated

pathways [34]. Moreover, L1-CAM, has been implicated not only

in cell interactions during nervous system development, but also in

synaptic plasticity and memory formation in the adult brain [35–

42]. During early development it promotes neurite outgrowth and

fasciculation [43], axon pathfinding [44] and myelination [45].

Recent clinical and preclinical work has highlighted L1-CAM as

particularly susceptible to showing alterations in stress-related

disorders and depression [46–50].

Here, we studied the effects of stress during juvenility (JS) (27–29

days of age) on emotional and cognitive aspects of depressive-like

behavior in adulthood, on HPA axis reactivity and on the

expression of L1-CAM. Furthermore, we combined it with the

examination of potential reversibility by enriched environment of

JS – induced disturbances of emotional and cognitive aspects of

behavior in adulthood.

Alterations in expression level of L1-CAM was checked in the

prefrontal cortex (PFC), basolateral amygdala (BLA), dorsal cornu

ammonis (CA) area 1 (dCA1) and thalamus (TL). This areas were

chosen because they share extensive anatomic connections [51]

and found to be affected by early life stress [52–54].

The exposure to JS resulted in both mood and anxiety

symptoms. Furthermore, EE could reverse most of the effects of

JS, at the behavioral, endocrine and at the biochemical levels.

Results

Body weightSignificant differences (p,0.05) were observed between the

body weight gain of the three groups (Naıve, JS and JS+EE).

Repeated measure analysis for lingering body weight gain

revealed a significant main effect for each measure

[WL = 0.005; F(1,174) = 17355.34; p,0.001] and for groups

[F(2,114) = 3.84; p,0.024]. Post-hoc Tukey analysis at 30 post

natal day (PND) indicated that in comparison with Naıve

(unexposed) rats, juvenile-stressed rats (from both groups JS and

JS+EE) exhibited less body weight gain when examined 24 h

after the exposure to stress. However, one week later (38 PND)

this difference was observed only for JS group, there was no

difference between Naıve and JS+EE groups. However, later on

during the maturation process (at 45, 52, and 59 PND), this

difference was no more evident. These results indicate that

though the stressor affected body weight gain in the short run, in

the long run juvenile-stressed rats (from both groups JS and

JS+EE) continued to develop normally in terms of their body

weight gain (Figure 1).

Behavioral Assessments in AdulthoodAnimals were tested in the open field (OF) and elevated plus-

maze (EPM) at 60 PND, after 1 month in different housing

environments. Significant differences in behavioral parameters

(activity and anxiety-like behavior) were observed. At the next, 61

PND, day animals were subjected to the TWS avoidance task.

Learning abilities of the animals were also affected by the

manipulations.

Open FieldOne-way ANOVA revealed a significant effect of group on time

spent in the open arena of the OF [F(2,51) = 12.75, p,0.001].

Post-hoc Tukey testing indicated that the time spent in the open

arena of the OF of the JS group was significantly lower than that of

the Naıve and JS+EE groups. The time spent in the open arena of

the JS+EE group was significantly higher than that of the Naıves

and JS. The time spent in the open arena of the Naıve group was

significantly higher than that of the JS, while being significantly

lower than that of the JS+EE group (Figure 2A).

One-way ANOVA revealed a significant effect of group on the

number of center square crossing [F(2,51) = 16.54, p,0.001].

Post-hoc Tukey testing indicated that the number of center square

crossing of the JS group was significantly lower than that of the

Naıve and JS+EE groups. The number of center square crossing of

the JS+EE group was significantly higher than that of the Naıves

and JS. The number of center square crossing of the Naıve group

was significantly higher than that of the JS, while being

significantly lower than that of the JS+EE group (Figure 2B).

One-way ANOVA for the number of periphery square crossing

showed no significant effect for groups [F(2,51) = 1.87, N.S.]

(Figure 2B).

One-way ANOVA for the locomotor activity (total number of

squares crossed) showed no significant effect for groups

[F(2,51) = 0.92, N.S.] (Figure 2B).

Elevated Plus MazeOne-way ANOVA revealed a significant effect of group on time

spent in the open arms of the EPM [F(2,87) = 8.82, p,0.001].

Post-hoc Tukey testing indicated that the time spent in the open

arms of the JS group was significantly lower than that of the Naıve

and JS+EE groups. There was no significant difference between

JS+EE and Naıve groups (Figure 3A).

One-way ANOVA revealed a significant effect of group on line

crossing in the open arms of the EPM [F(2,87) = 4.32, p,0.016].

Post-hoc Tukey testing indicated that the line crossing in the open

arms of the JS group was significantly lower than that of the

JS+EE group (Figure 3B).

One-way ANOVA revealed a significant effect of group on line

crossing in the closed arms of the EPM [F(2,87) = 10.07,

p,0.001]. Post-hoc Tukey testing indicated that the line crossing

in the closed arms of the JS+EE group was significantly higher

than that of the Naıve and JS groups (Figure 3B).

One-way ANOVA revealed a significant effect of group on total

line crossing in the EPM [F(2,87) = 8.89, p,0.001]. Post-hoc

Tukey testing indicated that the total line crossing of the JS+EE

group was significantly higher than that of the Naıve and JS

groups (Figure 3B).

Novel-setting explorationOne-way ANOVA revealed a significant effect of group on

novel setting exploration [F(2,116) = 12.32, p,0.001]. Post-hoc

Tukey testing indicated that the exploratory behavior of the JS

group was significantly lower than that of the Naıve and JS+EE

groups. The exploratory behavior of the JS+EE group was

significantly higher than that of the Naıves and JS. The

exploratory behavior of the Naıve group was significantly higher

than that of the JS, while being significantly lower than that of the

JS+EE group (Figure 4).

Two-way shuttle (TWS) avoidance taskAvoidance responses. One-way ANOVA revealed a

significant effect of group on percent of avoidance responses

Reversing Juvenile Stress


during TWS avoidance task [F(2,24) = 3.79, p,0.037]. Post-hoc

Tukey testing indicated that percent of avoidance responses of the

JS+EE group was significantly higher than that of the Naıve and

JS groups (Figure 5A).

Escape responses. One-way ANOVA revealed a significant

effect of group on percent of escape responses during TWS

avoidance task [F(2,24) = 3.96, p,0.033]. Post-hoc Tukey testing

indicated that percent of escape responses of the JS+EE group was

significantly lower than that of the Naıve and JS groups

(Figure 5B).

No Escape responses. One-way ANOVA revealed a

significant effect of group on percent of no escape responses

during TWS avoidance task [F(2,24) = 6.75, p,0.005]. Post-hoc

Tukey testing indicated that percent of no escape responses of the

JS group was significantly higher than that of the Naıve and

JS+EE groups (Figure 5C).

Endocrine and Molecular Assessments in AdulthoodAt PND 60, between 10:00 and 12:00 h, Naıve, JS and JS+EE

groups of animals, without previous history of testing, were taken

directly from their home-cages for brain and trunk blood

collection.

Concentrations of corticosteroneOne-way ANOVA revealed a significant effect of group on

basal CORT concentration [F(2,23) = 6.14, p,0.007]. Post-hoc

Tukey testing indicated basal CORT concentration of the JS

group was significantly higher than that of the Naıve and JS+EE

groups (Figure 6).

L1-CAM expressionL1-CAM expression was measured at 60 PND in the prefrontal

cortex (PFC), basolateral amygdala (BLA), dorsal cornu ammonis

(CA) area 1 (dCA1) and thalamus (TL) (Figure 7). Expression levels

are depicted as the ratio between the total L1-CAM expression

level and b-actin levels in each brain area (i.e. L1-CAM/b-actin),

normalized to the Naıve group.

In the PFC. One-way ANOVA for L1-CAM expression levels

in the PFC showed no significant effect for the group

[F(2,26) = 0.17, N.S.].

In the BLA. One-way ANOVA for L1-CAM expression levels

in the BLA revealed significant effect for the group [F(2,23) = 4.22,

p,0.027]. Post-hoc Tukey testing indicated that L1-CAM

expression levels of the JS group was significantly higher than

that of the Naıve group.

Figure 1. Body weight. JS (n = 39) and JS+EE (n = 39) exhibited less body weight gain when examined 24 h after the exposure to stress comparedto Naıve group (n = 48). However, one week later (38 PND) this difference was observed only for the JS group. There was no difference between Naıveand JS+EE groups. Later on during the maturation process (at 45, 52, and 59 PND), this difference was no longer evident. *JS significantly differentfrom Naıve (p,0.05); #JS+EE significantly different from Naıve (p,0.05).doi:10.1371/journal.pone.0004329.g001



In the dCA1. One-way ANOVA for L1-CAM expression

levels in the dCA1 revealed significant effect for the group

[F(2,23) = 4.30, p,0.026]. Post-hoc Tukey testing indicated that

L1-CAM expression levels of the JS+EE group was significantly

higher than that of the Naıve group.

In the TL. One-way ANOVA for L1-CAM expression levels

in the TL revealed significant effect for the group [F(2,25) = 6.09,

p,0.007]. Post-hoc Tukey testing indicated that L1-CAM

expression levels of the JS group was significantly higher than

that of the Naıve and JS+EE groups.

Discussion

Social/Environmental stress in early life such as maternal

separation, isolation, poverty, etc. is not avoidable in many children.

Cognitive deficits progressively emerging with development are the

results of complex interactions between genetic and environmental

factors [55–57], and evidence suggests that EE experience can

attenuate or reverse a variety of cognitive deficits [58].

This study was designed to experimentally investigate the effects

of EE during adolescents on JS rats. We found that, JS rats showed

anxiety- and depressive-like behaviors in adulthood, altered HPA

axis activity and L1-CAM expression pattern through limbic

system areas and the thalamus. Furthermore, we found that EE

could reverse most of the effects of JS, both at the behavioral,

endocrine and at the biochemical levels.

Behavioral Assessments in AdulthoodOur JS protocol resulted in a variety of behavioral changes in

the rodents that might be regarded as behavioral correlates of

depressive-like symptoms in humans. In our experiment rats were

compared in the following tests: (1) OF and EPM tests for the

assessment of anxiety level as one of the possible components of

depressive state [24,59]; (2) novel-setting exploration as motiva-

tional/hedonic state measure [60–62]; (3) TWS avoidance task for

revealing a possible sign for cognitive disturbances or learned

helplessness (LH) behavior, as an analogue of impaired coping

state in depression [63,64].

Exposure to the JS transiently delayed body weight gain.

In comparison with Naıve rats, both juvenile-stressed rats

(JS and JS+EE) exhibited less body weight gain when

examined 24 h after the exposure (at 30 PND), indicating that

Figure 2. Open Field (OF) Test. (A) Time spent in the open arena. Time spent in the open arena of the OF of the JS (n = 17) group was significantlyshorter than that of the Naıve (n = 20) and JS+EE (n = 17) groups. Time spent in the open arena of the JS+EE group was significantly longer than thatof the Naıves and JS. Time spent in the open arena of the Naıve group was significantly longer than that of the JS, while being significantly shorterthan that of the JS+EE group. (B) The locomotor activity in the OF. The number of center square crossing of the JS group was significantly lower thanthat of the Naıve and JS+EE groups. The number of center square crossing of the JS+EE group was significantly higher than that of the Naıves and JS.The number of center square crossing of the Naıve group was significantly higher than that of the JS, while being significantly lower than that of theJS+EE group. There was no difference between the groups in the number of periphery square crossing and total locomotor activity (total number ofsquares crossed) in the OF. *significantly different from all other groups (p,0.05).doi:10.1371/journal.pone.0004329.g002



Figure 3. Elevated Plus Maze (EPM). (A) Time spent in the open arms. Time spent in the open arms of the JS (n = 29) group was significantlyshorter than that of the Naıve (n = 31) and JS+EE (n = 30) groups. There was no significant difference between JS+EE and Naıve groups. (B) Thelocomotor activity in the EPM. The line crossing in the open arms of the JS group was significantly lower than that of the JS+EE group. Line crossing inthe closed arms of the JS+EE group was significantly higher than that of the Naıve and JS groups. Total line crossing of the JS+EE group wassignificantly higher than that of the Naıve and JS groups. *significantly different from all other groups (p,0.05); &significantly different from JS+EEgroup (p,0.05); #significantly different from Naıve group (p,0.05).doi:10.1371/journal.pone.0004329.g003

Figure 4. Exploration of a novel setting. Exploratory behavior of the JS (n = 38) group was significantly lower than that of the Naıve (n = 41) andJS+EE (n = 40) groups. The exploratory behavior of the JS+EE group was significantly higher than that of the Naıves and JS. The exploratory behaviorof the Naıve group was significantly higher than that of the JS, while being significantly lower than that of the JS+EE group. *significantly differentfrom all other groups (p,0.05).doi:10.1371/journal.pone.0004329.g004



both groups were affected in the same way by the stressors.

However, one week later, 37 PND, JS rats continued to show

less body weight gain than Naıve animals, while JS+EE rats

were no longer different from Naıves. This finding indicates that

the EE protocol started to have an impact already from the first

week.

However, later, during the maturation process, the body

weight gain difference was no longer evident, indicating that our

Figure 5. Two-Way Shuttle (TWS) Avoidance learning. (A) Avoidance responses. Percent of avoidance responses of the JS+EE (n = 8) group wassignificantly higher than that of the Naıve (n = 10) and JS (n = 8) groups. (B) Escape responses. Percent of escape responses of the JS+EE group wassignificantly lower than that of the Naıve and JS groups. (C) No Escape responses. Percent of no escape responses of the JS group was significantlyhigher than that of the Naıve and JS+EE groups. *significantly different from all other groups (p,0.05); #significantly different from Naıve group(p,0.05).doi:10.1371/journal.pone.0004329.g005



variable stress procedure affected body weight gain in the

short run; in the long run both JS groups (JS and JS+EE)

continued to develop normally in terms of their body weight

gain. The same pattern of weight changes was found also by

Brunson et. al. [65].

The emotional consequence of exposure to JS was examined in

both the OF and the EPM. High anxiety levels of the JS group

were found in both tests. In the OF there was a decrease in the

time spent in the central arena and center square crossing by JS

rats compared to Naıve and JS+EE rats. These findings confirm

previous findings of high anxiety level of JS rats even 1 month after

the exposure to the stress protocol [19,20]. EE not only reversed

high anxiety levels of JS, but reduced them even below the levels of

Naıve animals.

Figure 6. Serum corticosterone concentration. Basal CORT concentration of the JS (n = 9) group was significantly higher than that of the Naıve(n = 10) and JS+EE (n = 8) groups. *significantly different from all other groups (p,0.05).doi:10.1371/journal.pone.0004329.g006

Figure 7. L1-CAM expression at post natal day 60. (A) L1-CAM expression was measured at 60 PND in the PFC, BLA, dCA1 and TL (Naıve(n = 10); JS+EE (n = 8); JS+EE (n = 8)). Expression levels are depicted as the ratio between the total L1-CAM expression level and b-actin levels in eachbrain area (i.e. L1-CAM/b-actin), normalized to the Naıve group. In the PFC: no difference between the groups for L1-CAM expression levels. In theBLA: L1-CAM expression levels of the JS group was significantly higher than that of the Naıve group. In the dCA1: L1-CAM expression levels of theJS+EE group was significantly higher than that of the Naıve group. In the TL: L1-CAM expression levels of the JS group was significantly higher thanthat of the Naıve and JS+EE groups. (B) L1-CAM representative immunoblots. Bottom rows: b-actin; Top Rows: L1-CAM. *significantly different from allother groups (p,0.05); #significantly different from Naıve group (p,0.05).doi:10.1371/journal.pone.0004329.g007



Similar results were obtained in the EPM. JS reduced time spent

in open arms (as indicated both by time in open arms and open

arms crossing). Exposure of animals to the EE condition during 1

month after JS completely reversed this effect. In fact, EE in JS

animals increased total exploratory behavior even beyond that of

Naıve animals.

The initial activity of a rat placed in a novel surrounding (e.g.

novel setting exploration) can be taken as an indicator of its

emotional and motivational state [60]. It is assumed that a novel

context/situation reflects both the stress and the rewarding

component of novelty. It has been proposed that reduced

sensitivity to rewards in rodents might be homologous to human

anhedonia [66]. In rats, decreased exploratory activity in a novel

environment might reflect decreased motivation or drive, a

behavior representing ‘‘refractory loss of interest’’ [60,67] and

may also be related to an hedonic deficit, since novelty is

rewarding [61,62]. In our model JS rats exhibited reduced novel-

setting exploration compared to Naıve and JS+EE rats. This

reduced exploratory activity may represent the loss of interest in

new stimulating situations and may imply the presence of

motivational deficits. In contrast, novel-setting exploration of the

JS+EE rats not only was higher from JS rats but was also higher

from Naıve rats.

During TWS avoidance task JS rats were not different from

Naıve-controls in the total number of avoidance or escape

responses, but showed significantly more No Escape responses.

The increased rates of escape failure (no escape responses) during

this task that we found among JS rats may also imply an emotional

disruption. Such increases in escape failures were suggested to

correspond to learned helplessness, representing, in animals,

depressive symptoms of non-responsiveness [68]. EE completely

reversed this effect. Furthermore, EE increased total number of

avoidance responses even beyond that of Naıve-controls. Im-

proved learning and memory by EE is one of the most consistent

findings in the literature [69,70]. The present results confirm this

finding and extend its validity by showing that this effect even

overcomes the effects of JS.

Furthermore, exposure to stressors during juvenility affected the

HPA axis baseline activity. Analysis of basal circulating CORT

levels revealed elevated levels in the JS group, as compared to

Naıve and JS+EE groups. Serum corticosterone was used as the

traditional anxiety/stress marker [71]. This result provides

independent support to indicate that the JS group indeed

experienced significantly higher levels of anxiety than either of

the other groups. This finding is in agreement with reported

physiological abnormality of resting level titers of the hormone in

depressed humans [68,72,73]. EE reversed also this effect of JS.

Overall, JS appears to trigger anxiety- and depressive-like

behaviors; EE was found to be able to reverse these effects.

Moreover, EE not only reversed most of JS-induced disruptions

but rather, in some parameters made the animals less anxious,

more motivated and with better learning abilities compared also

with Naıve animals.

L1-CAM, together with other members of the L1 subfamily, is

critical for several early development processes like axon

outgrowth, fasciculation, neuronal migration and survival

[46,74–76]. Furthermore, L1-CAM restriction throughout post-

weaning and to adulthood developmental phase also affected stress

responsiveness and cognitive functions in adulthood [77],

suggesting a key role for L1-CAM in development related

processes during adolescence.

In the current study, exposure to stressors during juvenility

altered the expression levels of L1-CAM throughout the

monitored brain regions. Increased expression of L1-CAM was

evident among JS rats in the BLA and thalamus. In the thalamus,

EE completely reversed this effect, while in the BLA it only

reduced it.

Exposure to stressors during juvenility affected the HPA axis

baseline activity as was indicated by elevated basal CORT levels.

The amygdala shares extensive anatomic connections with the

thalamus [51]. Both these areas serve as feedback sites of HPA

regulation in stressed animals [78], so that alteration of L1-CAM

expression by JS and by EE could be related to the alterations

found in CORT levels under these conditions.

Individual variations in L1-CAM mRNA levels were positively

correlated with plasma CORT concentrations and anxiety-like

behaviors [38]. It was suggested that chronic-stress induced

increased L1-CAM levels may contribute to the chronic stress-

associated emotional and cognitive impairments [39,80]. In

addition, in the adult brain, L1-CAM regulation is affected by

continuous increased CORT levels or chronic stress exposure

[39]. Thus, elevated basal levels of CORT could explain the

observed amygadalar and thalamic L1-CAM alterations.

Since L1-CAM was implicated in repair processes in the adult

lesioned CNS [81–83], chronic-stress induced increased L1-CAM

levels were suggested to represent the activation of a neuropro-

tective mechanism [39,41,84]. However, early life stress could

disrupt the information processing in the cortex and thalamus of

the developing brain, and limbic system particularly, of juvenile

rats leading to cognitive and affective disorders. Controversially,

the limbic system is most probably modified by EE [24]. Thus, EE

experience could rescue the early life induced development

disruptions by triggering the release of nerve growth factors,

activating neurotransmitter receptors, or enhancing neurogenesis

[69,85,86].

The interaction between JS and EE resulted in an increased

expression of L1-CAM in dCA1, beyond that of Naıve-controls

and JS rats. EE has been found to have profound and long-lasting

neural and physiological consequences on the hippocampus. EE

has been shown to induce higher hippocampal expression of

glucocorticoid type II receptor mRNA [87]; enhanced hippocam-

pal field potentials [88,89]; and hippocampal neurogenesis in adult

animals [32,69]. Thus, L1-CAM increased levels in the dCA1 area

could reflect neuroprotective mechanism and the neurogenesis

that occurred through interaction between JS and EE.

EE was also found to improve the acquisition and long-term

retention of a two-way active avoidance [70]. These changes could

be correlated with the behavioral effects of EE compared to

controls. It is thus tempting to suggest that these alterations in an

area (CA1) associated with the behavior are relevant to the

behavioral effects of EE. Further experiments are required to

clarify this possibility

In conclusion, our data show that JS applied in rats induces a

broad spectrum of behavioral changes reminiscent of depressive

symptoms in humans. These results may be helpful for elucidating

cellular and molecular mechanisms involved in cognitive deficits

and affective disorders caused by early life stress. On the other

hand, our findings suggest that EE may be useful to prevent these

devastating effects in young adults following childhood stress.

Methods

SubjectsMale Sprague Dawley rats (SD), 22 days old, weighing 35–49 g

were purchased from Harlan (Jerusalem, Israel) and habituated in

the Brain and Behavior Research laboratory facilities for five days.

Three animals were housed per cage in 75655615 cm Plexiglas

cages in temperature-controlled (2361uC) animal quarters on a



12:12 light-dark cycle (lights on 07:00–19:00 hours). They had ad

libitum access to standard Purina rodent chow pellets and water.

Ethical approvalAll procedures and tests were approved by the Institutional

Animal Care Committee and adhered to the guidelines of the US

Institute of Laboratory Animal Research’s Guide for the Care and

Use of Laboratory Animals.

Three groups of SD rats were used

1. JS subjected to variable stress at post natal days (PND) 27–29

2. JS+EE subjected to variable stress at 27–29 PND and at 30

PND were transfered to EE housing conditions.

3. Naıve rats.

Juvenile stress procedureWe have designed a juvenile short-term variable stressor

protocol [20] in which rats were exposed to a different stressor

every day for three days (see below). Stress exposure took place

during juvenility (ages 27–29 days) at approximately midday

(12:00–14:00) in designated experimental rooms (a different room

each day) away from the vivarium.

N Day 1. (aged 27 d) Forced swim: 10 min forced swim in an

opaque circular water tank (diameter 0.5 m; height: 0.5 m;

water depth 0.4 m), water temperature 2262uC (adapted from

Avital et. al. [90]).

N Day 2. (aged 28 d) Elevated platform: three 30 min trials; ITI

(Inter-Trial Interval): 60 min in the home cage. Elevated

platform: 12612 cm at a height of 70 cm above floor level,

located in the middle of a small closet-like room (adapted from

[91].

N Day 3 (aged 29 d) Restraint stress: Rats were placed in a metal

mesh restraining box (116564 cm) that prevented forward-

backward movement and limited side-to-side mobility, but did

not discomfort the animal in any other way. Rats remained in

the restraining box for 2 hrs at 25uC under dim illumination.

Protocols were applied in parallel to rats in the stress groups, so

as not to isolate any rat in its home cage. Upon completion of the

each of the stress procedures, rats were returned to their home

cage.

Environmental Enrichment procedureEnriched Environment was defined in terms of combination of

physical environment and partially social housing conditions.

Therefore, animals were housed in larger and higher cages

provided with differently shaped plastic containers, colored

platforms and suspended objects. The objects were changed twice

a week. Once a week all animals from this group were taken

together to another enriched box with different objects, wheel, one

apple, carrot, cucumber and 50 g of granola.

For both housing conditions: standard and EE, the sawdust of

the cage was changed once a week in association with

measurement of animals body weight. Rats were put in standard

and EE cages at the age 30 PND and maintained in their housing

conditions throughout all the experimental assessment.

Experimental designIn the present study in order to prevent the tests from

influencing one another, different rats were used for each of the

following experiments: (1) behavioral measurements; (2) cortico-

sterone concentrations and L1-CAM expression.

Behavioral Assessments in AdulthoodIn adulthood, 60–61 PND, coping and stress responses were

examined using the open field test, elevated plus-maze test, the

novel-setting exploratory behavior, two-way shuttle (TWS)

avoidance task.

Open field test (OF). The apparatus is a quadrant box,

90 cm length with 30 cm wall, divided into 15615 cm squares.

Animal was placed in the center of the field and the following

variables were recorded for 5 min: the number of squares crossed

and center square entries. The open field was cleaned after each

rat. The test room had a dim illumination (40 W) for decreasing

the aversiveness of the test.

Elevated plus-maze test (EPM). The apparatus is elevated

80 cm above a floor and exposed to dim illumination. It consists of

two opposite open arms (45610 cm) and two opposite closed arms

of the same size with walls 10 cm high. The arms are connected by

a central square (10610 cm). Each rat was placed on the central-

platform facing an open arm and was allowed to explore the maze

for 5 min. Each test was videotaped and scored by an independent

observer. Arm entry was defined as entering an arm with all four

paws. The following terms were used: durations in open arms,

open and closed arm crossing and total crossing of all arms.

Novel-setting exploration. Rats were placed in the two-way

shuttle avoidance apparatus described below, although it was in an

inoperative mode, and were allowed to explore both

compartments for a total of 10 min. Crossingover between

compartments provided an index of exploratory behavior.

Two-way shuttle (TWS) avoidance task. Immediately after

the exploratory behavior assessment a training session began.

Apparatus: The TWS box, placed in a dimly-lit, ventilated, sound-

attenuated cupboard, is a rectangular chamber (60626628 cm)

divided by an opaque partition with a small flap passage

(1068 cm) that connects two equal sized, side-by-side, cube-

shaped compartments. Both metal grid floors of the compartments

are weight sensitive and electrifiable. Micro-switches transmit

information about the location of the rat to a computer control

and data collection program. This program controls both

conditioned stimulus (CS) presentations (a tone produced by

loudspeakers located on the distal walls of the compartments) and

unconditioned stimulus (US) – electric shock deliveries (to the

animals’ feet through the compartment floor, by a Solid State

Shocker/Distributor, Coulbourn Instruments Inc. Lehigh Valley,

PA, USA). The TWS avoidance task: One session comprises of 80

‘‘trace conditioning’’ trials. CS: 10 s tone presentation; US:

immediately following the termination of the CS an electric

shock (1.2 mA) will be delivered for a maximum of 10 s; ITI:

(randomly varying) 30612 s. Rats could perform one of the

following behaviors: (1) Avoidance - shuttling to the adjacent

chamber of the apparatus while the tone was on, thus avoiding the

shock altogether; (2) Escape - shuttling to the other compartment

after the shock began, thus reducing exposure to the shock; (3) NoEscape - not shuttling to the adjacent chamber, thus receiving the

full length of the shock.

Corticosterone (CORT) radioimmunoassayTrunk blood was collected into plastic tubes following

decapitation between 10:00 and 12:00 h. Samples were centri-

fuged at 3000 rpm for 20 min at 4uC. Approximately 1 ml of

serum from each rat was collected into 1.5 ml Eppendorf tubes

and stored at 280uC. The tubes were numbered, but not labeled,

so that analysis of CORT levels was blind to the experimental



procedure followed. CORT levels were assessed using DSL/10/

81000 ELISA kits (DSL, Texas). The sensitivity of the CORT

assay was 12.5 mg/L. Within-assay variation was less than 10% at

100 mg/L, and between-assay variation was less than 15% at

100 mg/L.

The CORT serum concentrations were used to further

corroborate basal stress levels.

Brain extractionAt the PND 60 animals were taken from their home cages and

sacrificed, their brains was extracted, immediately frozen in

isoproponol and stored at 280uC. Bilateral tissue punches with

seventeen-gauge needle of prefrontal cortex (PFC), basolateral

amygdala (BLA), dorsal cornu ammonis (CA) area 1 (dCA1) and

thalamus (TL) were obtained from ,1.5 mm coronal sections cut

in a cryostat at 220uC. The coronal sections were approximately

+4.0 (PFC), 21.8 (BLA) and 22.0 (dCA1 and TL) from bregma,

respectively [92].

The tissues were immediately homogenized in an ice-cold glass/

Teflon homogenizer (885502-0019; KONTES GLASS COMPA-

NY, Vineland, NJ, USA) using 50 Teflon/glass mortar strokes in

300 ml of ice-cold NP-40 lysis buffer (20 mM Tris HCl, 20 mM

EDTA, 1% NP-40, 137 mM NaCl, 10% glycerol, pH 8), with

freshly added with the following protease inhibitors: 0.1 mM

sodium orthovanadate, 1 mg/ml leupeptine, 1.6 mg/ml aprotinin

and 5 mM NaF and 1 mg/ml protease inhibitor cocktail P2714

(from Sigma). 30 ml of each lysate were saved for further protein

concentration by Bradford analysis. The regions were immediately

homogenized with ice cold sodium dodecyl sulfate (SDS) sample

buffer (20% glycerol, 10% b-mercaptoethanol and 20% SDS,

2.33 gr bromophenol blue in 62.5 mM Tris-HCl, pH 6.8) was

added to each remaining lysate, thoroughly mixed and denatured

5 min at 95uC. The denatured proteins were stored at 280uC for

further analysis.

Immunoblot analysisProtein concentration was monitored using Bradford assay, and

equal amounts of loaded protein were verified using b-actin

staining (1:1000, II - a-Goat 1:10000, BIOCHEM; 10%

acrylamide). No differences were observed between the groups

in b-actin concentrations in any of the examined regions.

Individual samples from each region of each rat (20 mg) were

loaded onto 7.5% SDS-PAGE gels. Following electrophoresis gels

were transferred by wet transfer tanks to nitrocellulose membranes

and stained against L1-CAM: (a-NCAM-L1-(C-20) Santa Cruz-

SC-1508-1:1000, II - a-Goat 1:10000, BIOCHEM). The mem-

branes were developed using the enhanced chemiluminescence

light (ECL) (Amersham, Piscataway, NJ) reaction with a charge

coupled device (CCD) camera (XRS BioRad).

QuantificationDensitometric analysis of L1-CAM and b-actin immunoreac-

tivity was conducted using Quantity One 1-D Analysis software.

Each sample was measured relative to the background, and

expression levels were calculated as the Optical Density (OD) ratio

between the b-actin and L1-CAM of each sample.

The results were normalized to Naıve group values.

Statistical AnalysisThe results are expressed as means6SEM. For statistical

analysis, a one-way ANOVA test was applied. For post-hoc

comparisons, the Tukey contrast test was used with an a level of

0.05, unless otherwise noted.


Conceived and designed the experiments: GRL. Performed the experi-

ments: YI. Analyzed the data: YI GRL. Wrote the paper: YI GRL.

References

1. Nestler EJ, Barrot M, DiLeone RJ, Eisch AJ, Gold SJ, et al. (2002) Neurobiology

of depression. Neuron 34(1): 13–25.

2. de Kloet ER, Oitzl MS, Joels M (1999) Stress and cognition: are corticosteroids

good or bad guys? Trends Neurosci 22(10): 422–426.

3. Kim JJ, Diamond DM (2002) The stressed hippocampus, synaptic plasticity and

lost memories. Nat Rev Neurosci 3(6): 453–462.

4. Roozendaal B (2002) Stress and memory: opposing effects of glucocorticoids on

memory consolidation and memory retrieval. Neurobiol Learn Mem 78(3):

578–595.

5. Gold PW, Chrousos GP (2002) Organization of the stress system and its

dysregulation in melancholic and atypical depression: high vs low CRH/NE

states. Mol Psychiatry 7(3): 254–275.

6. McEwen BS (2001) Plasticity of the hippocampus: adaptation to chronic stress

and allostatic load. Ann N Y Acad Sci 933: 265–277.

7. Rygula R, Abumaria N, Flugge G, Fuchs E, Ruther E, et al. (2005) Anhedonia

and motivational deficits in rats: impact of chronic social stress. Behav Brain Res

162(1): 127–134.

8. Fuchs E (2005) Social stress in tree shrews as an animal model of depression: an

example of a behavioral model of a CNS disorder. CNS Spectr 10(3): 182–190.

9. Wilner W (2005) The lone ranger as a metaphor for the psychoanalytic

movement from conscious to unconscious experience. Psychoanal Rev 92(5):

759–776.

10. Diamond DM, Park CR, Heman KL, Rose GM (1999) Exposing rats to a

predator impairs spatial working memory in the radial arm water maze.

Hippocampus 9(5): 542–552.

11. Akirav I, Richter-Levin G (1999) Biphasic modulation of hippocampal plasticity

by behavioral stress and basolateral amygdala stimulation in the rat. J Neurosci

19(23): 10530–10535.

12. McEwen BS (1999) Stress and hippocampal plasticity. Annu Rev Neurosci 22:

105–122.

13. McEwen BS (2002) The neurobiology and neuroendocrinology of stress.

Implications for post-traumatic stress disorder from a basic science perspective.

Psychiatr Clin North Am 25(2): 469–494, ix.

14. McEwen BS (2002) Protective and damaging effects of stress mediators: the good

and bad sides of the response to stress. Metabolism 51(6 Suppl 1): 2–4.

15. Heim C, Nemeroff CB (2001) The role of childhood trauma in the neurobiology

of mood and anxiety disorders: preclinical and clinical studies. Biol Psychiatry

49(12): 1023–1039.

16. Heim C, Plotsky PM, Nemeroff CB (2004) Importance of studying the

contributions of early adverse experience to neurobiological findings indepression. Neuropsychopharmacology 29(4): 641–648.

17. Nemeroff CB (2004) Neurobiological consequences of childhood trauma. J Clin

Psychiatry 65 Suppl 1: 18–28.

18. Gutman DA, Nemeroff CB (2002) Neurobiology of early life stress: rodent

studies. Semin Clin Neuropsychiatry 7(2): 89–95.

19. Avital A, Richter-Levin G (2005) Exposure to juvenile stress exacerbates the

behavioural consequences of exposure to stress in the adult rat.

Int J Neuropsychopharmacol 8(2): 163–173.

20. Tsoory M, Richter-Levin G (2006) Learning under stress in the adult rat is

differentially affected by ‘juvenile’ or ‘adolescent’ stress. Int J Neuropsy-chopharmacol 9(6): 713–728.

21. Tsoory M, Cohen H, Richter-Levin G (2007) Juvenile stress induces apredisposition to either anxiety or depressive-like symptoms following stress in

adulthood. Eur Neuropsychopharmacol 17(4): 245–256.

22. Pryce C, Mohammed A, Feldon J (2002) Environmental manipulations in

rodents and primates. Insights into pharmacology, biochemistry and behaviour.Pharmacol Biochem Behav 73(1): 1–5.

23. Armstrong KR, Clark TR, Peterson MR (1998) Use of Corn-Husk Nesting Materialto Reduce Aggression in Caged Mice. Contemp Top Lab Anim Sci 37(4): 64–66.

24. Chapillon P, Manneche C, Belzung C, Caston J (1999) Rearing environmentalenrichment in two inbred strains of mice: 1. Effects on emotional reactivity.

Behav Genet 29(1): 41–46.

25. Engellenner WJ, Goodlett CR, Burright RG, Donovick PJ (1982) Environmental

enrichment and restriction: effects on reactivity, exploration and maze learningin mice with septal lesions. Physiol Behav 29(5): 885–893.

26. Hansen LT, Berthelsen H (2000) The effect of environmental enrichment on thebehaviour of caged rabbits (Oryctolagus cuniculus). Appl Anim Behav Sci 68(2):

163–178.

27. Sharp J, Zammit T, Azar T, Lawson D (2003) Stress-like responses to common

procedures in individually and group-housed female rats. Contemp Top Lab

Anim Sci 42(1): 9–18.



28. Chamove AS (1989) Cage design reduces emotionality in mice. Lab Anim 23(3):

215–219.

29. Van Loo PL, Van der Meer E, Kruitwagen CL, Koolhaas JM, Van Zutphen LF,

et al. (2004) Long-term effects of husbandry procedures on stress-related

parameters in male mice of two strains. Lab Anim 38(2): 169–177.

30. Huang-Brown KM, Guhad FA (2002) Chocolate, an effective means of oral

drug delivery in rats. Lab Anim (NY) 31(10): 34–36.

31. Bennett EL, Diamond MC, Krech D, Rosenzweig MR (1964) Chemical and

Anatomical Plasticity Brain. Science 146: 610–619.

32. Kempermann G, Kuhn HG, Gage FH (1997) More hippocampal neurons in

adult mice living in an enriched environment. Nature 386(6624): 493–495.

33. Patterson-Kane EG, Hunt M, Harper DN (1999) Behavioral indexes of poor

welfare in laboratory rats. J Appl Anim Welf Sci 2(2): 97–110.

34. Wakabayashi Y, Uchida S, Funato H, Matsubara T, Watanuki T, et al. (2008)

State-dependent changes in the expression levels of NCAM-140 and L1 in the

peripheral blood cells of bipolar disorders, but not in the major depressive

disorders. Prog Neuropsychopharmacol Biol Psychiatry 32(5): 1199–1205.

35. Dityatev A, Dityateva G, Schachner M (2000) Synaptic strength as a function of

post- versus presynaptic expression of the neural cell adhesion molecule NCAM.

Neuron 26(1): 207–217.

36. Welzl H, Stork O (2003) Cell adhesion molecules: key players in memory

consolidation? News Physiol Sci 18: 147–150.

37. Schachner M (1997) Neural recognition molecules and synaptic plasticity. Curr

Opin Cell Biol 9(5): 627–634.

38. Venero C, Tilling T, Hermans-Borgmeyer I, Herrero AI, Schachner M, et al.

(2004) Water maze learning and forebrain mRNA expression of the neural cell

adhesion molecule L1. J Neurosci Res 75(2): 172–181.

39. Sandi C (2004) Stress, cognitive impairment and cell adhesion molecules. Nat

Rev Neurosci 5(12): 917–930.

40. Knafo S, Barkai E, Libersat F, Sandi C, Venero C (2005) Dynamics of olfactory

learning-induced up-regulation of L1 in the piriform cortex and hippocampus.

Eur J Neurosci 21(2): 581–586.

41. Sandi C, Merino JJ, Cordero MI, Touyarot K, Venero C (2001) Effects of

chronic stress on contextual fear conditioning and the hippocampal expression of

the neural cell adhesion molecule, its polysialylation, and L1. Neuroscience

102(2): 329–339.

42. Merino JJ, Cordero MI, Sandi C (2000) Regulation of hippocampal cell

adhesion molecules NCAM and L1 by contextual fear conditioning is dependent

upon time and stressor intensity. Eur J Neurosci 12(9): 3283–3290.

43. Fischer G, Kunemund V, Schachner M (1986) Neurite outgrowth patterns in

cerebellar microexplant cultures are affected by antibodies to the cell surface

glycoprotein L1. J Neurosci 6(2): 605–612.

44. Cohen NR, Taylor JS, Scott LB, Guillery RW, Soriano P, et al. (1998) Errors in

corticospinal axon guidance in mice lacking the neural cell adhesion molecule

L1. Curr Biol 8(1): 26–33.

45. Wood PM, Schachner M, Bunge RP (1990) Inhibition of Schwann cell

myelination in vitro by antibody to the L1 adhesion molecule. J Neurosci 10(11):

3635–3645.

46. Kenwrick S, Watkins A, De Angelis E (2000) Neural cell recognition molecule

L1: relating biological complexity to human disease mutations. Hum Mol Genet

9(6): 879–886.

47. Fransen E, Van Camp G, Vits L, Willems PJ (1997) L1-associated diseases:

clinical geneticists divide, molecular geneticists unite. Hum Mol Genet 6(10):

1625–1632.

48. Laifenfeld D, Karry R, Klein E, Ben-Shachar D (2005) Alterations in cell

adhesion molecule L1 and functionally related genes in major depression: a

postmortem study. Biol Psychiatry 57(7): 716–725.

49. Laifenfeld D, Karry R, Grauer E, Klein E, Ben-Shachar D (2005)

Antidepressants and prolonged stress in rats modulate CAM-L1, laminin, and

pCREB, implicated in neuronal plasticity. Neurobiol Dis 20(2): 432–441.

50. Laifenfeld D, Klein E, Ben-Shachar D (2002) Norepinephrine alters the

expression of genes involved in neuronal sprouting and differentiation: relevance

for major depression and antidepressant mechanisms. J Neurochem 83(5):

1054–1064.

51. Price JL, Carmichael ST, Drevets WC (1996) Networks related to the orbital and

medial prefrontal cortex; a substrate for emotional behavior? Prog Brain Res

107: 523–536.

52. Leventopoulos M, Ruedi-Bettschen D, Knuesel I, Feldon J, Pryce CR, et al.

(2007) Long-term effects of early life deprivation on brain glia in Fischer rats.

Brain Res 1142: 119–126.

53. Ichise M, Vines DC, Gura T, Anderson GM, Suomi SJ, et al. (2006) Effects of

early life stress on [11C]DASB positron emission tomography imaging of

serotonin transporters in adolescent peer- and mother-reared rhesus monkeys.

J Neurosci 26(17): 4638–4643.

54. Becker K, Abraham A, Kindler J, Helmeke C, Braun K (2007) Exposure to

neonatal separation stress alters exploratory behavior and corticotropin releasing

factor expression in neurons in the amygdala and hippocampus. Dev Neurobiol

67(5): 617–629.

55. Matsumoto M, Togashi H, Konno K, Koseki H, Hirata R, et al. (2008) Early

postnatal stress alters the extinction of context-dependent conditioned fear in

adult rats. Pharmacol Biochem Behav 89(3): 247–252.

56. Mattson MP, Chan SL (2001) Dysregulation of cellular calcium homeostasis in

Alzheimer’s disease: bad genes and bad habits. J Mol Neurosci 17(2): 205–224.

57. Rice CJ, Sandman CA, Lenjavi MR, Baram TZ (2008) A novel mouse model foracute and long-lasting consequences of early life stress. Endocrinology.

58. Guilarte TR, Toscano CD, McGlothan JL, Weaver SA (2003) Environmentalenrichment reverses cognitive and molecular deficits induced by developmental

lead exposure. Ann Neurol 53(1): 50–56.

59. Hogg S (1996) A review of the validity and variability of the elevated plus-mazeas an animal model of anxiety. Pharmacol Biochem Behav 54(1): 21–30.

60. Katz RJ, Roth KA, Carroll BJ (1981) Acute and chronic stress effects on openfield activity in the rat: implications for a model of depression. Neurosci

Biobehav Rev 5(2): 247–251.

61. Bevins RA, Bardo MT (1999) Conditioned increase in place preference by access

to novel objects: antagonism by MK-801. Behav Brain Res 99(1): 53–60.

62. Bardo MT, Dwoskin LP (2004) Biological connection between novelty- anddrug-seeking motivational systems. Nebr Symp Motiv 50: 127–158.

63. Overmier JB, Seligman ME (1967) Effects of inescapable shock upon subsequentescape and avoidance responding. J Comp Physiol Psychol 63(1): 28–33.

64. Seligman ME, Maier SF (1967) Failure to escape traumatic shock. J Exp Psychol74(1): 1–9.

65. Brunson KL, Kramar E, Lin B, Chen Y, Colgin LL, et al. (2005) Mechanisms of

late-onset cognitive decline after early-life stress. J Neurosci 25(41): 9328–9338.

66. Willner P, Muscat R, Papp M (1992) Chronic mild stress-induced anhedonia: a

realistic animal model of depression. Neurosci Biobehav Rev 16(4): 525–534.

67. Roth KA, Katz RJ (1979) Stress, behavioral arousal, and open field activity–a

reexamination of emotionality in the rat. Neurosci Biobehav Rev 3(4): 247–263.

68. Pryce CR, Ruedi-Bettschen D, Dettling AC, Weston A, Russig H, et al. (2005)

Long-term effects of early-life environmental manipulations in rodents and

primates: Potential animal models in depression research. Neurosci BiobehavRev 29(4–5): 649–674.

69. van Praag H, Kempermann G, Gage FH (2000) Neural consequences ofenvironmental enrichment. Nat Rev Neurosci 1(3): 191–198.

70. Escorihuela RM, Tobena A, Fernandez-Teruel A (1994) Environmentalenrichment reverses the detrimental action of early inconsistent stimulation

and increases the beneficial effects of postnatal handling on shuttlebox learning

in adult rats. Behav Brain Res 61(2): 169–173.

71. Mello DM, Marcinichen DR, Madruga D, Branco R, Paschoalini MA, et al.

(2007) Involvement of NK1 receptors in metabolic stress markers after thecentral administration of substance P. Behav Brain Res 181(2): 232–238.


Neuropsychopharmacology 23(5): 477–501.

73. Wong ML, Kling MA, Munson PJ, Listwak S, Licinio J, et al. (2000) Pronounced

and sustained central hypernoradrenergic function in major depression withmelancholic features: relation to hypercortisolism and corticotropin-releasing

hormone. Proc Natl Acad Sci U S A 97(1): 325–330.

74. Maness PF, Schachner M (2007) Neural recognition molecules of the

immunoglobulin superfamily: signaling transducers of axon guidance and

neuronal migration. Nat Neurosci 10(1): 19–26.

75. Brummendorf T, Kenwrick S, Rathjen FG (1998) Neural cell recognition

molecule L1: from cell biology to human hereditary brain malformations. CurrOpin Neurobiol 8(1): 87–97.

76. Hortsch M (2000) Structural and functional evolution of the L1 family: are fouradhesion molecules better than one? Mol Cell Neurosci 15(1): 1–10.

77. Law JW, Lee AY, Sun M, Nikonenko AG, Chung SK, et al. (2003) Decreased

anxiety, altered place learning, and increased CA1 basal excitatory synaptictransmission in mice with conditional ablation of the neural cell adhesion

molecule L1. J Neurosci 23(32): 10419–10432.

78. Jaferi A, Bhatnagar S (2006) Corticosterone can act at the posterior

paraventricular thalamus to inhibit hypothalamic-pituitary-adrenal activity inanimals that habituate to repeated stress. Endocrinology 147(10): 4917–4930.

79. Roozendaal B, Brunson KL, Holloway BL, McGaugh JL, Baram TZ (2002)

Involvement of stress-released corticotropin-releasing hormone in the basolateralamygdala in regulating memory consolidation. Proc Natl Acad Sci U S A 99(21):

13908–13913.

80. Sandi C, Bisaz R (2007) A model for the involvement of neural cell adhesion

molecules in stress-related mood disorders. Neuroendocrinology 85(3): 158–176.

81. Styren SD, Miller PD, Lagenaur CF, DeKosky ST (1995) Alternate strategies in

lesion-induced reactive synaptogenesis: differential expression of L1 in two

populations of sprouting axons. Exp Neurol 131(2): 165–173.

82. Jucker M, D’Amato F, Mondadori C, Mohajeri H, Magyar J, et al. (1996)

Expression of the neural adhesion molecule L1 in the deafferented dentate gyrus.Neuroscience 75(3): 703–715.

83. Brook GA, Houweling DA, Gieling RG, Hermanns T, Joosten EA, et al. (2000)Attempted endogenous tissue repair following experimental spinal cord injury in

the rat: involvement of cell adhesion molecules L1 and NCAM? Eur J Neurosci

12(9): 3224–3238.

84. Venero C, Tilling T, Hermans-Borgmeyer I, Schmidt R, Schachner M, et al.

(2002) Chronic stress induces opposite changes in the mRNA expression of thecell adhesion molecules NCAM and L1. Neuroscience 115(4): 1211–1219.

85. Pham TM, Ickes B, Albeck D, Soderstrom S, Granholm AC, et al. (1999)

Changes in brain nerve growth factor levels and nerve growth factor receptors inrats exposed to environmental enrichment for one year. Neuroscience 94(1):

279–286.

86. Rampon C, Jiang CH, Dong H, Tang YP, Lockhart DJ, et al. (2000) Effects of

environmental enrichment on gene expression in the brain. Proc Natl AcadSci U S A 97(23): 12880–12884.



87. Mohammed AH, Henriksson BG, Soderstrom S, Ebendal T, Olsson T, et al.

(1993) Environmental influences on the central nervous system and theirimplications for the aging rat. Behav Brain Res 57(2): 183–191.

88. Sharp PE, McNaughton BL, Barnes CA (1985) Enhancement of hippocampal

field potentials in rats exposed to a novel, complex environment. Brain Res339(2): 361–365.

89. Sharp PE, Barnes CA, McNaughton BL (1987) Effects of aging onenvironmental modulation of hippocampal evoked responses. Behav Neurosci

101(2): 170–178.

90. Avital A, Richter-Levin G, Leschiner S, Spanier I, Veenman L, et al. (2001)

Acute and repeated swim stress effects on peripheral benzodiazepine receptors inthe rat hippocampus, adrenal, and kidney. Neuropsychopharmacology 25(5):

669–678.

91. Maroun M, Richter-Levin G (2003) Exposure to acute stress blocks theinduction of long-term potentiation of the amygdala-prefrontal cortex pathway

in vivo. J Neurosci 23(11): 4406–4409.92. Paxinos G, Watson C (1996) The Rat Brain in Stereotaxic Coordinate, Compact

Third Edition. New York: Academic Press.



Exercise Improves Cognitive Responses to PsychologicalStress through Enhancement of Epigenetic Mechanismsand Gene Expression in the Dentate GyrusAndrew Collins, Louise E. Hill, Yalini Chandramohan, Daniel Whitcomb, Susanne K. Droste,

Johannes M. H. M. Reul*

Henry Wellcome Laboratories for Integrative Neuroscience and Endocrinology, University of Bristol, Bristol, United Kingdom

Abstract

Background: We have shown previously that exercise benefits stress resistance and stress coping capabilities. Furthermore,we reported recently that epigenetic changes related to gene transcription are involved in memory formation of stressfulevents. In view of the enhanced coping capabilities in exercised subjects we investigated epigenetic, gene expression andbehavioral changes in 4-weeks voluntarily exercised rats.

Methodology/Principal Findings: Exercised and control rats coped differently when exposed to a novel environment.Whereas the control rats explored the new cage for the complete 30-min period, exercised animals only did so during thefirst 15 min after which they returned to sleeping or resting behavior. Both groups of animals showed similar behavioralresponses in the initial forced swim session. When re-tested 24 h later however the exercised rats showed significantly moreimmobility behavior and less struggling and swimming. If rats were killed at 2 h after novelty or the initial swim test, i.e. atthe peak of histone H3 phospho-acetylation and c-Fos induction, then the exercised rats showed a significantly highernumber of dentate granule neurons expressing the histone modifications and immediate-early gene induction.

Conclusions/Significance: Thus, irrespective of the behavioral response in the novel cage or initial forced swim session, theimpact of the event at the dentate gyrus level was greater in exercised rats than in control animals. Furthermore, in view ofour concept that the neuronal response in the dentate gyrus after forced swimming is involved in memory formation of thestressful event, the observations in exercised rats of enhanced neuronal responses as well as higher immobility responses inthe re-test are consistent with the reportedly improved cognitive performance in these animals. Thus, improved stresscoping in exercised subjects seems to involve enhanced cognitive capabilities possibly resulting from distinct epigeneticmechanisms in dentate gyrus neurons.

Citation: Collins A, Hill LE, Chandramohan Y, Whitcomb D, Droste SK, et al. (2009) Exercise Improves Cognitive Responses to Psychological Stress throughEnhancement of Epigenetic Mechanisms and Gene Expression in the Dentate Gyrus. PLoS ONE 4(1): e4330. doi:10.1371/journal.pone.0004330



Copyright: � 2009 Collins et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was supported by the Medical Research Council and the Neuroendocrinology Charitable Trust. The funders had no role in study design, datacollection and analysis, decision to publish, or preparation of the manuscript.



Introduction

It is now well established that regular physical exercise has a

positive impact on a range of biological systems, including the

brain [1–6]. The resulting antidepressant-like and anxiolytic

effects have led to exercise being proposed as an effective co-

treatment (i.e. in addition to drug and behavioral therapies) for

anxious and depressed patients [7–10]. Our previous work has

indicated that voluntarily exercised animals show improved stress-

coping in the face of physically demanding or psychological

challenges. These improved stress-coping responses and strategies

surfaced as more pertinent, adaptive responses of the hypotha-

lamic-pituitary-adrenocortical (HPA) axis [11–13], improved sleep

quality and enhanced stress resistance of sleep/EEG profiles [14],

and decreased anxiety-related behavior and impulsivity in

voluntary exercised mice and rats [15] relative to sedentary

control animals. It is thought that these physiological and

behavioral changes may well be of relevance for the clinical

effects of exercise in patients suffering from stress-related mental

disorders [7–9,16,17]. However, currently the underlying mech-

anisms of these beneficial effects of exercise are still largely

unknown. In the present study it was our aim to gain more insight

into the neurobiological basis of the enhanced stress-coping

capabilities shown by voluntarily exercised animals.

In particular, we investigated the role of epigenetic mechanisms

in the brain involved in transcriptional activation in coordinating

adaptive behavioral responses to stressful events. Epigenetic

mechanisms comprise of post-translational modifications of DNA

and histone proteins within the chromatin structure, such as the

methylation of DNA, and the acetylation, methylation, phosphor-

ylation and other modifications of the N-terminal tails of distinct

histone molecules [18]. Specifically of interest is the phosphory-

lation of serine-10 (Ser10) combined with the acetylation of lysine-

14 (Lys14) in the N-terminal tail of histone H3 (i.e. P(Ser10)-


Ac(Lys14)-H3) as this modification is thought to be involved in the

local opening of condensed chromatin, thereby allowing the

transcriptional induction of specific, hitherto silent genes [19,20].

We postulated that these specific chromatin modifications are

involved in triggering gene expression responses required for

physiological and functional adjustments in neurons involved in

the cognitive processing of stressful events [21].

Indeed, we could show that psychologically stressful events such

as forced swimming and exposure to a novel environment

enhances the phosphorylation and phospho-acetylation of histone

H3 in a distinct population of dentate gyrus granule neurons in the

hippocampus [21–24]. The response to stress was transient,

peaking at 1–2 hours and coincided with the induction of c-Fos

specifically in these neurons [23,24]. Previous in vitro work has

indeed shown that phospho-acetylation of histone H3 of the

promotor region of the c-Fos gene occurs at induction of this

immediate-early gene [19]. Furthermore, we obtained evidence

that the phospho-acetylation of histone H3 and the induction of c-

Fos is brought about by at least two, concurrently acting signaling

pathways being the glucocorticoid receptor (GR) and the NMDA-

R/ERK/MSK pathway (NMDA-R, N-methyl-D-aspartate recep-

tor; ERK, extracellular signal-regulated kinase; MSK, mitogen-

and stress-activated kinase) [21–24]. The epigenetic and gene

expression responses in the dentate gyrus are thought to be

involved in learning to cope with stressful, traumatic events as we

recently obtained substantial evidence that these mechanisms are

required for the formation of memories of the events

[21,22,24,25].

Here we investigated whether changes in histone H3 phospho-

acetylation and gene expression responses would be involved in

the enhanced stress coping capabilities seen in exercised subjects.

Therefore, we subjected exercised and control rats to novelty

exposure and forced swimming and investigated changes in

dentate gyrus histone H3 phospho-acetylation and c-Fos expres-

sion, and acute behavioral responses as well as memory formation

of the event.


AnimalsMale Sprague-Dawley rats (140–160 g; purchased from Harlan,

(Oxon, UK) were singly housed under standard lighting (14:10-

hour light/dark cycle), humidity (50–60%) and temperature (22–

23uC) conditions. Food and water were available ad libitum.

Voluntary Exercise ParadigmAfter habituation to the housing conditions for 5 days, the

experimental group was allowed free access to a running wheel

(diameter 34 cm) in their home cages for a period of four weeks.

The rats ran approximately 4–7 km per night which is in

agreement with other reports [13,26]. The housing of sedentary

(i.e. control) animals remained unchanged. All animal experiments

were approved by the UK Home Office. Voluntary wheel running

is not regarded as a form of stereotypic behavior [27] because,

unlike other reported locomotor stereotypes, it is not expressed at

the cost of resting behavior such as sleep [14] as is the case in other

reported locomotor stereotypies [28,29].

All experiments were carried out four weeks after voluntary

exercise (or non-exercise) and between 8:00 and 12:00 h. For

killing, individual rats were quickly anaesthetized (,15 sec) in a

glass jar containing isoflurane (Merial Animal Health Ltd., UK)

vapor, after which animals were decapitated immediately and

their whole brains removed, snap frozen in isopentane at 240uCand deep-frozen in dry ice. Brains were stored at 280uC.

Novel Environment ExposureTo induce novelty stress, as reported before [23,30,31] rats were

placed singly for 30 min in a new cage (i.e. a clean cage with new

sawdust but no food or water) in a separate room with identical

environmental conditions except for increased light intensity (500

lx, holding conditions:100 lx). Behavior of rats was recorded using

digital cameras and a hard disk recorder and later scored every

10 sec throughout the total 30 min duration of the test. The

following behaviors were scored: lying (includes sleeping), rearing,

stationary (standing or sitting), walking, grooming, scratching and

burrowing behavior. Thereafter, rats were returned to their home

cages and then placed in a recovery room (i.e. a room with

identical environment and light conditions as the original holding

room) until they were killed at 2 h after the onset of the novelty

challenge.

Forced SwimmingFor forced swimming, as reported before [22,24,32,33] rats

were placed in a glass beaker (height 35 cm, diameter 21.7 cm)

containing 25uC water (depth of 21 cm) for 15 min. Thereafter,

the animals were dried with a towel and returned to their home

cages and placed in a recovery room (see above) until they were

killed at 2 h after the start of the forced swimming procedure. For

both forced swimming and novelty stress, the 2 h time point was

chosen as it has been described previously as the peak of P(Ser10)-

Ac(Lys14)-H3 and c-Fos expression [23,24].

For determination of forced swimming-induced acquisition of

behavioral immobility, separate groups of rats were forced to swim

for 15 min (as described above) and 24 h later were subjected

again to forced swimming, in this case for 5 min (i.e. the ‘re-test’;

water at 25uC). Behavioral immobility (or floating) is a behavioral

state in which the animal retains an immobile posture displaying

only enough movement to keep the head above water. Behavior in

the initial test and re-test was recorded as described above and

scored at a later time point. Three distinct behaviors were scored

[immobility, struggling (also called climbing, in which the animal

makes vertical movements along the wall of the beaker) and

swimming (horizontal movements in the water)] every 10 s for the

entire duration of the test and the re-test. In all cases scoring was

conducted in a blinded fashion.

ImmunohistochemistryBrain tissues were cut into coronal sections using a cryostat and

mounted on glass slides (Superfrost, Fisher, Loughborough, UK)

previously coated with poly-L-lysine (Sigma). Sections of rat brain

were taken from the nucleus accumbens, PVN and the dorsal

hippocampus in accordance with the atlas of Paxinos and Watson

(1986) (AP co-ordinates: between AP 1.60 mm and 1.00 mm from

Bregma for nucleus accumbens; between 21.80 mm and

22.12 mm from Bregma for PVN; between 22.92 mm and

23.96 mm from Bregma for the dorsal hippocampus). Sections

were stored at 220uC until use. Storage at this temperature does

not affect levels of P(Ser10)-Ac(Lys14)-H3 or c-fos in sections.

Immunohistochemical staining using diaminobenzidine (DAB)

was conducted according to standard protocol as previously

described [23,24,34] In brief, the brain sections were fixed in 4%

paraformaldehyde in 16 phosphate-buffered saline (PBS) for

30 min. Thereafter, endogenous peroxidase activity was blocked

by a 30-min incubation in 0.6% H2O2. To improve antibody

penetration of the tissue, the sections were incubated for 1 h in

0.2% Triton X-100, followed by blocking of sections with 5% goat

serum in PBS to prevent non-specific binding. The primary

antibodies were diluted in 1.5% goat serum/PBS. Rabbit

polyclonal antibody against P(Ser10)-Ac(Lys14)-H3 (dilution

Exercise-Epigenetics-Behavior


1:1,000) was purchased from Upstate (Charlottesville, VA, USA)

and the rabbit anti-c-Fos antibody (used dilution 1:10,000) was

purchased from Calbiochem (Nottingham, UK). Incubation with

primary antibodies occurred overnight at room temperature.

Biotinylated secondary antibody, avidin-biotin-peroxidase com-

plex and DAB/Ni+ substrate (Elite-ABC and DAB detection kits;

Vector Laboratories, Burlingame, CA, USA) for the development

of immunostaining were used according to company instructions.

After dehydration in ethanol, sections were finally mounted using

Histomount (Fisher) and coverslipped.

Data AnalysisThe numbers of both P(Ser10)-Ac(Lys14)-H3+ neurons and c-

Fos-positive (c-Fos+) neurons in the dentate gyrus (six sections per

animal) were counted by an individual blind to the treatment. The

location of each positive neuron was distinguished between the

dorsal and ventral blade of the dentate gyrus. For analysis of the

nucleus accumbens, two sample areas dorso-lateral to the anterior

commissure were used to count positive neurons. For each

antibody and brain region at least 2 assays were performed. The

assays provided similar results and data of one assay is presented

here. The experimental data were statistically evaluated using

ANOVA and, if significant, followed by the post-hoc Bonferroni

test as appropriate. Behavioral data of the novelty challenge were

statistically tested using ANOVA with repeated measures followed

by Student’s t-test in appropriate cases. Forced swim test data were

evaluated using Student’s t-test. The experimental data were

considered to be statistically different from control data when

P,0.05.

Results

Novel environment-induced changes in P(Ser10)-Ac(Lys14)-H3+ and c-Fos+ neurons in the dentate gyrus ofexercised and control rats

We first examined whether, in terms of histone H3 phospho-

acetylation and c-Fos induction, exercised rats would respond

differently than sedentary control animals to exposure to a mild

psychological challenge such as a novel environment. Figure 1A–

D shows representative immunohistochemical images of the dorsal

blade of the dentate gyrus of control rats killed under baseline

conditions or 2 h after novelty stress. Nuclear staining of phospho-

acetylated histone H3 immunoreactivity can be seen in distinct

granule neurons but no immunostaining was observed in

hippocampal pyramidal neurons except for very few CA3

pyramidal neurons (data not shown). As reported before [22–

24], only few P(Ser10)-Ac(Lys14)-H3+ neurons were found

elsewhere in the brain. As previous in vitro [19] and in vivo

[23,24] research has shown that histone H3 phospho-acetylation is

associated with the induction of immediate-early gene products

such as c-Fos, we investigated whether exercised rats showed

distinct responses in this gene product to novelty stress as well. The

staining pattern for both P(Ser10)-Ac(Lys14)-H3+ and c-Fos+

neurons in the dentate gyrus was sparse which concurs with

previous observations [22–24]. Novelty stress resulted in an

increase in the number of both P(Ser10)-Ac(Lys14)-H3+ and c-

Fos+ neurons (Fig. 1C, D).

In terms of P(Ser10)-Ac(Lys14)-H3+ neurons in the whole

dentate gyrus, novel environment exposure resulted in a significant

increase in both the control and exercised animals, but the

increase in exercised rats was substantially greater than that in the

control animals (Fig. 2A). If the dorsal and ventral blade of the

dentate gyrus were considered separately, a different picture

emerged. Considering the dorsal blade separately, the response in

P(Ser10)-Ac(Lys14)-H3+ neurons to a novelty challenge was

similar in exercised and control animals (Fig. 2B). However,

analysis of the ventral blade showed that novelty stress resulted in

higher numbers of P(Ser10)-Ac(Lys14)-H3+ neurons in exercised

animals than in control animals (Fig. 2C). Apparently, these higher

numbers in the ventral blade of exercised rats were the principal

reason for the enhanced histone H3 phospho-acetylation response

observed in the whole dentate gyrus after novelty stress. In line

with previous findings [23], the P(Ser10)-Ac(Lys14)-H3+ neurons

were mainly found in the middle and superficial aspects of the

granular cell layer, with an overall greater abundance in the dorsal

blade (data not shown).

With regard to c-Fos, we found that considering the whole

dentate gyrus novelty stress only evoked a significant increase in c-

Fos+ neurons in the dentate gyrus of exercised animals (Fig. 2D).

Surprisingly, control animals did not show a significant increase

over baseline levels when total numbers in the dentate gyrus were

considered. However, when considering the dorsal and ventral

blade separately and comparing values to those observed under

baseline conditions, novelty stress evoked an increase in cFos+

neurons in the dorsal blade whereas decreased numbers were

observed in the ventral blade (Fig. 2E, F). A more uniform

response was observed in the exercised rats where the same

stressor caused an increase in the number of cFos+ neurons in the

dorsal blade as well as in the ventral blade (Fig. 2E, F).

Novel environment-induced changes in behavior ofexercising animals

To assess whether differential behavioral coping strategies are

involved in the distinct histone modification and gene expression

responses in exercised and control rats, we scored various

behaviors displayed by the animals in the novel environment

(Fig. 3). Rats showed mainly exploratory behaviors such as walking

(Fig. 3A), rearing (Fig. 3B) and burrowing (data not shown) when

introduced in the novel cage. There were no differences between

the experimental groups at this early stage. The exercising rats

showed more stationary behavior during the first 10 min in the

novel cage (Fig. 3C). This observation corresponds with our earlier

observations in exercised mice. When placed in an open field these

animals also initially show more stationary behavior as a result of

increased vigilance and decreased impulsiveness [15].

Over the course of time in the novel cage the exercised rats

showed significantly less walking and rearing behavior suggesting a

gradual decline in exploratory behavior in these animals (Fig. 3A,

B). Indeed, the exercised rats lay down much more than the

control animals during the second half of the novel cage test

(Fig. 3D). It appeared that some animals slept whilst lying down

but this cannot be assessed with certainty as sleep/EEG

measurements were not conducted. The control rats maintained

walking and stationary, and to some extent, rearing behavior

throughout and until the end of the novel cage test. Rats did not

show differences in grooming, scratching and burrowing behavior

(data not shown). Thus, whereas the control rats kept exploring the

novel environment for the complete exposure time, the exercised

animals appeared to lose interest over time and returned to their

normal behavior at that time of day, being resting or sleeping.

Forced swimming-induced changes in P(Ser10)-Ac(Lys14)-H3+ and c-Fos+ neurons in the dentate gyrus

In contrast to the novelty stress paradigm which largely allows

passive coping strategies, forced swimming is a challenge that

enforces active coping styles such as struggling and swimming as

well as adaptive coping strategies such as immobility or floating



behavior. Furthermore, whereas the novelty stress paradigm only

addresses acute behavioral responses to the situation, the forced

swim paradigm includes behavioral responses to the acute

situation (i.e. the initial forced swim test) as well as behavioral

responses upon re-exposure to the forced swim challenge (i.e. the

re-test). Previously, we have shown that histone H3 phospho-

acetylation and c-Fos induction in dentate granule neurons is

required for the acquisition of immobility behavior seen in the re-

test [22,24].

Figure 1 shows representative immunohistochemical images of

the dorsal blade of the dentate gyrus of rats killed under baseline

conditions or 2 h after forced swimming. Forced swimming indeed

evoked an increase in histone H3 phospho-acetylation and c-Fos in

dentate granule neurons as compared to control animals (Fig. 1A,

B, E, F). Furthermore, counting of the immuno-stained neurons

revealed that the increase in the number of P(Ser10)-Ac(Lys14)-

H3+ neurons after forced swimming was significantly higher in the

exercised rats than in the control animals (Fig. 4). Separate

analyses of the dorsal and ventral blades showed that the forced

swimming-induced response was confined to the dorsal blade,

where there were almost twice as many (Ser10)-Ac(Lys14)-H3+

neurons in the exercised animals than in the controls (Fig. 4B).

There was no effect of forced swimming on the number of

P(Ser10)-Ac(Lys14)-H3+ neurons in the ventral blade (Fig. 4C).

Analysis of c-Fos immunostaining demonstrated a pattern of

forced swimming-induced changes in control and exercised rats

that was largely similar to that found for P(Ser10)-Ac(Lys14)-H3,

at least if the whole dentate gyrus was considered (Fig. 4D). Thus,

forced swimming resulted in enhanced c-Fos expression in the

dentate gyrus of both control and exercised animals but the

response in the exercised group was significantly higher. However,

in the dorsal blade and in contrast to the P(Ser10)-Ac(Lys14)-H3

data, this enhanced response was not as apparent because the

difference between the stressed control and stressed exercised

animals only showed a trend (Fig. 4E). The c-Fos data for the

ventral blade were parallel to those found for the P(Ser10)-

Ac(Lys14)-H3 data (Fig. 4F).

Exercised rats show improved adaptive behavior afterforced swimming

Behavior of rats during the initial forced swim test and the re-

test were recorded and scored to assess whether changes in histone

Figure 1. Representative images of anti-P(Ser10)-Ac(Lys14) (left panels) and anti-c-Fos (right panels) immuno-staining in the dorsalblade of the dentate gyrus of control rats under baseline conditions or at 2 h after novelty exposure or forced swimming. Blackarrows indicate positive nuclear immuno-staining. Immunohistochemistry and the challenge tests were conducted as described in the Materials andMethods.doi:10.1371/journal.pone.0004330.g001



H3 phospho-acetylation and c-Fos expression in the exercised rats

were related to behavioral changes in this paradigm. In the initial

forced swim test, control and exercised rats showed largely similar

behaviors (Fig. 5A–C), with similar levels of immobility and

struggling behavior but slightly more swimming behavior in the

exercised animals than in the controls. However, in the retest 24 h

later, the exercised rats showed significantly higher immobility and

lower swimming and struggling scores than the controls (Fig. 5D–

Figure 2. Effect of novelty exposure on the number of P(Ser10)-Ac(Lys14)+ (Left panels A, B, and C) and c-Fos+ neurons (right panelsD, E and F) in the dentate gyrus of control, sedentary and 4-weeks exercised rats. A and D show data on total number of immuno-positiveneurons in the dentate gyrus whereas in B and E and in C and F data are depicted separately for the dorsal blade and the ventral blade, respectively.Rats were allowed to voluntarily exercise by giving them access to a running wheel in their home cage. Data are expressed as the number ofimmuno-positive neurons (mean6SEM, n = 6) in the dentate gyrus of a 10-mm section. For additional information, see Materials and Methods.Statistical analyses: Two-way ANOVA: A, Effect of exercise: F(1,20) = 12.823, P = 0.002, Effect of novelty: F(1,20) = 93.616, P,0.0005, Interaction exercisex novelty: F(1,20) = 4.808, P = 0.043; B, Effect of Novelty: F(1,20) = 163.33, P,0.0005; C, Effect of exercise: F(1,20) = 13.346, P = 0.002; D, Effect ofexercise: F(1,20) = 7.392, P = 0.013, Effect of novelty: F(1,20) = 15.921, P = 0.001, Interaction exercise x novelty: F(1,20) = 7.203, P = 0.014; E, Effect ofnovelty: F(1,20) = 33.302, P,0.0005; F, Effect of exercise: F(1,20) = 11.827, P = 0.003, interaction exercise x novelty: F(1,20) = 16.875, P = 0.001. *, P,0.05,compared to the respective Baseline group; +, P,0.05, compared to the respective Control group, post-hoc Bonferroni test.doi:10.1371/journal.pone.0004330.g002



F). Given that the immobility response in the re-test is regarded as

a reflection of the strength of the memories formed after the initial

forced swim event, the enhanced immobility response in the

exercised rats suggests that these animals have improved cognitive

and adaptive abilities to cope with psychologically stressful events.

Forced Swimming-induced changes in cFos+ neurons inthe nucleus accumbens

The nucleus accumbens is a mesolimbic brain region involved

in behavioral responses to stress. It has also been suggested to act

as a neuroanatomical substrate for immobility/floating behavior in

the forced swim test [35]. Therefore, we investigated whether

forced swimming would lead to a differential c-Fos induction in

the nucleus accumbens of exercised rats as compared to control

animals. These animals were the same as those used for the

dentate gyrus analyses. Figure 6A and B show representative

images of the nucleus accumbens of rats killed under baseline

conditions or at 2 h after forced swimming. We counted c-Fos+

neurons in an area of the nucleus accumbens showing highest

numbers of immuno-positive neurons, comprising parts of both

the core and shell region. Forced swimming induced a marked

increase in the number of c-Fos+ neurons in the nucleus

accumbens of both control and exercised animals. The increase,

however, was similar in both groups (Fig. 6C). The similar degree

of c-Fos induction in exercised and control animals suggests that

any differential behavioral responses to stress in the exercisers are

unlikely to be mediated by the nucleus accumbens.

We also analyzed the hypothalamic paraventricular nucleus

(PVN), a stress-sensitive nucleus that plays a principal role in

hypothalamic-pituitary-adrenocortical (HPA) axis regulation [17].

We found that the increases in c-Fos levels were similar in control

and exercised rats after forced swimming (data not shown).

Discussion

In the present study, we show that exercised rats present

improved coping responses and memory performance after

exposure to a novel environment or a forced swim test.

Furthermore, these behavioral changes were associated with

enhanced responses in histone H3 phospho-acetylation and c-

Figure 3. Behavior of control and exercised rats during exposure to a novel environment, i.e. a new cage in a brightly lit (500 lx)room. Changes in walking (A), rearing (B), stationary (C) and lying behavior (D) were scored every 10 sec throughout the 30-min novelty exposure.Data were binned in 5-min time bins and expressed as behavioral counts (mean6SEM, n = 6). Statistical analyses: Two-way ANOVA with repeatedmeasures: A, Effect of time: F(5,45) = 5.387, P = 0.001, Effect of exercise: (F1,9) = 4.739, P = 0.057, Interaction time x exercise: F(5,45) = 1.331, notsignificant; B, Effect of time: F(5,45) = 15.545, P,0.0005, Effect of exercise: F(1,9) = 14.263, P = 0.004, Interaction time x exercise: F(5,45) = 1.305, notsignificant; C, Effect of time: F(5,45) = 0.153, not significant, Effect of exercise: F(1,9) = 1.878, not significant, Interaction time x exercise: F(5,45) = 7.881,P,0.0005; D, Effect of time: F(5,45) = 11.529, P,0.0005, Effect of exercise: F(1,9) = 11.332, P = 0.008, Interaction time x exercise: F(5,45) = 9.130,P,0.0005. *, P,0.05, Student’s t-test.doi:10.1371/journal.pone.0004330.g003



Fos in dentate gyrus granule neurons. In contrast, the immediate

early gene responses in the nucleus accumbens and PVN were

similar in exercised and sedentary control animals. Therefore, it

appears that changes in histone H3 phospho-acetylation and gene

expression responses in the dentate gyrus are involved in the

enhanced stress-coping capabilities seen in exercised animals.

Exercised rats adopted a different coping strategy than control

animals when faced with the novelty challenge. They exhibited less

emotionality than their non-running counterparts, exploring their

new environment initially but then settling down. In contrast, the

non-running, sedentary rats remained active virtually for the full

30 min. These findings correspond with our previous work in

Figure 4. Effect of forced swimming on the number of P(Ser10)-Ac(Lys14)+ (Left panels A, B, and C) and c-Fos+ neurons (right panelsD, E and F) in the dentate gyrus of control, sedentary and 4-weeks exercised rats. A and D show data on total number of immuno-positiveneurons in the dentate gyrus whereas in B and E and in C and F data are depicted separately for the dorsal blade and the ventral blade, respectively.Data are expressed as the number of immuno-positive neurons (mean6SEM, n = 6) in the dentate gyrus of a 10-mm section. For additionalinformation, see Materials and Methods. Statistical analyses: Two-way ANOVA: A, Effect of exercise: F(1,24) = 3.495, P = 0.077, Effect of forcedswimming: F(1,24) = 32.292, P,0.0005, Interaction exercise x forced swimming: F(1,24) = 4.135, P = 0.056; B, Effect of exercise: F(1,24) = 21.144,P,0.0005, Interaction exercise x forced swimming: F(1,24) = 3.257, P = 0.087; D, Effect of exercise: F(1,24) = 5.598, P = 0.026, Effect of forced swimming:F(1,24) = 59.533, P,0.0005, Interaction exercise x forced swimming: F(1,24) = 4.993, P = 0.035; E, Effect of forced swimming: F(1,24) = 38.318, P,0.0005.*, P,0.05, compared to the respective Baseline group; +, P,0.05, compared to the respective Control group, post-hoc Bonferroni test.doi:10.1371/journal.pone.0004330.g004



which we showed that regular voluntary exercise reduces anxiety-

related behavior and novelty-induced glucocorticoid responses

[11,13,15]. Thus, exercised animals cope better with this mild

psychological challenge situation. As they basically stopped

exploring the novel environment after 15 min and returned to

their normal behavior of this time of the day (which is resting or

sleeping), it seems that exercised animals are much quicker than

their sedentary counterparts in assessing the novel situation. This

may relate to the enhanced cognitive abilities reported in exercised

animals [36].

Although the novelty situation appeared to have less influence

on the exercised rats, they showed significantly higher dentate

histone H3 phospho-acetylation and c-Fos responses than the

control animals. In both groups of rats the challenge mainly

impacted on the dorsal (or suprapyramidal) blade of the dentate

gyrus which is in agreement with our previous results [23].

Currently, very little is known with regard to neuroanatomical and

functional differences between the dorsal and the ventral

(infrapyramidal) blade. The dentate gyrus receives its major

afferent input from Layer II of the lateral and medial entorhinal

cortex but these afferents seem to be distributed equally in density

between the dorsal and the ventral blade [37]. At the receiving

side, however, it appears that the granule neurons of the dorsal

blade show more extensive arborizations than those of the ventral

blade. Regarding subcortical regions, the dentate gyrus receives

input from the septal nuclei, the locus coeruleus, the supramam-

millary area and the raphe nuclei [37–39]. The supramammillary

area is of special interest, as the dorsal blade receives double as

Figure 5. Behavior of control and exercised rats in the forced swim test. Rats were subjected to an initial test of 15 min in 25uC-waterfollowed by a 5-min re-test 24 h later. Immobility, struggling and swimming behavior during the test (left panels) and re-test (right panels) wasscored every 10 sec. Data are expressed as the accumulated behavioral scores (mean6SEM, n = 6). *, P,0.05, Student’s t-test.doi:10.1371/journal.pone.0004330.g005



many fibres from this area than the ventral blade [40] and is

implicated in hippocampus-regulated emotional and cognitive

functions [39,41]. Interestingly, the ventral blade of the exercised

rats showed a relatively higher response than that of the control

animals which may indicate an enhanced engagement of this part

of the dentate in the response to novelty after long-term exercise.

The reason for this enhanced participation remains presently

unclear.

In contrast to the novelty challenge test, the forced swim test is a

test to address changes in active coping styles such as struggling

and swimming as well as adaptive coping strategies such as

immobility or floating behavior. Apart from the moderate increase

in swimming behavior during the initial test in exercised rats, there

were no significant differences between the control and exercised

animals. Nevertheless, there were substantial differences regarding

the impact of the challenge on dentate gyrus granule neurons.

After forced swimming, the exercised rats, as compared to the

sedentary animals, showed a significantly higher number of

P(Ser10)-Ac(Lys14)-H3+ and c-Fos+ neurons in the dentate gyrus.

In conjunction with observations in the novelty paradigm, it seems

that the neuronal response does not directly relate to the

immediate behavioral reaction during the challenge but rather

relates to how the information is being processed by the dentate

gyrus in the hours after the challenge. The difference in processing

materialized the next day when animals were re-exposed to the

forced swim test. At this time, the exercised rats showed more

pronounced immobility behavior than the sedentary controls. We

reported before that the phospho-acetylation of histone H3 and c-

Fos induction in dentate granule cells seen after the initial forced

swim test is strongly associated with the immobility behavior

response observed 24 h later in the re-test [22,24]. This behavioral

response is increasingly regarded as a reflection of the strength of

memory of the first forced swim experience [21,42,43]. Moreover,

we recently stipulated that these mechanisms may also play a role

in the formation of traumatic, pathological memories as occurring

in post-traumatic stress disorder (PTSD; [25]). We found that any

interruption (due to pharmacological intervention or gene

deletion) of the signaling cascade initiating the histone modifica-

tions and immediate-early gene induction in dentate granule

neurons resulted in an impaired immobility response [22,24].

Earlier work provided evidence specifically pointing to a critical

role of the glucocorticoid receptor located in the dentate gyrus in

the forced swimming-induced immobility response [44]. There

have been reports about a role of the nucleus accumbens in

immobility behavior. This nucleus has been studied in relation to

immobility behavior interpreting this behavior as being an

indicator of learned helplessness or depressive behavior [35,45].

However, our data question this interpretation because our

exercised animals showed increased immobility behavior in the

re-test. Interpretation of this increased immobility as indicating

increased depressive behavior would be highly debatable given

that exercised animals are known to be less anxious and

cognitively better than sedentary control animals [15,36].

Evidence is accumulating that exercise is anxiolytic and antide-

pressant in humans [7–10]. Furthermore, since exercised and

control rats produced similar c-Fos responses in the nucleus

accumbens in the face of different immobility responses in the re-

test, it seems that this brain structure does not play a critical role in

the differential immobility responses in exercised and control rats.

Moreover, mitogen and stress-activated kinase 1/2 (MSK1/2)

double knockout mice showed highly impaired immobility

behavior in the re-test of the forced swim test in conjunction with

virtually absent histone H3 phospho-acetylation and c-Fos

responses to forced swimming in the dentate gyrus; elsewhere in

the brain (including the nucleus accumbens) c-Fos responses to

forced swimming were normal [24](Chandramohan Y and Reul

JMHM, unpublished observations). Collectively, these results point

to behavioral immobility reflecting an adaptive response in which

formation of memories of the initial swim event plays an important

role. In addition, these cognitive processes seem to involve distinct

epigenetic and gene expression mechanisms in dentate granule

neurons.

Our previous work has shown that the phospho-acetylation of

histone H3 in dentate neurons after forced swimming and novelty

is brought about by concurrent signaling via the GR and NMDA/

ERK/MSK pathways [21–25]. Currently the mechanisms

underlying the enhanced epigenetic and gene expression responses

in the exercised animals are unknown. Yet, a variety of possible

mechanisms contributing to the altered responses after exercise

can be identified. Changes in the two principal pathways identified

by us, i.e. NMDA receptors and GRs, may be involved. Recent

work has shown that exercise leads to changes in NMDA receptor

composition and NMDA receptor-related neuroplasticity process-

es [46–48]. Furthermore, we reported recently that the expression

of GRs is increased in the hippocampus of exercised rats resulting

most likely in an enhanced impact of stress-induced elevations in

glucocorticoid hormone levels [13,49]. Other neurotransmitter

systems possibly involved in the differential epigenetic and gene

expression responses in the exercised rats include the central

noradrenergic, serotonergic and GABAergic systems. These

systems are known to modulate dentate neuron excitability and

are known to be altered after voluntary exercise or during general

motor activity [11,33,50–52] (Papadopoulos A., Chandramohan

Y., Collins A., Droste S.K., Nutt D.J. and Reul J.M.H.M.,

unpublished observations). In addition, changes in intracellular

pathways such as the ERK/MSK pathway and/or histone acetyl

transferase (HAT) activities cannot be excluded. Finally, it also

Figure 6. Effect of forced swimming on c-Fos expression in thenucleus accumbens of control and exercised rats. A and B showrepresentative images of anti-c-Fos immuno-staining in an area of thenucleus accumbens dorso-lateral to the anterior commissure (AC). Thisarea comprises parts of both the core and shell regions. Black arrowsindicate positive nuclear immuno-staining. C shows the number of c-Fos+ neurons in this area of control and exercised rats under baselineconditions and at 2 h after forced swimming. Statistical analysis: Two-way ANOVA: Effect of forced swimming: F(1,16) = 38.157, P,0.0005. *,P,0.05, compared to the respective Baseline group, post-hocBonferroni test.doi:10.1371/journal.pone.0004330.g006



cannot be ruled out that changes in the cyto-architecture of the

dentate gyrus may have contributed to the enhanced responses in

histone H3 phospho-acetylation and c-Fos expression to forced

swimming and novelty in the exercised rats.

Here we showed that exercised rats have improved capabilities

to cope with psychologically stressful challenges. This enhanced

stress coping materialized during the exposure to a novel

environment and when re-submitted to a forced swim challenge.

This improved adaptive capacity may be the logical consequence

of the complex of elevated cognitive abilities, lowered anxiety

levels and decreased impulsiveness known of exercised subjects

[13,15,36]. The increased responses in histone H3 phospho-

acetylation and c-Fos induction in dentate granule neurons of

exercised rats strengthens our concept that these epigenetic and

gene expression responses are part of neuroplasticity processes in

the hippocampus aimed at establishing memories of the event in

case the event would re-occur in the future. Further investigation

of these mechanisms should be of great relevance for the

elucidation of stress-related psychiatric disorders such as major

depression and PTSD.


Conceived and designed the experiments: JMR. Performed the experi-

ments: AC LEH YC SKD JMR. Analyzed the data: AC DW SKD JMR.

Wrote the paper: AC JMR.

References

1. Friedman JE, Ferrara CM, Aulak KS, Hatzoglou M, McCune SA, et al. (1997)

Exercise training down-regulates ob gene expression in the genetically obese

SHHF/Mcc-fa(cp) rat. Horm Metab Res 29: 214–219.

2. Lambert GW, Jonsdottir IH (1998) Influence of voluntary exercise on

hypothalamic norepinephrine. J Appl Physiol 85: 962–966.

3. Kramer JM, Plowey ED, Beatty JA, Little HR, Waldrop TG (2000)

Hypothalamus, hypertension, and exercise. Brain Res Bull 53: 77–85.

4. Pedersen BK, Hoffman-Goetz L (2000) Exercise and the immune system:

regulation, integration, and adaptation. Physiol Rev 80: 1055–1081.

5. Jonsdottir IH (2000) Neuropeptides and their interaction with exercise and

immune function. Immunol Cell Biol 78: 562–570.

6. Gielen S, Schuler G, Hambrecht R (2001) Exercise training in coronary artery

disease and coronary vasomotion. Circulation 103: e1–e6.

7. Steptoe A, Edwards S, Moses J, Mathews A (1989) The effects of exercise

training on mood and perceived coping ability in anxious adults from the general

population. J Psychosom Res 33: 537–547.

8. Byrne A, Byrne DG (1993) The effect of exercise on depression, anxiety and

other mood states: a review. J Psychosom Res 37: 565–574.

9. Salmon P (2001) Effects of physical exercise on anxiety, depression, and

sensitivity to stress: a unifying theory. Clin Psychol Rev 21: 33–61.

10. Dimeo F, Bauer M, Varahram I, Proest G, Halter U (2001) Benefits from

aerobic exercise in patients with major depression: a pilot study. Br J Sports Med

35: 114–117.

11. Droste SK, Gesing A, Ulbricht S, Linthorst ACE, Reul JMHM (2003) Effects of

long-term voluntary exercise on the mouse hypothalamic-pituitary-adrenocor-

tical axis. Endocrinology 144: 3012–3023.

12. Droste SK, Schweizer MC, Ulbricht S, Reul JMHM (2006) Long-term

voluntary exercise and the mouse hypothalamic-pituitary-adrenocortical axis:

impact of concurrent treatment with the antidepressant drug tianeptine.

J Neuroendocrinol 18: 915–925.

13. Droste SK, Chandramohan Y, Hill LE, Linthorst AC, Reul JMHM (2007)

Voluntary exercise impacts on the rat hypothalamic-pituitary-adrenocortical axis

mainly at the adrenal level. Neuroendocrinol 86: 26–37.

14. Lancel M, Droste SK, Sommer S, Reul JMHM (2003) Influence of regular

voluntary exercise on spontaneous and social stress-affected sleep in mice.

Eur J Neurosci 17: 2171–2179.

15. Binder E, Droste SK, Ohl F, Reul JMHM (2004) Regular voluntary exercise

reduces anxiety-related behaviour and impulsiveness in mice. Behav Brain Res

155: 197–206.

16. Steckler T, Holsboer F, Reul JMHM (1999) Glucocorticoids and depression. In:

Holly JMP, ed. Bailliere’s Best Practice & Research. London: Bailliere Tindall.

pp 597–614.

17. Reul JMHM, Droste SK (2005) The hypothalamic-pituitary-adrenal axis as a

dynamically organized system: lessons from exercising mice. In: Steckler T,

Kalin NH, Reul JMHM, eds. Handbook of Stress and the Brain - Part 1: The

Neurobiology of Stress. Amsterdam: Elsevier. pp 95–112.

18. Cheung P, Tanner KG, Cheung WL, Sassone-Corsi P, Denu JM, et al. (2000)

Synergistic coupling of histone H3 phosphorylation and acetylation in response

to epidermal growth factor stimulation. Mol Cell 5: 905–915.

19. Clayton AL, Rose S, Barratt MJ, Mahadevan LC (2000) Phosphoacetylation of

histone H3 on c-fos-and c-jun-associated nucleosomes upon gene activation.

EMBO J 19: 3714–3726.

20. Nowak SJ, Corces VG (2000) Phosphorylation of histone H3 correlates with

transcriptionally active loci. Genes Dev 14: 3003–3013.

21. Reul JMHM, Chandramohan Y (2007) Epigenetic mechanisms in stress-related

memory formation. Psychoneuroendocrinol 32 Suppl 1: S21–S25.

22. Bilang-Bleuel A, Ulbricht S, Chandramohan Y, Droste SK, Reul JMHM (2005)

Psychological stress increases histone H3 phosphorylation in adult dentate gyrus

granule neurons: involvement in a glucocorticoid receptor-dependent behav-

ioural response. Eur J Neurosci 22: 1691–1700.

23. Chandramohan Y, Droste SK, Reul JMHM (2007) Novelty stress induces

phospho-acetylation of histone H3 in rat dentate gyrus granule neurons through

coincident signalling via the N-methyl-d-aspartate receptor and the glucocor-

ticoid receptor: relevance for c-fos induction. J Neurochem 101: 815–828.

24. Chandramohan Y, Droste SK, Arthur JS, Reul JMHM (2008) The forced

swimming-induced behavioural immobility response involves histone H3

phospho-acetylation and c-Fos induction in dentate gyrus granule neurons via

activation of the N-methyl-D-aspartate/extracellular signal-regulated kinase/

mitogen- and stress-activated kinase signalling pathway. Eur J Neurosci 27:

2701–2713.

25. Reul JMHM, Nutt DJ (2008) Glutamate and cortisol–a critical confluence in

PTSD? J Psychopharmacol 22: 469–472.

26. Reul JMHM, Chandramohan Y, Droste SK (2006) Marked changes in the

Hypothalamic-Pituitary-Adrenocortical axis of male Sprague-Dawley rats after

long-term voluntary exercise. Proceedings of the 6th International Congress of

Neuroendocrinology, June 19–22, 2006, Pittsburgh.

27. Harri M, Lindblom J, Malinen H, Hyttinen M, Lapvetelainen T, et al. (1999)

Effect of access to a running wheel on behavior of C57BL/6J mice. Laboratory

Animal Science 49: 401–405.

28. Cooper JJ, Nicol CJ (1991) Stereotypic behavior affects environmental

preferences in bank voles, Chlethrionomys glareolus. Anim Behav 41: 971–977.

29. Cooper JJ, Nicol CJ (1996) Stereotypic behavior in wild caught and laboratory

bred bank voles (Clethrionomus glareolus). Anim Welfare 5: 245–257.

30. Gesing A, Bilang-Bleuel A, Droste SK, Linthorst ACE, Reul JMHM (2001)

Psychological stress increases hippocampal mineralocorticoid receptor levels:

involvement of corticotropin-releasing hormone. J Neurosci 21: 4822–4829.

31. Droste SK, de Groote L, Atkinson HC, Lightman SL, Reul JMHM, et al. (2008)

Corticosterone levels in the brain show a distinct ultradian rhythm but a delayed

response to forced swim stress. Endocrinology 149: 3244–3253.

32. Bilang-Bleuel A, Rech J, Holsboer F, Reul JMHM (2002) Forced swimming

evokes a biphasic response in CREB phosphorylation in extrahypothalamic

limbic and neocortical brain structures. Eur J Neurosci 15: 1048–1060.

33. Linthorst ACE, Penalva RG, Flachskamm C, Holsboer F, Reul JMHM (2002)

Forced swim stress activates rat hippocampal serotonergic neurotransmission

involving a corticotropin-releasing hormone receptor-dependent mechanism.

Eur J Neurosci 16: 2441–2452.

34. Dijkstra I, Tilders FJH, Aguilera G, Kiss A, Rabadan-Diehl C, et al. (1998)

Reduced activity of hypothalamic corticotropin-releasing hormone neurons in

transgenic mice with impaired glucocorticoid receptor function. J Neurosci 18:

3909–3918.

35. Mague SD, Pliakas AM, Todtenkopf MS, Tomasiewicz HC, Zhang Y, et al.

(2003) Antidepressant-like effects of kappa-opioid receptor antagonists in the

forced swim test in rats. J Pharmacol Exp Ther 305: 323–330.

36. van Praag H, Christie BR, Sejnowski TJ, Gage FH (1999) Running enhances

neurogenesis, learning and long-term potentiation in mice. Proc Natl Acad Sci

USA 96: 13427–13431.

37. Witter MP (1993) Organization of the entorhinal-hippocampal system: a review

of current anatomical data. Hippocampus 3 Spec No: 33–44.

38. Swanson LW (1977) The anatomical organization of septo-hippocampal

projections. Ciba Found Symp 58: 25–48.

39. Pan WX, McNaughton N (2004) The supramammillary area: its organization,

functions and relationship to the hippocampus. Prog Neurobiol 74: 127–166.

40. Wyss JM, Swanson LW, Cowan WM (1979) A study of subcortical afferents to

the hippocampal formation in the rat. Neuroscience 4: 463–476.

41. Aranda L, Santin LJ, Begega A, Aguirre JA, Arias JL (2006) Supramammillary

and adjacent nuclei lesions impair spatial working memory and induce

anxiolitic-like behavior. Behav Brain Res 167: 156–164.

42. De Pablo JM, Parra A, Segovia S, Guillamon A (1989) Learned immobility

explains the behavior of rats in the forced swim test. Physiol Behav 46: 229–237.

43. Korte SM (2001) Corticosteroids in relation to fear, anxiety and psychopathol-

ogy. Neurosci Biobehav Rev 25: 117–142.

44. Korte SM, De Kloet ER, Buwalda B, Bouman SD, Bohus B (1996) Antisense to

the glucocorticoid receptor in hippocampal dentate gyrus reduces immobility in

forced swim test. Eur J Pharmacol 301: 19–25.



45. Porsolt RD, Le Pichon M, Jalfre M (1977) Depression: a new animal model

sensitive to antidepressant treatments. Nature 266: 730–732.

46. Farmer J, Zhao X, Van PH, Wodtke K, Gage FH, et al. (2004) Effects of

voluntary exercise on synaptic plasticity and gene expression in the dentate gyrus

of adult male Sprague-Dawley rats in vivo. Neuroscience 124: 71–79.

47. Dietrich MO, Mantese CE, Porciuncula LO, Ghisleni G, Vinade L, et al. (2005)

Exercise affects glutamate receptors in postsynaptic densities from cortical mice

brain. Brain Res 1065: 20–25.

48. Vasuta C, Caunt C, James R, Samadi S, Schibuk E, et al. (2007) Effects of

exercise on NMDA receptor subunit contributions to bidirectional synaptic

plasticity in the mouse dentate gyrus. Hippocampus 17: 1201–1208.

49. Reul JMHM, De Kloet ER (1985) Two receptor systems for corticosterone in rat

brain: microdistribution and differential occupation. Endocrinology 117:2505–2512.

50. Linthorst ACE, Reul JMHM (1998) Brain neurotransmission during peripheral

inflammation. Ann N Y Acad Sci 840: 139–152.51. Linthorst ACE, Flachskamm C, Holsboer F, Reul JMHM (1995) Intraperitoneal

administration of bacterial endotoxin enhances noradrenergic neurotransmissionin the rat preoptic area: relationship with body temperature and hypothalamic-

pituitary-adrenocortical axis activity. Eur J Neurosci 7: 2418–2430.

52. Hill LE, Droste SK, Nutt DJ, Linthorst ACE, Reul JMHM (2008) Voluntaryexercise alters GABAA receptor subunit and glutamic acid decarboxylase-67

gene expression in the rat forebrain. J Psychopharmacol; in press.



Metabolic Consequences and Vulnerability to Diet-Induced Obesity in Male Mice under Chronic Social StressAlessandro Bartolomucci1*, Aderville Cabassi2, Paolo Govoni3, Graziano Ceresini4, Cheryl Cero1, Daniela

Berra1, Harold Dadomo1, Paolo Franceschini1, Giacomo Dell’Omo5, Stefano Parmigiani1., Paola

Palanza1.

1 Department of Evolutionary and Functional Biology, University of Parma, Parma, Italy, 2 Department of Internal Medicine, Nephrology and Health Sciences, University of

Parma, Parma, Italy, 3 Department of Experimental Medicine, University of Parma, Parma, Italy, 4 Department of Internal Medicine and Biomedical Sciences, University of

Parma, Parma, Italy, 5 Ornis Italica, Rome, Italy

Abstract

Social and psychological factors interact with genetic predisposition and dietary habit in determining obesity. However,relatively few pre-clinical studies address the role of psychosocial factors in metabolic disorders. Previous studies from ourlaboratory demonstrated in male mice: 1) opposite status-dependent effect on body weight gain under chronicpsychosocial stress; 2) a reduction in body weight in individually housed (Ind) male mice. In the present study theseobservations were extended to provide a comprehensive characterization of the metabolic consequences of chronicpsychosocial stress and individual housing in adult CD-1 male mice. Results confirmed that in mice fed standard diet,dominant (Dom) and Ind had a negative energy balance while subordinate (Sub) had a positive energy balance. Locomotoractivity was depressed in Sub and enhanced in Dom. Hyperphagia emerged for Dom and Sub and hypophagia for Ind. Domalso showed a consistent decrease of visceral fat pads weight as well as increased norepinephrine concentration and smalleradipocytes diameter in the perigonadal fat pad. On the contrary, under high fat diet Sub and, surprisingly, Ind showedhigher while Dom showed lower vulnerability to obesity associated with hyperphagia. In conclusion, we demonstrated thatsocial status under chronic stress and individual housing deeply affect mice metabolic functions in different, sometimeopposite, directions. Food intake, the hedonic response to palatable food as well as the locomotor activity and thesympathetic activation within the adipose fat pads all represent causal factors explaining the different metabolic alterationsobserved. Overall this study demonstrates that pre-clinical animal models offer a suitable tool for the investigation of themetabolic consequences of chronic stress exposure and associated psychopathologies.

Citation: Bartolomucci A, Cabassi A, Govoni P, Ceresini G, Cero C, et al. (2009) Metabolic Consequences and Vulnerability to Diet-Induced Obesity in Male Miceunder Chronic Social Stress. PLoS ONE 4(1): e4331. doi:10.1371/journal.pone.0004331



Copyright: � 2009 Bartolomucci et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: Supported by The University of Parma. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of themanuscript.

Competing Interests: GD supported the development of the automated system for activity measurement in collaboration with Technosmart.


. These authors contributed equally to this work.

Introduction

The chronic activation of the stress response has been associated

with metabolic disorders and altered energy homeostasis [1,2].

Acute increase of stress hormones, such as glucocorticoids (GCs),

catecholamines, etc. may determine the mobilization of fuel

molecules, stimulate or inhibit feeding, and oppose insulin action

[3–6]. However, sustained concentrations of GCs as observed

under chronic stress can also increase the salience of pleasurable or

compulsive activities (ingesting sucrose, fat, and drugs, or wheel-

running). This, in synergy with insulin, may increase ingestion of

‘‘comfort food’’ and systemically increase abdominal fat depots

[1,6,7]. Experimental studies in humans have demonstrated that

perturbations of the hypothalamus-pituitary-adrencortical (HPA)

axis function relate with abdominal obesity [8] and that stress

perception strongly associates with a higher waist-to-hype-ratio

and body mass index (BMI) [9,10]. In addition, in patients

depression has also been associated with the metabolic syndrome

and obesity [1], with pre-existing differences in BMI predicting the

direction of changes in energy balance determined by job stress

[11]. Finally, in a cohort of Finnish twins discordant for adult

BMI, the obese co-twins showed the highest index of psychosocial

stress perception when compared to the lean co-twins [12].

Differently from humans, experimental models in animals offer

the advantage to allow an easier manipulation of key experimental

variables for the investigation of psychosocial factors affecting

vulnerability to stress exposure [7,13–16]. In particular, animal

models of social stress appear to have a high validity as models of

human psychopathologies [13–18]. Unfortunately, until recently

there was a paucity of animal models in which stress exposure was

associated with body weight gain. Indeed, animal models of

chronic stress, including chronic subordination, have repeatedly

been associated with a reduction in body weight and a generalized

catabolic state [19–24]. This clear-cut effect is not present in the

human literature and the DSM-IV defines weight gain or loss as a

diagnostic criterion for major depression [25]. Recently, our and

other laboratories described animal models for chronic stress-

induced increase in body weight and adiposity [26–29] and


vulnerability to diet induced obesity [28,30,31]. In addition, recent

studies have showed neuroendocrine evidences of metabolic

syndrome in defeated rats fed high fat diet but not a standard

diet [32]. Furthermore, there is evidence that social status in

models of chronic stress might differentially affect stress-induced

metabolic effects: Bartolomucci et al [26], Moles et al [28] and

Solomon et al [29] using similar experimental models in mice and

hamsters, reported that subordination can be reliably associated

with increased weight gain, whereas dominance is associated with

lower weight gain or weight loss. However, there are currently no

studies comparing different models of social stress that simulta-

neously determine behavioral, metabolic, biochemical and ana-

tomical alterations in the experimental animals. Thus, the aims of

the present study were: 1) to clarify the metabolic consequences of

social stress using two models, i.e. chronic psychosocial stress

distinguishing between dominants (Dom) and subordinates (Sub)

[26,33], and individual housing (Ind) [34]; 2) to characterize for

the first time sympathetic system related parameters within visceral

adipose fat pads in animals under chronic stress; 3) to determine

morphological changes in the adipose tissue; and finally 4) to

determine if the metabolic consequences of stress-exposure might

translate into altered vulnerability to high fat diet (HFD)-induced

obesity.

Results

Behavioral and endocrine consequences of chronicpsychosocial stress

According to our standard protocol [26], after a few days each

dyad was clearly biased into a stable dominant/subordinate

relationship, with Dom being the only mice showing aggressive

behavior (Figure 1A). Individual locomotor activity was scored in

the home cage by means of infrared sensors. The analysis revealed

that in the dark phase (the active period for mice), Dom showed an

increase in locomotor activity, while Sub showed a depression of

locomotor activity when compared with baseline values

(Figure 1B). A separate analysis of locomotor activity during the

light phase revealed that Dom showed a strong stress-associated

increase both before and after interaction. On the contrary, Sub

showed increased activity only before, but not after, the daily fight

which can be interpreted as an anticipation of the agonistic

interaction [35] and imply a disturbance of the normal sleep

pattern, i.e. reduced sleep during the early light phase (the normal

inactive period for mice). In Sub the post-interaction light phase

activity remained unaffected when compared with baseline but

was clearly lower when compared with Dom (Figure 1C). Finally,

both Dom and Sub showed increased basal corticosterone plasma

level after 21 days of chronic stress exposure (Figure 2).

Metabolic consequences of chronic psychosocial stress:social status effects

In agreement with our previous report [26], the growing curves

of Dom and Sub mice (Figure 3A) started to diverge soon after the

beginning of stress procedure with Dom gaining less weight and

Sub gaining more weight than control (Con) mice. The growing

curve of both Dom and Sub was reduced in the week preceding

the stress procedure onset and this might be attributed to

individual housing [34 and see below]. Importantly, stress-induced

hyperphagia emerged with both Dom and Sub mice that

significantly increased the kcal ingested when compared to

baseline (Figure 3C). As a result, both Dom and Sub ingested

more kcal than Con and Ind mice during the stress phase

(Figure 3C).

We dissected and weighted major visceral fat pads to determine

the metabolic consequences of chronic stress and associated

hyperphagia. Results proved that Dom but not Sub showed a

marked decrease in the weight of perigonadal and perirenal fat

pads while only a trend emerged for a lower retroperitoneal fat

pad (Figure 3D). The mesenteric and the mediastinic fat pads

remained unaffected. Overall Dom showed a lower content of

visceral fat than Con (Figure 3E).

At the cellular level, Dom showed lower mean perigonadal

adipocytes diameter when compared to both Sub and Con

(Figure 4A,B). Furthermore, a quantitative analysis of individual

adipocytes demonstrated that in Dom larger adipocytes (i.e. larger

than 71 mm) were almost completely absent while they represented

20–30% of the adipocytes population in the other groups (a

significant increase in 30–50 mm and a decrease in 71–90 mm sized

adipocytes was observed, U10,10 = 15, p,0.0001 and U10,10 = 16,

p,0.010 when compared to Con. Figure 4C). Furthermore,

although the effect is quantitatively small, Sub showed an increase

(from 0.5 to 1% in all groups to 5% in Sub) in very large adipocytes

(i.e. larger than 91 mm. Figure 4C). This analysis revealed that

dominant mice under chronic stress showed a clear adipocytes

remodeling thus suggesting that the reduction in body weight may

be due to sympathetic-driven lipolysis leading to overall reduction of

adipocytes size and adipose tissue weight. To shed light on this

hypothesis, we determined the enzymatic activity of tyrosine

hydroxylase (TH), the rate-limiting enzyme in the biosynthesis of

catecholamines, as well as norepinephrine (NE) concentration in

perigonadal fat pads. Dom showed high NE concentration and a

slight but not significant increase in TH activity while Sub showed

no change in the same parameters (Figure 5). Furthermore negative

correlations were found between final body weight gain and TH

activity (r = 20.48, p,0.05) and NE concentration (r = 20.45,

p = 0.05) as well as between NE concentration and perigonadal fat

pad weight (r = 20.45, p = 0.05).

Overall, data from the present experiment proved that despite

similar stress-induced hyperphagia Dom and Sub showed opposite

metabolic consequences, i.e. Dom showed negative energy balance

associated with increased sympathetic tone and locomotor activity

which apparently were able to counteract hyperphagia, while Sub

showed positive energy balance driven by hyperphagia and lower

activity and being, thus, at risk for weight gain and obesity.

Metabolic consequences of chronic individual housingInd mice showed a clear inhibition of weight gain when

compared to Con under standard diet (Figure 3B). In addition

when comparing the growing curve after the first seven days of

individual housing (Figure 3A) (in analogy with Dom and Sub

under chronic psychosocial stress), Ind mice only differed from

Sub (lower weight gain) but not from Dom or Con. Ind mice

ingested less kcal than Con mice for the duration of the whole

experimental phase with values reaching significance in the last

week (Figure 3C).

The weight of adipose tissue fat pads was generally reduced in

Ind mice when compared to Con, though this effect was significant

only for the perigonadal pad, while a trend emerged for the

perirenal pad and no overall reduction of visceral fat pad was

observed (Figure 3D and E). It must be noted, however, that in Ind

mice neither changes in perigonadal adipocytes diameter nor any

major change in the frequency of differentially sized adipocytes

was noticed (Figure 4). Similarly, no change in TH activity or NE

concentration in perigonadal fat pad was detected (Figure 5).

Therefore, in mice fed a standard diet, the effect of individual

housing on weight gain were similar to those observed in mice that

were maintaining dominance under chronic psychosocial stress.

Body Weight under Stress


However, at variance with Dom, Ind mice showed a reduction in

food intake, which seems to be largely responsible for the

metabolic effects observed in absence of a hyperactivity of

sympathetic-markers such as TH enzymatic activity and NE

concentration.

Finally, in agreement with our previous report [34], Ind mice

showed increased basal blood corticosterone concentration

(Figure 2).

High fat diet exposureThe observed status-dependent (Dom vs. Sub) and stress model-

dependent (psychosocial stress vs. individual housing) metabolic

consequences of stress suggest a possible differential vulnerability

of Dom, Sub and Ind mice to diet-induced obesity (DIO) [28,31].

To test this hypothesis, mice were challenged with a HFD that

provides 45% kcal from fat and 5.2 kcal per gram (compared to

the 6.5% and 3.9 values respectively of the standard chow)

beginning on the first day of stress procedure or after 7 days of

baseline (for Con and Ind). Based on the data obtained under

standard diet conditions, we predicted that Dom and Ind should

be less vulnerable, and Sub more vulnerable, to HFD-induced

obesity when compared to Con.

Indeed, results proved that Sub were more vulnerable and Dom

more resistant to DIO than Con (Figure 6). Interestingly, this

occurred despite Dom showing a 3 weeks-long hyperphagia while

Sub being hyperphagic only in the last 2 weeks (Sub clearly

ingested more kcal when compared to baseline throughout the 3

weeks period. Figure 6B). Contrary to our prediction, individual

housing also determined an increased vulnerability to DIO.

Indeed, Ind showed increased weight gain, hyperphagia and food

efficiency when compared to Con (Figure 6). Therefore, despite

Dom and Ind showing similar hyperphagia, the metabolic cost of

Figure 1. Behavioral consequences of chronic psychosocial social stress in mice. A) Aggressive behavior assessed on days 1 to 4, 10 and 20of the stress phase. Graph clearly shows how dominants (Dom) and subordinates (Sub) are non-overlapping behavioral categories. B) Locomotoractivity measured during baseline (4 days) and the stress phase (20 days). Dom showed increased and Sub showed decreased locomotor activity(F(1,18) = 21.9, p,0.01). C) Locomotor activity measured before and after the daily agonistic interaction. Dom showed increased activity both beforeand after the agonistic interaction while Sub showed increased activity before but not after the agonistic interaction (F(1,18) = 4.1, p = 0.054). *p,0.05 and ** p,0.001 vs. basal, # p,0.05 vs. Dom.doi:10.1371/journal.pone.0004331.g001



dominance (as described in the previous section), was able to

restrain food efficiency and avoid HFD-induced weight gain.

HFD resulted in a massive overall increase in adipose tissue

weight when compared with mice fed a standard diet (see

Figure 3E and Figure 6E). In particular, Dom showed lower

perigonadal, retroperitoneal and mesenteric fat mass weight as

well as overall visceral adipose tissue when compared with Sub.

However, while a trend for Dom showing lower and for Sub

showing higher fat mass than Con emerged, such effects did not

reach statistical significance. In Ind mice, the adipose tissue was

greatly enlarged, with perigonadal, retroperitoneal and mesenteric

fat pads showing a greater increase than Con (Figure 6D), which

also resulted in an overall increase in visceral fat mass (Figure 6E).

Finally, Dom but not Sub showed lower adipose fat mass weight

when compared with Ind (Figure 6E).

Overall, the data of Dom and Sub mice largely agreed with the

prediction that Sub would have been more, and Dom less,

vulnerable to HFD-induced obesity when compared to Con. In

particular, HFD exposure increased the difference in adiposity

between Dom and Sub, with Sub also showing slightly greater

adipose mass than Con.

Data also proved that Ind mice were remarkably vulnerable to

HFD-induced obesity and that exposure to hypercaloric and

highly palatable diet was able to reverse the effects observed under

standard diet, i.e. lower food intake and weight loss. The more

likely explanation is that individual housing determined an

increased hedonic response to high fat food and that: 1) the

compensatory inhibition of initial hyperphagia (observed in

controls) is disrupted in Ind mice (mechanism to be identified);

2) Ind mice are faced with a smaller metabolic cost than mice

subjected to chronic psychosocial stress.

Discussion

Social and psychological factors [36,37] interact with genetic

predisposition [38] and dietary habit [39] to determine the current

obesity pandemia, and a possible link between chronic social

stress, hedonism and vulnerability to obesity has been suggested

[7]. However, up to now few pre-clinical studies directly addressed

the role played by psychosocial factors and provided validated

experimental models for human stress-induced metabolic disor-

ders, which are very common, for example, in several psychiatric

conditions [1,4,8–12]. In the present study we provided a

comprehensive characterization of the metabolic consequences

of social status under chronic psychosocial stress and social

deprivation in male mice. Overall, our findings showed that in

mice fed standard diet: 1) psychosocial stress determined opposite

effects on energy balance, with Dom showing a negative and Sub a

positive effect; 2) individual housing determined a reduction in

weight gain; 3) hyperphagia emerged for Dom and Sub and

hypophagia for Ind; 4) Dom showed increased NE concentration

in fat tissue, lower perigonadal fat pad weight and smaller

adipocytes diameter than Con. On the contrary, under high fat

diet, Sub and, surprisingly, Ind showed higher, while Dom lower,

vulnerability to obesity than Con.

Given the remarkable difference among the different experi-

mental groups, data will be first discussed separately and then a

general perspective on social modulation of metabolic functions

will be provided.

Chronic psychosocial stress: subordinate mice showpositive energy balance and increased vulnerability todiet-induced obesity

Subordination-induced weight gain is not a common observation

in animal models of chronic social stress [19–24]. Indeed, we were

the first to describe a subordination-induced weight gain in mice

during the chronic psychosocial stress procedure [26], a finding that

has now been replicated by other groups using similar preclinical

animal models of social stress [27–30,40]. This discrepancy in

subordination-stress induced positive o negative weight changes

does not have a clear explanation at the moment. However, when

assessing the literature there are a number of factors that should be

taken into account. Firstly, changes in body weight are often the sole

metabolic parameter presented and it is difficult to interpret a

decrease in body weight without a control for feeding, locomotion

or energy expenditure. Secondly, it appears that the species and the

strain investigated may play a role, since most of the data showing

weight loss have been obtained with subordinate rats or tree shrews

and only a few with mice [41–43]. Among the mouse studies none

was performed with the CD-1 strain. Thus the results presented

here raise the possibility of a strain-associated vulnerability to stress-

induced weight gain. However, we recently obtained very similar

Figure 2. Hormonal consequences of social stress in mice. Basal plasma corticosterone collected in the early light phase, was increased insubordinates (Sub, U9,13 = 23, p,0.016), dominants (Dom, U9,12 = 12, p,0.016) and individually housed (Ind, U9,5 = 3, p,0.005) mice when comparedto Controls (Con). * p,0.016.doi:10.1371/journal.pone.0004331.g002



subordination-induced metabolic effects on inbred strains of mice

(Bartolomucci et al., unpublished observations) thus suggesting that

positive vs negative changes in energy balance is probably primarily

dependent on the model of stress used rather than on the strain.

Thirdly, the experimental animals are generally faced with an

unstable aversive environment with the experimental procedure

often requiring a brief daily move into the dominant home cage with

individual housing for the rest of the day [21,22]. In other studies

the subordinate is moved daily, or every second day, into different

dominant cages [20,41–43]. Thus other models of social stress may

determine a mixed subordination/individual housing/instability

effect with major inhibitory effects on feeding (see also below).

Finally, when data on feeding have been collected, weight loss in

subordinate rats was associated with a reduction in feeding

[24,44,45], while post-stress hyperphagia and weight gain has been

reported for subordinate rats in the visible burrow system [46].

In our experimental setup, body weight changes were associated

with hyperphagia in Sub mice, similarly to what has been previously

reported [27–29,40]. In addition, we have previously shown similar

food consumption in Dom and Sub under stress [26]. In agreement

with our previous report [47] Sub also showed a reduction in

locomotor activity during stress exposure, which is reminiscent of the

psychomotor impairments and reduced willingness to engage in daily

activities observed in depressed patient [15,44]. Therefore, results

from the present and previous studies, prove that positive energy

balance in Sub is associated with increased feeding and lower activity.

Surprisingly, increased body weight gain in Sub did not translate into

higher fat pad weight. This finding is in agreement with our previous

report [26] and suggests that alterations in subcutaneous adipose

tissue, water content or lean mass might be responsible for the

increased weight gain, but rules out a primary role for visceral adipose

tissue in explaining increased body weight. This lack of effect on

visceral adiposity is also surprising because Sub showed increased

circulating corticosterone which is know to be associated with

increased visceral adiposity [1,4,5]. However, it is of interest to note

that Sub mice showed an increased number (although not significant)

of very large sized adipocytes (i.e. larger than 91 mm in diameter) in

the perigonadal pad, which can be considered as an incipient

hypertrophic obesity [48,49] possibly leading to increased vulnera-

bility to cell death [50]. Finally, in our model Sub show a similar up-

Figure 3. Metabolic consequences of social stress in mice. A) Body weight changes in the baseline and in the stress phase. At baseline, allexperimental groups showed a trend for a lower body weight gain than controls (Con) (F(3,39) = 2.6, p = 0.06). In the stress phase, subordinates (Sub)showed a larger body weight gain when compared to all other groups, which were not different from each other (F(3,38) = 4.6, p,0.01). Figuredescribes only post hoc comparisons to controls, * p,0.05; 1 p = 0.06. B) Body weight changes from baseline in Con and individually housed (Ind)mice starting from the first day of baseline. Ind showed a lower growth curve when compared to Con over the whole testing phase (F(1,15) = 6.3,p,0.05. * p,0.05. C) Food intake. Sub and dominants (Dom) mice under stress where hyperphagic when compared to baseline, Con and Ind mice(treatment, F(3,33) = 7.4, p,0.001; treatment x weeks F(9,99) = 3.8, p,0.001). In addition, Ind mice showed an overall lower level of kcal ingestedwhen compared to controls. D) Visceral fat pads weight. Dom showed a smaller perigonadal (F(3,37 = 3.2, p,0.05), perirenal (F(3,37 = 3.2, p,0.05)and a trend for lower retroperitoneal (F(3,37 = 1.7, p = 0.1) pad weight than Con. * p,0.05, 1p,0.07 vs. Con. E) Cumulative weight of visceral fat mass.Dom showed a reduction of visceral fat when compared to Con (F(3,37) = 2.3, p,0.1). * p,0.05 vs. Con.doi:10.1371/journal.pone.0004331.g003



regulation of HPA axis as well as tachycardia than Dom [reviewed in

14,16; present data]. Accordingly, the adipose tissue is probably

exposed to opposing stimuli that may result in the lack of a net effect

on adipose fat pad weight.

On the contrary, when subordinate mice were fed HFD, the

result was an increase in weight gain in the late phase of the stress

procedure and a consistent increase in adiposity. HFD determined

a generalized hyperphagia in the second and third week of stress

likely explaining the delayed effect of HFD on weight gain.

Therefore, subordination under chronic stress may represent a

vulnerability factor for diet-induced obesity.

Overall, our data indicate that subordinate male mice under

chronic stress represent a valid model of stress-induced depression-

related disorders [15,16]. As well, our data also validate the

conclusion that chronic psychosocial stress represents a model of

stress induced weight gain and vulnerability to obesity. These data

find a parallel also in primate and human literature. In a recent

study with rhesus macaque, Wilson and coworkers [51] showed

that subordinates gained more weight and dominants gained less

weight than controls under both low and high fat dietary regimen

and that subordinates were hyperphagic. Finally, in the human

literature it has been repeatedly reported that psychosocial and

Figure 4. Effect of chronic stress on the histology of the perigonadal adipose tissue. A) Representative sections of perigonadal adiposetissue from individually housed (Ind), Control (Con), subordinate (Sub) and dominant (Dom) mice. B) Dom mice showed a significant smaller meanadipocytes diameter when compared to Con (U10,10 = 17, p,0.016), while all other groups remained unaffected. C) Categorized distribution ofindividual adipocytes diameters (see text for statistical details).doi:10.1371/journal.pone.0004331.g004



socio-economic challenges such as low income, low education and

divorce have been associated with perturbed cortisol secretion,

over-eating, metabolic syndrome and type 2 diabetes [1,52–54].

Chronic psychosocial stress: dominant mice shownegative energy balance, sustained sympathetic activityin the visceral adipose tissue and resistance to diet-induced obesity

In the present experimental context, Dom mice showed a

negative energy balance associated with hyperphagia. Evidence for

a high cost of dominance in our experimental protocol comes from

both behavioral and biochemical results. Indeed, Dom showed a

marked behavioral hyperactivity in the stress phase both in the

light and in the dark period. Previous studies also demonstrated

that Dom showed a strong increase of sympathetic function as

indicated by tachycardia, hyperthermia, and increased energy

expenditure [26,28] as well as hyperphagia [28]. In addition, Sakai

and co-workers [23,24], reported that dominant rats housed in the

visible burrow system model of chronic stress showed a slight

decrease in body weight and a reduction in adiposity, which was

associated with higher feeding than subordinate rats [46].

No study had previously investigated sympathetic system related

parameters in the adipose tissue of mice under chronic stress. The

white adipose tissue (WAT) is innervated by the sympathetic

nervous system and a direct role for WAT sympathetic

noradrenergic nerves in lipid mobilization has been demonstrated

[46,55–57]. Here we showed that perigonadal WAT NE

concentration and, to a lesser extent, also the activity of the rate

limiting catecholamine-synthesizing enzyme TH [58], were

increased in Dom. Increased sympathetic markers in the adipose

tissue have previously been associated with catabolic processes and

weight loss [48,56,59,60]. In agreement with a direct role of NE in

regulating the adipose organ, here we demonstrated that Dom

showed a decrease in perigonadal, perirenal and retroperitoneal,

but not in mesenteric and mediastinic fat pads, thus supporting a

strong regional difference in sympathetic nervous system activity

on adipose tissue [56,61]. In addition Dom also showed lower

mean adipocytes diameter, and a classification of perigonadal

adipocytes based on their diameter revealed that Dom showed an

apparent disappearance of large adipocytes (greater than 71 mm).

These findings, in addition to increased NE concentration in the

same fat pad, suggests that a sympathetic mediated lipolysis is the

primary cause of the reduction of fat mass in dominant mice under

chronic stress. In this respect, it is of interest to note that NE was

negatively correlated with final body weight gain and with

perigonadal fat mass. Finally, the sustained metabolic cost

associated with maintaining dominance under stressful conditions

also translated in a resistance to HFD-induced obesity. Dom

showed lower weight gain, and lower adipose weight associated

with remarkable hyperphagia, thus supporting the conclusion that

sustained behavioral and sympathetic activity might limit diet-

induced obesity.

In conclusion, present data further strengthen the conclusion

that maintaining dominance in stressful conditions is strongly

associated with a physiological cost [16,62–64]. Central pathways

determining sustained sympathetic stimulation have not been

determined in the present study but increased CRH/AVP

signaling and hyperactivity of the melanocortin system [65,66] is

fully compatible with both high aggressive level/dominance and

negative energy balance leading to lipolysis [67].

Individual housing: opposite feeding response andmetabolic consequences with standard or high-fat diet

Individual housing is often considered a model of social stress in

rodents because of the factual deprivation of social contacts

[34,68–70]. Previous reports from our [34] and other groups [71–

73] proved that individual housing is associated with a negative

energy balance with animals loosing weight or maintaining a lower

weight gain than group housed siblings. In this study, we provided

a detailed investigation of metabolic functions associated with

individual housing and proved that: 1) in mice fed a standard diet,

isolation is associated with a reduction in food intake and a

decrease in perigonadal fat pad. Reduced feeding, lack of social

facilitation of feeding [74], and unbalanced thermoregulatory

functions associated with lack of social contact [75,76] are the

likely factors responsible for the decrease in body weight; 2) Ind

Figure 5. Sympathetic system related parameters in mice adipose tissue. A) Perigonadal adipose tissue tyrosine hydroxylase (TH) enzymaticactivity assay revealed a small but not significant increase in the dominant (Dom) mice. B) Dom mice showed a higher perigonadal norepinephrine(NE) concentration than Controls (Con) (F(3,21) = 6.0, p,0.01). *p,0.05.doi:10.1371/journal.pone.0004331.g005



mice fed HFD responded with sustained hyperphagia and

increased vulnerability to diet-induced obesity resulting in 16%

weight gain and a massive increase in adipose fat pad weight.

Therefore, it is apparent that reduced food intake under standard

feeding regimen can be due to lower social facilitation to initiate the

feeding [74] rather than to an overall negative motivation to feed

[77]. Indeed, when mice were provided with a highly palatable diet

they responded with conspicuous overfeeding as previously

observed with cafeteria diet [78]. There are very few investigations

on the metabolic consequences of individual housing [34,71–73,78].

In a recent important study Nonogaki and coworkers [31] reported

an impressive strain difference in the vulnerability to weight gain

induced by social isolation. Indeed, the authors proved that: 1)

individual housing was associated with increased weight gain and

overfeeding in the KK strain and in KK mice carrying the ectopic

overexpression of agouti (KKAy); 2) the C57BL6/J strain showed no

effect of individual housing; 3) individually housed diabetic db/db

mice, carrying a mutated leptin receptor gene, showed lower body

weight and hypophagia when compared with group housed db/db.

Our model using an outbred strain may recapitulate the variability

described by Nonogaki and coworkers and suggests that at the

‘‘population’’ level, male mice are vulnerable to obesity only when

faced with HFD. This model also complements recent evidence [79]

showing that epigenetic mechanisms might be more important than

genomic differences in explaining a large proportion of individual

vulnerability to obesity.

Figure 6. Vulnerability to high fat diet-induced obesity. A) Body weight changes in the baseline and in stress phase. At baseline, when micewere fed standard diet, all experimental groups showed a decrease in body weight, while controls (Con) showed a slight increase (F(3,23) = 3.2,p,0.05). In the stress phase subordinates (Sub) and individually housed (Ind) mice were more, and dominant (Dom) were less, vulnerable to weightgain than Con (F(3,23) = 5.3, p,0.01). In the graph only statistical comparison with Con are shown. In addition, both Sub and Ind mice differed fromDom (p,0.001) and Sub differed from Ind on day 14 only (p,0.05). B) Food intake. When animals were fed a high fat diet they showed a markedincrease in kcal ingested. However a clear difference emerged between experimental groups (F(6,32) = 2.9, p,0.05) with Dom and Ind showingsustained hyperphagia when compared to Con along the entire experiment. Sub were hyperphagic only in the third week while showing a trend inthe second week of the stress phase. Finally Sub also differed from Ind and Dom in the first week of the stress phase (p,0.01). C) Food efficiencyanalysis revealed that while Con were able to maintain a balance trough the changing dietary environment, Sub and Ind but not Dom significantlyincreased food efficiency with HFD (F(9,69) = 5.1, p,0.0001). D) Visceral fat pad weight. Dom showed an overall lower amount of perigonadal(F(3,23) = 9.2, p,0.001), perirenal (F(3,23) = 2.5, p,0.08), retroperitoneal (F(3,23) = 3.7, p,0.05) and mesenteric (F(3,23) = 7.2, p,0.005) but notmediastinic fat pad weight when compared to Sub. Ind showed a robust increase in perigonadal, retroperitoneal and mesenteric adipose fat padswhich was significant versus Con and Dom but not versus Sub. E) Cumulative weight of visceral fat mass. Dom showed lower overall visceral adiposetissue than Sub. On the contrary Ind differed from Con and Dom but not from Sub (F(3,23) = 8.4, p,0.001). * p,0.05 and **p,0.01 vs. Controls, 1p,0.07 vs. Controls, c p,0.01 vs. Con and Dom. #p,0.05 and ## p,0.01 vs. Basal level for each group. Arrows describe the change from standardto high fat diet.doi:10.1371/journal.pone.0004331.g006



An overall view: social stress affects metabolic function inmale mice

In the present study we directly compared different models of

social stress and described major metabolic alterations associated

with dominance, subordination and individual housing (Figure 7).

Overall data proved that: 1) subordinate mice under chronic stress

showed increased weight gain without increased visceral adiposity

under standard diet and increased vulnerability to obesity with

HFD; 2) dominant mice under chronic stress showed lower weight

gain and reduced adipose tissue independently from the feeding

regimen; 3) individual housing resulted in lower weight gain and

adiposity with standard chow and massive vulnerability to obesity

with HFD; 4) group housed sibling mice (our control group)

showed large fat mass under standard diet but lower vulnerability

to HFD-induced obesity when compared to Sub and Ind. The

latter result is important because it demonstrates that although

CD-1 are among the heavier laboratory strain of mice,

psychosocial stress exposure is sufficient to increase vulnerability

to HFD-induced obesity.

Our data also provide direct confirmation to a model linking

allostatic load to metabolic disorders recently proposed by Van

Dijk and Buwalda [32]. This model states that metabolic

syndrome and obesity can develop in presence of a high fat

regimen only when an environmental threat prevents active

coping (fight/flight) but permits only a passive strategy. Indeed, in

our experimental model both Dom and Sub are faced with a

threatening situation, and show similar overactive HPA axis and

cardiac hyperactivity as well as hyperphagia, while: a) dominants

responded with an active coping style associated with sympathetic

overactivity in metabolic tissues that limited the development of

obesity despite overfeeding; b) subordinates instead responded

with a passive helplessness strategy and, particularly when faced

with a high fat diet, developed weight gain and obesity. Indirect

confirmation comes from the profile of Ind mice (considered a

model of mild depression [34,69–72]) which showed lower feeding

and body weight gain in the absence of any sympathetic

hyperactivation when fed chow diet while becoming hyperphagic

and obese in the presence of HFD.

Although the molecular and endocrine mechanisms responsible

for metabolic disorders are currently unknown, present data clarify

the role of social factors in modulating the individual vulnerability

to weight gain and offer an important experimental tool for the

investigation of the mechanisms linking stress and psychological

disorders to metabolic dysfunctions.

Methods

Overview of the experimental procedureAdult male mice were individually housed (Ind), group housed

in groups of 3 siblings (here considered as the control group, (Con)

[33,80]) or were submitted to chronic psychosocial stress [26,32]

and identified as dominant (Dom) or subordinates (Sub) by

behavioral observations. The experimental phase consisted in a

baseline phase and in a stress phase (were animals were fed

standard or high fat diet). Body weight, food intake and locomotor

activity (in Sub and Dom only) were determined (see below).

Subsequently on day 20 mice were behaviorally tested in the

modified open-field test and the following morning sacrificed.

After termination, adipose fat pad weight, tyrosine hydroxylase

(TH) activity and norepinephrine (NE) concentration in the

perigonadal fat pad along with histological determination of

adipocytes diameter were obtained. Finally, plasma level of

corticosterone was determined.

AnimalsSubjects were adult male Swiss CD-1 mice from an outbreed

stock originally obtained from Charles River Italia (Calco, Italy).

Mice were born and reared in a colony room at the University of

Parma at 2262uC in a 12-hr light–dark cycle (lights on 0700-

1900). After weaning (25–28 days of age) they were housed in

same-sex- groups of siblings (4–7 per cage) in Plexiglas cages

(38620618 cm) with wood shaving bedding changed weekly. All

animal experimentation was conducted in accordance with the

European Communities Council Directive of 24 November 1986

(86/EEC) and approved by the Ethical committees of the

University of Parma and the Italian Institute of Health.

Figure 7. Overview of the metabolic effects induced by chronic psychosocial stress and individual housing. The graph shows variation(versus the mean value of the control group-housed mice) for body weight changes, food intake and total visceral adipose fat mass weight, understandard or high fat diet. Individual housing (Ind) determined negative or positive energy balance depending on the diet being standard or high fatdiet respectively. Dominance (Dom) determined a similar negative energy balance with both standard and high fat diet. Subordination (Sub)determined similar positive energy balance with both diets. However, body weight gain and feeding were similarly affected under standard and highfat diets while visceral fat pad mass increased with high fat diet only.doi:10.1371/journal.pone.0004331.g007



Chronic Psychosocial stressThe procedure has been originally described by Bartolomucci et

al. [33] and was used here with minor changes to adapt to specific

requirement of metabolic studies. Three-months old male mice to

be used as residents or intruders, were individually housed in

Plexiglas cages (38620618 cm) for a 7 days baseline phase. To

allow recording of baseline individual locomotor activity, after day

1 a wire-mesh partition bisecting the cage longitudinally was

introduced. This restricts the access to only half the cage to mimic

the conditions of the stress phase (detailed below). On day 6 of the

baseline phase, the wire-mesh partition was removed to give the

animal access to the entire cage thus allowing re-establishing of

individual territory in the whole cage. Baseline, body weight and

food intake were monitored at the beginning and the end of the 7

days. On day 7 the 21 days stress phase begun and each resident

mouse received an unfamiliar same-sex weight-matched intruder

mouse and the two animals were allowed to freely interact for

10 minutes. In order to prevent injuries, the social interaction was

interrupted if fighting escalated (when the dominant persistently

bit the opponent). After the interaction, the two animals were

separated by means of a wire-mesh partition, which allowed

continuous sensory contact but no physical interaction. The

partition bisected the cage longitudinally in two symmetrical

compartments. Between 10:00 and 12:00 hours the partition was

removed daily for 10 min. Throughout the stress phase body

weight was monitored weekly, food intake was monitored daily

and locomotor activity was monitored continuously except during

the aggressive interaction. Throughout the study food and water

were available ad libitum to all experimental mice.

During the social interaction offensive behaviors of the animals

were manually recorded and mice social status was determined as

follows: the chasing and biting animal was defined as ‘Dominant’,

while the mouse displaying upright posture flight behavior and

squeaking vocalization was the ‘Subordinate’. The numbers of

attack bouts performed by each animal were quantified during the

first four days than again at day 10 and 20 by direct observation.

When the fight has to be interrupted before the 10 min, the number

of attacks was computed proportionally. Four behavioral categories

were distinguished within the stress group: (i) resident dominant, (ii)

resident subordinate (RS), (iii) intruder dominant, (iv) intruder

subordinate (InS). Previous studies showed minor differences in the

metabolic functions of RS and InS mice and no difference between

the two dominant categories [16]. Although RS had the largest

effects in terms of body weight gain and adiposity [26], there was no

statistical difference between the two groups (which on the contrary

largely differ in immune function [16]). In addition, the present

study confirms no significant difference between RS and InS (data

not shown). Therefore RS and InS were pooled in the group ‘‘Sub’’

and the two dominant categories in the group ‘‘Dom’’.

Age-matched mice, housed in groups of 3 siblings, were

included as the non-stressed control group (Con). This choice

was based on previous observations showing no metabolic,

immune-endocrine and behavioral evidence of stress activation

or anxiety in group-housed siblings (see [33,34,80] for details).

Within each control group, the hierarchical status of the animals

was determined according to [33], and then the dominant and one

of the two subordinate mice (randomly chosen) were used for

experimental measurements. Data from this experiment confirmed

absence of status-associated effects between dominant and

subordinate mice in groups of siblings (data not shown).

Individual housingThree-months old male mice were individually housed in

Plexiglas cages (38620618 cm). Body weight was monitored

weekly and food intake daily. Controls were the same age-matched

mice housed in groups of 3 siblings described above.

Home cage locomotor activityThe assessment of individual daily activity was carried out by

means of an automated system that use small passive infrared

sensors positioned on the top of each cage (TechnoSmart, Rome,

Italy). To avoid interference between the movement of a resident

and an intruder mouse in the same cage the two individual sensors

were separated by a Plexiglas partition which completely blocks

infrared waves. The system was set-up prior to the beginning of

the experimental procedure to verify absence of false signals across

adjacent sensors (data not shown). Locomotor activity was

continuously monitored throughout the whole experiment includ-

ing 4 days of baseline phase and 20 days of stress phase. Recording

was interrupted only during the daily agonistic interaction.

Modified open field testThe test was performed between 16:00 and 19:00 of day 20, in

agreement with Berton et al [81] with minor changes. Each

experimental mouse was introduced into a squared open field

(54654 cm) for two consecutive sessions of 2.5 min. During the

first session (T1, ‘‘target cage empty’’) the open field contained an

empty wire mesh target cage (10 cm diameter) located at one end

of the field. During the second session (T2, ‘‘intruder mouse

present’’), the conditions were identical except that a social target

animal (a same age unfamiliar CD-1 male mouse) had been

introduced into the cage. Between the 2 sessions, the experimental

mouse was removed from the arena, and was placed back into its

home cage for approximately one minute. Mouse behavior was

scored with Ethovision (Noldus, the Netherlands). Within the

arena the following area were identified and time, frequency and

latency determined: ‘‘target zone’’ (an 8 cm wide corridor

surrounding the target cage); the ‘‘far corners’’ of the open field

opposite to the location of the cage; the four corners. All CD-1

mice independently from the experimental treatment spent around

70–80% of the time in the target zone (data not shown) with no

group difference in avoidance/approach time ratio spent in the

target area between T1 and T2. On the contrary, using C57BL6/J

mice the procedure determined similar response as described by

Berton et al [81] (Bartolomucci et al., unpublished). This finding

highlights a major strain difference (C57BL6/J vs. CD-1) in the

behavioral response to an object located within the arena. Thus, a

procedural modification is needed to investigate the behavior of

CD-1 mice in this behavioral test. Because of this limitation data

from this test are not presented. Nevertheless, the test is discussed

here because previous data from our group revealed that

corticosterone level in Ind mice are particularly sensitive to the

acute exposure to an open field [33].

DietMice were fed a standard (6.55% kcal from fat and 3.9 kcal/g;

4RF21, Mucedola, Italy) or a custom pelletted high fat diet (45%

kcal from fat and 5.2 kcal/g manufactured by Mucedola)

modifying the formula of the standard diet 4RF21.

Adipose organ parametersAdipose fat pads (perigonadal, perirenal, retroperitoneal,

mesenteric and mediastinic [82]) were manually dissected and

weighted. Perigonadal pads were split in two parts and one half

was snap frozen in liquid nitrogen and stored at 280uC for later

measurement of sympathetic related parameters (see below). The

second half was immerged in a ice-cold solution of 4%



paraformaldeyd, stored at 24uC overnight and processed for

histological analyses (see below).

Norepinephrine concentration and tyrosine hydroxylaseactivity

TH activity in adipose tissue was analyzed by the method of

Naoi et al [83]. Biopsies were homogenized and incubated at 37uCfor 10 minutes with 140 mmol/L L-tyrosine in 880 mL of sodium

acetate-acetic acid buffer (100 mmol/L, pH = 6.0) containing

1.4 mmol/L (6R)-5,6,7,8-tetrahydrobiopterin, 10 mg of catalase,

and 0.7 mmol/L 4-bromo-3-hydroxybenzyloxyamine (NSD1055,

an inhibitor of aromatic L-amino acid decarboxylase). The

incubation was stopped by the addition of 0.1 mmol/L perchloric

acid containing 0.4 mmol/L sodium metabisulphite and

0.1 mmol/L disodium EDTA. After vortexing, the sample was

allowed to stand in an ice bath for 10 minutes and then

centrifuged at 1000 g for 10 minutes. The supernatant was

injected in a HPLC-ECD system for L-3,4-dihydroxyphenylala-

nine (L-DOPA) analysis. TH activity was calculated as the amount

of L-DOPA generated from L-tyrosine per minute per milligram

of tissue. NE was measured by HPLC using electrochemical

detection, as previously described [84].

Histological analysisSpecimens of perigonadal adipose tissue from different mice

were carefully removed, weighted and immersed in 4% parafor-

maldehyde, dehydrated in ethanol, transitioned in xylene, and

embedded in paraffin. Five-micrometer-thick sections cut with a

cryostat were stained with hematoxylin and eosin. Optical

microscopy images (Nikon Microscope Eclipse 80i) were digitally

captured with NIS-Elements imaging software F 2.20, and the

diameter of 200 adipocytes for each mouse was measured with

ImageJ software (Image Processing and Analysis in Java).

Analysis of CorticosteroneTrunk blood was collected in heparinized tubes, centrifuged at

4,000 RPM for 10 min and plasma was frozen at 220uC for later

analysis. Level of circulating corticosterone was measured in

duplicate with a commercially available RIA kit (Diagnostic

Systems Laboratories, Inc., USA) with a sensitivity of 0.06 ng/ml.

The intraassay variability was 3.4%. To avoid the interassay

variability, all samples were run in a single assay.

Statistical analysisData were checked for agreement with parametric assumption

and analyzed with ANOVA followed by Tukey’s HSD post hoc or

Mann-Whitney U test with the Bonferroni correction when

appropriate. Correlations were performed with parametric

Pearson test.

Acknowledgments

Prof. Elena Choleris (University of Guelph, Canada), Prof. Martin

Kavaliers (University of Western Ontario, Canada) and Dr. Anna Moles

(CNR, Rome, Italy) are acknowledged for helpful comments and

suggestions.


Conceived and designed the experiments: AB SP PP. Performed the

experiments: AB AC PG GC CC DB HD PF. Analyzed the data: AB AC

PG GC. Contributed reagents/materials/analysis tools: GD. Wrote the

paper: AB.

References

1. Bjorntorp P (2001) Do stress reactions cause abdominal obesity and

comorbidities? Obes Rev 2: 73–86.

2. McEwen BS (1998) Protective and damaging effects of stress mediators.

N Engl J Med 338: 171–9.

3. Sapolsky RM, Romero LM, Munck AU (2000) How do glucocorticoids

influence stress responses? Integrating permissive, suppressive, stimulatory, and

preparative actions. Endocr Rev 21: 55–89.

4. Dallman MF, la Fleur SE, Pecoraro NC, Gomez F, Houshyar H, et al. (2004)

Glucocorticoids-food intake, abdominal obesity, and wealthy nations in 2004.

Endocrinology 145: 2633–8.

5. Dallman MF, Pecoraro N, Akana SF, La Fleur SE, Gomez F, et al. (2003)

Chronic stress and obesity: a new view of ‘‘comfort food’’. Proc Natl Acad

Sci U S A 100: 11696–701.

6. Dallman MF, Strack AM, Akana SF, Bradbury MJ, Hanson ES, et al. (1993)

Feast and famine: critical role of glucocorticoids with insulin in daily energy flow.

Front Neuroendocrinol 14: 303–47.

7. Coccurello R, D’Amato FR, Moles A (2008) Chronic social stress, hedonism and

vulnerability to obesity: Lessons from Rodents. Neurosci Biobehav Rev; in press.

8. Rosmond R, Dallman MF, Bjorntorp P (1998) Stress-related cortisol secretion in

men: relationships with abdominal obesity and endocrine, metabolic and

hemodynamic abnormalities. J Clin Endocrinol Metab 83: 1853–9.

9. Epel ES, McEwen B, Seeman T, Matthews K, Castellazzo G, et al. (2000) Stress

and body shape: stress-induced cortisol secretion is consistently greater among

women with central fat. Psychosom Med 62: 623–32.

10. Smith AW, Baum A, Wing RR (2005) Stress and weight gain in parents of

cancer patients. Int J Obes 29: 244–50.

11. Kivimaki M, Head J, Ferrie JE, Shipley MJ, Brunner E, et al. (2006) Work stress,

weight gain and weight loss: evidence for bidirectional effects of job strain on

body mass index in the Whitehall II study. Int J Obes (Lond) 30: 982–7.

12. Marniemi J, Kronholm E, Aunola S, Toikka T, Mattlar CE, et al. (2002)

Visceral fat and psychosocial stress in identical twins discordant for obesity.

J Intern Med 251: 35–43.

13. Huhman KL (2006) Social conflict models: can they inform us about human

psychopathology? Horm Behav 50: 640–6.

14. Bartolomucci A (2007) Social stress, immune functions and disease in rodents.

Front Neuroendocrinol 28: 28–49.

15. Bartolomucci A (2005) Resource loss and stress-related disease: is there a link?

Med Sci Monit 11: RA147–154.

16. Bartolomucci A, Palanza P, Sacerdote P, Panerai AE, Sgoifo A, et al. (2005)

Social factors and individual vulnerability to chronic stress exposure. Neurosci

Biobehav Rev 29: 67–81.

17. Buwalda B, Kole MH, Veenema AH, Huininga M, de Boer SF, et al. (2005)

Long-term effects of social stress on brain and behavior: a focus on hippocampal

functioning. Neurosci Biobehav Rev 29: 83–97.

18. Tamashiro KL, Nguyen MM, Sakai RR (2005) Social stress: from rodents to

primates. Front Neuroendocrinol 26: 27–40.

19. Baran SE, Campbell AM, Kleen JK, Foltz CH, Wright RL, et al. (2005)

Combination of high fat diet and chronic stress retracts hippocampal dendrites.

Neuroreport 16: 39–43.

20. Czeh B, Michaelis T, Watanabe T, Frahm J, de Biurrun G, et al. (2001) 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 98: 12796–801.

21. Rygula R, Abumaria N, Flugge G, Fuchs E, Ruther E, et al. (2005) Anhedonia

and motivational deficits in rats: impact of chronic social stress. Behav Brain Res

162: 127–34.

22. Meerlo P, Overkamp GJ, Daan S, Van Den Hoofdakker RH, Koolhaas JM

(1996) Changes in behaviour and body weight following a single or double social

defeat in rats. Stress 1: 21–32.

23. Tamashiro KL, Hegeman MA, Nguyen MM, Melhorn SJ, Ma LY, et al. (2007)

Dynamic body weight and body composition changes in response to

subordination stress. Physiol Behav 91: 440–8.

24. Tamashiro KL, Hegeman MA, Sakai RR (2006) Chronic social stress in a

changing dietary environment. Physiol Behav 89: 536–42.


Mental Disorders, (4th ed.), Text Revised (DSM-IV-TR). Washington, DC:

American Psychiatric Association.

26. Bartolomucci A, Pederzani T, Sacerdote P, Panerai AE, Parmigiani S, et al.

(2004) Behavioral and physiological characterization of male mice under chronic

psychosocial stress. Psychoneuroendocrinology 29: 899–910.

27. Foster MT, Solomon MB, Huhman KL, Bartness TJ (2006) Social defeat

increases food intake, body mass, and adiposity in Syrian hamsters. Am J Physiol

Regul Integr Comp Physiol 290: R1284–93.

28. Moles A, Bartolomucci A, Garbugino L, Conti R, Caprioli A, et al. (2006)

Psychosocial stress affects energy balance in mice: modulation by social status.

Psychoneuroendocrinology 31: 623–33.



29. Solomon MB, Foster MT, Bartness TJ, Huhman KL (2007) Social defeat and

footshock increase body mass and adiposity in male Syrian hamsters.Am J Physiol Regul Integr Comp Physiol 292: R283–90.

30. Kuo LE, Kitlinska JB, Tilan JU, Li L, Baker SB, et al. (2007) Neuropeptide Y

acts directly in the periphery on fat tissue and mediates stress-induced obesityand metabolic syndrome. Nat Med 13: 803–11.

31. Nonogaki K, Nozue K, Oka Y (2007) Social isolation affects the development ofobesity and type 2 diabetes in mice. Endocrinology 148: 4658–66.

32. van Dijk G, Buwalda B (2008) Neurobiology of the metabolic syndrome: an

allostatic perspective. Eur J Pharmacol 585: 137–46.33. Bartolomucci A, Palanza P, Gaspani L, Limiroli E, Panerai AE, et al. (2001)

Social status in mice: behavioral, endocrine and immune changes are contextdependent. Physiol Behav 73: 401–10.

34. Bartolomucci A, Palanza P, Sacerdote P, Ceresini G, Chirieleison A, et al. (2003)Individual housing induces altered immuno-endocrine responses to psycholog-

ical stress in male mice. Psychoneuroendocrinology 28: 540–58.

35. Ferrari PF, van Erp AM, Tornatzky W, Miczek KA (2003) Accumbal dopamineand serotonin in anticipation of the next aggressive episode in rats. Eur J Neurosci

17: 371–8.36. Wells JC (2006) The evolution of human fatness and susceptibility to obesity: an

ethological approach. Biol Rev 81: 183–205.

37. Speakman JR (2007) A nonadaptive scenario explaining the genetic predispo-sition to obesity: the ‘‘predation release’’ hypothesis. Cell Metab 6: 5–12.

38. Farooqi S, O’Rahilly S (2006) Genetics of obesity in humans. Endocr Rev 27:710–18.

39. Wardle J (2007) Eating behaviour and obesity. Obes Rev 8 Suppl 1: 73–5.40. Lutter M, Sakata I, Osborne-Lawrence S, Rovinsky SA, Anderson JG, et al.

(2008) The orexigenic hormone ghrelin defends against depressive symptoms of

chronic stress. Nat Neurosci 11: 752–3.41. Kudryavtseva NN, Bakshtanovskaya IV, Koryakina LA (1991) Social model of

depression in mice of C57BL/6J strain. Pharmacol Biochem Behav 38: 315–20.42. Veenema AH, Meijer OC, de Kloet ER, Koolhaas JM (2003) Genetic selection

for coping style predicts stressor susceptibility. J Neuroendocrinol 15: 256–67.

43. Boer U, Alejel T, Beimesche S, Cierny I, Krause D, et al. (2007) CRE/CREB-driven up-regulation of gene expression by chronic social stress in CRE-

luciferase transgenic mice: reversal by antidepressant treatment. PLoS ONE 2:e431.

44. Melhorn SJ, Krause EG, Johnson JD, Sakai RR (2007) Meal patterns andhypothalamic gene expression following chronic social stress. Appetite 49: 313.

45. van Kampen M, Kramer M, Hiemke C, Flugge G, Fuchs E (2002) The chronic

psychosocial stress paradigm in male tree shrews: evaluation of a novel animalmodel for depressive disorders. Stress 5: 37–46.

46. Tamashiro KL, Nguyen MM, Ostrander MM, Gardner SR, Ma LY, et al.(2007) Social stress and recovery: implications for body weight and body

composition. Am J Physiol Regul Integr Comp Physiol 293: R1864–74.

47. Bartolomucci A, Palanza P, Costoli T, Savani E, Laviola G, et al. (2003) Chronicpsychosocial stress persistently alters autonomic function and physical activity in

mice. Physiol Behav 80: 57–67.48. Shi H, Song CK, Giordano A, Cinti S, Bartness TJ (2005) Sensory or

sympathetic white adipose tissue denervation differentially affects depot growthand cellularity. Am J Physiol Regul Integr Comp Physiol 288: R1028–37.

49. Jernas M, Palming J, Sjoholm K, Jennische E, Svensson PA, et al. (2006)

Separation of human adipocytes by size: hypertrophic fat cells display distinctgene expression. FASEB J 20: 1540–2.

50. Monteiro R, de Castro PM, Calhau C, Azevedo I (2006) Adipocyte size andliability to cell death. Obes Surg 16: 804–6.

51. Wilson ME, Fisher J, Fischer A, Lee V, Harris RB, et al. (2008) Quantifying food

intake in socially housed monkeys: Social status effects on caloric consumption.Physiol Behav 94: 586–94.

52. Rosmond R, Dallman MF, Bjorntorp P (1998) Stress-related cortisol secretion inmen: relationships with abdominal obesity and endocrine, metabolic and

hemodynamic abnormalities. J Clin Endocrinol Metab 83: 1853–9.

53. Van Strien T, Rookus MA, Bergers GP, Frijters JE, Defares PB (1986) Lifeevents, emotional eating and change in body mass index. Int J Obes 10: 29–35.

54. Greeno CG, Wing RR (1994) Stress-induced eating. Psychol Bull 115: 444–464.55. Bartness TJ, Song CK (2007) Sympathetic and sensory innervation of white

adipose tissue. J Lipid Res 48: 1655–72.56. Giordano A, Frontini A, Murano I, Tonello C, Marino MA, et al. (2005)

Regional-dependent increase of sympathetic innervation in rat white adipose

tissue during prolonged fasting. J Histochem Cytochem 53: 679–87.57. Giordano A, Morroni M, Santone G, Marchesi GF, Cinti S (1996) Tyrosine

hydroxylase, neuropeptide Y, substance P, calcitonin gene-related peptide and

vasoactive intestinal peptide in nerves of rat periovarian adipose tissue: an

immunohistochemical and ultrastructural investigation. J Neurocytol 25:

125–36.

58. Flatmark T (2000) Catecholamine biosynthesis and physiological regulation in

neuroendocrine cells. Acta Physiol Scand 168: 1–17.

59. Cabassi A, Coghi P, Govoni P, Barouhiel E, Speroni E, et al. (2005) Sympathetic

modulation by carvedilol and losartan reduces angiotensin II-mediated lipolysis

in subcutaneous and visceral fat. J Clin Endocrinol Metab 90: 2888–97.

60. Cabassi A, Vinci S, Cantoni AM, Quartieri F, Moschini L, et al. (2002)

Sympathetic activation in adipose tissue and skeletal muscle of hypertensive rats.

Hypertension 39: 656–61.

61. Bartness TJ, Hamilton JM, Wade GN, Goldman BD (1989) Regional differences

in fat pad responses to short days in Siberian hamsters. Am J Physiol 257:

R1533–40.

62. Henry JP, Stephens PM (1977) Stress, Health and the Social Environment.

Berlin: Springer Verlag.

63. Sapolsky RM (2005) The influence of social hierarchy on primate health.

Science 308: 648–52.

64. Kaplan JR, Manuck SB (1999) Status, stress, and atherosclerosis: the role of

environment and individual behavior. Ann N Y Acad Sci 896: 145–61.

65. Schwartz MW, Woods SC, Porte D Jr, Seeley RJ, Baskin DG (2000) Central

nervous system control of food intake. Nature 404: 661–71.

66. Gao Q, Horvath TL (2007) Neurobiology of feeding and energy expenditure.

Annu Rev Neurosci 30: 367–98.

67. Nogueiras R, Wiedmer P, Perez-Tilve D, Veyrat-Durebex C, Keogh JM, et al.

(2007) The central melanocortin system directly controls peripheral lipid

metabolism. J Clin Invest 117: 3475–88.

68. Katz RJ (1981) Animal models and human depressive disorders. Neurosci

Biobehav Rev 5: 231–46.

69. Valzelli L (1973) The ‘‘isolation syndrome’’ in mice. Psychopharmacologia 31:

305–20.

70. Palanza P (2001 May) Animal models of anxiety and depression: how are

females different? Neurosci Biobehav Rev 25(3): 219–33.

71. Meisel RL, Hays TC, Del Paine SN, Luttrell VR (1990) Induction of obesity by

group housing in female Syrian hamsters. Physiol Behav 47: 815–7.

72. van Leeuwen SD, Bonne OB, Avraham Y, Berry EM (1997) Separation as a new

animal model for self-induced weight loss. Physiol Behav 62: 77–81.

73. Yamada K, Ohki-Hamazaki H, Wada K (2000) Differential effects of social

isolation upon body weight, food consumption, and responsiveness to novel and

social environment in bombesin receptor subtype-3 (BRS-3) deficient mice.

Physiol Behav 68: 555–61.

74. Galef BG, Giraldeau L-A (2001) Social influences on foraging in vertebrates:

causal mechanisms and adaptive functions. Anim Behav 61: 3–15.

75. Cichon M, Chadzinska M, Ksiazek A, Konarzewski M (2002) Delayed effects of

cold stress on immune response in laboratory mice. Proc Biol Sci 269:

1493–1497.

76. Gordon CJ, Becker P, Ali JS (1998) Behavioral thermoregulatory responses of

single- and group-housed mice. Physiol Behav 65: 255–262.

77. Cross-Mellor SK, Kent WD, Ossenkopp KP, Kavaliers M (1999) Differential

effects of lipopolysaccharide and cholecystokinin on sucrose intake and

palatability. Am J Physiol 277: R705–15.

78. Sahakian BJ, Burdess C, Luckhurst H, Trayhurn P (1982) Hyperactivity and

obesity: the interaction of social isolation and cafeteria feeding. Physiol Behav

28: 117–24.

79. Koza RA, Nikonova L, Hogan J, Rim JS, Mendoza T, et al. (2006) Changes in

gene expression foreshadow diet-induced obesity in genetically identical mice.

PLoS Genet 2: e81.

80. Bartolomucci A, Palanza P, Parmigiani S (2002) Group housed mice are they

really stressed? Ethol Ecol Evol 14: 341–350.

81. Berton O, McClung CA, Dileone RJ, Krishnan V, Renthal W, et al. (2006)

Essential role of BDNF in the mesolimbic dopamine pathway in social defeat

stress. Science 311: 864–8.

82. Cinti S (1999) The adipose organ. Milano: Kurtis.

83. Naoi M, Takahashi T, Nagatsu T (1988) Simple assay procedure for tyrosine

hydroxylase activity by high-performance liquid chromatography employing

coulometric detection with minimal sample preparation. J Chromatography 427:

229–238.

84. Cabassi A, Vinci S, Calzolari M, Bruschi G, Borghetti A (1998) Regional

sympathetic activity in pre-hypertensive phase of spontaneously hypertensive

rats. Life Sci 62: 1111–1118.



Expression of the Axonal Membrane Glycoprotein M6a IsRegulated by Chronic StressBen Cooper1¤, Eberhard Fuchs1,2,3*, Gabriele Flugge1,3

1 Clinical Neurobiology Laboratory, German Primate Center, Leibniz Institute for Primate Research, Gottingen, Germany, 2 Department of Neurology, Medical School,

University of Gottingen, Gottingen, Germany, 3 DFG Research Center Molecular Physiology of the Brain (CMPB), University of Gottingen, Gottingen, Germany

Abstract

It has been repeatedly shown that chronic stress changes dendrites, spines and modulates expression of synapticmolecules. These effects all may impair information transfer between neurons. The present study shows that chronic stressalso regulates expression of M6a, a glycoprotein which is localised in axonal membranes. We have previously demonstratedthat M6a is a component of glutamatergic axons. The present data reveal that it is the splice variant M6a-Ib, not M6a-Ia,which is strongly expressed in the brain. Chronic stress in male rats (3 weeks daily restraint) has regional effects: quantitativein situ hybridization demonstrated that M6a-Ib mRNA in dentate gyrus granule neurons and in CA3 pyramidal neurons isdownregulated, whereas M6a-Ib mRNA in the medial prefrontal cortex is upregulated by chronic stress. This is the first studyshowing that expression of an axonal membrane molecule is differentially affected by stress in a region-dependent manner.Therefore, one may speculate that diminished expression of the glycoprotein in the hippocampus leads to altered output inthe corresponding cortical projection areas. Enhanced M6a-Ib expression in the medial prefrontal cortex (in areas prelimbicand infralimbic cortex) might be interpreted as a compensatory mechanism in response to changes in axonal projectionsfrom the hippocampus. Our findings provide evidence that in addition to alterations in dendrites and spines chronic stressalso changes the integrity of axons and may thus impair information transfer even between distant brain regions.

Citation: Cooper B, Fuchs E, Flugge G (2009) Expression of the Axonal Membrane Glycoprotein M6a Is Regulated by Chronic Stress. PLoS ONE 4(1): e3659.doi:10.1371/journal.pone.0003659



Copyright: � 2009 Cooper et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was in part funded by the DFG (German Science Foundation) Research Center Molecular Physiology of the Brain (CMPB), University ofGottingen. DFG and CMPB had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.



¤ Current address: Max-Planck Institute for Experimental Medicine, Gottingen, Germany

Introduction

The membrane glycoprotein M6a is the only member of the

proteolipid protein family of tetraspan proteins to be expressed

exclusively by neurons in the central nervous system [1,2]. Non-

neuronal expression of M6a in peripheral tissues is restricted to the

apical membranes of polarized epithelial cells within the choroid

plexus and proximal renal tubules [3]. Neuronal M6a was

formerly suspected to play a role in the formation of nerve cell

processes since in cultured cerebellar neurons treated with

monoclonal M6a antibody, neurite formation was severely

impaired [4]. Moreover, targeted depletion of endogenous M6a

expression with small inhibitory RNA (siRNA) attenuated neurite

outgrowth and impaired synapse formation [5]. On the other

hand, overexpression of M6a in cultured primary hippocampal

neurons promoted neurite outgrowth and the formation of

filopodial protrusions [5]. However, in a previous publication we

showed that the membrane glycoprotein is not present in

dendrites, but only in axons of glutamatergic neurons [6]. In the

present study, we analyzed the relative abundance of M6a splice

variants Ia and Ib in the rat brain and their regulation by chronic

stress exposure.

M6a initially attracted attention as a gene downregulated by

stress in the hippocampal formation [7,8]. In humans, chronic

stress-induced perturbations of the central nervous system

including structural changes in neurons have the potential to lead

to psychopathologies [9,10]. Stress-induced changes in the

expression of M6a, a structural protein of axonal membranes,

are therefore of particular interest. Stress-induced downregulation

of hippocampal M6a has been confirmed in several species using

quantitative real-time RT-PCR, a method that allows quantifica-

tion of mRNA expression levels in homogenates from defined

brain regions [8,11]. In the present study, using in situ

hybridization with emulsion autoradiography, we quantified

M6a mRNA levels after chronic stress in neurons from distinct

hippocampal subregions. Silver grains representing M6a mRNA

transcripts were counted in dentate gyrus granule neurons, the

cells that extend mossy fiber projections to the hippocampal region

CA3. Moreover, we analyzed M6a mRNA expression in the CA3

pyramidal neurons of stressed rats and controls. To induce stress,

male rats were submitted to three weeks of daily restraint (6 hr/

day) according to established protocols [12,13].

In addition to the hippocampal pyramidal neurons that respond

to chronic stress by retracting their dendrites [14,15] pyramidal

neurons in the medial prefrontal cortex (mPFC) are also sensitive

to stress [16–18]. Chronic restraint stress in male rats reduced the

length of apical dendrites of layer III pyramidal neurons in the

right prelimbic cortex (PL) and eliminated inter-hemispheric

differences in dendritic length in PL and infralimbic cortex (IL),

both of which represent sub-areas of the mPFC [19,20]. In the


cbaker

Cross-Out

cbaker

Inserted Text

30

present study we quantified M6a mRNA expression in cells of the

three mPFC sub-areas, PL, IL and anterior cingulated cortex

(ACx) to detect whether chronic stress might also have an effect on

axons of prefrontocortical neurons.


AnimalsAdult male Sprague Dawley rats (Harlan-Winkelmann, Borchen,

Germany) weighing 250–300 g on arrival were housed in groups of

three animals per cage with food and water ad libitum in

temperature-controlled rooms (2161uC) under an inverse light

cycle (lights off at 07:00, lights on at 19:00). All handling procedures

including stress exposure were performed in the morning under dim

red light (see below). Animal experiments were performed in

accordance with the European Communities Council Directive of

November 24, 1986 (86/EEC) and the US National Institutes of

Health Guide for the Care and Use of Laboratory Animals, and

were approved by the Lower Saxony Federal State Office for

Consumer Protection and Food Safety, Germany.

Quantitative Real-time RT-PCRCloning of rat M6a cDNA has been previously described [6].

To isolate RNA for RT-PCR, animals were decapitated and

brains quickly dissected. Hippocampal formation, prefrontal

cortex and cerebellum were dissected and kidneys were also

sampled. Total RNA was immediately isolated from the individual

tissue samples using the Trizol method (Life Technologies,

Rockville, MD, USA) according to the manufacturer’s instructions

with some modifications. Modifications improving the yield of

isolated RNA included a 30 sec sonification step and the addition

of linear acrylamide (5 mg/ml) to Trizol homogenates. DNase I

digestion was performed and total RNA was purified using

phenol/isoamyl/chloroform and subsequent isopropyl/sodium ac-

etate precipitation [21]. The integrity and quantity of purified RNA

was assessed by spectrophotometry and subsequently confirmed with

RNA 6000 Nano Labchip technology (Agilent Technologies Sales,

Waldbronn, Germany). Complementary DNA (cDNA) was synthe-

sized from mRNA transcripts using oligo (dT)12–18 primers and

Superscript II reverse transcriptase (Invitrogen, Karlsruhe, Ger-

many) according to manufacturers’ instructions. Primer Express

software v2.0 (Applied Biosystems; Darmstadt, Germany) was used

to design gene-specific primers with amplicons ranging between 50–

150 bp in length. The intron-exon organisation of murine M6a and

M6b genes has been previously described [2]. M6a isoform Ia

encodes a short N-terminal domain, whereas M6a isoform Ib

encodes a longer N-terminal domain containing a putative PKC

phosphorylation site. In the present study rat ESTs corresponding to

M6a isoforms Ia (Genebank Acc: DV216104) and Ib (Genebank

Acc: CO401660) were identified in the NCBI database and intron-

exon boundaries were mapped according to genomic rat DNA

(Contig Accession: NW_001084718). Thus, three types of primers

were synthesized for real-time RT-PCR analysis; i) primers

recognizing the 39-UTR region of M6a (common to all isoforms of

M6a); ii) primers specific for M6a isoform Ia; and iii) primers specific

for M6a isoform Ib (Table 1; Fig. 1).

A quantitative analysis of gene expression was performed using

the 7500 Real-time PCR (Applied Biosystems, Darmstadt,

Germany) in combination with Quantitect SYBR green technol-

ogy (Qiagen, Hilden, Germany). The light cycler was programmed

to the following conditions: an initial PCR activation step of

10 min at 95uC, followed by cycling steps; denaturation for 15 sec

at 95uC, annealing for 30 sec at 60uC, and elongation for 60 sec at

72uC; these steps were repeated for 40 cycles. Details of the

quantitative RT-Real time PCR have been described before [21].

Dissociation curves were generated for all PCR products to

confirm that SYBR green emission is detected from a single PCR

product [22]. The relative abundance of M6a mRNA transcripts

was calculated in reference to the mRNA levels of the internal

reference gene cyclophilin as described before [8].

In situ hybridizationFresh frozen brains from adult rats were cut on a cryostat and

10 mm cryosections were thaw-mounted on gelatine-coated slides.

Sections were dried at room temperature for 20 min, fixed in 4%

buffered paraformaldehyde (PFA, pH 7.2), rinsed in phosphate-

buffered saline (PBS; 0.1 mM phosphate buffer, 0.9% NaCl,

pH 7.2), dehydrated through graded alcohols, air dried and frozen

at 280uC. Prior to hybridization, sections were rehydrated

through graded alcohols, fixed in 4% PFA, washed in PBS,

acetylated (0.1 M triethanolamine, 0.25% acetic anhydride),

washed in PBS and dehydrated once again through graded

alcohols. M6a plasmid DNA [6] was linearized and riboprobes

were synthesized with T7 and SP6 RNA polymerases (Promega,

Madison, WI, USA) for the antisense and sense probe,

respectively, in the presence of 9.25 MBq of 33P-UTP (ICN;

specific activity 3000 Ci/mmol) for 1 h at 37uC. Probes were

purified with Microspin S-400 HR columns (Amersham Pharma-

cia, Freiburg, Germany) and hybridization buffer (50% deionised

formamide, 10% dextran sulphate, 0.3 M NaCl, 1 mM EDTA,

10 mM Tris-HCl, ph 8.0, 500 mg/ml tRNA, 0.1 M dithiothreitol,

and 16 Denhardt’s solution) was added to give a final probe

activity of 56104 CPM. The hybridization mixture containing the

probe was denatured at 70uC for 10 min, cooled to 55uC, and

pipetted directly onto sections (80 ml/section). Hybridization was

performed for 18 hrs at 43uC. Sections were subsequently washed

in 46 SSC (0.6 M NaCl, 0.06 M citric acid), 26 SSC, and 0.56SSC for 10 min each at 37uC. Following 1 hr incubation at 70uCin 0.26 SSC, sections were washed twice in 0.16 SSC, once at

37uC and again at room temperature, for 10 min each. Sections

were dehydrated through graded alcohols, air dried, and exposed

to Bio-Max MR film (Amersham Pharmacia, Freiburg, Germany)

Table 1. Primer used for quantitative RT-PCR.

Primer Pairs Forwards Reverse

M6a 39UTR 59-TTCAACGTGTGGACCATCTGC 59-AGAGATTTGCTCCCTCCACGAG

M6a Isoform Ia 59-GCCTGCCTGGTCTTTACACTTC 59-CACTCAAAACACCCCATATCCA

M6a Isoform Ib 59-CCTGAAGAAAGGTAGCCATGGA 59-GCAGCACTCAAAACACCCTTTT

Cyclophilin 59-CAAATGCTGGACCCAACACA 59-TGCCATCCAACCACTCAGTCT

doi:10.1371/journal.pone.0003659.t001

Stress Regulates M6a in Axons


for 4 days at 4uC. Films were developed and fixed with GBX

(Kodak, Rochester, NJ, USA).

Quantitative in situ hybridizationRat brains were prepared for cryosectioning under RNAse-free

conditions as previously described [21]. Serial, anatomically

matched cryosections from both control (n = 9) and stress (n = 9)

animals were thaw-mounted on gelatin-coated slides from the level

of the prefrontal cortex (bregma position 4.2 to 2.2) [23] and

hippocampus (bregma position 22.8 to 24.3). Hippocampal

cryosections were mounted in pairs (one control, one stress section

per slide) and prefrontal cortical sections in groups of four (two

control and two stress sections per slide). Individual slides thus held

sections from each experimental group to minimize variations in

hybridization conditions between experimental groups. Following

hybridization (as described above), sections were coated with

photoemulsion (Kodak NBT) at 42uC, dried for 90 min at RT,

and stored for 7 weeks at 4uC in a light-proof container. Exposed

slides were developed at 15uC for 5 min (Kodak developer D-19),

rinsed twice briefly in H20, fixed 5 min at RT (fixer, Kodak

Polymax). Sections were counterstained with methyl-green (M-

8884, Sigma), cleared in xylol, and coverslipped with mounting

medium (Eukitt, Kindler, Freiburg, Germany). Hybridized

sections were visualized with a 406 objective (NA = 1.4; Zeiss,

Jena, Germany) under a light microscope (Axioscope, Zeiss) and

silver grain quantification was performed on a cell by cell basis

using the silver grain count function of MCID Basic software

(Imaging Research Inc., St. Catherines, Ontario, Canada). ROD

(relative optical density) threshold intensities were optimized to

exclusively detect exposed silver grains: background interference

from methyl-green was eliminated by the introduction of a green

filter during quantification. The number of pixels contained within

an individual silver grain was determined and used in subsequent

calculations to extrapolate the number of silver grains within the

area of interest. Circular counting masks of 125 pixel diameter

were used to estimate silver grain number in hippocampal region

CA3 and in prefrontal pyramidal neurons, whereas a smaller

counting mask of 100 pixel diameter corresponding approximately

to the size of a granule neuron cell body was used in the dentate

gyrus to account for the tight packing of neurons within the

granule cell layer. Boundaries delineating cortical laminae and the

sub-areas of the prefrontal cortex were determined according to

the published anatomical findings of Gabbott et al. [24]. Silver

grain estimates were calculated from 2 sections per animal and 100

neurons per section within the dentate gyrus, CA3 pyramidal cell

layer, anterior cingulate cortex, prelimbic cortex, and infralimbic

cortex, respectively. For statistical analysis, the mean number of

silver grains/brain area/rat was calculated and the individual data

from stressed animals and controls were compared with the

Student’s t-test. Differences were regarded significant at p#0.05.

Immunocytochemistry for light microscopyAnimals received a lethal dose of ketamine, 50 mg/ml; xylazine,

10 mg/ml; atropine, 0.1 mg/ml) and were transcardially perfused

first with saline (0.9% NaCl, for 2 min) and then with 4%

paraformaldehyde in PBS (pH 7.2; for 10 min). Brains were

removed, washed overnight in PBS and immersed in cryoprotec-

tant (2% DMSO, 20% glycerol in 0.125 M PBS, pH 7.2) until

Figure 1. Expression of M6a isoforms Ia and Ib in rat brain and kidney. A: The M6a gene comprises 7 exons (E I to E VII). Because of alternatetranscription start sites, two N-terminus variants of M6a transcripts are generated with sequences corresponding to either exon Ia (E Ia) whichencodes a short N-terminal domain, or exon Ib (E Ib) which encodes a longer N-terminal domain. B: Constitutive expression of M6a isoforms Ia and Ibin rat brain regions and kidney. For each sample, M6a mRNA transcripts were quantified by RT-PCR and expressed as a percentage of mRNA for theinternal reference gene cyclophilin. Data are mean6SEM (standard error of the mean) from n = 9 animals. Statistical differences according toStudent’s t-test: *, p,0.05; **, p.0.01. UTR, untranslated region.doi:10.1371/journal.pone.0003659.g001



saturation. Coronal cryosections (40 mm) were collected through

prefrontal and hippocampal regions, washed briefly in PBS and

quenched of endogenous peroxidase activity by 30 min incubation

at room temperature (RT) in 0.5% H202 in distilled water.

Sections were washed in 0.5% Triton X-100 (TX-100) in PBS,

blocked for 1 hr at RT (5% normal rabbit serum and 0.5% TX-

100 in PBS), incubated 48 hr at 4uC with monoclonal anti-M6a

rat IgG (Medical & Biological Laboratories Co., Ltd, Japan;

1:1000 dilution in 1% normal rabbit serum and 0.5% TX-100 in

PBS), and washed again. Sections were then incubated in blocking

solution (5% normal rabbit serum and 0.5% TX-100 in PBS) for

1 hr at RT, incubated with biotin-conjugated rabbit anti-rat IgG

(DAKO, Hamburg, Germany; 1:400 dilution in 1% normal rabbit

serum and 0.5% TX-100 in PBS) for 4 hr at RT, then washed

overnight at 4uC. The sections were treated with streptavidin-

HRP (DAKO; 1:200 dilution in 1% normal rabbit serum and

0.5% TX-100 in PBS) for 2 hr at RT, washed in PBS and then

again in 0.05 M Tris/HCl (pH 7.2) prior to DAB detection (DAB

detection was performed according to the manufacturer’s

instructions; DAB-Kit, Vector Laboratories, USA). Sections were

washed in 0.05 M Tris/HCl (pH 7.6) and again in 0.1 M PBS

prior to xylol clearance, dehydration, and coverslipping with

Eukitt mounting medium (Kindler).

Immunofluorescence and confocal microscopyAntibodies used in double-labelling experiments were applied

sequentially and blocking steps were performed using normal

serum of host species from which respective secondary antibodies

were derived. Cryostat sections (40 mm) from prefrontal cortex

and hippocampus were rinsed in normal PBS and non-specific

antibody binding sites were blocked with 3% normal serum, 0.3%

TX-100 in PBS, for 1 hr at 4uC. Sections were then incubated in

rat monoclonal anti-M6a (1/1500; in 3% normal serum, and 0.3%

TX-100 in PBS) for 24 hr at 4uC, washed, and incubated in

secondary antiserum (Alexa 594-coupled donkey anti-rat (Molec-

ular Probes, Invitrogen, Leiden, the Netherlands) dilution 1/300

for 2 hr in a light proof container. Sections were washed and

incubated in either rabbit anti-synaptophysin (Synaptic Systems,

Gottingen, Germany), dilution 1/1000, or in mouse monoclonal

anti microtubule-associated protein (MAP-2; Sigma), dilution 1/

2000 in 3% normal serum, 0.5% TX-100 in PBS over night.

Sections were then washed and incubated 2 hr at 4uC in

secondary antiserum diluted 1/300 in 0.5% TX-100 in PBS:

Alexa 488-coupled goat anti-rabbit IgG or Alexa 488-coupled goat

anti-mouse IgG (Molecular Probes), respectively. Thereafter,

sections were washed in PBS and floated/mounted on Histobond

slides in PBS, allowed to dry overnight at 4uC and coverslipped

with mounting medium (DakoCytomation, DAKO, Glostrup,

Denmark).

Confocal microscopy was performed with a laser scanning

microscope (LSM 5 Pascal, Zeiss, Gottingen, Germany) with an

argon 488 nm and a helium/neon 543 nm laser. Analysis was

performed in multiple tracking mode to avoid bleed-through

between channels. The 543-nm laser was always used with a

smaller detection pinhole diameter than the 488-nm laser to

obtain the same optical slice thickness (slice thickness typically

between 0.5–1.0 mm). High magnification images were obtained

with an Apochromat 636 oil objective (NA = 1.4) and immersion

oil (Immersol, Zeiss; refractive index = 1.518).

Chronic restraint stressFor the experiment, male rats were housed individually in

separate cages. Animals were randomly divided into two groups

(n = 9/group) and allowed to habituate to the housing conditions

and to daily handling for 10 days prior to the onset of

experimentation. To expose rats to stress, we used a modified

protocol of an established restraint stress paradigm [12,13].

Accordingly, animals of the ‘Stress’ group were restrained daily

for six hours (from 10:00 to 16:00, that is during the dark phase)

for a total of 21 days in well-ventilated polypropylene tubes

without access to food and water. Food was withheld from control

animals during the restraint period to ensure that any effect on

body weight gain was not simply a result of limited food

availability. During restraint, animals were not physically

compressed and did not experience pain. Bodyweights were

recorded daily prior to the onset and during the entire period of

daily restraint. For statistical evaluation, a day-by-day comparison

of body weights was performed with paired t-tests using GraphPad

Prism 4.03 (GraphPad Software, Inc., La Jolla, CA, USA).

Differences were regarded significant at p#0.05.

At the end of the experiment, 24 hrs following the last restraint,

all animals were weighed and subsequently sacrificed. Brains were

quickly dissected and adrenal glands were removed and weighed

for analysis of relative adrenal weight.

Results

M6a splice variants Ia and IbA comparative real-time RT-PCR analysis of M6a transcript

expression was performed in the brain and kidneys using primers

specific for M6a isoforms Ia and Ib, and for the 39-UTR region of

the M6a transcript which is common to both isoforms (Fig. 1A).

The results indicate that N-terminus variants of M6a are

differentially expressed in central and peripheral tissues. M6a

isoform Ia was found to be ubiquitously expressed at a low level in

both brain and kidney, whereas variant Ib was identified as the

predominant isoform expressed in the brain, especially in the

hippocampal formation (Fig. 1B).

M6a expression in hippocampal formation and prefrontalcortex

We visualized M6a mRNA expression in the hippocampal

formation and in the medial prefrontal cortex (mPFC) using in situ

hybridization. The gene is strongly expressed in the pyramidal

neurons of all hippocampal subfields (CA1–CA4) and in the

granule cells of the dentate gyrus (Fig. 2a). Whereas M6a mRNA is

concentrated in the cell bodies of the principal neurons, M6a

protein is found in processes of those cells. Immunocytochemistry

reveals that all hippocampal layers containing dense fiber networks

are stained (Fig. 3). Strong M6a immunoreactivity is especially

found in the stratum lucidum, the area where mossy fibers

originating from the dentate gyrus granule neurons synapse on

dendrites of CA3 pyramidal neurons (Fig. 3B). Immunofluores-

cence reveals that the membrane protein is concentrated in the

mossy fiber axons (Fig. 4A). The giant mossy fiber terminals of

these glutamatergic axons are strongly stained with the antibody

against the synaptic vesicle protein synaptophysin. Co-staining

with MAP-2 antibody which labels neuronal dendrites and cell

bodies of pyramidal neurons reveals that M6a is not present in

dendrites and cell bodies (Fig. 4C).

Moderate M6a mRNA expression is found in the three mPFC

sub-areas, anterior cingulate, prelimbic and infralimbic cortex

(Fig. 2B). In the mPFC, M6a immunoreactivity of cross cut axons

appears as puncta which surround the somata of pyramidal

neurons that are not stained with the synaptophysin antibody

(Fig. 4B) but with MAP-2 antibody (Fig. 4D). These data confirm

our previous results showing that M6a is a component of the

membrane of glutamatergic axons but not of dendrites [6].



Figure 2. Autoradiograms showing M6a expression in the hippocampal formation (A) and the prefrontal cortex (B) as revealed byin situ hybridization. Abbreviations: ACx, anterior cingulated cortex; CA1, hippocampal region CA1; CA3, hippocampal region CA3; CA4,hippocampal region CA4; DG, dentate gyrus; gcl, granule cell layer; IL, infralimbic cortex; PL, prelimbic cortex; pyr, pyramidal cell layer.doi:10.1371/journal.pone.0003659.g002

Figure 3. Immunocytochemical detection of M6a expression in the hippocampus. (A) shows no immunoreactivity in the granule cell layer(gcl) whereas the hilus (h) is strongly stained. A laminated pattern of immunoreactivity is detected in the molecular layer (ml) of the dentate gyrus, instratum radiatum (rad) and stratum oriens (or) of region CA1. B (enlarged area from the box in A), mossy fibers terminating in the stratum lucidum(str.luc.) are strongly labeled by the M6a antibody whereas pyramidal neurons (pyr) are not stained.doi:10.1371/journal.pone.0003659.g003



Physiological effects of chronic stressCoinciding with what has been shown previously [19] chronic

restraint stress reduces body weight in male rats. Body weight

gain differed significantly between rats submitted to daily

restraint stress and controls (p,0.001, Student’s t-test; Fig. 5A).

At the end of the 3 weeks period of daily immobilization,

the weight of adrenal glands relative to body weight was

significantly increased compared to controls reflecting enhanced

activity of the hypothalamus-pituitary-adrenal axis in stressed

animals (Fig. 5B).

M6a transcript expression after chronic stressThe effect of 21 days chronic restraint stress on M6a expression

in specific brain regions was quantified with real-time PCR (Fig. 6).

M6a 39-UTR primers revealed a significant down-regulation of

M6a transcripts (65% of controls, p,0.01) in the hippocampus of

stressed animals. Subsequent analyses with isoform-specific

primers demonstrated that isoform Ib (73% of controls, p,0.05),

but not isoform Ia, is significantly reduced by stress in the

hippocampal formation. RT-PCR detected no significant effect of

stress on M6a expression in the prefrontal cortex, however, both

isoforms Ia and Ib showed a tendency towards upregulation by

stress, but failed to reach significance.

Quantitative in situ hybridization was performed to investigate

the effects of chronic restraint stress on M6a expression in neurons

of the hippocampal subfields and of the prefrontal cortex.

Hybridization signals represented by silver grains appear as black

puncta clustered over cells which were counterstained with

methyl-green appearing blue (Fig. 7, bottom). M6a expression in

granule neurons of dentate gyrus and CA3 pyramidal neurons was

reduced to 85% (p,0.05) of controls. No effect of stress on M6a

expression was detected in the anterior cingulate cortex, however,

significant increases to approximately 112% (p,0.05) of controls

were detected in the prelimbic and infralimbic cortex (Fig. 7).

Discussion

It has been shown in the past that stress alters the structural

organization of dendrites and of synapses on pyramidal neurons.

Moreover, it has been concluded that such stress-induced changes

would affect information transfer between the cells that communi-

cate via axo-dendritic synapses [25]. The present data show that also

axons are affected by stress. Daily restraint stress for three weeks

reduces M6a expression in glutamatergic neurons of the hippocam-

pal formation and may thus affect the structural integrity of the

axons of those neurons. Since projections from CA3 hippocampal

pyramidal neurons comprise a subset of axonal inputs to nuclei

within the medial prefrontal cortex [26–27] our findings indicate

that stress may affect communication between brain regions.

M6a isoforms and axonal localizationThe proteolipids including M6a, M6b, and PLP (major myelin

proteolipid protein) are among the most abundantly expressed

genes in the brain [28,29]. The present quantitative RT-PCR

analysis reveals that N-terminus variants of M6a are constitutively

expressed at different levels within central and peripheral tissue:

M6a isoform Ia is expressed at low levels in brain and kidney

epithelia, whereas isoform Ib is highly expressed in the brain, but

at very low levels in the kidneys. Previous studies addressing the

function of M6a have suggested a role in ion transport, a

hypothesis based initially on the immunolocalization of M6a to

neuronal plasma membranes and the apical surface of polarized

epithelia, both of which rely heavily on the coordinated transport

of ions across membranes [4]. The findings of the present study

reveal that distinct isoforms of M6a are differentially localised to

neuronal and epithelial membranes, suggesting that splice variants

of M6a may serve different functions in central and peripheral

tissues.

We have previously shown by immunocytochemistry that M6a

is present in axons of glutamatergic neurons with the strongest

immunoreactivity being detected within the hippocampal forma-

tion [6]. The present data further confirm this: the membrane

glycoprotein is located in the mossy fibers that originate in the

dentate gyrus granule neurons and synapse on dendrites of CA3

pyramidal neurons. The giant axon terminals themselves are

largely not labelled by the M6a antibody but are strongly stained

by the antibody that binds to the synaptic vesicle marker protein

synaptophysin. Colocalization is only observed as a result of close

proximity between synaptophysin-immunoreactive vesicles and

M6a as a component of the terminal membrane.

Hippocampal pyramidal neurons were also found to express

M6a, with no apparent difference in expression levels observed

between subfields of the cornu ammonis (CA). Axonal projections

from CA3 pyramidal neurons within the hippocampal formation

include Schaffer collaterals terminating on the dendrites of CA1

Figure 4. Immunofluorescence showing M6a immunoreactivityin the stratum lucidum (str.luc.) of the hippocampus (A, C) andin the infralimbic cortex (B, D) as revealed by confocalmicroscopy. A: M6a (red) and synaptophysin immunoreactivity(green) in the stratum lucidum. Note that the giant mossy fiberterminals (mft) which are strongly synaptophysin positive surround theunlabeled dendrites (d) of the CA3 pyramidal neurons. B: M6a (red) andsynaptophysin immunoreactivity (green) in the infralimbic cortex; axons(ax) are M6a immunopositive, dendrites (d) are unlabeled. Note thatthere is occasional colocalization (yellow) of M6a and synaptophysin inthe axonal terminals that surround the unlabeled soma of thepyramidal neuron (pyr). C: M6a (red) and MAP-2 immunoreactivity(green) in the stratum lucidum. Note that the dendrites (d) of CA3pyramidal neurons which are strongly MAP-2 positive are not labelledby the M6a antibody. D: M6a (red) and MAP-2 immunoreactivity (green)in the infralimbic cortex. Note that there is no colocalization of M6a andMAP-2 in the axons/axonal terminals (ax) that surround the MAP-2immunopositive soma of the pyramidal neurons (pyr).doi:10.1371/journal.pone.0003659.g004



pyramidal neurons within stratum radiatum, and associational

projections terminating on the apical dendrites of CA3 pyramidal

neurons within stratum radiatum. Schaffer collaterals diverge

extensively throughout the longitudinal axis of the hippocampal

formation [30] and are therefore not visualized as a coherent fiber

pathway. Instead, M6a targeted to the terminal regions of CA3

projections is primarily detected as synaptic puncta within the

stratum radiatum.

M6a is downregulated by chronic stress in thehippocampus

As determined by quantitative real-time RT-PCR chronic

restraint selectively downregulates neuronally expressed M6a

isoform Ib in the hippocampus, but not isoform Ia. M6a was

initially identified by subtractive hybridization as a glucocorticoid

responsive gene in tree shrews chronically treated with cortisol [7],

suggesting stress-induced reductions in hippocampal M6a expression

may occur via glucocorticoid-regulated repression of transcription

[31]. Downregulation of M6a mRNA in the hippocampal formation

of chronically restrained rats is consistent with previous data

demonstrating reduced M6a expression in the hippocampus of

chronically restrained mice [8]. Moreover, M6a mRNA was also

downregulated in the hippocampal formation of psychosocially

stressed tree shrews [11] indicating that the effect of stress on M6a

expression is robustly conserved across species and is reproducible

with different stress paradigms.

M6a in the prefrontal cortexThe medial prefrontal cortex comprises functionally distinct

sub-areas of which the PL and IL are particularly involved in the

integration of autonomic and cognitive information ultimately

contributing to the perception of stress [32–37]. In situ

hybridization with emulsion autoradiography showed that M6a

is abundantly expressed in all mPFC layers in large neurons

bearing the morphological characteristics of pyramidal neurons.

Moreover, pyramidal neurons within layers II/III exhibited

comparable levels of expression in all sub-areas examined, ACx,

PL and IL, as determined by quantitative in situ hybridization.

PFC pyramidal neurons receive synaptic inputs in an organized

fashion: Distal portions of the apical dendritic tree (cortical layer I)

receive inputs primarily from extracortical regions, such as the

medial dorsal thalamic nuclei and hippocampal CA3 subfield

[38,39], whereas proximal portions of apical and basilar dendrites

receive inputs primarily from local cortical circuits [40].

Figure 5. Effects of chronic restraint stress on body weight gain and adrenal weight. Left: Animals exposed to chronic restraint stressexhibited reduced body weight gain coinciding with the onset of restraint; data are from 11 controls and 10 stressed animals. Right: Adrenal weightof stressed rats is significantly increased compared to controls (10 animals/group). Data are expressed as mean6SEM (standard error of the mean).Significant differences between groups as determined by Student’s t-test: *, p,0.05.doi:10.1371/journal.pone.0003659.g005

Figure 6. Quantitative real-time PCR showing M6a expression inthe hippocampus (upper pannel) and prefrontal cortex (lowerpannel). Analysis with primers recognizing the 39-UTR of M6a, which iscommon to all isoforms of M6a, revealed a significant stress-induceddownregulation in the hippocampus. Subsequent analysis with isoform-specific primers showed that M6a isoform Ib, but not Ia, is regulated bystress. In the prefrontal cortex, both isoforms show a tendency towardsupregulation but fail to reach significance. Data are expressed aspercentage of the mean control6SEM (standard error of the mean),n = 9/group. Significant differences between groups as determined byStudent’s t-test: *, p,0.05, **, p,0.01. UTR, untranslated region.doi:10.1371/journal.pone.0003659.g006



Quantitative RT-PCR permits the detection of changes in gene

expression with high sensitivity, however, the anatomical specific-

ity of data generated with this method relies on the ability to

precisely excise the tissue/cells of interest. As described, we

detected a tendency towards increased M6a expression in

chronically restrained rats, but this tendency failed to reach

significance. Since significant changes in gene expression within

the mPFC may be masked in a combined analysis of all sub-areas,

in situ hybridization was performed which allows a semi-

quantitative evaluation of mRNA transcript abundance in single

neurons. From a methodological perspective, in situ hybridization

demonstrated less sensitivity to stress-induced changes in hippo-

campal M6a expression compared to quantitative RT-PCR

analyses, but enabled expression to be quantified within specific

neurons. In the mPFC, three weeks restraint increased M6a

expression in pyramidal neurons (layers II/III) of PL and IL

whereas no change of expression was observed in the ACx.

In previous studies, dendritic remodelling observed in pyramidal

neurons within layers II/III of the mPFC was interpreted to represent

an adaptive response to altered synaptic input from extracortical

sources such as the CA3 region of the hippocampus [41]. It is

conceivable that the increased M6a expression observed in pyramidal

neurons of PL/IL reflects an adaptive mechanism designed to

strengthen associational contacts and in doing so, to sensitize

pyramidal neurons to weakened inputs form the hippocampus.

Possible implications of M6a regulation in glutamatergicaxons

As shown in our previous study [6] M6a is present in axonal

membranes and may as such play an important role in the

structural integrity of axons. Since the membrane glycoprotein is

in particular strongly expressed in the mossy fibers one has to

assume that stress changes the integrity of those axonal projections

from the granule cells to CA3 pyramidal neurons. Indeed, three

weeks of daily restraint changed the morphology of the giant

mossy fiber terminals in the stratum lucidum [25]. Maladaptive

changes in mossy fiber terminal morphology induced by stress are

likely to have a profound impact on transmission within the

Figure 7. M6a expression in neurons of the hippocampal formation and the mPFC; quantitative in situ hybridization with emulsionautoradiography. Upper panel: Numbers of silver grains per neuron reveal reduced M6a mRNA expression after stress in dentate gyrus granuleneurons and in CA3 pyramidal neurons, but enhanced M6a mRNA expression in neurons of the prelimbic and infralimbic cortex. Lower panel:Examples of sections from the dentate gyrus (left) and the infralimbic cortex (right) showing silver grains over cells that were counter stained withmethyl-green (cyan). Significant differences between groups as determined by Student’s t-test: *, p,0.05, **, p,0.01.doi:10.1371/journal.pone.0003659.g007



hippocampal circuits and may contribute to perturbations in

glutamatergic transmission previously associated with chronic

stress [42–43]. Moreover, also stress-induced changes in other

hippocampal subregions such as CA1 may be related to reduced

M6a expression [44]. Altogether, these changes may contribute to

the inhibition of long-term potentiation that has been recorded in

the hippocampus after stress [45–46].


Conceived and designed the experiments: EF GF. Performed the

experiments: BC. Analyzed the data: BC. Contributed reagents/materi-

als/analysis tools: EF. Wrote the paper: GF.

References

1. Yan Y, Narayanan V, Lagenaur C (1996) Expression of members of theproteolipid protein gene family in the developing murine central nervous system.

J.Comp Neurol 370: 465–478.2. Werner H, Dimou L, Klugmann M, Pfeiffer S, Nave KA (2001) Multiple splice

isoforms of proteolipid M6B in neurons and oligodendrocytes. Mol.Cell

Neurosci 18: 593–605.3. Yan Y, Lagenaur C, Narayanan V (1993) Molecular cloning of M6:

identification of a PLP/DM20 gene family. Neuron 11: 423–431.4. Lagenaur C, Kunemund V, Fischer G, Fushiki S, Schachner M (1992)

Monoclonal M6 antibody interferes with neurite extension of cultured neurons.

J Neurobiol 23: 71–88.5. Alfonso J, Fernandez ME, Cooper B, Flugge G, Frasch AC (2005) The stress-

regulated protein M6a is a key modulator for neurite outgrowth and filopodium/spine formation. Proc Natl Acad Sci U.S.A 102: 17196–17201.

6. Cooper B, Werner HB, Flugge G (2008) Glycoprotein M6a is present inglutamatergic axons in adult rat forebrain and cerebellum. Brain Res 1197:

1–12.

7. Alfonso J, Aguero F, Sanchez DO, Flugge G, Fuchs E, et al. (2004a) Geneexpression analysis in the hippocampal formation of tree shrews chronically

treated with cortisol. J Neurosci Res 78: 702–710.8. Alfonso J, Frick LR, Silberman DM, Palumbo ML, Genaro AM, et al. (2006)

Regulation of hippocampal gene expression is conserved in two species subjected

to different stressors and antidepressant treatments. Biol Psychiatry 59: 244–251.9. McEwen BS (1998) Stress, adaptation, and disease. Allostasis and allostatic load.

Ann NY Acad Sci 840: 33–44.10. Kendler KS, Karkowski LM, Prescott CA (1999) Causal relationship between

stressful life events and the onset of major depression. Am J Psychiatry 156:837–841.

11. Alfonso J, Pollevick GD, Van Der Hart MG, Flugge G, Fuchs E, et al. (2004b)

Identification of genes regulated by chronic psychosocial stress and antidepres-sant treatment in the hippocampus. Eur J Neurosci 19: 659–666.

12. Magarinos AM, McEwen BS (1995) Stress-induced atrophy of apical dendritesof hippocampal CA3c neurons: comparison of stressors. Neuroscience 69:

83–88.

13. McLaughlin KJ, Gomez JL, Baran SE, Conrad CD (2007) The effects of chronicstress on hippocampal morphology and function: an evaluation of chronic

restraint paradigms. Brain Res 1161: 56–64.14. Watanabe Y, Gould E, McEwen BS (1992) Stress induces atrophy of apical

dendrites of hippocampal CA3 pyramidal neurons. Brain Res 588: 341–345.15. Magarinos AM, McEwen BS, Flugge G, Fuchs E (1996) Chronic psychosocial

stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in

subordinate tree shrews. J Neurosci 16: 3534–3540.16. Brown SM, Henning S, Wellman CL (2005) Mild, short-term stress alters

dendritic morphology in rat medial prefrontal cortex. Cereb Cortex 15:1714–1722.

17. Cook SC, Wellman CL (2004) Chronic stress alters dendritic morphology in rat

medial prefrontal cortex. J Neurobiol 60: 236–248.18. Radley JJ, Sisti HM, Hao J, Rocher AB, McCall T, et al. (2004) Chronic

behavioral stress induces apical dendritic reorganization in pyramidal neurons ofthe medial prefrontal cortex. Neuroscience 125: 1–6.

19. Perez-Cruz C, Muller-Keuker JI, Heilbronner U, Fuchs E, Flugge G (2007)

Morphology of pyramidal neurons in the rat prefrontal cortex: lateralizeddendritic remodeling by chronic stress. Neural Plast 46276: PMID 18253468.

20. Czeh B, Perez-Cruz C, Fuchs E, Flugge G (2008) Chronic stress-induced cellularchanges in the medial prefrontal cortex and their potential clinical implications:

does hemisphere location matter? Behav Brain Res 190: 1–13.21. Abumaria N, Ribic A, Anacker C, Fuchs E, Flugge G (2008) Stress upregulates

TPH1 but not TPH2 mRNA in the rat dorsal raphe nucleus: identification of

two TPH2 mRNA splice variants. Cell Mol Neurobiol 28: 331–42.22. Ririe KM, Rasmussen RP, Wittwer CT (1997) Product differentiation by

analysis of DNA melting curves during the polymerase chain reaction. AnalBiochem 245(2): 154–160.

23. Paxinos G, Watson C (1986) The rat brain in stereotaxic coordinates. Second

Edition. New York: Academic Press.

24. Gabbott PL, Dickie BG, Vaid RR, Headlam AJ, Bacon SJ (1997) Local-circuitneurones in the medial prefrontal cortex (areas 25, 32 and 24b) in the rat:

morphology and quantitative distribution. J Comp Neurol 377: 465–499.25. Magarinos AM, Verdugo JM, McEwen BS (1997) Chronic stress alters synaptic

terminal structure in hippocampus. Proc Natl.Acad Sci USA 94: 14002–14008.

26. Groenewegen HJ, Wright CI, Uylings HB (1997) The anatomical relationshipsof the prefrontal cortex with limbic structures and the basal ganglia.

J Psychopharmacol 11: 99–106.27. Vertes RP (2006) Interactions among the medial prefrontal cortex, hippocampus

and midline thalamus in emotional and cognitive processing in the rat.

Neuroscience 142: 1–20.28. Nave KA, Lai C, Bloom FE, Milner RJ (1987) Splice site selection in the

proteolipid protein (PLP) gene transcript and primary structure of the DM-20protein of central nervous system myelin. Proc Natl Acad Sci USA 84:

5665–5669.29. Huminiecki L, Lloyd AT, Wolfe KH (2003) Congruence of tissue expression

profiles from Gene Expression Atlas, SAGEmap and TissueInfo databases.

BMC.Genomics 4: 31.30. Witter MP (1993) Organization of the entorhinal-hippocampal system: a review

of current anatomical data. Hippocampus 3 Spec No: 33–44.31. Meijer OC (2002) Coregulator proteins and corticosteroid action in the brain.


32. Terreberry RR, Neafsey EJ (1987) The rat medial frontal cortex projects directlyto autonomic regions of the brainstem. Brain Res Bull 19: 639–49.

33. Diorio D, Viau V, Meaney MJ (1993) The role of the medial prefrontal cortex(cingulate gyrus) in the regulation of hypothalamic-pituitary-adrenal responses to

stress. J Neurosci 13: 3839–3847.34. Sullivan RM, Gratton A (1999) Lateralized effects of medial prefrontal cortex

lesions on neuroendocrine and autonomic stress responses in rats. J Neurosci 19:

2834–2840.35. Cardinal RN, Parkinson JA, Hall J, Everitt BJ (2002) Emotion and motivation:

the role of the amygdala, ventral striatum, and prefrontal cortex. NeurosciBiobehav Rev 26: 321–352.

36. Sullivan RM (2004) Hemispheric asymmetry in stress processing in rat prefrontal

cortex and the role of mesocortical dopamine. Stress 7: 131–143.37. Gabbott PL, Warner TA, Jays PR, Salway P, Busby SJ (2005) Prefrontal cortex

in the rat: projections to subcortical autonomic, motor, and limbic centers.J Comp Neurology 492: 145–177.

38. Swanson LW, Cowan WM (1977) An autoradiographic study of theorganization of the efferent connections of the hippocampal formation in the

rat. J Comp Neurol 172: 49–84.

39. Groenewegen HJ (1988) Organization of the afferent connections of themediodorsal thalamic nucleus in the rat, related to the mediodorsal-prefrontal

topography. Neuroscience 24: 379–431.40. Scheibel ME, Scheibel AB (1970) Of pattern and place in dendrites. Int Rev

Neurobiol 13: 1–26.

41. Wellman CL (2001) Dendritic reorganization in pyramidal neurons in medialprefrontal cortex after chronic corticosterone administration. J Neurobiol 49:

245–253.42. Lowy MT, Gault L, Yamamoto BK (1993) Adrenalectomy attenuates stress-

induced elevations in extracellular glutamate concentrations in the hippocam-

pus. J Neurochem 61: 1957–1960.43. Sousa N, Lukoyanov NV, Madeira MD, Almeida OF, Paula-Barbosa MM

(2000) Reorganization of the morphology of hippocampal neurites and synapsesafter stress-induced damage correlates with behavioral improvement. Neurosci-

ence 97: 253–266.44. Donohue HS, Gabbott PL, Davies HA, Rodriguez JJ, Cordero MI, et al. (2006)

Chronic restraint stress induces changes in synapse morphology in stratum

lacunosum-moleculare CA1 rat hippocampus: a stereological and three-dimensional ultrastructural study. Neuroscience 140: 597–606.

45. Xu L, Anwyl R, Rowan MJ (1997) Behavioural stress facilitates the induction oflong-term depression in the hippocampus. Nature 387: 497–500.

46. Pavlides C, Nivon LG, McEwen BS (2002) Effects of chronic stress on

hippocampal long-term potentiation. Hippocampus 12: 245–257.



Opposite Effects of Early Maternal Deprivation onNeurogenesis in Male versus Female RatsCharlotte A. Oomen, Carlos E. N. Girardi, Rudy Cahyadi, Eva C. Verbeek, Harm Krugers, Marian Joels,

Paul J. Lucassen*

SILS Centre for Neuroscience, University of Amsterdam, Amsterdam, The Netherlands

Abstract

Background: Major depression is more prevalent in women than in men. The underlying neurobiological mechanisms arenot well understood, but recent data shows that hippocampal volume reductions in depressed women occur only whendepression is preceded by an early life stressor. This underlines the potential importance of early life stress, at least inwomen, for the vulnerability to develop depression. Perinatal stress exposure in rodents affects critical periods of braindevelopment that persistently alter structural, emotional and neuroendocrine parameters in adult offspring. Moreover,stress inhibits adult hippocampal neurogenesis, a form of structural plasticity that has been implicated a.o. in antidepressantaction and is highly abundant early postnatally. We here tested the hypothesis that early life stress differentially affectshippocampal structural plasticity in female versus male offspring.

Principal Findings: We show that 24 h of maternal deprivation (MD) at PND3 affects hippocampal structural plasticity atPND21 in a sex-dependent manner. Neurogenesis was significantly increased in male but decreased in female offspring afterMD. Since no other structural changes were found in granule cell layer volume, newborn cell survival or proliferation rate,astrocyte number or gliogenesis, this indicates that MD elicits specific changes in subsets of differentiating cells anddifferentially affects immature neurons. The MD induced sex-specific effects on neurogenesis cannot be explained bydifferences in maternal care.

Conclusions: Our data shows that early environment has a critical influence on establishing sex differences in neuralplasticity and supports the concept that the setpoint for neurogenesis may be determined during perinatal life. It istempting to speculate that a reduced level of neurogenesis, secondary to early stress exposure, may contribute tomaladaptation of the HPA axis and possibly to the increased vulnerability of women to stress-related disorders.

Citation: Oomen CA, Girardi CEN, Cahyadi R, Verbeek EC, Krugers H, et al. (2009) Opposite Effects of Early Maternal Deprivation on Neurogenesis in Male versusFemale Rats. PLoS ONE 4(1): e3675. doi:10.1371/journal.pone.0003675



Copyright: � 2009 Oomen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: PJL is supported by the Netherlands Brain Foundation. The funders had no role in study design, data collection and analysis, decision to publish, orpreparation of the manuscript.



Introduction

A multitude of studies have implicated alterations in the

hypothalamo-pituitary-adrenal (HPA) axis in the vulnerability to

develop stress-related disorders like major depression [1–7]. One of

the brain areas sensitive to stress and stress hormones is the

hippocampus, a region involved in learning and memory, behavioral

adaptation and HPA-axis regulation [8–10] and richly endowed with

glucocorticoid (GR) and mineralocorticoid receptors (MR), [11].

Chronic exposure to stress can affect both hippocampal

function and structure. Consistent reductions in hippocampal

volume have e.g. been reported in major depression, as a

predictive factor rather than a consequence of the disorder [12].

Although the functional implications and the biological substrates

that underlie hippocampal volume reductions are ill understood,

animal studies have shown that chronic stress can induce cellular

and dendritic atrophy, alter glia cell numbers and reduce adult

neurogenesis [13–20]. Adult hippocampal neurogenesis represents

a form of structural plasticity that has been implicated a.o., in

hippocampal function [21–25] and the efficacy of antidepressants

[17,26–30]. Stress and glucocorticoids potently inhibit neurogen-

esis in adult animals [18–20,31–33].

Major depression is more prevalent in women than in men [34–

38]. The neurobiological mechanisms that could account for this

difference are not well identified, but recent data show that

hippocampal volume reductions in depressed women occur only

when depression is preceded by an early life stressor [39]. This

underlines the potential importance of early life stress, at least in

women, for vulnerability to develop depression.

In rodents, exposure of the pregnant dam to stress affects critical

periods of fetal brain development that can persistently alter

structural, emotional and neuroendocrine parameters in the

offspring [40–49]. This results e.g. in altered anxiety-like behavior,

increased hypothalamic-pituitary-adrenal (HPA) axis reactivity

and memory deficits in adult life [50,51]. Prenatal restraint stress

was further shown to reduce dentate granular cell number, but

only in female offspring [52], while also other sex-related

behavioral and structural differences have been reported [48].


Postnatally, several manipulations have been shown to shape

various stress and HPA-axis parameters [53–56]. In rats, from

postnatal day (PND)3–4 to PND14 of life, basal ACTH and

corticosterone levels are kept low and the response to most

stressors is suppressed [57,58]. Escape from this suppression is seen

after maternal deprivation (MD), an early life stressor during

which pups are separated from their mother. Single, 24 h

maternal deprivation in rats resulted in increased basal cortico-

sterone levels in young [59], but not older rats [53,60,61]. Also,

stress-induced ACTH [59,61] and corticosterone responses were

increased in 24 h MD rats [53,60] while similar patterns are found

in repeated separation paradigms [56,62–65]. Most of these

endocrinological analyses were performed on male offspring, but

female specific effects occur as well [66–69].

In addition, postnatal stress affects hippocampal structure [70–

73]. Maternal deprivation or low levels of maternal care reduces

hippocampal neurogenesis in some [73,74], but not all [75]

studies. As the dentate gyrus of the hippocampus is largely formed

postnatally [76,77], effects on neurogenesis and structural

plasticity are potentially more pronounced and longer lasting

when stress is applied early in life [45,78].

Since early life stressors are important, at least in women, for the

development of depression, we here tested the hypothesis that 24 h

of MD at PND3 differentially affects hippocampal structural

plasticity in female versus male offspring. Postnatal day 3 was

chosen as a timepoint for maternal deprivation, since this day

represents the start of the development of the tertiary matrix of

granular cells during dentate gyrus formation [76,77,79]. This

tertiary matrix produces the inner shell of the granular cell layer,

i.e. the future site of adult neurogenesis. We expect this timepoint

will have considerable impact on dentate gyrus neurogenesis,

structure and possibly also function later in life. We stereologically

analyzed newborn cell proliferation and survival as well as

neurogenesis at PND21, an age at which we expect both short-

lasting as well as chronic effects of MD to be detectable. Given

recent studies showing stress effects on glia cell numbers [13], we

also analyzed the total number of GFAP-positive astrocytes and

the extent of astrogliogenesis.


Animals and breeding procedureAll experimental procedures were approved by the local animal

committee of the University of Amsterdam. To standardize the

perinatal environment, rats were bred inhouse. Thirteen male and

26 female Wistar rats (3 months old) were purchased from Harlan

(Zeist, the Netherlands) and habituated to the animal facilities for

10 days. Animals were housed in pairs with food and water

available ad libitum. During the entire experiment, rats were put

on a 12 h light/dark cycle (lights on at 8.00 a.m.), at 20uC with

40–60% humidity. After habituating, breeding started. Two

females were housed together with 1 male for one week after

which the male was separated from the females. Females were

then housed together for another week after which they were

separated. Females were daily observed for pups and when a litter

was found before 9.00 a.m., the previous day was designated as the

day of birth or postnatal day (PND) 0. Litters were left undisturbed

until PND3 and were then randomly assigned to one of the four

groups, taking into account that litters from one male were not

included in the same experimental group.

Groups and experimental designMaternal deprivation for 24 h does not only result in the

absence of maternal care, but also in lack of nutrition, which is

considered a physiological stressor. To control for this, additional

experimental groups were included in which deprived pups were

injected with glucose (maternally deprived+glucose; MDG). This

has been previously described to delay the onset of the HPA-axis

activation in mice during an MD procedure of 8 h [80]. To

control for the stress of glucose injections, two other groups were

sham-injected (control sham; CONS and maternally deprived

sham; MDS) in addition to a non-injected control group (control

undisturbed; CONU).

Pups from all four experimental groups (CONU, CONS, MDS,

MDG) were sacrificed on PND21 and brains were used to study

hippocampal neurogenesis. In order to asses effects of MD on

newborn cell survival, bromodeoxyuridine (BrdU; 75 mg/kg;

subcutaneous) was injected on PND3 (see below).

Maternal deprivationAll four groups were left undisturbed until PND3. On the

morning of PND3, one hour after the onset of the light-phase, the

dam was taken from the nest and placed in a clean cage. To

minimize stress for the dam, her cage was returned to the same

room. To avoid disturbance by the vocalization of the pups [81],

the home-cage with the litter was moved to another room and

placed on a heating pad and the litter was kept on a temperature

of 32uC for the rest of the 24 h. During deprivation, glucose

(200 mg/kg bodyweight, in a volume of 5 ml per gram body-

weight) was administered to MDG pups three times, starting at

2 hours after the onset of deprivation (11.00 am), and additionally

at 17.00 and 22.00h to compensate for the lack of nutrition during

the full 24 h. The timepoint of first injection was used to

additionally inject the birth date marker BrdU in a similar

injection volume for all four groups. Control litters were left either

undisturbed (with the exception of a BrdU injection), (CONU), or

received sham injections at all timepoints (CONS). Sham-

injections to CONS animals were performed in a minimum

amount of time, during which the mother was briefly placed in

another cage. All injections were given subcutaneously with a 50

microliter Hamilton syringe (33 Gauge, Hamilton, Switzerland) in

the skin of the neck.

The following day at 9.00 am, MDS and MDG animals were

weighed, culled to 4 males and 4 females per litter and placed back

with their mother. The CONS litters were also weighed and culled

to 8 animals, but the CONU group was left undisturbed.

Acute effects of MDTo determine the acute effects of MD on corticosterone levels,

surplus animals from culling and a subset of four litters were

sacrificed by rapid decapitation at 9.00 am on PND4. Blood

samples were collected in EDTA-containing tubes, placed on ice

and subsequently centrifuged at 5000 rpm for 20 minutes after

which the supernatant was stored at 220uC. Plasma corticoste-

rone levels were measured by means of a radioimmunoassay (MP

Biomedicals., Amsterdam, the Netherlands). In addition to blood

samples, PND4 brains were taken out rapidly and fixed by

immersion fixation in 4% paraformaldehyde in phosphate buffer

(PB 0,1 M; pH 7.4) for 48 hours. Afterwards, brains were stored

on PB with azide until further processing.

Maternal care observationsPreviously, maternal care has been shown to be increased

towards the male pups [82] and to be affected by early handling

and deprivation [83,84]. As neurogenesis could be possibly

affected by changes in maternal care [74], we therefore assessed

whether changes in the amount of maternal care directed to either

male or female pups after MD had occurred. To differentiate

Early Stress and Neurogenesis


between male and female pups, all litters were marked on a daily

basis. Since marking itself represents an additional handling

procedure, offspring of these experiments was not used for further

analysis. Two of the previously described experimental groups

were included in this experiment: the 24 h deprived group

receiving sham-injections (DS, n = 6 litters) and the control group

receiving sham injections (CS, n = 5 litters). All experimental

procedures were the same as in the original experiment, except

that all pups were labeled with a non-toxic surgical marker

(Codman) daily on either the upper part of the body or the lower

part to separate the sexes. To exclude any effects of marking either

the lower or the upper part of the body, this was alternated for

males and females between litters.

Maternal care observations were performed as described earlier

[85]. Maternal behavior of each dam was observed for five 60-

minutes observation periods per day, from PND1 until PND7.

Observation periods took place at 7.00 am, 10.00 am, 1.00 pm,

5.00 pm, and 8.00 pm, resulting in two dark-phase observations

and three light-phase observations. On each day, marking of the

pups took place within 10 minutes per litter, immediately

following the first (7.00 am) observation period. During each

60 minute observation period, behavior was scored every

3 minutes, resulting in 20 observations per period and 700

observations in total (PND1–PND7) for the CS-group and 600

observations in total for the DS-group, since PND3–4 was lacking.

Licking and grooming (with or without nursing) was scored for

male and female pups. The scoring of other behaviors, however,

was done for the litter as a whole and included: a) arch back

nursing (defined as the dam displaying an obvious arc in her back

while nursing), b) blanket nursing (dam lies flat on her pups while

nursing), c) passive nursing (dam lies on her side), d) self grooming

of the dam and e) time away from the litter.

On PND3, six litters were deprived from their mother for 24 h,

according to the procedures described above. The control (CS)

litters remained with their mother. Both CS and DS groups

received sham injections at 11.00 am, 5.00 pm and 10.00 pm to

replicate the procedures of the groups studied for neurogenesis as

closely as possible. This resulted for the CS group in two

observation periods that preceded injection times (11.00 am and

10.00 pm) and in one observation period that followed injection

time (5.00 pm). On the following morning, i.e. on PND4, DS litter

were placed back with their mothers at 9.00 am, which was

followed by the 10.00 am observation period.

Perfusion and tissue processingOn PND21 rats were anaesthesized in the morning by an

injection of pentobarbital sodium salt (Nembutal, 1 mg/kg

bodyweight; A.U.V. Cuijk, The Netherlands) and perfused

transcardially with saline followed by 4% paraformaldehyde in

phosphate buffer (PB; 0.1 M; pH 7.4). To prevent pressure

artefacts, brains were additionally postfixed overnight within the

skull at 4uC, washed and cryoprotected in 20% sucrose in PB.

Frozen sections (30 mm thick) were cut using a sliding microtome

and collected in PB with azide.

ImmunohistochemistryDifferent stages of neurogenesis were studied as described

previously [17]. Immunohistochemistry for BrdU (monoclonal

murine anti-BrdU, Roche Diagnostics, Netherlands, 1:2000) was

used to assess newborn cell survival, Ki-67 (polyclonal rabbit a-Ki-

67, Novocastra, New Castle, UK, 1:2000) to assess proliferation,

and doublecortin (DCX; polyclonal goat a-DCX, Santa Cruz,

1:800) to assess the number of immature neurons. To analyze

astrocyte numbers and astrogliogenesis in the dentate gyrus,

immunohistochemistry for GFAP (polyclonal goat anti-GFAP,

DAKO 1:10000) was done as well. The primary antibody was

amplified by biotinylated rabbit anti goat (Vector); avidin-biotin

enzyme complex (ABC kit; Elite Vectastain, Brunschwig Chemie,

Amsterdam, 1:1000) and developed with diaminobenzidine (DAB;

20 mg/100 ml tris buffer; TB, 0.01% H2O2). For BrdU/GFAP

double-labeling, protocols were combined. After first developing

BrdU immunoreactive signal with nickel ammonium sulphate

(0.02%) added to the diaminobenzidine (DAB; 20 mg/100 ml TB,

0.01% H2O2), sections were subsequently incubated in GFAP

primary antibody (polyclonal goat anti-GFAP, DAKO 1:10000)

overnight. The next day, GFAP antibody was amplified according

to standard protocols using secondary biotinylated rabbit anti goat

(Vector) and avidin-biotin complex (ABC kit; Elite Vectastain,

Brunschwig Chemie, Amsterdam, 1:1000). For GFAP, chromogen

development was by diaminobenzidine (DAB; 20 mg/100 ml TB,

0.01% H2O2) alone.

StereologyGFAP+, DCX+ and BrdU+ cells were quantified stereologically

by systematic random sampling in every 10th section using the

StereoInvestigator system (Microbrightfield, Germany) in a total of

9 sections per animal. Because of the occurrence of cell clusters

when using Ki-67, all individual positive cells were counted by

means of a modified stereological procedure, manually in every

10th hippocampal section (Zeiss microscope 2006 magnification)

and multiplied by 10 to estimate the total number of Ki-67+ cells

per hippocampus.

To determine the percentage of BrdU-labeled astrocytes as part

of the whole BrdU cell population, random sampling was done in

six hippocampal sections per animal. In each section, two

randomly selected parts of the dentate gyrus were sampled

throughout the granular cell layer, the subgranular zone and the

hilus, and the ratio of BrdU/GFAP double-positive over the

number of BrdU-single positive cells was expressed as a

percentage. Dentate gyrus granular cell layer volume measure-

ments were performed by using the Cavalieri estimator in every

10th section in a total of 9 sections per animal.

StatisticsData are presented as mean+SEM. All statistics were performed

by SPSS16 for Mac. Immunohistochemical data were initially

compared using a two-factor ANOVA to study the main effect of

sex and treatment. In case of a significant interaction between sex

and treatment, a one-way ANOVA over the four treatment groups

was performed, separately for male and female data. If the one-

way ANOVA revealed significance, a pair wise comparison was

performed with a post-hoc LSD test. Maternal care data was

analyzed with a one-way ANOVA per day in case of individual

licking and grooming; for arch-back nursing an ANOVA for

repeated measures was used.

Results

Acute effects of 24 h MD on PND4Body weights and corticosterone levels. Animals were not

labeled to avoid any disturbance in the control groups. Therefore,

individual bodyweight data are unavailable. However, from P3 to

P4, both MD groups experienced on average a 4% weight-loss, as

opposed to an average weight gain of 7% in the CONS group.

The undisturbed groups were not weighed on PND4.

Maternally deprived pups in both MD-groups had significantly

higher morning corticosterone levels when compared to controls

(F(3,55) = 10.30; p,0.0001; post-hoc: CONU = CONS,MDS =



MDG at least p,0.05; see figure 1). As there were no differences

between the sexes in corticosterone level at this age, data were

pooled for males and females.

Overall, MD had appreciable immediate effects on body weight

and corticosterone level. Importantly though, the sham-injected

control pups responded in a comparable manner as the pups that

were left undisturbed. Moreover, repeated administration of

glucose during the 24 MD period did not ‘rescue’ the phenotype.

Dentate gyrus cell proliferation. In a subset of animals,

dentate gyrus cell proliferation was measured immediately after

24 h MD, i.e. on PND4. For this experiment CONS males and

females and MDS males and females were used. Despite the effect

of MD on body weight and circulating corticosteroid levels, no

acute effects of treatment were observed on dentate gyrus cell

proliferation, as measured by total Ki-67+ cell population

(F(3,21) = 1.13; p = 0.30), see figure 2. A two-factor ANOVA

revealed no effect of sex on proliferation (F(2,21) = 0.04; p = 0.95)

or an interaction between sex and treatment (F(3,21) = 2.45;

p = 0,14).

Effects of early MD on structural parameters at PND21Body weights. There was a significant difference in body

weight between the 4 experimental groups at PND21 (F(3.91) = 50;

p,0.0001). In the control undisturbed group (CONU), body

weights were the lowest (mean+SE: 45+0.6). CONS animals were

heavier (57+0.9), compared to all the other groups, and MDS and

MDG animals had lower body weights than the CONS group but

did not differ from each other (52+0.7 and 49+1.2 respectively).

Post-hoc analysis revealed that CONS animals were significantly

heavier than the MDS (p,0.05), as were the MDG compared to

CONU rats. The two MD groups did not differ from each other.

Granular cell layer volume. The volume of the granular

cell layer was not affected by MD with or without glucose

administration, neither in males (F(3.28) = 0.41; p = 0.75) nor in

females (F(3.28) = 0.54; p = 0.65) as shown in figure 3. A two-factor

ANOVA revealed a significant effect of sex on granular cell layer

volume (F(1.51) = 122.4; p,0.0001), female offspring showed a

lower average granular cell layer volume than males. No

significant interaction of treatment and sex was found

(F(3.51) = 0.95; p = 0.42).

Proliferation, newborn cell survival and

neurogenesis. A two-factor ANOVA revealed no main effect

of treatment on proliferation (F(3,56) = 0.72; p = 0.55) but a

significant effect of sex (F(1,56) = 8.56; p = 0.005) and a significant

interaction (sex6treatment) was found, indicating that the effects

of MD are different in males than in females (F(3,56) = 2.88;

p = 0.045). Females have overall lower numbers of Ki-67 positive

cells. A one-way ANOVA per sex revealed that MD affected

dentate cell proliferation in males (F(3,28) = 3.2; p = 0.043) but not

in females (F(3,28) = 0.41; p = 0.75, see figure 4). Only the MDG

males had significantly decreased numbers of Ki-67 positive cells

(post-hoc comparison, p,0.05).

On PND21, survival of newborn cells in the dentate gyrus was

not affected by MD in males (F(3,28) = 0.40; p = 0.75), or females

(F(3,28) = 1.1; p = 0.37), see figure 5. There was a significant effect

of sex on cell survival. Overall, female offspring had a lower

number of 17 day old BrdU+ cells in the dentate gyrus than males

(F(1,51) = 29.8; p,0.0001). No significant interaction of treatment

and sex was found (F(3.51) = 0.93; p = 0.43). During counting the

location of the cells within the dentate gyrus, i.e. subgranular zone,

granular cell layer or hilus, was taken into account, but this did not

yield any subregion-specific effects. When corrected for granular

Figure 1. Basal corticosterone levels on PND4. A significant increase is found in corticosterone levels after maternal deprivation (MDs, shaminjected) compared to controls (both undisturbed CONU and sham-injected CONS). Additional glucose treatment (MDG) failed to normalize this(F(3,55) = 10.30; p,0.0001; post-hoc: CONU = CONS,MDS = MDG at least p,0.05).doi:10.1371/journal.pone.0003675.g001



cell layer volume, all of the above effects persisted (data not

shown).

The most prominent effects of MD were observed at PND21 in

neurogenesis, as measured by the young neuronal marker

doublecortin (DCX; figure 6), A two-factor ANOVA revealed no

main effect of treatment (F(7,48) = 1.50; p = 0.23), but an effect of

sex, indicating that females have in general less DCX-positive cells

(F(1,48) = 65.80; p,0.0001). Interestingly, a significant interaction

between sex and treatment was found, indicating that the effects of

MD on neurogenesis were different for males than females,

(treatment6sex; F(3,48) = 8.04; p,0.0001). A one-way ANOVA in

males, revealed a significant increase in the total number of DCX+cells due to maternal deprivation (F(3,28) = 4.3; p = 0.018; post-hoc

LSD: CONU = CONS,MDS = MDG, at least p,0.05, see

Figure 2. Dentate gyrus cell proliferation (Ki-67) on PND4 in CONS and MDS animals. There was no significant effect of sex (F(2,21) = 0.04;p = 0.95) or treatment (F(3,21) = 1.13; p = 0.30) and no interaction between the two (F(3,21) = 2.45; p = 0,14).doi:10.1371/journal.pone.0003675.g002

Figure 3. Granular cell layer (GCL) volume on PND21. There was no effect of any of the treatments on granular cell layer volume. However,significantly lower GCL volumes were found in females (F(1.51) = 122.4; p,0.0001).doi:10.1371/journal.pone.0003675.g003



figure 7A). In females, MD was found to lead to lower total DCX+cell numbers, (F(3,28) = 4.65; p = 0.013; post-hoc LSD: CONU = -

CONS.MDS = MDG, at least p,0.05 see figure 7B). When

corrected for granular cell layer volume, these effects persisted

(data not shown).

Astrocyte numbers and gliogenesis. The total number of

GFAP+ astrocytes was measured in a stereologically sampled series

using the StereoInvestigator throughout the entire dentate gyrus;

in the hilus, granule cell layer as well as molecular layer. No effects

of early MD were found at PND21 on the total number of GFAP+cells in males (F(3,28) = 1.3; p = 0.28), nor in females (F(3,28) = 0.05;

p = 0.99, see figure 8). A two-factor ANOVA revealed a significant

effect of sex (F(3,56) = 25.01; p,0.0001) but no interaction of

treatment and sex (F(3,56) = 0.23; p = 0.88). Females show a

significant higher number of GFAP-positive cells compared to

males.

Quantification of gliogenesis was done by analyzing BrdU and

BrdU/GFAP double stained cells (for an example, see figure 9).

Random sampling of a minimum of 200 BrdU positive cells per

animal in six hippocampal sections was done, and the percentage

double labeling with GFAP is shown in figure 10. In this

experiment, only CONS and MDS males (n = 6) and females

(n = 6) were analyzed. There was no effect of MD on the

percentage of double labeled cells (F(3,24) = 0.40; p = 0.54) no effect

Figure 4. Dentate gyrus cell proliferation (Ki-67) on PND21. A. In males, MD affected dentate cell proliferation rate at PND21 (F(3,28) = 3,2,p = 0.043). Post-hoc analysis revealed a decrease in MD-glucose injected animals (p,0.05). B. In females, there was no effect of treatment(F(3,28) = 0.41 p = 0.75), although a significant lower number of Ki-67 positive cells was found in females when compared to males (F(1,56) = 8.56;p = 0.005).doi:10.1371/journal.pone.0003675.g004

Figure 5. Number of newborn surviving cells (BrdU+) in the dentate gyrus on PND21. There was a significant effect of sex, but not oftreatment on BrdU positive cell numbers. A. Maternal deprivation did not alter BrdU+ cell numbers in males (F(3,28) = 0.40, p = 0.75), or (B) females(F(3,28) = 1.1, p = 0.37). Irrespective of MD, females had an overall lower number of BrdU-positive cells (F(1,51) = 29.8; p,0.0001).doi:10.1371/journal.pone.0003675.g005



of sex (F(1,24) = 0.56; p = 0.46) and no interaction between sex and

treatment was found (F(3,24) = 0.6; p = 0.45).

Maternal careAs neurogenesis could possibly be affected by maternal care

[74], we therefore assessed whether possible changes after MD in

the amount of maternal care directed to either male or female

pups could explain the outcome of this study. As expected [84],

maternal care was affected 24 h after MD (figure 11) and dams

provided more care towards the pups when returned to the nest, as

illustrated by the increased percentage of time spent on licking and

grooming (LG). Both MD-males and MD-females received

significantly more care (LG) on PND4 compared to the respective

control animals. In total, males received more individual care than

females (males: 5.6%60,4; females 3.2%60.3; T-test, p = 0.008).

This effect was also seen on PND4, because the percentage of LG

towards male and female pups increased similarly, which brought

the MD females to the level of non-deprived males (PND4 LG-

scores: males F(3,21) = 7.31; p = 0.002; post-hoc: MD males.CON

males = MD females.CON females, p,0.05, see figure 11).

When comparing the amount of active nursing by the dam

towards the whole litter, as measured by arch-back nursing (ABN),

MD resulted in a significant increase of ABN after PND3

(repeated measures ANOVA; F = 7.89, p = 0.02, see figure 12).

Discussion

We show that 24 h of maternal deprivation at PND3 alters

hippocampal structural plasticity in a sex-dependent manner.

Although newborn cell survival and proliferation rate were not

altered by MD, neurogenesis in the dentate gyrus was increased in

male, but decreased in female offspring. Since no such differential

changes were found in granular cell layer volume, astrocyte

number or astrogliogenesis, this indicates that instead of altering

granule cell numbers, MD-induced stress elicits specific changes in

subsets of the differentiating cell population and e.g. impacts only

the immature, DCX positive cells.

Neurogenesis specific effects of MDGiven the extensive neurogenesis during gestation and the early

postnatal period [86,87], it is not surprising that early life stress

affects structural brain development. Indeed, long-lasting reduc-

tions in neurogenesis and hippocampal functions after both pre- as

well as postnatal stressors have been reported in most

[41,45,46,73,88], but not all [89] studies. Stress-induced increases

in maternal and offspring plasma corticosterone levels during a

sensitive time window of brain development appear to be critical

in mediating these long-lasting effects [51,90,91]. The timepoint of

early life stress as was used in the present model coincides with the

formation of the inner shell of the dentate granular cell layer i.e.

the future site of adult neurogenesis [76,77,79]. The present results

show that for the rat dentate gyrus, the sensitive time window

during which brain development can be affected, appears to

extend at least into the early postnatal period.

Despite the significantly different corticosterone levels at the end

of the 24 h MD period, no changes were found in the numbers of

Ki-67+ proliferating cells at PND4. In adult rats, stress frequently

reduces proliferation [19,92–94] but clear exceptions have also

been reported [20,95–97] that may depend on the experimental

design and type of stress [33]. If any reduction was induced during

the 24 hours of MD, recovery was fast and compensated for, or

normalized, rapidly. However, in MD males that were additionally

given a glucose injection, an unexpected reduction in proliferation

was found. It can be speculated that a protective response of the

pup during MD is disturbed by metabolic activation and a

subsequent insulin response. Although there is considerable

evidence for a strong and complex interplay between the

metabolic system and the (development of) the HPA-axis

[80,98], the exact mechanisms can not be derived from the

current data-set and await future studies. Given the very large

numbers of newly generated cells at these young ages, also

stochastic differences within the population could have occurred

that may have been missed with Ki-67 immunohistochemistry as

this antibody labels all cells engaged in cell cycle.

Even though the sex differences in hippocampal neurogenesis

could be attributable to differential precursor kinetics, this has not

led to reductions in granular cell layer volume between MD and

control females. Also, no changes were found in the number of

BrdU+ surviving cells. Various stress protocols have been shown to

reduce single or multiple phases of the neurogenic process, but

only in a few instances, and when stress or corticosterone exposure

was chronic, did this actually lead to reductions in DG granule cell

number [99]. In addition to the MD-induced differences between

male and females, also differences between non-deprived control

Figure 6. Doublecortin (DCX) -positive neurons. A. Photo of thedentate gyrus of a 21 day old CONS male showing extensiveimmunostaining of DCX in the subgranular zone (sgz) and the firstthird of the granular cell layer (GCL) with dendrites extending throughthe granular cell layer (GCL) into the molecular layer (ML). B. Highpower photomicrograph showing details of the DCX+ cell bodieslocated in the SGZ and GCL, with extending dendrites in the GCL.doi:10.1371/journal.pone.0003675.g006



male and females were present. So far, not much is known about

sex-differences per se in DG neurogenesis measured around the

age of weaning and thus before the onset of the oestrus cycle.

However, our results show that control males generally have a

higher proliferation rate, an increased survival of newborn cells,

more young neurons (DCX) but less astrocytes then control

females. In addition, a larger granular cell layer volume was found

in males.

Furthermore, an extensive amount of literature has shown

differential regulation by gonadal hormones [92,100–107],

resulting in established sex differences in proliferation and survival.

Also, a differential response to chronic stress in males and females

was found [92,106]. Typically, under basal conditions, both

estradiol and testosterone enhance neurogenesis [105,108,109]. In

adult female rats, an increased proliferation rate occurs during

pro-oestrus [105] due to higher estradiol levels. Males were found

to have more doublecortin positive cells [47]. During develop-

ment, sex steroids are also able to modulate astrocytes [110] and a

higher number of hippocampal astrocytes in females was found in

some [111,112], but not all [113] studies. In summary, our data

Figure 7. Doublecortin (DCX) -positive neuron numbers on PND21. A significant treatment6sex interaction revealed a differential effect ofMD on males versus females (F(3,48) = 8.04; p,0.0001). A. An increase in DCX+ cell number was found in deprived males (F(3,28) = 4.3, p = 0.018; post-hoc: CONU = CONS,MDS = MDG, at least p,0.05) and a decrease in deprived females (F(3,28) = 4.65, p = 0.013; post-hoc: CONU = CONS.MDS = MDG,at least p,0.05) when compared to controls. A significant effect of sex indicates a general lower amount of DCX+ cells in females (F(1,48) = 65.80;p,0.0001).doi:10.1371/journal.pone.0003675.g007

Figure 8. Astrocyte numbers. GFAP-positive astrocyte numbers determined stereologically in the entire hippocampal dentate gyrus on PND21. A.Maternal deprivation did not affect GFAP+ cell number in males (F(3,28) = 1.3, p = 0.28), B. nor females (F(3,28) = 0.05, p = 0.99). However, a significanteffect of sex on GFAP+ cell number was found (F(3,56) = 25.01; p,0.0001).doi:10.1371/journal.pone.0003675.g008



indicate that sex-differences in dentate gyrus structure and

neurogenesis are already present before sexual maturity is reached.

In contrast to proliferation and newborn cell survival that

remained unchanged, neuronal differentiation was differentially

altered between males and females, as based on DCX immuno-

histochemistry. One possibility to explain these differential

changes in DCX could be a change in cell-fate determination.

In deprived males, more newborn cells may have differentiated

into a neuronal phenotype, as compared to deprived females, e.g.

at the cost of gliogenesis. The latter option is unlikely, since no

compensatory changes in either total astrocyte cell numbers or

astrogliogenesis were found. An additional option is a shift towards

oligodendrogliosis, but although these numbers were not deter-

mined, the proportion of newborn cells that differentiate into

oligodendrocytes is generally very small [114].

Another possibility is that the differential changes in DCX+ cell

numbers are related to the time window of DCX expression. DCX

is a microtubule associated protein (MAP) expressed by migratory

and immature neurons from PND 4 till 14 and accurately reflects

neurogenesis in the adult hippocampus [115,116]. In theory, MD

may shorten the time window during which DCX is expressed,

which would imply that when shut-off prematurely, it could lead to

an early arrest in DG granule cell development and lower DCX+cell numbers in females, which would result in less complex

granular cells. The finding that the number of BrdU-+ cells did not

differ between male and female groups, indicates that the DCX

changes are at least not accompanied by a changed survival rate or

different developmental kinetics of the newborn cell population.

Instead of altering neuron number, MD may specifically impact

the population of immature DCX-IR cells. Preliminary data

indicate that in adult females, the complexity of the individual

granule cell is indeed diminished (unpublished results). Whether

this has consequences for functional properties of the hippocampal

circuit awaits future research. Following the same line of

reasoning, males could benefit from maternal deprivation on

PND3, as this resulted in an increase in neurogenesis. It is known

that different early life experiences can cause differential responses

in stress reactivity in adulthood. Whether consequences are

Figure 9. GFAP/BrdU double labeling. Immunohistochemicaldouble labeling for GFAP and BrdU shows single GFAP+ astrocytes inthe hilus with their processes occasionally extending into the sgz.GFAP+ cells reveal brown DAB-staining in their processes andcytoplasm whereas the nucleus is devoid of staining. In the GCL, BrdU+single cells are stained black by DAB-nickel as indicated (BrdU+) in thegranular cell layer (GCL). Shown on the left is a BrdU/GFAP doublelabeled cell (arrow).doi:10.1371/journal.pone.0003675.g009

Figure 10. Percentage of GFAP/BrdU double-labeled cells in the dentate gyrus. There was no effect of MD on the percentage of doublelabeled cells in both males and females (F(3,24) = 0.40; p = 0.54). No effect of sex was found (F(1,24) = 0.56; p = 0.46).doi:10.1371/journal.pone.0003675.g010



detrimental or more beneficial, depends to a large extend on the

context the adult subject encounters [117]. It not only remains to

be shown in future studies whether the present differences in

neurogenesis are transient, but also whether MD-induced changes

in neurogenesis correlate with later structural and functional

parameters in a more positive way in adult male offspring.

Sex differences in the effects of MD on neurogenesisRecent studies have revealed sex-dependent alterations in DCX

expression [47] after prenatal restraint stress exposure. Similar to

our study, higher DCX expression was found in males. However,

in this study, no effects of prenatal stress on total DCX cell number

were found in females. The interaction of stress hormones with

gonadal steroids during gestation may explain these results [48]. In

the present study, however, both male and female offspring

experienced an increase in corticosterone not until PND4 and

gonadal interactions during pregnancy are therefore unlikely.

However, also in adulthood, gonadal steroids affect hippocampal

plasticity to a great extent. For example, stress experienced in

adulthood decreases proliferation and neurogenesis in males, but

not females, and estrogen is thought to protect against stress-

induced reductions in dentate gyrus proliferation [92,105,106].

Estrogens can exert non-genomic effects directly and indirectly on

newly generated cells in neonatal and adult rat dentate gyrus while

specific estrogen receptors are found on DCX+ cells, which is

interesting in this respect [118]. Testosterone on the other hand,

promotes neurogenesis and survival but not differentiation of the

newborn cells [101,119]. Although gonadal steroids may hence

contribute to the development of sex differences in neurogenesis

per se, it awaits to be determined whether they are also implicated

in the differential effects of MD in PND21 animals.

Body weight and basal corticosterone levels were affected by

MD but, interestingly, these measures were not altered by glucose

supplementation, nor by the multiple injections (of glucose or

vehicle) associated with the treatment. Importantly, though,

corticosteroid levels were not different between males and females.

This does not exclude, however, that other factors determining

corticosteroid functionality change in a sex-dependent manner.

Thus, sex-specific effects were reported for MR and GR

expression after 24 h MD on PND3 [120]. In males, the same

MD design reduced GR and MR binding whereas in females, GR

was upregulated and MR was unaffected [120]. Selective

upregulation of hippocampal GR after MD in females could

sensitize young neurons to the actions of circulating corticosterone,

and may e.g. result in a premature shut-down of DCX expression.

Also, MD may differentially affect early HPA axis (re)activity. The

DG may be particularly vulnerable then as it undergoes rapid

postnatal development during the first two weeks of life [76,77,87].

Figure 11. Licking and grooming on PND1–7 in control and MD litters. A significant increase in LG was found on PND4 both in males andfemales compared to their same-sex controls. (PND4 LG-scores: males F(3,21) = 7.31; p = 0.002; post-hoc: MD males.CON males = MD females.CONfemales, p,0.05).doi:10.1371/journal.pone.0003675.g011

Figure 12. Arch back nursing. Arch back nursing towards the entirelitter in control and MD litters on PND1–7. There is a significant increaseafter PND3 (F = 7.89, p = 0.02).doi:10.1371/journal.pone.0003675.g012



Whether or not the newborn cells are actually sensitive to

glucocorticoid action depends on the GR and MR expression of

the individual newborn cell. As documented elsewhere [121],

newly formed, proliferating cells only express low and variable

levels of GRs, which may explain the lack of effect of CORT on

proliferation. Considerable expression levels of both GR and MR

on the individual newborn cells are only reached in young neurons

[121], consistent with the present DCX findings.

Finally, since maternal behavior shapes hippocampal properties

later in life [83,117], we investigated whether sex-specific

differences in maternal care had been instrumental in the changed

pattern of neurogenesis. Maternal behavior regulates maturation

of offspring HPA activity [57,83] while stress affects maternal-

offspring interactions [84]. Consistent with earlier findings [82],

we also found a higher level of licking and grooming towards male

than female pups. Individual maternal care after MD was

increased on PND4, but this increase was comparable for male

and female pups and returned to control levels from PND5

onwards. Maternal care through arch back nursing was indeed

enhanced by MD but this was towards the entire litter. Therefore,

it is unlikely that the male-female differences in neurogenesis can

be explained by sex-specific changes in LG or other maternal

behavioral components induced by MD.

Taken together, the present data support the concept that the

setpoint for neurogenesis may be determined during perinatal life

and illustrate the critical influence of early environment on

establishing sex differences in neural plasticity. They expand our

understanding of the mechanisms underlying sex differences and

highlight the critical role early stress can play in determining the

structural make up of the hippocampus in adulthood. Given their

specific properties [122], newborn cells can make disproportion-

ately large contributions to overall DG composition, average age

and output of the DG cells, which will have considerable

consequences for hippocampal function [21,123]. It is tempting

to speculate that a reduced level of neurogenesis, secondary to e.g.

early stress exposure, may contribute to maladaptation of

hippocampal function and possibly to increased vulnerability of

women to stress-related disorders.

Acknowlegdments

We thank Maaike van der Mark (LACDR, LeidenUniversity) for

performing the RIA, Seymour Levine{, Melly Oitzl and Danielle

Champagne (LACDR) for valuable methodological advise, Felisa van

Hasselt and Heleen Soeters (UVA) for help with the maternal care

experiment, Jose Wouda and Edwin Jousma (UVA) for histotechnical

assistance.


Conceived and designed the experiments: CAO HK MJ PJL. Performed

the experiments: CAO CENG RC ECV. Analyzed the data: CAO.

Contributed reagents/materials/analysis tools: HK MJ PJL. Wrote the

paper: CAO HK MJ PJL.

References

1. Belanoff JK, Flores BH, Kalezhan M, Sund B, Schatzberg AF (2001) Rapid

reversal of psychotic depression using mifepristone. J Clin Psychopharmacol 21:

516–521.

2. Belanoff JK, Rothschild AJ, Cassidy F, DeBattista C, Baulieu EE, et al. (2002)

An open label trial of C-1073 (mifepristone) for psychotic major depression.



dysregulation in melancholic and atypical depression: high vs low CRH/NE

states. Mol Psychiatry 7: 254–275.

4. Heim C, Nemeroff CB (2001) The role of childhood trauma in the

neurobiology of mood and anxiety disorders: preclinical and clinical studies.


5. Heim C, Newport DJ, Heit S, Graham YP, Wilcox M, et al. (2000) Pituitary-

adrenal and autonomic responses to stress in women after sexual and physical

abuse in childhood. Jama 284: 592–597.


Neuropsychopharmacology 23: 477–501.

7. Reus VI, Wolkowitz OM (2001) Antiglucocorticoid drugs in the treatment of

depression. Expert Opin Investig Drugs 10: 1789–1796.

8. Hollup SA, Kjelstrup KG, Hoff J, Moser MB, Moser EI (2001) Impaired

recognition of the goal location during spatial navigation in rats with

hippocampal lesions. J Neurosci 21: 4505–4513.

9. Morris R (1984) Developments of a water-maze procedure for studying spatial

learning in the rat. J Neurosci Methods 11: 47–60.

10. Herman JP, Mueller NK (2006) Role of the ventral subiculum in stress

integration. Behav Brain Res 174: 215–224.

11. de Kloet ER, Joels M, Holsboer F (2005) Stress and the brain: from adaptation

to disease. Nat Rev Neurosci 6: 463–475.

12. Gilbertson MW, Shenton ME, Ciszewski A, Kasai K, Lasko NB, et al. (2002)

Smaller hippocampal volume predicts pathologic vulnerability to psychological

trauma. Nat Neurosci 5: 1242–1247.

13. Czeh B, Simon M, Schmelting B, Hiemke C, Fuchs E (2006) Astroglial

plasticity in the hippocampus is affected by chronic psychosocial stress and

concomitant fluoxetine treatment. Neuropsychopharmacology 31: 1616–

1626.

14. Kole MH, Czeh B, Fuchs E (2004) Homeostatic maintenance in excitability of

tree shrew hippocampal CA3 pyramidal neurons after chronic stress.

Hippocampus 14: 742–751.

15. Magarinos AM, McEwen BS, Flugge G, Fuchs E (1996) Chronic psychosocial

stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons

in subordinate tree shrews. J Neurosci 16: 3534–3540.

16. Magarinos AM, Verdugo JM, McEwen BS (1997) Chronic stress alters synaptic

terminal structure in hippocampus. Proc Natl Acad Sci U S A 94:

14002–14008.

17. Oomen CA, Mayer JL, Joels M, de Kloet ER, Lucassen PJ (2007) Brief

treatment with the GR antagonist mifepristone normalizes the reduction in

neurogenesis after chronic stress. submitted.

18. Gould E, McEwen BS, Tanapat P, Galea LA, Fuchs E (1997) Neurogenesis in

the dentate gyrus of the adult tree shrew is regulated by psychosocial stress and

NMDA receptor activation. J Neurosci 17: 2492–2498.

19. Heine VM, Maslam S, Zareno J, Joels M, Lucassen PJ (2004) Suppressed

proliferation and apoptotic changes in the rat dentate gyrus after acute and

chronic stress are reversible. Eur J Neurosci 19: 131–144.

20. Pham K, Nacher J, Hof PR, McEwen BS (2003) Repeated restraint stress

suppresses neurogenesis and induces biphasic PSA-NCAM expression in the

adult rat dentate gyrus. Eur J Neurosci 17: 879–886.

21. Abrous DN, Koehl M, Le Moal M (2005) Adult neurogenesis: from precursors

to network and physiology. Physiol Rev 85: 523–569.

22. Dobrossy MD, Drapeau E, Aurousseau C, Le Moal M, Piazza PV, et al. (2003)

Differential effects of learning on neurogenesis: learning increases or decreases

the number of newly born cells depending on their birth date. Mol Psychiatry

8: 974–982.

23. Dupret D, Fabre A, Dobrossy MD, Panatier A, Rodriguez JJ, et al. (2007)

Spatial learning depends on both the addition and removal of new

hippocampal neurons. PLoS Biol 5: e214.

24. Kee N, Teixeira CM, Wang AH, Frankland PW (2007) Preferential

incorporation of adult-generated granule cells into spatial memory networks

in the dentate gyrus. Nat Neurosci 10: 355–362.

25. Leuner B, Gould E, Shors TJ (2006) Is there a link between adult neurogenesis

and learning? Hippocampus 16: 216–224.

26. Banasr M, Soumier A, Hery M, Mocaer E, Daszuta A (2006) Agomelatine, a

new antidepressant, induces regional changes in hippocampal neurogenesis.


27. Dranovsky A, Hen R (2006) Hippocampal neurogenesis: regulation by stress

and antidepressants. Biol Psychiatry 59: 1136–1143.

28. Mayer JL, Klumpers L, Maslam S, de Kloet ER, Joels M, et al. (2006) Brief

treatment with the glucocorticoid receptor antagonist mifepristone normalises

the corticosterone-induced reduction of adult hippocampal neurogenesis.


29. Santarelli L, Saxe M, Gross C, Surget A, Battaglia F, et al. (2003) Requirement

of hippocampal neurogenesis for the behavioral effects of antidepressants.

Science 301: 805–809.

30. Thomas RM, Peterson DA (2008) Even neural stem cells get the blues:

evidence for a molecular link between modulation of adult neurogenesis and

depression. Gene Expr 14: 183–193.

31. Wong EY, Herbert J (2006) Raised circulating corticosterone inhibits neuronal

differentiation of progenitor cells in the adult hippocampus. Neuroscience 137:

83–92.



cbaker

Cross-Out

cbaker

Inserted Text

dg

32. Mirescu C, Gould E (2006) Stress and adult neurogenesis. Hippocampus 16:233–238.

33. Lucassen PJ, Oomen CA, Van Dam A, Czeh B (2008) Regulation of

hippocampal neurogenesis by systemic factors including stress, glucocorticoids,sleep, and inflammation.; Kempermann G, Gage FH, Song H, eds. Cold

Spring Harbor. New York: Cold Spring Harbor Laboratory Press. pp 363–395.

34. Bland RC (1997) Epidemiology of affective disorders: a review. Can J Psychiatry

42: 367–377.

35. Kessler RC, McGonagle KA, Swartz M, Blazer DG, Nelson CB (1993) Sex anddepression in the National Comorbidity Survey. I: Lifetime prevalence,

chronicity and recurrence. J Affect Disord 29: 85–96.

36. Ustun TB (2000) Cross-national epidemiology of depression and gender. J Gend

Specif Med 3: 54–58.

37. Weiss EL, Longhurst JG, Mazure CM (1999) Childhood sexual abuse as a riskfactor for depression in women: psychosocial and neurobiological correlates.

Am J Psychiatry 156: 816–828.

38. Young EA (1998) Sex differences and the HPA axis: implications for psychiatricdisease. J Gend Specif Med 1: 21–27.

39. Vythilingam M, Heim C, Newport J, Miller AH, Anderson E, et al. (2002)Childhood trauma associated with smaller hippocampal volume in women with

major depression. Am J Psychiatry 159: 2072–2080.

40. Brummelte S, Pawluski JL, Galea LA (2006) High post-partum levels ofcorticosterone given to dams influence postnatal hippocampal cell proliferation

and behavior of offspring: A model of post-partum stress and possibledepression. Horm Behav 50: 370–382.

41. Coe CL, Kramer M, Czeh B, Gould E, Reeves AJ, et al. (2003) Prenatal stress

diminishes neurogenesis in the dentate gyrus of juvenile rhesus monkeys. BiolPsychiatry 54: 1025–1034.

42. Darnaudery M, Maccari S (2008) Epigenetic programming of the stressresponse in male and female rats by prenatal restraint stress. Brain Res Rev 57:

571–585.

43. Koehl M, Lemaire V, Vallee M, Abrous N, Piazza PV, et al. (2001) Long termneurodevelopmental and behavioral effects of perinatal life events in rats.

Neurotox Res 3: 65–83.

44. Kofman O (2002) The role of prenatal stress in the etiology of developmentalbehavioural disorders. Neurosci Biobehav Rev 26: 457–470.

45. Lemaire V, Koehl M, Le Moal M, Abrous DN (2000) Prenatal stress produceslearning deficits associated with an inhibition of neurogenesis in the

hippocampus. Proc Natl Acad Sci U S A 97: 11032–11037.

46. Lemaire V, Lamarque S, Le Moal M, Piazza PV, Abrous DN (2006) Postnatalstimulation of the pups counteracts prenatal stress-induced deficits in

hippocampal neurogenesis. Biol Psychiatry 59: 786–792.

47. Mandyam CD, Crawford EF, Eisch AJ, Rivier CL, Richardson HN (2008)

Stress experienced in utero reduces sexual dichotomies in neurogenesis,

microenvironment, and cell death in the adult rat hippocampus. Dev Neurobiol68: 575–589.

48. Weinstock M (2007) Gender differences in the effects of prenatal stress on braindevelopment and behaviour. Neurochem Res 32: 1730–1740.

49. Welberg LA, Seckl JR (2001) Prenatal stress, glucocorticoids and the

programming of the brain. J Neuroendocrinol 13: 113–128.

50. Henry C, Kabbaj M, Simon H, Le Moal M, Maccari S (1994) Prenatal stress

increases the hypothalamo-pituitary-adrenal axis response in young and adultrats. J Neuroendocrinol 6: 341–345.

51. Vallee M, Mayo W, Dellu F, Le Moal M, Simon H, et al. (1997) Prenatal stress

induces high anxiety and postnatal handling induces low anxiety in adultoffspring: correlation with stress-induced corticosterone secretion. J Neurosci

17: 2626–2636.

52. Schmitz C, Rhodes ME, Bludau M, Kaplan S, Ong P, et al. (2002) Depression:reduced number of granule cells in the hippocampus of female, but not male,

rats due to prenatal restraint stress. Mol Psychiatry 7: 810–813.

53. Lehmann J, Russig H, Feldon J, Pryce CR (2002) Effect of a single maternal

separation at different pup ages on the corticosterone stress response in adult

and aged rats. Pharmacol Biochem Behav 73: 141–145.

54. Levine S, Chevalier JA, Korchin SJ (1956) The effects of early shock and

handling on later avoidance learning. J Pers 24: 475–493.

55. Pryce CR, Feldon J (2003) Long-term neurobehavioural impact of the postnatal

environment in rats: manipulations, effects and mediating mechanisms.

Neurosci Biobehav Rev 27: 57–71.

56. Plotsky PM, Meaney MJ (1993) Early, postnatal experience alters hypothalamic

corticotropin-releasing factor (CRF) mRNA, median eminence CRF contentand stress-induced release in adult rats. Brain Res Mol Brain Res 18: 195–200.

57. Levine S (1994) The ontogeny of the hypothalamic-pituitary-adrenal axis. The

influence of maternal factors. Ann N Y Acad Sci 746: 275–288; discussion 289–293.

58. Walker CD, Perrin M, Vale W, Rivier C (1986) Ontogeny of the stress responsein the rat: role of the pituitary and the hypothalamus. Endocrinology 118:

1445–1451.

59. Workel JO, Oitzl MS, Fluttert M, Lesscher H, Karssen A, et al. (2001)Differential and age-dependent effects of maternal deprivation on the

hypothalamic-pituitary-adrenal axis of brown norway rats from youth tosenescence. J Neuroendocrinol 13: 569–580.

60. Penke Z, Felszeghy K, Fernette B, Sage D, Nyakas C, et al. (2001) Postnatal

maternal deprivation produces long-lasting modifications of the stress response,feeding and stress-related behaviour in the rat. Eur J Neurosci 14: 747–755.

61. Workel JO, Oitzl MS, Ledeboer A, De Kloet ER (1997) The Brown Norwayrat displays enhanced stress-induced ACTH reactivity at day 18 after 24-h

maternal deprivation at day 3. Brain Res Dev Brain Res 103: 199–203.

62. Ladd CO, Owens MJ, Nemeroff CB (1996) Persistent changes in corticotropin-

releasing factor neuronal systems induced by maternal deprivation. Endocri-nology 137: 1212–1218.

63. Liu D, Caldji C, Sharma S, Plotsky PM, Meaney MJ (2000) Influence of

neonatal rearing conditions on stress-induced adrenocorticotropin responsesand norepinepherine release in the hypothalamic paraventricular nucleus.


64. Macri S, Mason GJ, Wurbel H (2004) Dissociation in the effects of neonatalmaternal separations on maternal care and the offspring’s HPA and fear

responses in rats. Eur J Neurosci 20: 1017–1024.

65. Plotsky PM, Thrivikraman KV, Nemeroff CB, Caldji C, Sharma S, et al. (2005)Long-term consequences of neonatal rearing on central corticotropin-releasing

factor systems in adult male rat offspring. Neuropsychopharmacology 30:

2192–2204.

66. Desbonnet L, Garrett L, Daly E, McDermott KW, Dinan TG (2008) Sexually

dimorphic effects of maternal separation stress on corticotrophin-releasing

factor and vasopressin systems in the adult rat brain. Int J Dev Neurosci 26:259–268.

67. Kalinichev M, Easterling KW, Plotsky PM, Holtzman SG (2002) Long-lasting

changes in stress-induced corticosterone response and anxiety-like behaviors asa consequence of neonatal maternal separation in Long-Evans rats. Pharmacol

Biochem Behav 73: 131–140.

68. Slotten HA, Kalinichev M, Hagan JJ, Marsden CA, Fone KC (2006) Long-lasting changes in behavioural and neuroendocrine indices in the rat following

neonatal maternal separation: gender-dependent effects. Brain Res 1097:123–132.

69. Wigger A, Neumann ID (1999) Periodic maternal deprivation induces gender-

dependent alterations in behavioral and neuroendocrine responses to emotional

stress in adult rats. Physiol Behav 66: 293–302.

70. Brunson KL, Kramar E, Lin B, Chen Y, Colgin LL, et al. (2005) Mechanisms

of late-onset cognitive decline after early-life stress. J Neurosci 25: 9328–9338.

71. Fabricius K, Wortwein G, Pakkenberg B (2008) The impact of maternalseparation on adult mouse behaviour and on the total neuron number in the

mouse hippocampus. Brain Struct Funct 212: 403–416.

72. Huot RL, Plotsky PM, Lenox RH, McNamara RK (2002) Neonatal maternal

separation reduces hippocampal mossy fiber density in adult Long Evans rats.Brain Res 950: 52–63.

73. Mirescu C, Peters JD, Gould E (2004) Early life experience alters response of

adult neurogenesis to stress. Nat Neurosci 7: 841–846.

74. Bredy TW, Grant RJ, Champagne DL, Meaney MJ (2003) Maternal careinfluences neuronal survival in the hippocampus of the rat. Eur J Neurosci 18:

2903–2909.

75. Greisen MH, Altar CA, Bolwig TG, Whitehead R, Wortwein G (2005)Increased adult hippocampal brain-derived neurotrophic factor and normal

levels of neurogenesis in maternal separation rats. J Neurosci Res 79: 772–778.

76. Altman J, Bayer SA (1990) Migration and distribution of two populations ofhippocampal granule cell precursors during the perinatal and postnatal periods.

J Comp Neurol 301: 365–381.

77. Altman J, Bayer SA (1990) Mosaic organization of the hippocampalneuroepithelium and the multiple germinal sources of dentate granule cells.

J Comp Neurol 301: 325–342.

78. Fenoglio KA, Brunson KL, Baram TZ (2006) Hippocampal neuroplasticityinduced by early-life stress: functional and molecular aspects. Front

Neuroendocrinol 27: 180–192.

79. Kempermann G (2006) Adult neurogenesis, Stem Cells and NeuronalDevelopment in the Adult Brain. Oxford: University Press.

80. Schmidt MV, Levine S, Alam S, Harbich D, Sterlemann V, et al. (2006)

Metabolic signals modulate hypothalamic-pituitary-adrenal axis activationduring maternal separation of the neonatal mouse. J Neuroendocrinol 18:

865–874.

81. Hofer MA, Brunelli SA, Shair HN (1994) Potentiation of isolation-induced

vocalization by brief exposure of rat pups to maternal cues. Dev Psychobiol 27:503–517.

82. Moore CL, Morelli GA (1979) Mother rats interact differently with male and

female offspring. J Comp Physiol Psychol 93: 677–684.

83. Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D, et al. (1997) Maternalcare, hippocampal glucocorticoid receptors, and hypothalamic-pituitary-

adrenal responses to stress. Science 277: 1659–1662.

84. Pryce CR, Bettschen D, Feldon J (2001) Comparison of the effects of earlyhandling and early deprivation on maternal care in the rat. Dev Psychobiol 38:

239–251.

85. Champagne FA, Francis DD, Mar A, Meaney MJ (2003) Variations inmaternal care in the rat as a mediating influence for the effects of environment

on development. Physiol Behav 79: 359–371.

86. Heine VM, Maslam S, Joels M, Lucassen PJ (2004) Prominent decline ofnewborn cell proliferation, differentiation, and apoptosis in the aging dentate

gyrus, in absence of an age-related hypothalamus-pituitary-adrenal axisactivation. Neurobiol Aging 25: 361–375.

87. Schlessinger AR, Cowan WM, Gottlieb DI (1975) An autoradiographic study

of the time of origin and the pattern of granule cell migration in the dentate

gyrus of the rat. J Comp Neurol 159: 149–175.



88. Lucassen PJ, Bosch OJ, Jousma E, Kromer S, Andrews R, et al. (2008) Prenatal

stress reduces postnatal neurogenesis in rats selectively bred for heigh, but notlow anxiety: possible key role of induction of placental 11B-hydroxysteroid

dehydrogenase type 2. Eur J Neurosci in press.

89. Tauber SC, Bunkowski S, Schlumbohm C, Ruhlmann M, Fuchs E, et al.(2008) No long-term effect two years after intrauterine exposure to

dexamethasone on dentate gyrus volume, neuronal proliferation anddifferentiation in common marmoset monkeys. Brain Pathol 18: 497–503.

90. Barbazanges A, Piazza PV, Le Moal M, Maccari S (1996) Maternal

glucocorticoid secretion mediates long-term effects of prenatal stress.J Neurosci 16: 3943–3949.

91. Szuran TF, Pliska V, Pokorny J, Welzl H (2000) Prenatal stress in rats: effectson plasma corticosterone, hippocampal glucocorticoid receptors, and maze

performance. Physiol Behav 71: 353–362.92. Falconer EM, Galea LAa (2003) Sex differences in cell proliferation, cell death

and defensive behavior following acute predator odor stress in adult rats. Brain

Res 975: 22–36.93. Hill MN, Kambo JS, Sun JC, Gorzalka BB, Galea LA (2006) Endocannabi-

noids modulate stress-induced suppression of hippocampal cell proliferationand activation of defensive behaviours. Eur J Neurosci 24: 1845–1849.

94. Malberg JE, Duman RS (2003) Cell proliferation in adult hippocampus is

decreased by inescapable stress: reversal by fluoxetine treatment. Neuropsy-chopharmacology 28: 1562–1571.

95. Nacher J, Gomez-Climent MA, McEwen B (2004) Chronic non-invasiveglucocorticoid administration decreases polysialylated neural cell adhesion

molecule expression in the adult rat dentate gyrus. Neurosci Lett 370: 40–44.96. Thomas RM, Hotsenpiller G, Peterson DA (2007) Acute psychosocial stress

reduces cell survival in adult hippocampal neurogenesis without altering

proliferation. J Neurosci 27: 2734–2743.97. Thomas RM, Urban JH, Peterson DA (2006) Acute exposure to predator odor

elicits a robust increase in corticosterone and a decrease in activity withoutaltering proliferation in the adult rat hippocampus. Exp Neurol 201: 308–315.

98. Dallman MF, Akana SF, Laugero KD, Gomez F, Manalo S, et al. (2003) A

spoonful of sugar: feedback signals of energy stores and corticosterone regulateresponses to chronic stress. Physiol Behav 79: 3–12.

99. Jayatissa MN, Bisgaard CF, West MJ, Wiborg O (2008) The number of granulecells in rat hippocampus is reduced after chronic mild stress and re-established

after chronic escitalopram treatment. Neuropharmacology 54: 530–541.100. Galea LA, McEwen BS (1999) Sex and seasonal differences in the rate of cell

proliferation in the dentate gyrus of adult wild meadow voles. Neuroscience 89:

955–964.101. Galea LA, Spritzer MD, Barker JM, Pawluski JL (2006) Gonadal hormone

modulation of hippocampal neurogenesis in the adult. Hippocampus 16:225–232.

102. Miranda P, Williams CL, Einstein G (1999) Granule cells in aging rats are

sexually dimorphic in their response to estradiol. J Neurosci 19: 3316–3325.103. Shors TJ, Chua C, Falduto J (2001) Sex differences and opposite effects of stress

on dendritic spine density in the male versus female hippocampus. J Neurosci21: 6292–6297.

104. Sutcliffe JS, Marshall KM, Neill JC (2007) Influence of gender on working andspatial memory in the novel object recognition task in the rat. Behav Brain Res

177: 117–125.

105. Tanapat P, Hastings NB, Reeves AJ, Gould E (1999) Estrogen stimulates atransient increase in the number of new neurons in the dentate gyrus of the

adult female rat. J Neurosci 19: 5792–5801.106. Westenbroek C, Den Boer JA, Veenhuis M, Ter Horst GJ (2004) Chronic stress

and social housing differentially affect neurogenesis in male and female rats.

Brain Res Bull 64: 303–308.

107. Galea LA (2008) Gonadal hormone modulation of neurogenesis in the dentate

gyrus of adult male and female rodents. Brain Res Rev 57: 332–341.

108. Lee SJ, McEwen BS (2001) Neurotrophic and neuroprotective actions of

estrogens and their therapeutic implications. Annu Rev Pharmacol Toxicol 41:

569–591.

109. Frye CA, McCormick CM (2000) Androgens are neuroprotective in the

dentate gyrus of adrenalectomized female rats. Stress 3: 185–194.

110. Garcia-Segura LM, Suarez I, Segovia S, Tranque PA, Cales JM, et al. (1988)

The distribution of glial fibrillary acidic protein in the adult rat brain is

influenced by the neonatal levels of sex steroids. Brain Res 456: 357–363.

111. Llorente R, Llorente-Berzal A, Petrosino S, Marco EM, Guaza C, et al. (2008)

Gender-dependent cellular and biochemical effects of maternal deprivation on

the hippocampus of neonatal rats: A possible role for the endocannabinoid

system. Dev Neurobiol 68: 1334–1347.

112. Mouton PR, Long JM, Lei DL, Howard V, Jucker M, et al. (2002) Age and

gender effects on microglia and astrocyte numbers in brains of mice. Brain Res

956: 30–35.

113. Conejo NM, Gonzalez-Pardo H, Cimadevilla JM, Arguelles JA, Diaz F, et al.

(2005) Influence of gonadal steroids on the glial fibrillary acidic protein-

immunoreactive astrocyte population in young rat hippocampus. J Neurosci

Res 79: 488–494.

114. Steiner B, Kronenberg G, Jessberger S, Brandt MD, Reuter K, et al. (2004)

Differential regulation of gliogenesis in the context of adult hippocampal

neurogenesis in mice. Glia 46: 41–52.

115. Rao MS, Shetty AK (2004) Efficacy of doublecortin as a marker to analyse the

absolute number and dendritic growth of newly generated neurons in the adult

dentate gyrus. Eur J Neurosci 19: 234–246.

116. Brandt MD, Jessberger S, Steiner B, Kronenberg G, Reuter K, et al. (2003)

Transient calretinin expression defines early postmitotic step of neuronal

differentiation in adult hippocampal neurogenesis of mice. Mol Cell Neurosci

24: 603–613.

117. Champagne DL, Bagot RC, van Hasselt F, Ramakers G, Meaney MJ, et al.

(2008) Maternal care and hippocampal plasticity: evidence for experience-

dependent structural plasticity, altered synaptic functioning, and differential

responsiveness to glucocorticoids and stress. J Neurosci 28: 6037–6045.

118. Herrick SP, Waters EM, Drake CT, McEwen BS, Milner TA (2006)

Extranuclear estrogen receptor beta immunoreactivity is on doublecortin-

containing cells in the adult and neonatal rat dentate gyrus. Brain Res 1121:

46–58.

119. Spritzer MD, Galea LA (2007) Testosterone and dihydrotestosterone, but not

estradiol, enhance survival of new hippocampal neurons in adult male rats. Dev

Neurobiol 67: 1321–1333.

120. Sutanto W, Rosenfeld P, de Kloet ER, Levine S (1996) Long-term effects of

neonatal maternal deprivation and ACTH on hippocampal mineralocorticoid

and glucocorticoid receptors. Brain Res Dev Brain Res 92: 156–163.

121. Garcia A, Steiner B, Kronenberg G, Bick-Sander A, Kempermann G (2004)

Age-dependent expression of glucocorticoid- and mineralocorticoid receptors

on neural precursor cell populations in the adult murine hippocampus. Aging

Cell 3: 363–371.

122. Schmidt-Hieber C, Jonas P, Bischofberger J (2004) Enhanced synaptic

plasticity in newly generated granule cells of the adult hippocampus. Nature

429: 184–187.

123. Joels M, Karst H, Krugers HJ, Lucassen PJ (2007) Chronic stress; implications

for neuron morphology, function and neurogenesis. Frontiers in Neuroendo-

crinology.



Rhythmicity in Mice Selected for Extremes in StressReactivity: Behavioural, Endocrine and Sleep ChangesResembling Endophenotypes of Major DepressionChadi Touma1.*, Thomas Fenzl1., Jorg Ruschel1, Rupert Palme2, Florian Holsboer1, Mayumi Kimura1,

Rainer Landgraf1

1 Max Planck Institute of Psychiatry, Munich, Germany, 2 Department of Biomedical Sciences, University of Veterinary Medicine, Vienna, Austria

Abstract

Background: Dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, including hyper- or hypo-activity of the stresshormone system, plays a critical role in the pathophysiology of mood disorders such as major depression (MD). Furtherbiological hallmarks of MD are disturbances in circadian rhythms and sleep architecture. Applying a translational approach,an animal model has recently been developed, focusing on the deviation in sensitivity to stressful encounters. This so-called‘stress reactivity’ (SR) mouse model consists of three separate breeding lines selected for either high (HR), intermediate (IR),or low (LR) corticosterone increase in response to stressors.

Methodology/Principle Findings: In order to contribute to the validation of the SR mouse model, our study combined theanalysis of behavioural and HPA axis rhythmicity with sleep-EEG recordings in the HR/IR/LR mouse lines. We found thathyper-responsiveness to stressors was associated with psychomotor alterations (increased locomotor activity andexploration towards the end of the resting period), resembling symptoms like restlessness, sleep continuity disturbancesand early awakenings that are commonly observed in melancholic depression. Additionally, HR mice also showedneuroendocrine abnormalities similar to symptoms of MD patients such as reduced amplitude of the circadianglucocorticoid rhythm and elevated trough levels. The sleep-EEG analyses, furthermore, revealed changes in rapid eyemovement (REM) and non-REM sleep as well as slow wave activity, indicative of reduced sleep efficacy and REM sleepdisinhibition in HR mice.

Conclusion/Significance: Thus, we could show that by selectively breeding mice for extremes in stress reactivity, clinicallyrelevant endophenotypes of MD can be modelled. Given the importance of rhythmicity and sleep disturbances asbiomarkers of MD, both animal and clinical studies on the interaction of behavioural, neuroendocrine and sleep parametersmay reveal molecular pathways that ultimately lead to the discovery of new targets for antidepressant drugs tailored tomatch specific pathologies within MD.

Citation: Touma C, Fenzl T, Ruschel J, Palme R, Holsboer F, et al. (2009) Rhythmicity in Mice Selected for Extremes in Stress Reactivity: Behavioural, Endocrine andSleep Changes Resembling Endophenotypes of Major Depression. PLoS ONE 4(1): e4325. doi:10.1371/journal.pone.0004325


Received September 11, 2008; Accepted November 26, 2008; Published January 29, 2009

Copyright: � 2009 Touma et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: Funding for this study was provided by the Max Planck Society (MPS). The MPS had no role in study design, data collection and analysis, decision topublish, or preparation of the manuscript.



. These authors contributed equally to this work.

Introduction

The rotation of the earth exposes all organisms to a daily change

in light intensity and virtually all species have adapted their lifestyles

to cycles of 24 hours [1]. These daily rhythms are endogenously

generated and are synchronised to external time cues in order to

ensure that bodily processes are carried out at the appropriate,

optimal time of day or night [2,3]. In mammals, the suprachias-

matic nuclei in the anterior-ventral hypothalamus are the principal

oscillator coordinating many physiological and behavioural func-

tions, including the circadian rhythms of body temperature,

hormone secretion (e.g. melatonin, luteinising hormone, growth

hormone) and sleep-wake behaviour [2–4]. The activity of the

hypothalamic-pituitary-adrenal (HPA) axis is also characterised by a

prominent circadian rhythm with peak glucocorticoid (GC)

secretion occurring shortly before the onset of an animal’s activity

period and trough levels during the beginning of the resting period

[4,5]. This daily variation of GC concentration is critical for

homeostatic regulation of metabolic, cardiovascular and neural

processes, and a bidirectional interaction between sleep and the

HPA system has been well established [6–8].

Interestingly, the sleep-endocrine regulation is critically influ-

enced by brain areas, which also play an important role in the

pathophysiology of affective disorders such as major depression

(MD) [6–10]. These include the hypothalamus, particularly the

paraventricular nucleus, but also limbic areas such as the

hippocampus and the amygdala, the prefrontal cortex as well as

afferent brain nuclei, in particular the locus coeruleus and the

raphe nuclei [6–11]. Therefore, it is not surprising that sleep

disturbances are among the most common symptoms of MD


[6,7,9–12]. Compared to healthy subjects, electroencephalogram

(EEG) recordings, which allow to objectively assess sleep

alterations, revealed that MD patients often suffer from insomnia

and sleep fragmentation (i.e. increased time to fall asleep, frequent

awakenings and early morning awakenings, some hours earlier

than desired, with difficulty returning to sleep). They also show a

reduced latency to the first episode of rapid eye movement (REM)

sleep, an increased proportion of REM sleep (increased REM

density and intensity) and reduced slow-wave activity (SWA)

during non-REM (NREM) sleep [6,7,11,12].

Another biological hallmark of MD is the dysregulation of the

HPA axis (hyper- or hypo-activity), largely involving pathological

alterations in the corticotrophin-releasing hormone (CRH) system

(for reviews see [6,9,10,13–20]). Therefore, sleep-EEG and stress

hormone alterations were also among the first biological changes

reported in MD [21–23]. Common neuroendocrine symptoms of

severely depressed patients include a flattened diurnal rhythm of

GC secretion (in particular, elevated trough levels have been

observed [24–27], elevated plasma and 24-h urinary GC

concentrations (hypercortisolism) and adrenal hyperplasia. Fur-

thermore, dysfunctional GC receptor-mediated negative feedback

regulation of the HPA axis and changes in vasopressin and CRH

responsiveness have frequently been described [9,13,15–20].

However, it is increasingly acknowledged that the diagnosis of

MD encompasses patients who do not necessarily share the same

disease biology, supporting the concept of different subtypes of

depression [7,10,15,17,28]. For instance, HPA axis overdrive,

related to an enhanced secretion of CRH and an impaired

negative feedback via GC receptors, is most consistently observed

in patients with melancholic depression. These patients also show

the most pronounced sleep-EEG alterations, including disrupted

sleep, decreased SWA, short REM sleep latency and high REM

sleep density. In contrast, patients presenting with the so-called

atypical subtype of depression are characterised by markedly

reduced activity of the HPA axis, while sleep-EEG data suggest

that SWA is not reduced and REM sleep parameters are not

considerably altered in these patients [7,15,17,28,29].

Based on the vital link between stress sensitivity and the

development of MD [9,13,16,18,20], a new, genetic animal model

has been recently established at the Max Planck Institute of

Psychiatry, focusing on alterations in HPA axis reactivity [30].

This so-called ‘stress reactivity’ (SR) mouse model consists of three

separate breeding lines selected for either high (HR), intermediate

(IR), or low (LR) corticosterone increase in response to a moderate

psychological stressor (15-min restraint). Significant differences in

the reactivity of the HPA axis between HR, IR and LR mice were

already found in the first generation of the selective breeding

process and proved to be a highly heritable trait, i.e. the respective

phenotype was confirmed across all subsequent generations and

could even be increased by assortative breeding [30]. Moreover,

results of an extensive behavioural test battery applied to the

selected mouse lines as well as neuroendocrine characterisation

experiments revealed several phenotypic similarities with changes

observed in depressive patients [30]. In general, HR animals were

relatively hyperactive in some behavioural paradigms, resembling

symptoms of restlessness and agitation often seen in melancholic

depression. LR mice, on the other hand, showed more passive-

aggressive coping styles, corresponding to signs of retardation and

retreat observed in atypical depression.

As outlined above, HPA axis functioning plays a critical role for

the regulation of sleep and activity rhythms. Therefore, the aim of

this study was to combine the analysis of behavioural and HPA

axis rhythmicity with sleep-EEG recordings in the HR/IR/LR

mouse lines, in order to provide a more comprehensive picture of

endophenotypes associated with increased or decreased stress

reactivity. Thus, we intended to further contribute to the

validation of the SR mouse model as a promising tool to elucidate

molecular genetic, neuroendocrine and behavioural parameters

associated with altered HPA axis reactivity.

Methods

Animals and general housing conditionsAll animals used in this study derived from the seventh

generation (Gen VII) of the ‘stress reactivity’ (SR) mouse model.

As outlined above, this model consists of three independent mouse

lines selectively bred for either high (HR), intermediate (IR) or low

(LR) reactivity of the HPA axis (for a detailed description of the

model see [30]).

Details about housing conditions, age, and the number of mice

used in each experiment are given in the respective sections (see

below). In general, from weaning at the age of about four weeks all

animals were housed in same-sex groups of two to four mice in

transparent polycarbonate cages (standard Macrolon cages type

III, 38622615 cm3) with wood chips as bedding and wood

shavings as nesting material (Product codes: LTE E-001 and NBF

E-011, ABEDD - LAB and VET Service GmbH, Vienna, Austria).

The animal housing room as well as the experimental rooms were

maintained under standard laboratory conditions (light-dark cycle:

12 : 12 h, lights on at 8 a.m.; temperature: 2261uC; relative

humidity: 55610%). Commercial mouse diet (Altromin No. 1324,

Altromin GmbH, Lage, Germany) and bottled tap water were

available ad libitum.

The presented work complies with current regulations covering

animal experimentation in Germany and the EU (European

Communities Council Directive 86/609/EEC). All experiments

were announced to the appropriate local authority and were

approved by the ‘Animal Welfare Officer’ of the Max Planck

Institute of Psychiatry (Az. 55.2-1-54-2531-64-07 and Az. 209.1/

211-33/04).

Stress reactivity testing and selection of experimentalanimals

Routinely, all animals of each breeding generation of the SR

mouse model are subjected to a so-called ‘stress reactivity test’

(SRT) performed at around eight weeks of age. Details about the

test procedure and subsequent analyses are described by Touma

and colleagues [30]. Briefly, the SRT consists of a 15-min restraint

period and two tail blood samplings immediately before and after

exposure to the stressor. All animals are tested in the first hours of

the light phase (between 9 a.m. and 11 a.m.), i.e. during the trough

of the circadian rhythm of GC secretion. From the collected

‘initial’ and ‘reaction’ blood samples, corticosterone concentra-

tions are determined by radioimmunoassay, quantifying the

reactivity of the HPA axis as corticosterone increase in response

to a moderate psychological stressor.

According to the outcome of this SRT, 36 male mice (12 of each

breeding line) of Gen VII were selected as experimental animals,

showing a high, intermediate or low corticosterone increase,

characteristic of the neuroendocrine stress response phenotype of

the HR, IR and LR breeding line, respectively [30].

Behavioural activity rhythmsTo monitor the undisturbed behavioural activity rhythm of the

animals, we used a self-made device (‘System for Automatic

Measurement of Laboratory Animals’ Behaviour’ SAMLAB) to

automatically track resting versus motor activity, explorative

behaviours as well as feeding and drinking activities in the home

Rhythmicity & HPA Reactivity


cage. This was achieved by placing two computer-connected metal

frames equipped with 48 infrared light barriers (32616 photo

electric sensors) positioned at a distance of about 1 cm around

each cage (standard Macrolon cage type III with a special bedding

material: Rehofix maize granulate, MK 2000, round particle size:

1.7–2.2 mm, absorption capacity: 2.0 l/kg), with the first frame 3

cm above ground level and the other 5 cm higher. Thereby, the

cage floor as well as the upper level of the cage were divided into a

grid of 1.2561.5 cm2, enabling the detection of the position of the

mouse and the calculation of its movements according to light

beam breaks using a customized software (construction and

programming of SAMLAB by Oleg Dolgov). Motor activity was

defined as non-resting, i.e. when the mouse was moving and

activating changing sets of light barriers. Explorative behaviours

such as rearing and climbing on the cage lid could also be

detected, i.e. when the mouse activated light barriers at the ground

and upper level simultaneously (rearing) or only showed

movement in the upper grid of light barriers (climbing). In order

to avoid potentially confounding influences of human activities in

the housing room on the animals activity rhythm, the whole

activity monitoring setup was built into a soundproof cabinet

equipped with an autonomic ventilation, temperature, humidity

and light control system (set to the same conditions as in the

housing room; see above). Glass doors allowed inspections of the

test animals without disturbing the measurements.

Each mouse (12 HR, IR and LR males, respectively, about

10 weeks of age; see above) was single housed for at least two

weeks before being put into the activity monitoring device and was

kept in the light barrier monitored cages for one week with

minimal disturbance from outside. The first three days in the

apparatus were regarded as a habituation period and, therefore,

only the data of the last four days were analysed. The ‘time spent

active’ (motor activity and explorative behaviours; see definitions

above) was continuously recorded during the entire 24-hour light-

dark cycle and was averaged for each individual in hourly intervals

over the four recording days, resulting in a mean activity pattern

for each mouse.

Diurnal rhythm of glucocorticoid secretionIn order to accurately follow the natural diurnal rhythm of

glucocorticoid secretion in HR, IR and LR mice without

interfering with the activation of the HPA axis by repeated

handling and blood sampling, a non-invasive technique to monitor

adrenocortical activity by measuring corticosterone metabolites

(CM) in the faeces of mice was applied [31,32]. This technique of

glucocorticoid metabolite quantification in faecal samples has been

established in a large number of species (for review see [33]) and

has been extensively validated for laboratory mice [31,32].

The same 36 HR/IR/LR males (12 of each breeding line) that

were characterised with respect to their behavioural activity rhythms

(see above) were also used in this experiment (after two weeks of

normal housing in standard cages). For a period of 48 hours, faecal

samples were collected quantitatively in short sampling intervals of

two hours and stored at 220uC until analysis of CM (see below). To

facilitate individual sampling and quantitative collection of all voided

faeces without handling the animal, the method described by Touma

and colleagues was used [31,32]. Briefly, the mice were housed

individually in stainless steel wire cages (38622615 cm3), which were

placed in standard Macrolon cages of the same size. All excreta

dropped through the bars of the wire cage and could easily be

collected from the floor of the lower cage, which was completely

covered with filter paper that immediately absorbed the urine. To

habituate the mice to this sampling procedure and to being housed in

wire cages, the animals were already placed into this housing system

three days prior to the beginning of the experiment and samples were

collected in 12 hour intervals during this time. Since mice are

nocturnal animals and their steroid excretion pattern is known to be

influenced by their activity [32], all sample collections performed

during the dark phase of the light-dark cycle were conducted under

dimmed lighting conditions (less than 5 lux) to avoid disturbing the

animals’ natural activity pattern.

The collected faecal samples were analyzed for immunoreactive

CM using a 5a-pregnane-3b,11b,21-triol-20-one enzyme-immu-

noassay (EIA). Details regarding development, biochemical

characteristics, and biological validation of this assay are described

by Touma and colleagues [31,32]. Before EIA analysis, the faecal

samples were homogenized and aliquots of 0.05 g were extracted

with 1 ml of 80% methanol. A detailed description of the assay

performance has been published elsewhere [32]. Briefly, the EIA

used a double-antibody technique and was performed on anti-

rabbit-IgG-coated microtitre plates. After overnight incubation (at

4uC) of standards (range: 0.8–200 pg/well) and samples with

steroid antibody and biotinylated label, the plates were emptied,

washed and blotted dry, before a streptavidin horseradish

peroxidase conjugate was added. After 45 minutes incubation

time, plates were emptied, washed, and blotted dry. The substrate

(tetramethylbenzidine) was added and incubated for another

45 minutes at 4uC before the enzymatic reaction was stopped with

1 mol/l sulphuric acid. Then, the optical density (at 450 nm) was

recorded with an automatic plate reader and the hormone

concentrations were calculated. The intra- and inter-assay

coefficients of variation were 8.8% and 13.4%, respectively.

For each individual, CM concentrations of the two correspond-

ing sampling intervals during the 48 hour sampling period were

averaged, yielding a mean diurnal pattern of glucocorticoid

secretion for each mouse.

Sleep recordingsTo study the sleep patterns and quality of HR, IR and LR mice,

EEG recordings were performed with a subset of animals (N = 8

males per line) from the experiments described above. All animals

were housed individually in customized recording cages (26626635

cm3) located in sound-attenuated chambers kept at constant

laboratory conditions (22uC 61uC, 12 : 12 h light-dark cycles,

lights on at 10 a.m.). Food and water were available ad libitum.

Surgical procedures were performed under isoflurane/oxygen

anaesthesia using a custom-made vaporizing device. At the

beginning of the surgery, each animal also received atropinesulfate

(0.05 mg/kg BW) and meloxicam (0.5 mg/kg BW) subcutaneously

for cardiovascular stabilisation and analgesia, respectively. The

animals were positioned in a stereotactic frame and four epidural

EEG and two intramuscular electromyogram (EMG) electrodes

were implanted. Briefly, the skin and muscles overlaying the skull

were cut rostro-caudally along the midline, drawn to the sides and

kept in place using small retractors. To insert the EEG electrodes,

four small holes (diameter: 200 mm) were drilled into the skull.

Two electrodes were placed bilaterally at the frontal region of the

cortex, one reference-electrode was placed at the right parietal

area, and the ground electrode was inserted at the left parietal

area. All four electrodes were fixed with dental cement to the skull.

Additionally, two bilateral EMG electrodes were embedded

laterally of the spine into the neck muscles. All electrodes were

composed of gold wire with ball-shaped endings, soldered to a

small standard printed circuit board connector. In order to

provide more stability to the assembly on the skull for chronic

recording, two additional small holes were drilled for mounting

jeweller’s screws that were also framed with dental cement and

glued together with the electrodes and the connector.



After surgery, the animals were allowed to recover for two weeks

in the recording cages before two successive 23-hour recording

sessions of EEG and EMG signals were performed. During

recovery and recording each mouse was attached to a recording

cable, which was connected to a swivel system allowing relatively

free movement of the animals.

EEG and EMG signals were fed online into a preamplifier

(10006, custom made) and a main amplifier (106, custom made).

The EEG signals were analogue band-pass filtered (0.5–29 Hz,

filter frequency roll off 48 dB/octave) and digitized at a sampling

rate of 64 Hz (AD board, NI PCI-6070, National Instruments,

Austin, USA). Root mean square was applied to all non-filtered

EMG signals before its digital conversion (64 Hz). The vigilance

states ‘wake’, ‘non-rapid eye movement sleep’ (NREM sleep) and

‘rapid eye movement sleep’ (REM sleep) were scored on a

LabVIEW-based scoring program (SEA, Koln, Germany) semi-

automatically with a Fast Fourier Transformation algorithm

spectral analysis and could be corrected manually, if necessary

(the scoring technique was validated beforehand). The frequency

bands were as follows: d (0.5–5 Hz), h (6–9 Hz), a (10–15 Hz), g(16–22.5 Hz) and b (23–31.75 Hz). A detailed description of the

scoring procedure is described elsewhere [34]. Slow wave activity

(SWA, NREM sleep frequency bands: 0.5–15 Hz; SWA frequency

bands: 0.5–4 Hz in 0.5 Hz steps) was calculated from the total

amount of NREM sleep across the 23-hour recording time in one

hour means.

Statistical AnalysisSince a normal distribution and variance homogeneity of the

data could not always be assumed, analyses were exclusively

performed using non-parametric statistics [35]. All tests were

applied two-tailed and were calculated using the software package

SPSS (version 16.0). ANOVA on ranks (Friedman-test) was used to

evaluate differences between more than two dependent (related)

samples. Two independent samples were compared using the

Mann-Whitney U-test (MWU-test), while differences between

more than two independent samples were calculated with the

Kruskal-Wallis H-test (KWH-test). In the case of significant

variation proved by the KWH-test, post-hoc pairwise comparisons

between the groups were done using multiple MWU-tests.

Spearman’s rank-order correlation was calculated to elucidate

the degree of association between two variables. As nominal level

of significance a= 0.05 was accepted and corrected for post-hoc

tests according to the sequential Bonferroni technique [36].

Results

HPA axis reactivityAs expected, the experimental animals selected from generation

VII of the HR, IR and LR breeding lines differed significantly

regarding their corticosterone increase in the SRT (KWH-test:

N = 12 for each line, H = 31.1, df = 2, p,0.001; see Fig. 1). HR

mice showed a very much exaggerated stress response, while

compared to IR animals the secretion of corticosterone was

strongly reduced in LR mice (see Fig. 1).

Behavioural activity rhythmsThe behavioural activity rhythms also differed significantly

between the three breeding lines. Although a clear pattern of

increased motor activity during the dark phase and less activity

during the light phase could be observed in all mouse lines

(Friedman-tests: N = 12 for each line, Chir2 = 134.1–142.7,

Figure 1. Corticosterone increase in the stress reactivity test (SRT) of the experimental animals selected from the seventhgeneration of the high (HR), intermediate (IR) and low (LR) reactivity mouse lines. Data are given as box plots showing medians (lines inthe boxes), 25% and 75% percentiles (boxes) as well as 10% and 90% percentiles (whiskers). Statistical differences between the three lines (KWH-test,for details see text) are given at the top of the panel and results of the pairwise group comparisons (post-hoc MWU-tests) are indicated below(Bonferroni corrected p,0.001 ***).doi:10.1371/journal.pone.0004325.g001



df = 23, all p,0.001; see Fig. 2), at several time points during the

24-hour light-dark cycle significant differences in motor activity

were found between HR, IR and LR animals (KWH-tests: N = 12

for each line, experimental time points: 4, 7, 8, 9, 16, 20, H = 6.4–

8.9, df = 2, all p,0.05; see Fig. 2). In particular, in the second half

of the light phase, i.e. some hours before the light-dark transition,

HR mice were clearly more active than IR and LR mice, while the

latter did not differ significantly from each other in their overall

activity pattern (see Fig. 2).

A similar picture emerged for the distribution of explorative

behaviours such as rearing and climbing. Again, all three mouse

lines showed a significant variation over the day regarding the time

spent exploring the cage (Friedman-tests: N = 12 for each line,

Chir2 = 135.5–149.5, df = 23, all p,0.001; see Fig. 3). However,

the increase of explorative activities shortly before the light-dark

transition was much more pronounced and advanced by some

hours in HR mice compared to the other two lines, resulting in

significant differences at several time points during the light phase

(KWH-tests: N = 12 for each line, experimental time points: 6, 7,

8, 9, 10, H = 6.1–12.6, df = 2, all p,0.05; see Fig. 3).

Diurnal rhythm of glucocorticoid secretionRegarding the diurnal rhythm of glucocorticoid secretion, all

three mouse lines showed a significant variation of CM

concentrations over the 24-hour light-dark cycle (Friedman-tests:

N = 12 for each line, Chir2 = 100.9–110.5, df = 12, all p,0.001;

see Fig. 4). Overall, highest concentrations were measured during

the dark phase (peaking around midnight), while relatively low

CM levels were observed during the light phase. Comparing the

concentrations of excreted CM between HR, IR and LR animals

across the day, however, revealed significant differences at several

sampling time points during the light as well as during the dark

phase (KWH-tests: N = 12 for each line, experimental time points:

0, 2, 4, 6, 8, 10, 20, 24, H = 8.4–18.2, df = 2, all p,0.05; see Fig. 4).

In general, HR mice showed distinctly and significantly higher

CM concentrations than IR and LR animals, in particular during

the light phase, but not so much at the beginning of the dark phase

(see Fig. 4), resulting in a flattened diurnal rhythm of CM

excretion (difference between the maximum and minimum CM

concentration across the day = Delta means: HR = 53.4,

IR = 69.9, LR = 68.2, KWH-test: N = 12 for each line, H = 7.5,

df = 2, p,0.05). Furthermore, the area under the curve (AUC) and

the mean location (ML) of CM concentrations was significantly

higher in HR animals compared to the other two lines, which did

not differ significantly from each other (AUC means:

HR = 1678.3, IR = 1280.1, LR = 1133.5; ML means: HR = 68.9,

IR = 52.2, LR = 46.1; KWH-tests: N = 12 for each line, H = 8.2

and 8.4, df = 2, both p,0.05).

Sleep recordingsThe results of the sleep recording experiment are presented in

figure 5. Similar to the behavioural activity patterns described

above, the distribution of wakefulness and sleep varied significantly

during the course of the day in all three mouse lines, with a larger

amount of time spent sleeping in the light phase than in the dark

phase. However, at several time points across the light-dark cycle,

the relative amount of time the animals spent in either vigilance

state (wake, NREM sleep or REM sleep) differed significantly

between HR, IR and LR mice.

The amount of wakefulness was higher in HR mice during the

second half of the light phase, but significant effects were also

found at the end of the dark phase (KWH-tests: N = 8 for each

line, experimental time points: 7, 12, 23, H = 6.0–8.4, df = 2, all

p,0.05; see Fig. 5A). Similarly, the lines differed significantly in

the total amount of NREM sleep (KWH-tests: N = 8 for each line,

experimental time points: 7, 12, 14, H = 6.2–9.9, df = 2, all

Figure 2. Distribution of motor activity over the 24-h light-dark cycle in high (HR), intermediate (IR), and low (LR) reactivity malesfrom generation VII. Data are given as means6SEM for each line. Statistical differences between the three lines are indicated by asterisks (KWH-tests, for details see text, p,0.05 *). The dark phase of the light-dark cycle is indicated by the shaded area.doi:10.1371/journal.pone.0004325.g002



Figure 3. Distribution of explorative behaviour over the 24-h light-dark cycle in high (HR), intermediate (IR), and low (LR) reactivitymales from generation VII. Data are given as means6SEM for each line. Statistical differences between the three lines are indicated by asterisks(KWH-tests, for details see text, p,0.05 *, p,0.01 **). The dark phase of the light-dark cycle is indicated by the shaded area.doi:10.1371/journal.pone.0004325.g003

Figure 4. Diurnal variation of immunoreactive corticosterone metabolites (CM) in faecal samples of high (HR), intermediate (IR),and low (LR) reactivity males from generation VII over the 24-h light-dark cycle. Data are given as means6SEM for each line. Statisticaldifferences between the three lines are indicated by asterisks (KWH-tests, for details see text, p,0.05 *, p,0.01 **, p,0.001 ***). The dark phase ofthe light-dark cycle is indicated by the shaded area.doi:10.1371/journal.pone.0004325.g004



p,0.05; see Fig. 5B). Post-hoc pairwise comparisons revealed that

HR mice spent less time in NREM sleep than LR mice (MWU-

tests: N = 8 for each line, experimental time points: 7, 12, U = 6–

10, all Bonferroni corrected p,0.05; see Fig. 5B). The most

pronounced differences between the three lines, however, were

found regarding the amount of REM sleep (KWH-tests: N = 8 for

each line, experimental time points: 1, 4, 6, 8, 9, 21, 23, H = 8.4–

15.8, df = 2, all p,0.05; see Fig. 5C). During the majority of time

points in the light phase as well as towards the end of the dark

phase, HR mice spent much more time in REM sleep than LR

(MWU-tests: N = 8 for each line, experimental time points: 1, 4, 6,

8, 9, 21, 23, U = 0–8, all Bonferroni corrected p,0.05; see Fig. 5C)

and IR animals (MWU-tests: N = 8 for each line, experimental

time points: 1, 4, 9, 21, 23, U = 1–10, all Bonferroni corrected

p,0.05; see Fig. 5C).

Focusing on the SWA within NREM sleep episodes also

revealed distinct differences between the three mouse lines at

virtually every time point across the light-dark cycle (KWH-tests:

N = 8 for each line, experimental time points: 1, 2, 3, 4, 5, 6, 7, 8,

9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, H = 4.6–15.0,

df = 2, all p,0.05; see Fig. 6). Post-hoc tests confirmed that HR

mice showed a clearly and significantly decreased SWA,

particularly when compared to LR males (MWU-tests: N = 8 for

each line, experimental time points: 1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,

13, 15, 16, 17, 19, 20, 21, 23, U = 0–5, all Bonferroni corrected

p,0.05, experimental time points: 2, 18, 22, U = 6–8, all

Bonferroni corrected p,0.1; see Fig. 6), but also in comparison

to IR animals (MWU-tests: N = 8 for each line, experimental time

points: 1, 3, 4, 5, 6, 7, 8, 9, 10, 12, 13, 17, 20, 21, 22, 23, U = 0–13,

all Bonferroni corrected p,0.05, experimental time points: 15, 16,

18, 19, U = 8–14, all Bonferroni corrected p,0.1; see Fig. 6). At

some experimental time points, the amount of SWA was also

significantly higher in LR than in IR animals (MWU-tests: N = 8

for each line, experimental time points: 13, 19, 20, 21, U = 2–8, all

Bonferroni corrected p,0.05, experimental time point: 12,

U = 12, Bonferroni corrected p,0.1; see Fig. 6).

Correlation analysis further revealed in HR mice significant

associations between stress reactivity (corticosterone increase in the

Figure 5. Distribution of vigilance states over the 24-h light-dark cycle in high (HR), intermediate (IR), and low (LR)reactivity males from generation VII. The relative amount ofwakefulness, non-rapid eye movement (NREM) sleep and rapid eyemovement (REM) sleep are plotted in panel A, B and C, respectively.Data are given as means6SEM for HR and LR mice and as SEM-area forthe IR mouse line. Statistical differences between the three lines areindicated by asterisks (KWH-tests, for details see text, p,0.05 *). Thedark phase of the light-dark cycle is indicated by the shaded area.doi:10.1371/journal.pone.0004325.g005

Figure 6. Distribution of the relative amount of slow waveactivity (SWA) over the 24-h light-dark cycle in high (HR),intermediate (IR), and low (LR) reactivity males from genera-tion VII. Data are given as means6SEM for HR and LR mice and asSEM-area for the IR mouse line. Statistical differences between the threelines are indicated by asterisks (KWH-tests, for details see text, p,0.05*). The dark phase of the light-dark cycle is indicated by the shadedarea.doi:10.1371/journal.pone.0004325.g006



SRT) and the AUC of excreted CM (rs = 0.783, Bonferroni

corrected p,0.01), the proportion of REM sleep (rs = 0.810,

Bonferroni corrected p,0.05) and the amount of SWA

(rs = 20.738, Bonferroni corrected p,0.05) across the day. That

is, animals with a greater corticosterone response in the SRT

showed a higher CM excretion profile, a greater increase in REM

sleep and a stronger decrease in SWA. For IR and LR animals,

however, no such correlations were found.

Discussion

Clinical studies provide clear evidence for a critical role of

circadian rhythm and sleep disturbances in the pathophysiology of

mood disorders, which are also closely linked to another biological

marker of MD, the dysregulation of the HPA axis (for reviews see

[6,9,10,13–20]. Applying a selective breeding approach, we

developed an animal model that resembles the deviation in

sensitivity to stressful encounters [30]. The aim of the present

study was to investigate this ‘stress reactivity’ mouse model with

respect to the clinically relevant endophenotypes of rhythmicity

and sleep disturbances.

We found significant differences between HR, IR and LR mice

regarding their circadian rhythm of psychomotor activity and GC

secretion as well as pronounced alterations in their sleep-EEG

profiles. HR mice for instance showed increased wakefulness,

locomotor activity and exploratory behaviours towards the end of

the resting period. Moreover, the amplitude of the circadian GC

rhythm was reduced due to elevated trough levels and the

proportion of REM sleep was clearly increased in these animals.

NREM sleep and SWA on the other hand were reduced in

comparison to the other two lines. No major rhythmicity

differences were found between IR and LR mice, except for a

significantly higher proportion of slow wave sleep across the day in

LR animals.

In the experiments addressing the behavioural activity rhythms

of the animals our results revealed significant differences in the

diurnal activity patterns of the three mouse lines. In general, as

expected for nocturnal rodents, all animals were more active

during the dark phase than during the light phase, but compared

to the other two lines, HR mice showed a marked increase in

activity towards the end of the light phase, i.e. some hours before

the light-dark transition. This increased psychomotor activity

during the resting period was found in the analysis of locomotion

(see Fig. 2) as well as exploratory behaviours (see Fig. 3) and can be

interpreted as resembling the symptoms of sleep fragmentation

and early morning awakenings often seen in melancholically

depressed patients [6,7,12]. This interpretation is also supported

by our sleep-EEG data, including a detailed event related analysis

(see discussion below). The fact that LR mice did not differ

considerably from IR animals with respect to their behavioural

activity rhythm is also in accordance with clinical findings, as MD

patients with atypical features are not reported to suffer from sleep

continuity disturbances or restlessness [7,37].

Concerning the diurnal variation of HPA axis activity, i.e. the

circadian rhythm of GC secretion, similar differences between the

three mouse lines were found, as observed for the behavioural

rhythms. Again, a clear pattern of increased GC concentrations

(measured as faecal CM) during the activity period and relatively

low levels during the resting period (including a trough at the

beginning of the light phase) were observed in all animals (see

Fig. 4). This typical cycle of nadir and peak concentrations is very

much in accordance with published data on laboratory rats

(plasma samples [38,39]; faecal samples [40]) and mice (plasma

samples [41–43]; faecal samples [31,44,45]). However, compared

to the other two lines, HR mice showed clearly elevated

concentrations of faecal CM during the entire light phase as well

as at the end of the dark phase, resulting in a markedly flattened

diurnal rhythm (see Fig. 4). IR and LR mice, on the other hand,

did not differ very much, although LR animals tended to have

lower CM levels across the 24-h light-dark cycle (see Fig. 4). These

findings further support the close association between HPA axis

activity/reactivity and disturbances of neuroendocrine rhythms, as

for example very similar alterations, including a reduced

amplitude in circadian cortisol secretion patterns, elevated trough

cortisol levels and increased 24-h means, have been found in

patients suffering from melancholic or psychotic depression, both

of which are characterized by a strong increase in HPA axis drive

[24–27]. Interestingly, data available for atypical depression

suggest no change or a slight decrease in trough cortisol levels

[7], indicating similarities with the phenotype observed in the LR

mouse line (see also [30]). Although our findings match reasonably

with these clinical observations, it should be highlighted that in

rodents, the entire human syndrome of MD cannot be modelled,

but they may share core symptoms of the disease, including the

molecular pathways underlying key endophenotypes.

Potential mechanisms that might be involved in bringing about

the described alterations in the circadian GC rhythm of our mouse

lines include variations in the activity of neural networks

(assessable as brain glucose metabolism differences across times

of day) as well as abnormal levels or patterns of noradrenalin and

melatonin secretion [6–8,12]. Furthermore, neurodegenerative

processes, particularly in structures participating in the regulation

of the HPA axis such as the hippocampus, might be an important

factor, as similar disturbances in the diurnal variation of GC have

been reported in Alzheimer’s and Parkinson’s disease patients as

well as in experimental models of prion disease [45–48]. The

deterioration of the circadian rhythm is interestingly often

observed before other clinical symptoms are manifested and can

be indicative of a relapse in the case of MD. Therefore alterations

of the circadian rhythm appear to be closely linked to the body’s

stress system and might have a significant impact for a number of

pathologies, including MD (for reviews see [16,49,50]).

Genotyping efforts as well as studies addressing changes in brain

neurotransmitter and neuromodulator systems (including CRH,

serotonin and noradrenalin) are currently underway, shedding

light on the molecular underpinnings of the endophenotypes

observed in the HR/IR/LR mouse lines. Potentially, this pre-

clinical research will also yield novel insights into the fundamental

mechanisms involved in the pathophysiology of human diseases.

As outlined above, sleep abnormalities are very common

symptoms of MD patients and have been in the focus of

researchers for several decades (for reviews see [6,7,12,51–54]).

Sleep is typically divided into NREM sleep and REM sleep

episodes; the former can be further subdivided into sleep stages I–

IV in humans. Stage I sleep, the transition from wakefulness with

its mixed frequency activity and dominant alpha waves (8–12 Hz)

to shallow sleep, is marked with dominant EEG frequencies of 4–7

Hz (theta waves). Sleep spindles with frequencies of 12–15 Hz and

K-complexes are hallmarks of stage II sleep [55]. In sleep stage III,

delta waves with a frequency of around 1-3(4) Hz, so-called SWA,

are present and become increasingly dominant in stage IV sleep

(referred to as slow wave sleep). REM sleep on the other hand is

characterised by a desynchronised EEG (similar to wakefulness)

and episodic erratic movements of the eyes together with low

amplitude electromyogram activity [55].

In healthy adults, NREM sleep and REM sleep normally

alternate periodically through the night starting with around 90

min of NREM sleep, followed by a short REM sleep period of



approximately 10 min. This cycle is then repeated four to six times

during the night, with decreasing portions of sleep stages III and

IV and increasing durations of the successive REM sleep periods

towards the end of the night [56]. In depressed patients, however,

increased stage I sleep, decreased stage III and stage IV sleep,

shorter NREM sleep duration, insomnia (involving difficulties

falling asleep, sleep fragmentation and early morning awakenings)

are often reported [52,53,57,58]. In addition, common sleep-EEG

alterations include decreased REM sleep latency, increased REM

density [59,60] and increased total time spent in REM sleep [61].

It has to be noted, however, that these sleep alterations are not

uniformly found across all MD patients. In particular, when the

different subtypes of melancholic and atypical depression are

considered, the emerging picture is different. Melancholic

depression, for instance, is characterised by the aforementioned

alterations, including poor sleep quality and decreased amounts of

sleep, whereas in atypical depression poor sleep quality is rather

associated with an overall increased amount of sleep and fatigue-

like behaviour during the day [7,15,37].

Interestingly, our findings from the sleep-EEG recordings in

HR, IR and LR mice also support this dichotomy of symptom

clusters linked with diametral differences in HPA axis reactivity.

HR mice were found to have more bouts of wakefulness during the

normal resting period of the animals (see Fig. 2 and Fig. 5A) and

also showed a significant reduction in the amount of NREM sleep

at several experimental time points (see Fig. 5B). An extensive

event related analysis (applying the ‘event-history-analysis pro-

gram’ developed by Alexander Yassouridis [62]) additionally

supports the notion of a shallower and more fragmented sleep in

HR mice, as the number of awakenings and stage shifts,

particularly from REM sleep to wake, was clearly increased

during the light as well as during the dark phase in this mouse line

(Fenzl and Touma et al., in preparation). These differences in sleep

architecture might be attributed to the increased activation of the

HPA axis across the day in the HR mouse line (see discussion

above and Fig. 4). CRH is known to impair sleep and enhance

vigilance, thereby suggesting a causal relationship between shallow

sleep and the hyperactivity of the HPA system in melancholic

depression [6–8,15,51]. Other preclinical studies also support this

view. In rats, after intracerebroventricular administration of CRH,

waking was enhanced, whereas alpha-helical CRH (a specific

CRH receptor antagonist) reduced spontaneous waking [51,63].

The most pronounced differences between HR, IR and LR

mice, however, were found regarding the amount of REM sleep.

At the majority of time points during the animals’ normal resting

period, HR mice spent much more time in REM sleep than the

other two lines (see Fig. 5C). Human sleep data suggest that

changes in REM sleep, mediated by the noradrenergic, seroto-

nergic and cholinergic systems, are not only a consequence of

depression, but can be seen as true endophenotype of the disease

(reviewed in [64]). Interestingly, in a transgenic mouse model

overexpressing CRH in the entire brain, REM sleep was also

significantly enhanced [65], along with a clearly increased

responsiveness of the HPA axis to stressors and alterations in

emotional behaviour [66], hence largely overlapping with our

observations in HR mice (see also [30]). Other animal studies as

well as clinical findings further support the notion that CRH

promotes REM sleep [6,67,68], although the effect of CRH on

REM sleep seems to be site- and dose-dependent [8]. Moreover,

our findings are in line with results of sleep investigations

performed in different animal models of depression such as

exposure to chronic unpredictable stress [69] and selection for

increased ‘helplessness’ in the tail suspension test [70,71]. These

studies revealed very similar alterations in sleep/wake patterns,

distribution of sleep stages and facilitation of REM sleep as we saw

in the HR mouse line, again underlining the significant impact of

stress responsiveness on sleep architecture.

In similarity to REM sleep, significant differences between HR,

IR and LR mice were found in the proportion of slow wave sleep.

Virtually across the entire light-dark cycle, HR mice showed a

dramatically lowered level of SWA, while higher SWA was

observed in LR animals (see Fig. 6). Sleep deprivation studies

indicate that SWA reflects sleep intensity, as it was clearly

increased as a function of waking [72]. In other words, SWA can

serve as a distinct marker for homeostatic sleep pressure [73]. The

regulation of SWA itself was proposed to be a function of the ‘Two

Process Model’ [74], depending on the interaction of processes S

(sleep dependent) and C (circadian). Sleep propensity, increasingly

depending on extended time spent awake, is reflected by process S.

In this model, the sleep intensity (process S) is at its maximum at

sleep onset, declining during consecutive sleep. It is beyond the

scope of this study to reveal whether the decreased amounts of

NREM sleep can be attributed to attenuated levels of SWA, but

this would implicate that reduced SWA is an intrinsic sleep-

physiological feature of the HR mouse line, which might be

brought about by a chronic activation of the CRH system [6,75].

Interestingly, clinical studies also report a reduction in SWA in

depressed patients [6,7,11,12,76], although slow wave sleep is not

reduced and REM sleep parameters seem to be less consistently

altered in patients with atypical depression [7,17].

Taken together, our study provides clear evidence for a critical

interaction between HPA axis dysregulation and rhythmicity

disturbances, including changes in behavioural activity patterns,

circadian GC secretion and sleep architecture. In our mouse model,

hyper-responsiveness to stressors was associated with psychomotor

activity alterations, resembling the restlessness, sleep discontinuity

and early awakenings commonly observed in melancholic depres-

sion. Furthermore, HR mice also showed neuroendocrine abnor-

malities such as reduced amplitude of the circadian GC rhythm and

elevated trough levels, potentially mimicking similar symptoms in

MD patients. The sleep-EEG analyses revealed changes in NREM

and REM sleep as well as SWA in HR mice, indicative of reduced

sleep efficacy and REM disinhibition, which reasonably overlap

with observations in melancholically depressed patients. Thus, by

selectively breeding mice for extremes in stress reactivity, clinically

relevant endophenotypes of MD can be modelled, presumably

including the symptomatology and pathophysiology of specific

subtypes of depression.

It should be emphasized, however, that animal models will only

be able to mimic certain aspects of the human disease biology

rather than the entire clinical syndrome and that not all features of

our SR model match with findings in MD patients. Limitations to

the clinical relevance of the HR/IR/LR mouse lines for instance

include that HPA axis dysregulation is currently not one of the

critical diagnostic criteria for MD and that it is a genetic model,

i.e. the differences in stress responsiveness are already present early

in life, thereby potentially influencing developmental processes

that shape the respective endophenotypes. On the other hand, also

in humans, the latter mechanisms (driven by both genetic and

environmental factors) might represent key variables underlying

individual vulnerability to psychiatric disorders [9,16,77,78].

Therefore, we are convinced that elucidating similar aspects of

biological alterations in animal models and human patients can be

a major progress and that translational approaches using

appropriate animal models can substantially further our under-

standing of how organisms respond to stress and the nature of

inter-individual differences in the stress response. Given the

importance of rhythmicity and sleep disturbances as biomarkers of



MD, both animal and clinical studies on the interaction of

behavioural, neuroendocrine and sleep parameters may reveal

molecular pathways that ultimately lead to the discovery of new

targets for antidepressant drugs tailored to match specific

pathologies within MD.

Acknowledgments

The authors like to thank Cornelia Flachskamm, Edith Klobetz-Rassam,

Markus Nussbaumer, Wolfgang Plendl and Marina Zimbelmann for

excellent technical assistance, Alana Knapman for critical reading of the

manuscript and Alexander Yassouridis for statistical advice.


Conceived and designed the experiments: CT TF. Performed the

experiments: CT TF JR. Analyzed the data: CT TF JR. Contributed

reagents/materials/analysis tools: RP FH MK RL. Wrote the paper: CT

TF.

References

1. Devlin PF, Kay SA (2001) Circadian photoperception. Annu Rev Physiol 63:

677–694.

2. Buijs RM, Kalsbeek A (2001) Hypothalamic Integration of Central andPeripheral Clocks. Nature Rev Neurosci 2: 521–526.

3. Hastings MH, Reddy AB, Maywood ES (2003) A clockwork web: circadiantiming in brain and periphery, in health and disease. Nat Rev Neurosci 4:

649–661.

4. Kriegsfeld LJ, Silver R (2006) The regulation of neuroendocrine function:Timing is everything. Horm Behav 49: 557–574.

5. Engeland WC, Arnhold MM (2005) Neural circuitry in the regulation of adrenal

corticosterone rhythmicity. Endocrine 28: 325–332.

6. Steiger A (2002) Sleep and the hypothalamo-pituitary-adrenocortical system.Sleep Med Rev 6: 125–138.

7. Antonijevic I (2008) HPA axis and sleep: Identifying subtypes of major

depression. Stress 11: 15–27.

8. Pawlyk AC, Morrison AR, Ross RJ, Brennan FX (2008) Stress-induced changes in

sleep in rodents: models and mechanisms. Neurosci Biobehav Rev 32: 99–117.

9. Nestler EJ, Barrot M, DiLeone RJ, Eisch AJ, Gold SJ, Monteggia LM (2002)Neurobiology of Depression. Neuron 34: 13–25.

10. Hasler G, Drevets WC, Manji HK, Charney DS (2004) Discovering

endophenotypes for major depression. Neuropsychopharmacol 29: 1765–1781.

11. Steiger A (2007) Neurochemical regulation of sleep. J psychiat Res 41: 537–552.

12. Germain A, Kupfer DJ (2008) Circadian rhythm disturbances in depression.

Hum Psychopharmacol.

13. Holsboer F (2000) The Corticosteroid Receptor Hypothesis of Depression.

Neuropsychopharmacol 23: 477–501.

14. Wong ML, Licinio J (2001) Research and treatment approaches to depression.Nat Rev Neurosci 2: 343–351.


dysregulation in melancholic and atypical depression: high vs low CRH/NEstates. Mol Psychiatry 7: 254–275.



17. Antonijevic IA (2006) Depressive disorders – is it time to endorse differentpathophysiologies? Psychoneuroendocrinol 31: 1–15.

18. Bale TL (2006) Stress sensitivity and the development of affective disorders.Horm Behav 50: 529–533.

19. Stone EA, Lin Y, Quartermain D (2008) A final common pathway for

depression? Progress toward a general conceptual framework. NeurosciBiobehav Rev 32: 508–524.

20. Thomson F, Craighead M (2008) Innovative approaches for the treatment of

depression: targeting the HPA axis. Neurochem Res 33: 691–707.

21. Kupfer DJ (1976) REM latency: a psychobiologic marker for primary depressivedisease. Biol Psychiatry 11: 159–174.

22. Sachar EJ (1967) Corticosteroids in depressive illness. II. A longitudinalpsychoendocrine study. Arch Gen Psychiatry 17: 554–567.

23. Sachar EJ, Mackenzie JM, Binstock WA, Mack JE (1967) Corticosteroid

responses to psychotherapy of depressions. I. Evaluations during confrontationof loss. Arch Gen Psychiatry 16: 461–470.

24. Deuschle M, Schweiger U, Weber B, Gotthardt U, Korner A, Schmider J,

Standhardt H, Lammers CH, Heuser I (1997) Diurnal activity and pulsatility ofthe hypothalamus-pituitary-adrenal system in male depressed patients and

healthy controls. J Clin Endocrinol Metab 82: 234–238.

25. Wong ML, Kling MA, Munson PJ, Listwak S, Licinio J, Prolo P, Karp B,McCutcheon IE, Geracioti TD Jr., DeBellis MD, Rice KC, Goldstein DS,

Veldhuis JD, Chrousos GP, Oldfield EH, McCann SM, Gold PW (2000)Pronounced and sustained central hypernoradrenergic function in major

depression with melancholic features: relation to hypercortisolism and

corticotropin-releasing hormone. Proc Natl Acad Sci U S A 97: 325–330.

26. Keller J, Flores B, Gomez RG, Solvason HB, Kenna H, Williams GH,

Schatzberg AF (2006) Cortisol circadian rhythm alterations in psychotic major

depression. Biol Psychiatry 60: 275–281.

27. Sachar EJ, Hellman L, Roffwarg HP, Halpern FS, Fukushima DK,

Gallagher TF (1973) Disrupted 24-Hour Patterns of Cortisol Secretion in

Psychotic Depression. Arch Gen Psychiatry 28: 19–24.

28. Angst J, Gamma A, Sellaro R, Zhang H, Merikangas K (2002) Toward

validation of atypical depression in the community: results of the Zurich cohort

study. J Affect Disord 72: 125–138.

29. Hasler G, Pine DS, Gamma A, Milos G, Ajdacic V, Eich D, Rossler W, Angst J(2004) The associations between psychopathology and being overweight: a 20-

year prospective study. Psychol Med 34: 1047–1057.

30. Touma C, Bunck M, Glasl L, Nussbaumer M, Palme R, Stein H,

Wolferstatter M, Zeh R, Zimbelmann M, Holsboer F, Landgraf R (2008) Miceselected for high versus low stress reactivity: a new animal model for affective

disorders. Psychoneuroendocrinol 33: 839–862.

31. Touma C, Palme R, Sachser N (2004) Analyzing corticosterone metabolites in

fecal samples of mice: a noninvasive technique to monitor stress hormones.

Horm Behav 45: 10–22.

32. Touma C, Sachser N, Mostl E, Palme R (2003) Effects of sex and time of day on

metabolism and excretion of corticosterone in urine and feces of mice. GenComp Endocrinol 130: 267–278.

33. Touma C, Palme R (2005) Measuring fecal glucocorticoid metabolites inmammals and birds: the importance of validation. Ann N Y Acad Sci 1046:

54–74.

34. Fenzl T, Romanowski CP, Flachskamm C, Honsberg K, Boll E, Hoehne A,

Kimura M (2007) Fully automated sleep deprivation in mice as a tool in sleep

research. J Neurosci Methods 166: 229–235.

35. Siegel S, Castellan NJ (1988) Nonparametric Statistics for the Behavioral

Sciences. New York: McGraw-Hill Book Company.

36. Rice WR (1989) Analyzing Tables of Statistical Tests. Evolution 43: 223–225.

37. Quitkin FM, Rabkin JG, Stewart JW, McGrath PJ, Harrison W, Davies M,Goetz R, Puig-Antich J (1985) Sleep of atypical depressives. J Affect Disord 8:

61–67.

38. Kalsbeek A, van Heerikhuize JJ, Wortel J, Buijs RM (1996) A diurnal rhythm of

stimulatory input to the hypothalamo-pituitary-adrenal system as revealed bytimed intrahypothalamic administration of the vasopressin V1 antagonist.

J Neurosci 16: 5555–5565.

39. Spiga F, Harrison LR, Wood SA, Atkinson HC, MacSweeney CP, Thomson F,

Craighead M, Grassie M, Lightman SL (2007) Effect of the glucocorticoid

receptor antagonist Org 34850 on basal and stress-induced corticosteronesecretion. J Neuroendocrinol 19: 891–900.

40. Lepschy M, Touma C, Hruby R, Palme R (2007) Non-invasive measurement ofadrenocortical activity in male and female rats. Lab Anim 41: 372–387.

41. Halberg F, Albrecht PG, Bittner JJ (1959) Corticosterone rhythm of mouseadrenal in relation to serum corticosterone and sampling. Am J Physiol 197:

1083–1085.

42. Dalm S, Enthoven L, Meijer OC, van der Mark MH, Karssen AM, de Kloet ER,

Oitzl MS (2005) Age-related changes in hypothalamic-pituitary-adrenal axis

activity of male C57BL/6J mice. Neuroendocrinol 81: 372–380.

43. Malisch JL, Breuner CW, Gomes FR, Chappell MA, Garland T Jr (2008)

Circadian pattern of total and free corticosterone concentrations, corticosteroid-binding globulin, and physical activity in mice selectively bred for high voluntary

wheel-running behavior. Gen Comp Endocrinol 156: 210–217.

44. Dallmann R, Touma C, Palme R, Albrecht U, Steinlechner S (2006) Impaired

daily glucocorticoid rhythm in Per1 ( Brd ) mice. J Comp Physiol A Neuroethol

Sens Neural Behav Physiol 192: 769–775.

45. Voigtlander T, Unterberger U, Touma C, Palme R, Polster B,

Strohschneider M, Dorner S, Budka H (2006) Prominent corticosteroiddisturbance in experimental prion disease. Eur J Neurosci 23: 2723–2730.

46. Hartmann A, Veldhuis JD, Deuschle M, Standhardt H, Heuser I (1997) Twenty-Four Hour Cortisol Release Profiles in Patients with Alzheimer’s and Parkinson’s

Disease Compared to Normal Controls: Ultradian Secretory Pulsatility andDiurnal Variation. Neurobiol Aging 18: 285–289.

47. Deuschle M, Gotthardt U, Schweiger U, Weber B, Korner A, Schmider J,Standhardt H, Lammers C-H, Heuser I (1997) With Aging in Humans the

Activity of the Hypothalamus-Pituitary-Adrenal System Increases and its

Diurnal Amplitude Flattens. Life Sci 61: 2239–2246.

48. Magri F, Locatelli M, Balza G, Molla G, Cuzzoni G, Fioravanti M, Solerte SB,

Ferrari E (1997) Changes in Endocrine Circadian Rhythms as Markers ofPhysiological and Pathological Brain Aging. Chronobiol Internat 14: 385–396.

49. Sapolsky RM (2000) Glucocorticoids and Hippocampal Atrophy in Neuropsy-chiatric Disorders. Arch Gen Psychiatry 57: 925–935.

50. Swaab DF, Bao AM, Lucassen PJ (2005) The stress system in the human brain indepression and neurodegeneration. Ageing Res Rev 4: 141–194.

51. Chang FC, Opp MR (2001) Corticotropin-releasing hormone (CRH) as a

regulator of waking. Neurosci Biobehav Rev 25: 445–453.



52. Benca RM, Obermeyer WH, Thisted RA, Gillin JC (1992) Sleep and psychiatric

disorders. A meta-analysis. Arch Gen Psychiatry 49: 651–668.53. Armitage R, Hoffmann R (1997) Sleep electrophysiology of major depressive

disorders. Curr Rev Mood Anx Disord 1: 139–151.

54. Siegel JM (2005) Clues to the functions of mammalian sleep. Nature 437:1264–1271.

55. Rechtschaffen A, Kales A (1968) A Manual of Standardized Terminology,Techniques and Scoring System for Sleep Stages of Human Subjects.

Washington: U.S. Government Printing Office.

56. Kleitman N (1963) Sleep and Wakefulness. Chicago: University of ChicagoPress.

57. Kupfer DJ (1995) Sleep research in depressive illness: clinical implications–atasting menu. Biol Psychiatry 38: 391–403.

58. Reynolds CF, Kupfer DJ (1987) Sleep Research in Affective Illness: State of theArt circa 1987. Sleep 10: 199–215.

59. Buysse DJ, Frank E, Lowe KK, Cherry CR, Kupfer DJ (1997) Electroenceph-

alographic sleep correlates of episode and vulnerability to recurrence indepression. Biol Psychiatry 41: 406–418.

60. Steiger A, Guldner J, Colla-Muller M, Friess E, Sonntag A, Schier T (1994)Growth hormone-releasing hormone (GHRH)-induced effects on sleep EEG

and nocturnal secretion of growth hormone, cortisol and ACTH in patients with

major depression. J Psychiatr Res 28: 225–238.61. Shaffery J, Hoffmann R, Armitage R (2003) The Neurobiology of Depression:

Perspectives from Animal and Human Sleep Studies. The Neuroscientist 9:82–98.

62. Yassouridis A, Steiger A, Klinger A, Fahrmeir L (1999) Modelling and exploringhuman sleep with event history analysis. Journal of Sleep Research 8: 25–36.

63. Chang FC, Opp MR (1998) Blockade of corticotropin-releasing hormone

receptors reduces spontaneous waking in the rat. Am J Physiol 275: R793–R802.64. Modell S, Lauer CJ (2007) Rapid eye movement (REM) sleep: an

endophenotype for depression. Curr Psychiatry Rep 9: 480–485.65. Kimura M, Mueller-Preuss P, Wiesner E, Lu A, Deussing JM (2006) Site-specific

CRH overexpression alters sleep in transgenic mice. J Sleep Res 15 (Suppl. 1):

101.66. Lu A, Steiner MA, Whittle N, Vogl AM, Walser SM, Ableitner M, Refojo D,

Ekker M, Rubenstein JL, Stalla GK, Singewald N, Holsboer F, Wotjak CT,Wurst W, Deussing JM (2008) Conditional mouse mutants highlight mechanisms

of corticotropin-releasing hormone effects on stress-coping behavior. Mol

Psychiatry.67. Gonzalez MM, Valatx JL (1997) Effect of intracerebroventricular administration

of alpha-helical CRH (9–41) on the sleep/waking cycle in rats under normal

conditions or after subjection to an acute stressful stimulus. J Sleep Res 6:164–170.

68. Gonzalez MM, Valatx JL (1998) Involvement of stress in the sleep reboundmechanism induced by sleep deprivation in the rat: use of alpha-helical CRH (9–

41). Behav Pharmacol 9: 655–662.

69. Cheeta S, Ruigt G, van Proosdij J, Willner P (1997) Changes in sleeparchitecture following chronic mild stress. Biol Psychiatry 41: 419–427.

70. El Yacoubi M, Bouali S, Popa D, Naudon L, Leroux-Nicollet I, Hamon M,Costentin J, Adrien J, Vaugeois J-M (2003) Behavioral, Neurochemical, and

Electophysiological Characterization of a Genetic Mouse Model of Depression.PNAS 100: 6227–6232.

71. Popa D, El Yacoubi M, Vaugeois JM, Hamon M, Adrien J (2005) Homeostatic

Regulation of Sleep in a Genetic Model of Depression in the Mouse: Effects ofMuscarinic and 5-HT1A Receptor Activation. Neuropsychopharmacol 31:

1637–1646.72. Borbely AA, Tobler I, Hanagasioglu M (1984) Effect of sleep deprivation on

sleep and EEG power spectra in the rat. Behav Brain Res 14: 171–182.

73. Finelli LA, Baumann H, Borbhly AA, Achermann P (2000) Dual electroen-cephalogram markers of human sleep homeostasis: correlation between theta

activity in waking and slow-wave activity in sleep. Neuroscience 101: 523–529.74. Borbely AA (1982) A two process model of sleep regulation. Hum Neurobiol 1:

195–204.75. Holsboer F, von Bardeleben U, Steiger A (1988) Effects of intravenous

corticotropin-releasing hormone upon sleep-related growth hormone surge and

sleep EEG in man. Neuroendocrinol 48: 32–38.76. Armitage R, Hoffmann R, Trivedi M, Rush AJ (2000) Slow-wave activity in

NREM sleep: sex and age effects in depressed outpatients and healthy controls.Psychiatry Res 95: 201–213.

77. Heim C, Newport DJ, Mletzko T, Miller AH, Nemeroff CB (2008) The link

between childhood trauma and depression: Insights from HPA axis studies inhumans. Psychoneuroendocrinol 33: 693–710.

78. Holsboer F (2008) How can we realize the promise of personalizedantidepressant medicines? Nat Rev Neurosci 9: 638–646.



Chronic Mild Stress (CMS) in Mice: Of Anhedonia,‘Anomalous Anxiolysis’ and ActivityMartin C. Schweizer1*, Markus S. H. Henniger1, Inge Sillaber1,2

1 Affectis Pharmaceuticals AG, Martinsried, Germany, 2 Max-Planck-Institute of Psychiatry, Munich, Germany

Abstract

Background: In a substantial proportion of depressed patients, stressful life events play a role in triggering the evolution ofthe illness. Exposure to stress has effects on different levels in laboratory animals as well and for the rat it has been shownthat chronic mild stress (CMS) can cause antidepressant-reversible depressive-like effects. The adoption of the model to themouse seems to be problematic, depending on the strain used and behavioural endpoint defined. Our aim was to evaluatethe applicability of CMS to mice in order to induce behavioural alterations suggested to reflect depression-like symptoms.

Methodology/Principal Findings: A weekly CMS protocol was applied to male mice of different mouse strains (D2Ola, BL/6Jand BL/6N) and its impact on stress-sensitive behavioural measures (anhedonia-, anxiety- and depression-related parameters)and body weight was assessed. Overnight illumination as commonly used stressor in CMS protocols was particularlyinvestigated in terms of its effect on general activity and subsequently derived saccharin intake. CMS application yieldedstrain-dependent behavioural and physiological responses including ‘paradox’ anxiolytic-like effects. Overnight illuminationwas found to be sufficient to mimic anhedonic-like behaviour in BL/6J mice when being applied as sole stressor.

Conclusions/Significance: The CMS procedure induced some behavioural changes that are compatible with the commonexpectations, i.e. ‘anhedonic’ behaviour, but in parallel behavioural alterations were observed which would be described as‘anomalous’ (e.g. decreased anxiety). The results suggest that a shift in the pattern of circadian activity has a particular highimpact on the anhedonic profile. Changes in activity in response to novelty seem to drive the ‘anomalous’ behaviouralalterations as well.

Citation: Schweizer MC, Henniger MSH, Sillaber I (2009) Chronic Mild Stress (CMS) in Mice: Of Anhedonia, ‘Anomalous Anxiolysis’ and Activity. PLoS ONE 4(1):e4326. doi:10.1371/journal.pone.0004326


Received September 15, 2008; Accepted November 14, 2008; Published January 29, 2009

Copyright: � 2009 Schweizer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The research was supported in part by the Bundesministerium fur Bildung und Forschung (FKZ 313685). There was no role of any sponsors in thedesign, data collection, analysis, interpretation, etc. of the study.



Introduction

Animal models are important research tools in psychiatry and

should mimic some of the human conditions of interest. According

to DSM-IV [1], major depression is characterised by either

depressed mood or anhedonia, in combination with four

additional symptoms related to weight changes, sleep disturbances,

psychomotor agitation or retardation, fatigue, feelings of worth-

lessness or guilt, diminished cognitive functioning, or recurrent

thoughts of death. The presence of some of these symptoms can be

defined operationally (e.g., loss of appetite and weight, sleep

disturbances, cognitive and psychomotor changes), and thus can

be assessed in laboratory animals.

Stressful experiences have been reported to favour the develop-

ment of depression in humans [e.g. 2,3]. Therefore, in order to

provoke depressive-like behavioural changes, some animal models for

this phenotype are generated by exposing them to stressful situations

[4–8]. In rats, application of chronic mild stress (CMS) procedures

resulted in a variety of behavioural, neurochemical, neuroendocrine

and neuroimmune alterations resembling some of the dysfunctions

observed in human depression [9–16]. Therefore, the CMS model,

developed by Willner and colleagues [4], has attracted a lot of interest

due to its potential of combining several validity criteria requested for

an animal model of depression [16,17]. In terms of symptoms evoked

by CMS, the induction of anhedonia was the primary focus in this

model [4]. Anhedonia, a core symptom of depression, was modelled

by inducing a decrease in responsiveness to rewards reflected by a

reduced consumption and/or preference of sweetened solutions. In a

more recent review, Willner [17] summarises results of positive

reproduction of CMS-induced anhedonia as well as ‘anomalous’

findings. In this context the attributes ‘anomalous’ or ‘paradox’ refer

to findings that include CMS-induced anxiolysis and hyperlocomo-

tion and apparently are contrary to ‘classic’ comorbidities of

depression-related behaviour such as increased anxiety and reduced

locomotor activity [17,18]. Most studies were performed in rats but

the few that used mice generally point towards the applicability of a

non-standardised CMS procedure to induce anhedonic behaviour in

mice as well. One reason for the adoption of this approach to mice

was the introduction of genetic mouse models as tools in psychiatric

disorder research. Using mice with specific genetic modifications in

suspected vulnerability genes in combination with CMS as

environmental factor will allow scrutinising the role of both, the

candidate gene and stress in a genetically predisposed animal, as risk

factors for depression.

Studies published on the effects of CMS on mouse behaviour

agreed on the importance of the strain used and showed that some


strains are responsive to the CMS procedure, a finding which

strongly depends on the respective endpoint defined [19–24]. The

co-occurrence of CMS-induced alterations in palatable liquid

consumption and other behavioural and physiological parameters,

suggested to reflect depressive-like behaviour, has rarely been

studied in mice. Pothion and colleagues [21] investigated three

different mouse strains in terms of sucrose consumption, body

weight changes, coat state, spontaneous alternations in the Y-maze

and spatial learning in the Morris water maze test. Only in one

strain (CBA/H) a decrease in sucrose consumption was found

along with changes in physiological or behavioural parameters.

Therefore, a CMS-induced reduction of sucrose consumption was

not predictive of the occurrence of any other parameter assessed in

the study. Applying a different chronic stress procedure in BL/6N

mice, comprising more severe stressors presented in a more

sequential than unpredictable manner, Strekalova and colleagues

[25] described an association of a stress-provoked decrease in

sucrose preference and behavioural measures that were restricted

to increased passivity in the forced swim test (FST) as well as

decreased novel object exploration. An increase in anxiety-related

behaviour and locomotor disturbances were found to be induced

by the stress procedure independent of the induction of

‘‘anhedonic’’ behaviour [25]. In other studies using the CMS

procedure, an increase in immobility in the FST [26–28],

decreased grooming [23], altered performance in cognitive tasks

[29] as well as changes in anxiety-like behaviour [20,24] were

observed. These behavioural measures were considered to reflect

additional depression-related symptoms [17]. Unfortunately con-

sumption of sweetened solutions was not assessed in all these

studies.

A CMS model with a decrease in intake and preference of

sweetened solutions as central readout presents specific advantag-

es. Contrary to operant behaviour it can be implemented without

cost-intensive apparatus, it is non-invasive (in contrast to e.g.

intracranial self-stimulation), and the progressive evolution of a

depression-related symptom as well as the time course of

antidepressant-induced resolution of the symptom can be traced

by repeated measurements. The latter is rather difficult for a

variety of test paradigms used to assess, for example, anxiety-like

behaviour. As stated by Anisman and Matheson [5], an

appropriate model of depression requires that multiple behaviour-

al tests be employed to approximate the range of symptoms that

characterise depressive illness. Therefore, the aim of our studies

was to further investigate the feasibility to establish a CMS model

which includes the measurement of intake and preference of a

sweetened liquid and addresses additional indicators of depressive-

like behaviour. We used a stress-procedure which included most of

the commonly used ‘‘mild’’ stressors [30] but was devoid of food or

water deprivation. Measurements were performed 2–3 times per

week with the aim to possibly gain information on a direct effect of

the precedent stressor on ‘‘hedonic’’ behaviour. CMS effects on

intake of a 0.2% saccharin solution and body weight gain were

assessed in different mouse strains as the vulnerability for stress-

induced changes is supposed to be genetically determined. In

selected mouse strains, we investigated the influence of CMS on

anxiety-related behaviour and on passive behaviour in the FST.

The fluctuations in saccharin intake observed in the first

experiments showed a tendency of association to the precedent

stressor which according to our interpretation had some impact on

nocturnal activity of the mice. Therefore, an independent

experiment aimed at detecting the influence of overnight

illumination as particular stressor on subsequent saccharin intake

and activity in parallel.


Animals and housing conditionsMale mice from seven strains (Balb/cOla, C57BL/6JOla,

C57BL/6N, DBA/2Ola, DBA/2JIco, FVB/N, NMRI) - referred

to as Balb/c, BL/6J, BL/6N, D2Ola, D2JIco, FVB and NMRI

respectively - were purchased from Harlan-Winkelmann GmbH

(Borchen, Germany; Balb/c, BL/6J, D2Ola, FVB and NMRI)

and Charles River GmbH (Sulzfeld, Germany; BL/6N and

D2JIco). Upon arrival, the animals were singly housed in

Macrolon Type II cages under standard laboratory conditions

(temperature 2161uC, rel. humidity 40–60%, 12h:12h light/dark

cycle, lights on 6 A.M.), had free access to food and water and

were allowed to habituate to the novel environment for at least

2 weeks. Afterwards, the basal consummatory behaviour for water

and either saccharin or sucrose (intake and preference) was

determined. Based on these parameters and on body weight, mice

were matched and assigned to stress and control groups. Animal

experiments were performed in accordance with the NIH Guide

for the Care and Use of Mammals in Neuroscience and

Behavioural Research and the Guide for the Care and Use of

Laboratory Animals of the Government of Bavaria, Germany.

CMS ParadigmThe CMS procedure followed a fixed weekly schedule of

commonly used mild stressors such as repeated cold stress (4uC),

space reduction in the homecage, changed cages within CMS

group, cage tilt, empty cage, intermittent air puff, wet bedding,

white noise, overnight illumination and social interaction with

other animals of the CMS group. The particular context (stressor

applied during preceding dark phase) of the 2 hrs liquid

consumption measurement intervals involved overnight illumina-

tion (Sunday–Monday), wet cage and cage tilt (Tuesday–

Wednesday and Thursday–Friday, long-term CMS in BL6/J

and D2Ola mice) and changed cages (Wednesday–Thursday,

shorter CMS in BL6/J, BL6/N and the other strains), respectively.

For details see Table 1.

Behavioural TestsConsummatory Behaviour. Extensive preliminary tests for

preference of cage side, saccharin preference over water, saccharin

concentration and one- versus two-bottle paradigm conducted

before the CMS period in the long-term CMS paradigm resulted

in the following measurement protocol. Liquid intake was

determined 3 times a week (Monday, Wednesday, Friday) in a

two-bottle paradigm by weighing the bottles before and after the

first 2 hours of the dark phase ( = measurement interval of liquid

consumption) on the basis of D’Aquila’s studies [31]. Sweet

solutions were offered on the preferred right cage side, as a shifting

from the preferred solution at the preferred cage side to water at

the other side might more closely model anhedonic behaviour.

With the exception of the experiment in BL/6N mice, a 0.2%

saccharin solution was presented as palatable liquid. Using BL/6N

mice and applying a different stress procedure Strekalova et al.

[25] reported a decrease in sucrose preference coupled with other

behavioural alterations. In order to test whether our CMS

procedure induces a profile in BL/6N mice comparable to the

reported one we stayed with the presentation of sucrose and

changed the stress protocol only.

In the other experiments saccharin was chosen to avoid a caloric

impact of the sweetened liquid consumption on the CMS effects.

Since saccharin presentation 3 times a week might have worn

down the hedonic value of saccharin also in control mice, in the

CMS-Anhedonia-Activity


shorter experiments sweet solutions were presented only twice a

week on Monday and Thursday.

Dark/Light Box (DaLi). Anxiety-related behaviour was

assessed in a DaLi (dark compartment 15620625 cm, light

compartment 30620625 cm, connected by a 4 cm long tunnel).

The light compartment was illuminated with 700 lx (for D2Ola

and BL/6J mice) and 50 lx (for BL/6N animals) cold light,

whereas in the dark compartment the illumination level was 5 lx.

Animals were placed in the dark compartment and the time spent

in, the latency to first entry (all four paws) and the number of

entries into each compartment was recorded for 5 min using the

ANY-maze software (Stoelting Co., Wood Dale, IL).

Forced Swim Test (FST). Each animal was placed into a

beaker (diameter 12 cm, height 24 cm) filled with water

(temperature 25226uC) to a height of 12 cm for a test period of

5 min. The parameters floating (immobility with only small

movements to keep balance), swimming and struggling (vigorous

attempts to escape) were scored throughout the 5 min test period

by a trained observer blind to the treatment and recorded using

the ANY-maze software.

Modified Hole Board Test (mHb). The mHb test (for

details see [32]) was performed at 6 P.M. at 60 lx (long-term CMS)

and 30 lx (BL/6N) respectively. The apparatus consisted of a dark

grey PVC box (100650 cm, 50 cm height) containing a board

(7062060.5 cm) in the centre which was equipped with 12

cylinders (‘‘holes’’, 2 cm height63 cm diameter). Behavioural

parameters such as time spent on board (with all four paws),

rearing (standing in an upright position on the hindpaws) and

locomotor activity (total distance travelled) were recorded during a

5 min session using the ANY-maze software.

Experimental DesignAssessment of CMS effects on consummatory

behaviour. a) Long-term CMS in BL/6J and D2Ola mice:

Animals (n = 24 per strain) underwent preliminary tests to develop

the liquid intake measurement protocol. Subsequently, at an age of

36 weeks, the CMS regimen was applied for 13 weeks (for each

strain control and stress group n = 12). During the CMS period the

saccharin consumption was measured 3 times a week (Monday,

Wednesday and Friday) for 10 weeks.

b) Short-term CMS in other mouse strains: An independent

batch of BL/6J animals (aged 12 weeks at the beginning of CMS

application, n = 10/group) was tested to assess reproducibility of

CMS-effects on consummatory behaviour. BL/6N mice (aged

17 weeks at the beginning of the stress period, 9 control and 20

CMS animals) were offered a 2% sucrose solution for 4 weeks of

CMS.

Consummatory behaviour was accordingly assessed also in

additional mouse strains during 3–4 weeks of CMS. The animals

were aged 12 weeks (Balb/c, n = 10/group), 7 weeks (D2JIco,

n = 24/group) and 21 weeks (FVB and NMRI, n = 8/group),

respectively, at the beginning of the stress period.

Assessment of CMS effects on anxiety- and depression-

related behaviour. a) Long-term CMS in BL/6J and D2Ola

mice: The long-term (13 weeks) stressed BL/6J and D2Ola mice

were tested in the DaLi after 11 weeks and the FST after 12 weeks

of CMS exposure. The testing times were 6 P.M. and 8 A.M.,

respectively. The week after, mice were tested in the mHb (6 P.M.,

60 lx).

b) Short-term CMS in BL/6J and BL/6N mice: The short-term

(4 weeks) stressed BL/6J mice were tested for anxiety-related

behaviour in the DaLi after 2 weeks and for depression-related

behaviour in the FST after 4 weeks of CMS. The FST was

conducted from 9 A.M. to 12 P.M. under standard laboratory

conditions. Before being tested for CMS effects on sucrose

consumption as described above, the BL/6N mice were

characterised in their basal ( = non-stressed) behaviour in several

paradigms (DaLi, elevated plus maze, open field, resident-intruder

test), then following the CMS procedure for 4 weeks were re-tested

in the above mentioned paradigms (plus FST) and subjected to

additional 3 weeks of CMS. Stress effects on behaviour were

finally assessed in the mHb. All behavioural tests were conducted

during the first 2 hours of the dark phase (corresponding to the

consumption measurement interval) either at reduced illumination

(mHb: 30 lx) or under red light conditions (all other tests).

Effects of overnight illumination as sole stressor in BL/6J

mice. For 4 weeks, BL/6J mice (40 stressed and 36 control

animals aged 7 weeks at the beginning of the stress period) were

exposed to overnight illumination (light stress, LS) twice a week

(Sunday and Wednesday) and the saccharin consumption and

homecage activity was measured during the following dark period

(Monday/Tuesday and Thursday/Friday, respectively). No

further stressors were applied. To assess their general activity

pattern, mice were monitored in their homecages in side-view

using small CCD cameras and the ANY-maze software.

Parameters such as locomotor activity, rearing and climbing

behaviour were determined. Data were analysed for the 2 hrs

consumption measurement interval as used in the CMS

experiments and for the subsequent 22 hours to yield a 24 hrs

consumption/activity profile.

Table 1. Weekly CMS schedule

LIGHT PHASE DARK PHASE

First half Second half First 2 hrs Remaining 10 hrs

Mon Repeated Cold Stress (2630 min) Cold Stress (30 min) 2 hrs liquid intake measurement* Homecage Space Reduction

Tue Changed Room Air Puff (363 intermittent) Wet Cage Wet Cage

Wed Wet Cage Social Interaction Foreign Cage Foreign Cage

Thu Foreign Cage Social Interaction 2 hrs liquid intake measurement* Cage Tilt

Fri Empty Cage Changed Room White Noise&Strobe White Noise&Strobe

Sat White Noise Pause/Changed Room Pause/Changed Room Pause/Changed Room

Sun Pause/Changed Room Pause/Changed Room Overnight Illumination Overnight Illumination

*Exemplified for short-term CMS experiments. During long-term CMS liquid intake was assessed on Mon, Wed and Fri in the same interval.doi:10.1371/journal.pone.0004326.t001



Statistical AnalysisData were analysed using STATISTICA Software 6.0 (Statsoft

Inc., Tulsa, US). A two-sided Student’s t-test was applied for

comparison of two experimental groups. For time series analysis

(consummatory behaviour and locomotor activity) repeated

measures ANOVA with Greenhouse-Geisser correction for non-

sphericity was used. For traceability reasons uncorrected degrees

of freedom plus the respective correction factor e and corrected F

values are shown. When repeated measures ANOVA revealed a

significant interaction effect of the factors time and group, data

were further analysed with pairwise comparisons for each time

point (Student’s t-test). The inter-strain comparison of saccharin

intake (see supplementary material) was analysed using one-way

ANOVA followed by Tukey post-hoc testing. For the analysis of

body weight gain averages were calculated for the CMS period

and compared using a two-sided Student’s t-test. Effects were

considered significant when p,0.05.

Results

CMS effects on consummatory behaviour in differentmouse strains

For saccharin intake of long-term CMS exposed D2Ola mice

repeated measurement ANOVA yielded no significant effect for

the factor ‘‘group’’ ( = CMS). After beginning of the CMS

period, at the times marked in Figure 1A, differences in

saccharin consumption in the experimental groups could be

detected (interaction group6time: F31,682 = 9.29; p,0.01;

e= 0.359), albeit stating in some measurements an elevated

intake in CMS mice compared to the control group. CMS had

no effect (factor group) on the preference for saccharin in D2Ola

mice (Figure 1A), no significant interaction resulted from the

factors group and time.

In BL/6J mice, statistical analysis revealed a reduced saccharin

intake (factor group: F1,22 = 12.30; p,0.01). This difference also

appeared in the interaction of the factors group6time

(F30,660 = 14.27; p,0.01; e= 0.366), whose details are illustrated

in Figure 1A. CMS led to a decreased saccharin preference (factor

group: F1,22 = 4.92; p,0.05; interaction group6time

F30,660 = 2.57; p,0.01; e= 0.298) of BL/6J animals in the long-

term CMS paradigm (see Figure 1A). After four weeks CMS

exposure, these findings could be replicated for an independent

batch of BL6/J mice (data not shown) regarding their saccharin

intake (factor group: F1,18 = 24.98; p,0.01; interaction group6time: F5,90 = 4.33; p,0.01; e= 0.584) but not in respect of their

saccharin preference.

Further, as illustrated in Figure S1, the CMS regimen affected

saccharin intake of Balb/c mice (factor group: F1,18 = 6.03;

p,0.05; interaction group6time: n.s.), FVB animals (factor group:

F1,14 = 14.09; p,0.01; interaction group6time: F8,112 = 5.63;

p,0.01; e= 0.536), NMRI mice (factor group: F1,14 = 6.75;

p,0.05; interaction group6time: n.s.) and D2JIco mice (factor

group: F1,45 = 12.90; p,0.01; interaction group6time: n.s.). Inter-

strain comparison for basal saccharin consumption and percental

change induced by CMS is provided in Table S1.

In BL/6N mice, repeated measurements ANOVA yielded

no CMS effect on sucrose intake in terms of factor group but an

interaction effect of factors group6time (F7,245 = 5.89; p,0.01;

e= 0.618). Sucrose preference was significantly decreased by

CMS (factor group: F1,35 = 13.48; p,0.01; interaction factors

group6time: F7,245 = 3.32; p,0.01; e= 0.629). The sucrose

consumption of BL/6N mice is illustrated in Figure 1B.

CMS effects on anxiety- and depression-relatedbehaviour and body weight gain

BL/6J mice were selected to be further investigated as this

mouse strain is commonly involved in the generation of transgenic

mice. We were interested in the D2Ola mice as we observed in

previous studies that D2Ola compared to BL/6J responded with a

reduced inhibitory HPA axis feedback to a single stress exposure

[33]. BL/6N mice were integrated in our studies as Strekalova et

al. [25] showed an association of anhedonic and depressive-like

behaviour for a subgroup of this mouse strain after chronic stress.Dark/Light Box (DaLi). Long-term CMS application

caused D2Ola mice to spend significantly more time in the lit

compartment compared to controls (t21 = 4.55; p,0.01), whereas

stressed BL/6J mice of the same experiment did not show an

altered behaviour in the DaLi paradigm (Figure 2A). The latter

finding was in line with results obtained using the same strain in

the DaLi test after 2 weeks of CMS (Figure 2B). In BL/6N mice,

the time spent in the lit compartment was not affected by 4 weeks

of CMS pre-experience (Figure 2C).Modified Hole Board (mHb). Behaviour of the different

strains in the mHb is displayed in Figure 3. Long-term stressed

D2Ola animals spent more time on the board (t20 = 2.78; p,0.05),

while rearing behaviour and locomotor activity tended to result in

a slight increase due to CMS. BL/6J animals did not show any

behavioural changes induced by CMS in the mHb test. However,

mHb testing revealed strong effects of 7 weeks of CMS on

behaviour of BL6/N mice: Time spent on board, number of

rearings and locomotor activity were significantly increased in

stressed animals (time board: t16 = 2.91; p,0.05, rearing:

t16 = 3.23; p,0.01, locomotion: factor group F1,16 = 14.82;

p,0.01; interaction group6time: F9,144 = 1.97; p,0.05).Forced Swim Test (FST). In none of the investigated mouse

strains an increased immobility in the FST due to CMS could be

observed (p.0.05) (Figure 4).Body weight gain. As illustrated in Figure 5A, CMS reduced

body weight gain in D2Ola and BL/6J mice during the 13 weeks

CMS schedule (t21 = 2.22; p,0.05 and t22 = 4.00; p,0.01,

respectively). This result could be confirmed for mice of the latter

strain in the shorter CMS experiment (t17 = 2.93; p,0.01), whereas

CMS had no effect on body weight gain in BL/6N animals

(Figure 5B).

Effects of overnight illumination as sole stressor in BL/6Jmice

Consummatory behaviour. Overnight illumination as sole

stressor was able to considerably decrease saccharin intake of BL/6J

animals during the 2 hrs measurement interval (factor group:

F1,74 = 185.96; p,0.01; interaction group6time: F7,518 = 6.22;

p,0.01; e= 0.715), an effect that almost retained significance

throughout the 24 hrs measurement interval (factor group:

F1,74 = 3.60; p = 0.06; interaction group6time: n.s.), for details see

Figure 6A.Homecage activity. BL/6J animals exposed to overnight

illumination as sole stressor showed reduced homecage activity in

the 24 hrs interval (factor group: F1,62 = 6.73; p,0.05),

whereupon localisation of differences revealed a distinctly

decreased activity during the first 3 hours of the dark phase

(interaction group6time: F22,1364 = 12.25; p,0.01; e= 0.149), i.e.

comprising the time span during which both 2 hrs saccharin

measurement took place and control animals peaked in homecage

activity (for details see Figure 6B). Overnight illumination delayed

(and decreased) peak activity in the presence of the saccharin

solution in LS animals by around 2–3 hours. Two-way ANOVA

of homecage activity during dark and light phase yielded an



interaction effect of light stress and daytime (F1,62 = 10.86;

p,0.01), whose details are shown in Figure 6B. Here, overnight

illumination could be shown to reduce homecage activity during

the dark phase (p,0.01) while light phase activity remained

unchanged. Correlation of homecage activity of both control and

CMS animals during the 2 hrs consumption measurement

interval with 2 hrs saccharin intake revealed Pearson’s r = 0.64

(p,0.01), see Figure 6B.

Figure 1. CMS effects on palatable liquid consumption per 2 hrs in D2Ola, BL/6J and BL/6N mice. (A) Effects of CMS on saccharin intakeand preference over a period of 10 weeks in D2Ola and BL/6J mice (measurement 3x/week). (B) CMS effects on sucrose intake and preference over aperiod of 4 weeks in BL/6N mice (measurement 2x/week). Saccharin concentration: 0.2%, sucrose concentration 2%. White circles: control group,black circles: CMS group. First 4 data points of each graph represent basal consumption. Data represent mean6SEM, n = 9–20/group. * p,0.05,** p,0.01 pairwise between-group comparisons (Student’s t-test).doi:10.1371/journal.pone.0004326.g001



Figure 2. CMS effects on time spent in the lit compartment of the DaLi. (A) In D2Ola mice long-lasting CMS exposure (11 weeks) led tobehavioural changes reflected by an increase in time spent in the lit compartment (700 lx) compared to the control group. In BL/6J mice with similarCMS experience no changes were observed. (B) BL/6J mice with CMS experience of 2 weeks were also unaffected in terms of time spent in the litcompartment (700 lx). (C) BL/6N mice with CMS experience of 4 weeks did not differ from control mice in the parameter time lit, when tested at amoderate illumination level of 50 lx in the lit compartment. Data represent mean + SEM, n = 9–20/group. White bars: control group, black bars: CMSgroup. ** p,0.01 (Student’s t-test).doi:10.1371/journal.pone.0004326.g002

Figure 3. CMS effects on behaviour in the mHb. In D2Ola mice 13 weeks of CMS led to an increase in time spent on the board, whereas BL/6Janimals of the same experiment did not show altered behaviour (time board, number rearings and distance travelled). Mice of the BL/6N strain(7 weeks of CMS experience) showed significant increases in the parameters time board, number of rearings and an increased locomotor activity.Illumination conditions: 60 lx (D2Ola, BL/6J) and 30 lx (BL/6N), respectively. Data represent mean6SEM, n = 9–12/group. White circles and bars:control group, black circles and bars: CMS group. * p,0.05, ** p,0.01 (Student’s t-test).doi:10.1371/journal.pone.0004326.g003



Discussion

The data presented in this study were derived during the

development of a CMS protocol focussed on consumption of sweet

solutions as central readout measure for hedonic/motivational

behaviour in mice. We used a stress-procedure which included

most of the commonly used ‘‘mild’’ stressors [16,17,30] but was

devoid of food or water deprivation. Five out of seven mouse

strains investigated responded with a decrease in the consumption

of the sweetened solution compared to the respective control

Figure 4. CMS effects on immobility in the FST. (A) 12 weeks of exposure to CMS had no effect on immobility time of D2Ola and BL/6J mice inthe FST. (B) 4 weeks of CMS exposure had no effect on immobility time of BL/6J mice. (C) After 4 weeks of CMS exposure, BL/6N mice were subjectedto the FST at the end of week 5. No changes in immobility time compared to the control group could be observed. Time of test: 8 A.M. (A,B) and 6P.M. (C), respectively. Data represent mean + SEM, n = 9–20/group. White bars: control group, black bars: CMS group.doi:10.1371/journal.pone.0004326.g004

Figure 5. CMS effects on body weight. (A) The illustrated time courses of body weights are characterised by a flattened mean slope in CMSanimals compared to controls, thus revealing a reduced body weight gain due to CMS in both D2Ola and BL/6J mice of the long-term CMSexperiment. (B) In the shorter CMS experiments a similar reduction of body weight gain due to CMS could be observed in younger BL/6J but not inBL/6N animals. Data represent mean6SEM, n = 9–20/group. White circles: control group, black circles: CMS group.doi:10.1371/journal.pone.0004326.g005



Figure 6. Effects of Light Stress (LS) on subsequent saccharin intake and homecage activity in BL/6J mice. (A) LS resulted in a reductionof saccharin intake during the first 2 hrs interval while there was no effect on overall 24 hrs saccharin intake after LS. The dashed horizontal lines aty = 0,72 mL and y = 4,80 mL, respectively, represent basal water intake when only water is available in a two-bottle paradigm during thecorresponding measurement interval. (B) Profiling of homecage activity for a total of 24 hrs (t0 = beginning of the dark phase) revealed a distinctdecrease during the first hours of the dark phase (which comprised the 2 hrs saccharin measurement interval, hatched area) that was positivelycorrelated with the respective 2 hrs saccharin intake (Pearson’s r = 0.64, p,0.01, n = 76, i.e. 40 LS and 36 control animals). Data represent mean6SEM,n = 36–40/group. White circles and bars: control group, grey circles and bars: light stress group. * p,0.05, ** p,0.01 pairwise between-groupcomparisons (Student’s t-test).doi:10.1371/journal.pone.0004326.g006



group. As summarised by Willner [17], a variety of chronic stress

protocols has been able to yield an anhedonic phenotype in

rodents regardless of particular content or timetabling of

microstressors in the respective CMS schedule. Strain-dependent

effects of CMS on consumption of palatable solutions of mice have

been reported earlier [20,21,24].

These general findings could be confirmed in our study for

stressed BL/6J, D2Ola, D2JIco, FVB and BL/6N mice despite

differences in the designs of the experiments. However, the effects

of CMS appeared not to be enduring, since the saccharin intake of

the respective control group decreased to the level of stressed

animals after 4–6 weeks in the long-term CMS experiment.

Trying to re-establish their working CMS protocol after moving to

a new laboratory, Willner and colleagues reported similar

problems of an initially obtained anhedonic CMS effect in rats

that vanished after several weeks of stress application [16]. The

discrepancy to previous studies was explained by a diurnal

variation of CMS effects that were found to be more robust when

the sucrose test was performed during the dark period [31]. In our

study in mice, we followed the suggestion of measuring the intake

behaviour during the first hours of the dark period, therefore we

should have picked the most sensitive time period for measuring a

CMS effect on saccharin intake. In a study of Pothion and

colleagues [21] the difference between control and CMS groups of

mice in a 24 hrs sucrose test was also only observed for the first

4 weeks after onset of stress. We observed strong fluctuations in

the consummatory data in our experiments. These might be a

consequence of (1) the short measurement interval and the

consequential low intake during that time, (2) the assessment of

consummatory behaviour more than once a week and (3) a

possibly disturbing effect on the intake of control animals since

CMS mice most of the time were kept and stressed in the same

room [17].

One aim of the study was to identify potential differences on the

impact of specific stressors contained in the CMS procedure on

saccharin consumption and therefore, the protocol followed a

fixed weekly schedule rather than being designed entirely in an

unpredictable way. The preceding stressor with the largest effect

on saccharin intake in the experiments was the wet cage. Our

interpretation was that under this condition it was hard for the

animals to rest. Indeed, keeping rats in a wet cage was used to

establish an animal model of fatigue [34]. Thus, we followed the

idea of a potential impact of mild disturbance of activity/sleep

rhythm on saccharin consumption. Overnight illumination was

applied as the sole stressor and measuring the subsequent

saccharin intake was paralleled by recording homecage activity.

As shown in Figure 6A, the volume of saccharin consumption

during the first 2 hours at the beginning of the dark period of

stressed animals dropped to a level similar to basal water

consumption. In parallel, a reduction of activity was observed

that was most prominent during the time when saccharin was

presented. This suggests that the reduced saccharin intake was not

only due to a shift of general consummatory behaviour but

combined with a parallel shift and reduction in activity as observed

during the same time interval. There was a moderate positive

correlation of activity and intake during the time interval of

saccharin presentation for LS as well as control animals.

Therefore, the levels of activity and of saccharin intake are closely

interwoven and the potential influence of a stressor preceding the

saccharin measurement on activity changes should be considered.

In the rat version of the CMS model, changes in diurnal

rhythms [35] and sleep architecture [36,37] were reported.

Further, D’Aquila et al. [31] showed the diurnal variation of

CMS-induced anhedonic behaviour with its presence mainly

during the active phase (dark period) of the animals. Papp and

colleagues [38] concluded that the procedure causes a generalised

disorganisation of internal rhythms which are postulated to play an

important role in the pathophysiology of depression [39]. In mice,

strain-dependent differences in locomotor activity rhythm and its

changes due to daylight reversal [40], as well as a relation of sleep

changes due to mild stressors - like environmental novelty - with

trait anxiety [41] have been described.

Taken together, in addition to strain-dependent intake of

sweetened liquids [21,42,43] and stress effects on the reward

system [42,44], the sensitivity to changes in activity/sleep due to

the CMS procedure contributes to the final decrease in

consumption behaviour. This could be of specific relevance for

those studies that apply mild stressors and determine intake or

preference of sweetened solutions following different stressors [21].

A consequence could be a higher variation in the anhedonic

profile as observed in our study.

Besides the measurement of saccharin intake our second focus

was to address additional indicators of anxiety- and depressive-like

behaviour. Therefore, animals of three strains of mice were

additionally characterised regarding CMS effects on anxiety-

related behaviour. Furthermore, the animals were investigated in

the FST, a test often used to evaluate antidepressant-like

properties of substances [45]. First, BL/6J and D2Ola mice were

tested in above mentioned paradigms after CMS experience of

more than 10 weeks. Second, a different batch of BL/6J mice was

investigated after 2 weeks of CMS experience when the decrease

in saccharin intake became apparent. Third, experiments with

CMS-experienced BL/6N mice also addressed the influence of

illumination during behavioural testing [18]. Testing BL/6J and

D2Ola mice after long-term CMS experience revealed that BL/6J

mice remained unaffected by CMS, whereas D2Ola mice of the

CMS group showed a decrease in anxiety-related behaviour. In

the second group of BL/6J mice, after 2 weeks of CMS, again no

difference in anxiety-like behaviour was seen compared to the

control group. The anxiolytic-like effect of CMS in D2Ola mice is

commonly appraised ‘anomalous’, yet has also been found in other

chronic stress paradigms involving rats, other mouse strains and

different anxiety tests [23,46–52]. Generally, these findings are

interpreted as either being due to blunted emotionality [52] or

caused by methodological differences [48] of CMS procedures that

yield a ‘classic’ anxiogenic stress response. Nevertheless, the

resilience of BL/6J mice against disturbing CMS effects on

anxiety-related behaviour as shown by Mineur et al. [24] could be

confirmed in our study.

According to Strekalova et al. [18] bright and even moderate

illumination conditions (.5 lx) can confound anxiety- and

depression-related behavioural readout in chronically stressed

BL/6N mice in common test paradigms by eliciting hyperlocomo-

tion. Furthermore, chronic stress in general is supposed to exert

stimulating effects on locomotor activity [53] that might mask

other stress effects in anxiety- as well as in depression-related test

paradigms [18]. For this reason, in our study involving BL/6N

mice, illumination levels of the test apparatus were attenuated and

testing was performed at the beginning of the dark phase. Under

these conditions, no CMS-induced changes were observed in the

DaLi. In the mHb test a CMS-induced hyperlocomotion and

increased vertical exploration was paralleled by an increased time

spent on the board, thus fitting into the picture drawn by

Strekalova et al [18]. Since hyperlocomotion still appeared under

red light conditions in the open field test in BL/6N mice (data not

shown), our CMS protocol seems to exert additional effects on

reactivity to a test situation that lie beyond an increased sensitivity

to even moderate illumination conditions. Possibly in line with this



concept are the results of all conducted FSTs in our study where

no significant CMS-induced change in immobility was observed.

According to the outlined concept, an increased immobility might

have been concealed by CMS-induced hyperarousal, rather

independent of test time and illumination.

Taken together, our findings on anxiety- and depression-related

behaviour point towards a higher reactivity to a novel situation in

response to CMS. This might imply the need to put the

‘anomalous’ behavioural profile in response to our CMS protocol

in perspective - although this conclusion is highly speculative in

respect to our limited data. Taking into account the above

described inefficiency of CMS to alter anxiety-related behaviour in

BL/6J mice, the presented data suggest that the occurrence of

anhedonic behaviour and changes in behaviour assessed in tests

for anxiety- or depression-related behaviour were uncoupled from

each other.

Strain effects of CMS are further diversified by different

responses of body weight gain. Whereas BL/6J and D2Ola mice

showed a reduced body weight gain due to CMS, BL/6N

remained entirely uninfluenced by CMS in this parameter.

Apparently stressor intensity was high enough to yield an

attenuation of body weight gain in D2Ola and BL/6J mice, while

BL/6N mice seem to require stronger stressor application to show

the same effect [18,25]. Particularly in the face of contradictory

findings reported in other CMS studies [20,21] we consider the

observed changes in body weight gain as a verification of CMS

effectiveness in general rather than associating them with the

phenomenology of major depression.

On the basis of the results obtained so far, our next steps in the

direction of a CMS model in mice will include some modifications

of the CMS procedure and the experimental design. Assessment of

consummatory behaviour will be accompanied by detailed

determination of homecage activity during the saccharin tests,

and finally, the effects of antidepressant treatment on the diverse

behavioural endpoints included in the present study will decide

upon the applicability of our CMS model in general and

additionally might reveal a possible relation of ‘anomalous’

behavioural changes to symptoms observed in human depression.

Supporting Information

Figure S1 Effects of short-term CMS on saccharin consumption

per 2 hrs in other mouse strains Effects of CMS on saccharin

intake over a period of 3-4 weeks in Balb/c, D2JIco, FVB and

NMRI mice (measurement 2x/week). White circles: control group,

black circles: CMS group. First 4 data points of each graph

represent basal consumption. Data represent mean6SEM, n = 8-

24/group. * p,0.05, ** p,0.01 pairwise between-group compar-

isons (Student’s t-test).

Found at: doi:10.1371/journal.pone.0004326.s001 (0.45 MB TIF)

Table S1 Inter-strain comparison of saccharin intake before and

after CMS application in stressed animals. Averaged saccharin

intake during 2 weeks under basal conditions, the same parameter

corrected for body weight and averaged intake during the first 3-

4 weeks of CMS (as percentage of basal intake). a-f indicate

significant differences to Balb/c, BL/6J, D2JIco, D2Ola, FVB and

NMRI mice, respectively (Tukey test). Data represent mean6-

SEM of CMS animals only, n = 8-24/group.

Found at: doi:10.1371/journal.pone.0004326.s002 (0.01 MB

RTF)

Acknowledgments

We thank Christine Bartl, Thomas Pohl and Florian Schleicher for their

enduring engagement in this work and Prof. Chris Turck for proof-reading

the manuscript.


Conceived and designed the experiments: MCS IS. Performed the

experiments: MCS MSHH IS. Analyzed the data: MCS. Contributed

reagents/materials/analysis tools: MCS. Wrote the paper: MCS MSHH

IS.

References


Mental Disorder (DSM-IV). Washington, DC.

2. Kendler KS, Karkowski LM, Prescott CA (1999) Causal relationship between

stressful life events and the onset of major depression. Am J Psychiatry 156:

837–841.

3. Kessler RC (1997) The effects of stressful life events on depression. Annu Rev

Psychol 48: 191–214.

4. Willner P, Towell A, Sampson D, Sophokleous S, Muscat R (1987) Reduction of

sucrose preference by chronic unpredictable mild stress, and its restoration by a

tricyclic antidepressant. Psychopharmacology (Berl) 93: 358–364.

5. Anisman H, Matheson K (2005) Stress, depression, and anhedonia: caveats

concerning animal models. Neurosci Biobehav Rev 29: 525–546.

6. Henn FA, Vollmayr B (2005) Stress models of depression: forming genetically

vulnerable strains. Neurosci Biobehav Rev 29: 799–804.

7. Fuchs E (2005) Social stress in tree shrews as an animal model of depression: an

example of a behavioral model of a CNS disorder. CNS Spectr 10: 182–190.

8. Schmidt MV, Sterlemann V, Ganea K, Liebl C, Alam S, et al. (2007) Persistent

neuroendocrine and behavioral effects of a novel, etiologically relevant mouse

paradigm for chronic social stress during adolescence. Psychoneuroendocrinol-

ogy 32: 417–429.

9. Anisman H, Merali Z, Hayley S (2008) Neurotransmitter, peptide and cytokine

processes in relation to depressive disorder: comorbidity between depression and

neurodegenerative disorders. Prog Neurobiol 85: 1–74.




Neuropsychopharmacology 23: 477–501.

12. Linthorst AC, Reul JM (2008) Stress and the brain: solving the puzzle using

microdialysis. Pharmacol Biochem Behav 90: 163–173.

13. McEwen BS (2005) Glucocorticoids, depression, and mood disorders: structural

remodeling in the brain. Metabolism 54: 20–23.

14. Nemeroff CB (1988) The role of corticotropin-releasing factor in the

pathogenesis of major depression. Pharmacopsychiatry 21: 76–82.

15. Sapolsky RM (2003) Stress and plasticity in the limbic system. Neurochem Res

28: 1735–1742.

16. Willner P (1997) Validity, reliability and utility of the chronic mild stress model

of depression: a 10-year review and evaluation. Psychopharmacology (Berl) 134:

319–329.

17. Willner P (2005) Chronic mild stress (CMS) revisited: consistency and

behavioural-neurobiological concordance in the effects of CMS. Neuropsycho-

biology 52: 90–110.

18. Strekalova T, Spanagel R, Dolgov O, Bartsch D (2005) Stress-induced

hyperlocomotion as a confounding factor in anxiety and depression models in

mice. Behav Pharmacol 16: 171–180.

19. Griffiths J, Shanks N, Anisman H (1992) Strain-specific alterations in

consumption of a palatable diet following repeated stressor exposure. Pharmacol

Biochem Behav 42: 219–227.

20. Ibarguen-Vargas Y, Surget A, et al. (2008) Multifaceted strain-specific effects in

a mouse model of depression and of antidepressant reversal. Psychoneuroendo-

crinology;doi:10 1016/j psyneuen 2008 07 010.

21. Pothion S, Bizot JC, Trovero F, Belzung C (2004) Strain differences in sucrose

preference and in the consequences of unpredictable chronic mild stress. Behav

Brain Res 155: 135–146.

22. Ducottet C, Aubert A, Belzung C (2004) Susceptibility to subchronic

unpredictable stress is related to individual reactivity to threat stimuli in mice.

Behav Brain Res 155: 291–299.

23. Ducottet C, Belzung C (2004) Behaviour in the elevated plus-maze predicts

coping after subchronic mild stress in mice. Physiol Behav 81: 417–426.

24. Mineur YS, Belzung C, Crusio WE (2006) Effects of unpredictable chronic mild

stress on anxiety and depression-like behavior in mice. Behav Brain Res 175: 43–50.

25. Strekalova T, Spanagel R, Bartsch D, Henn FA, Gass P (2004) Stress-induced

anhedonia in mice is associated with deficits in forced swimming and

exploration. Neuropsychopharmacology 29: 2007–2017.

26. Griebel G, Stemmelin J, Scatton B (2005) Effects of the cannabinoid CB1

receptor antagonist rimonabant in models of emotional reactivity in rodents. Biol




27. Elizalde N, Gil-Bea FJ, Ramirez MJ, Aisa B, Lasheras B, et al. (2008) Long-

lasting behavioral effects and recognition memory deficit induced by chronicmild stress in mice: effect of antidepressant treatment. Psychopharmacology

(Berl) 199: 1–14.

28. Griebel G, Simiand J, Serradeil-Le GC, Wagnon J, Pascal M, et al. (2002)Anxiolytic- and antidepressant-like effects of the non-peptide vasopressin V1b

receptor antagonist, SSR149415, suggest an innovative approach for thetreatment of stress-related disorders. Proc Natl Acad Sci U S A 99: 6370–6375.

29. Li S, Wang C, Wang W, Dong H, Hou P, et al. (2008) Chronic mild stress

impairs cognition in mice: from brain homeostasis to behavior. Life Sci 82:934–942.

30. Harkin A, Houlihan DD, Kelly JP (2002) Reduction in preference for saccharinby repeated unpredictable stress in mice and its prevention by imipramine.

J Psychopharmacol 16: 115–123.31. D’Aquila PS, Newton J, Willner P (1997) Diurnal variation in the effect of

chronic mild stress on sucrose intake and preference. Physiol Behav 62:

421–426.32. Ohl F, Holsboer F, Landgraf R (2001) The modified hole board as a differential

screen for behavior in rodents. Behav Res Methods Instrum Comput 33:392–397.

33. Thoeringer CK, Sillaber I, Roedel A, Erhardt A, Mueller MB, et al. (2007) The

temporal dynamics of intrahippocampal corticosterone in response to stress-related stimuli with different emotional and physical load: an in vivo

microdialysis study in C57BL/6 and DBA/2 inbred mice. Psychoneuroendo-crinology 32: 746–757.

34. Tanaka M, Nakamura F, Mizokawa S, Matsumura A, Nozaki S, et al. (2003)Establishment and assessment of a rat model of fatigue. Neurosci Lett 352:

159–162.

35. Gorka Z, Moryl E, Papp M (1996) Effect of chronic mild stress on circadianrhythms in the locomotor activity in rats. Pharmacol Biochem Behav 54:

229–234.36. Moreau JL, Scherschlicht R, Jenck F, Martin JR (1995) Chronic mild stress-

induced anhedonia model of depression; sleep abnormalities and curative effects

of electroshock treatment. Behav Pharmacol 6: 682–687.37. Cheeta S, Ruigt G, van PJ, Willner P (1997) Changes in sleep architecture

following chronic mild stress. Biol Psychiatry 41: 419–427.38. Papp M, Gruca P, Boyer PA, Mocaer E (2003) Effect of agomelatine in the

chronic mild stress model of depression in the rat. Neuropsychopharmacology28: 694–703.

39. Wehr TA, Wirz-Justice A (1982) Circadian rhythm mechanisms in affective

illness and in antidepressant drug action. Pharmacopsychiatria 15: 31–39.

40. Kopp C (2001) Locomotor activity rhythm in inbred strains of mice: implications

for behavioural studies. Behav Brain Res 125: 93–96.

41. Tang X, Xiao J, Parris BS, Fang J, Sanford LD (2005) Differential effects of two

types of environmental novelty on activity and sleep in BALB/cJ and C57BL/6J

mice. Physiol Behav 85: 419–429.

42. Cabib S, Puglisi-Allegra S (1996) Stress, depression and the mesolimbic

dopamine system. Psychopharmacology (Berl) 128: 331–342.

43. Bachmanov AA, Tordoff MG, Beauchamp GK (2001) Sweetener preference of

C57BL/6ByJ and 129P3/J mice. Chem Senses 26: 905–913.

44. Cabib S, Puglisi-Allegra S (1996) Different effects of repeated stressful

experiences on mesocortical and mesolimbic dopamine metabolism. Neurosci-

ence 73: 375–380.

45. Cryan JF, Mombereau C (2004) In search of a depressed mouse: utility of models

for studying depression-related behavior in genetically modified mice. Mol


46. D’Aquila PS, Brain P, Willner P (1994) Effects of chronic mild stress on

performance in behavioural tests relevant to anxiety and depression. Physiol

Behav 56: 861–867.

47. Gronli J, Murison R, Fiske E, Bjorvatn B, Sorensen E, et al. (2005) Effects of

chronic mild stress on sexual behavior, locomotor activity and consumption of

sucrose and saccharine solutions. Physiol Behav 84: 571–577.

48. Ducottet C, Griebel G, Belzung C (2003) Effects of the selective nonpeptide

corticotropin-releasing factor receptor 1 antagonist antalarmin in the chronic

mild stress model of depression in mice. Prog Neuropsychopharmacol Biol


49. Rossler AS, Joubert C, Chapouthier G (2000) Chronic mild stress alleviates

anxious behaviour in female mice in two situations. Behav Processes 49:

163–165.

50. Li S, Wang C, Wang M, Li W, Matsumoto K, et al. (2007) Antidepressant like

effects of piperine in chronic mild stress treated mice and its possible

mechanisms. Life Sci 80: 1373–1381.

51. Kopp C, Vogel E, Rettori MC, Delagrange P, Misslin R (1999) The effects of

melatonin on the behavioural disturbances induced by chronic mild stress in

C3H/He mice. Behav Pharmacol 10: 73–83.

52. Kompagne H, Bardos G, Szenasi G, Gacsalyi I, Harsing LG, et al. (2008)

Chronic mild stress generates clear depressive but ambiguous anxiety-like

behaviour in rats. Behav Brain Res 193: 311–314.

53. Cancela LM, Bregonzio C, Molina VA (1995) Anxiolytic-like effect induced by

chronic stress is reversed by naloxone pretreatment. Brain Res Bull 36: 209–213.



Timing Is Critical for Effective Glucocorticoid ReceptorMediated Repression of the cAMP-Induced CRH GeneSiem van der Laan, E. Ronald de Kloet, Onno C. Meijer*

Division of Medical Pharmacology, Leiden/Amsterdam Center for Drug Research&Leiden University Medical Center, Leiden, The Netherlands

Abstract

Glucocorticoid negative feedback of the hypothalamus-pituitary-adrenal axis is mediated in part by direct repression ofgene transcription in glucocorticoid receptor (GR) expressing cells. We have investigated the cross talk between the twomain signaling pathways involved in activation and repression of corticotrophin releasing hormone (CRH) mRNA expression:cyclic AMP (cAMP) and GR. We report that in the At-T20 cell-line the glucocorticoid-mediated repression of the cAMP-induced human CRH proximal promoter activity depends on the relative timing of activation of both signaling pathways.Activation of the GR prior to or in conjunction with cAMP signaling results in an effective repression of the cAMP-inducedtranscription of the CRH gene. In contrast, activation of the GR 10 minutes after onset of cAMP treatment, results in asignificant loss of GR-mediated repression. In addition, translocation of ligand-activated GR to the nucleus was found asearly as 10 minutes after glucocorticoid treatment. Interestingly, while both signaling cascades counteract each other onthe CRH proximal promoter, they synergize on a synthetic promoter containing ‘positive’ response elements. Since theorder of activation of both signaling pathways may vary considerably in vivo, we conclude that a critical time-window existsfor effective repression of the CRH gene by glucocorticoids.

Citation: van der Laan S, de Kloet ER, Meijer OC (2009) Timing Is Critical for Effective Glucocorticoid Receptor Mediated Repression of the cAMP-Induced CRHGene. PLoS ONE 4(1): e4327. doi:10.1371/journal.pone.0004327


Received October 1, 2008; Accepted November 19, 2008; Published January 29, 2009

Copyright: � 2009 van der Laan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The research described was funded by a ZON-MW Vidi grant (016.036.381) to OCM, and the Royal Dutch Academy of Arts and Sciences. The fundershad no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.



Introduction

Corticotropin releasing hormone (CRH) is a pivotal signaling

molecule in the regulation of the stress response. This neuropep-

tide is expressed at high levels in the hypothalamus, from where it

coordinates the hormonal and autonomic response to stress, and

the central nucleus of the amygdala, where it plays a crucial role in

regulating anxiety. Regulation of the expression of CRH is

therefore thought to be physiologically important in relation to

coping with stress. CRH gene regulation is a complex process that

involves multiple activating and repressing transcription factors

[1]. Among the often studied factors that can regulate CRH

expression are glucocorticoid hormones [2], which in a cell-

dependent manner either repress of stimulate the CRH gene. As

such, the CRH promoter can be considered a model gene for cell-

specific negative regulation of gene expression via glucocorticoids.

Cross-talk of intracellular signaling pathways is central to many

neuroendocrine control systems [3,4]. The expression and/or

secretion of the two main neuroendocrine secretagogues of the

hypothalamus-pituitary-adrenal axis (HPA axis) are both stimu-

lated by cAMP and suppressed by glucocorticoids, the end-

product of the HPA axis: hypothalamic CRH, as well as

adrenocorticotrophic hormone (ACTH) from anterior pituitary

corticotrophs [5–8]. At the molecular level, these signals are

represented by protein kinase A (PKA), the transcription factor

cAMP element-binding protein (CREB), and the GR, respectively.

The proximal promoter of the human corticotrophin releasing

hormone (hCRH) gene contains a canonical, functional cAMP

response element (CRE) and a negative glucocorticoid receptor

response element (nGRE) (fig. 1). Induction of the hCRH gene by

the PKA pathway is mediated by phosphorylation of CREB at

serine residue 133 [9,10]. In vivo, Wolfl et al. showed that binding of

CREB to the canonical CRE located at the nucleotide position

2224 (upstream of exon 1) was specifically induced after activation

of the PKA pathway with forskolin [11]. Additionally, Kovacs et al.

demonstrated that in the hypothalamic parvocellular neurons of

rodents subjected to ether stress, CREB phosphorylation was

induced in a time course that parallels the increase of CRH

heteronuclear RNA levels [12].

In vitro, the At-T20 cell-line is a well-established model system for

studying glucocorticoid-induced repression of the hCRH proximal

promoter. Nested deletions and site-specific point mutations of the

CRE located at nucleotide 2224 resulted in a significant loss of

induction by cAMP, demonstrating that CREB binding is necessary

for the stimulation of the gene [13]. In parallel, electrophoretic

mobility shift assays (EMSA) identified a GR-binding site at position

nt 2249 that was indispensable for GR-mediated repression of the

cAMP-induced promoter. Internal deletion of the entire nGRE and

specific point mutations resulted in a loss of repression by the ligand-

activated GR, indicating that DNA binding is essential for the

glucocorticoid-induced repression [14]. Of note: while we have

taken this nGRE-mode as working model, a separate series of

experiments did not find evidence for direct GR binding to the

CRH promoter, but rather suggested direct CREB-GR interactions

as the cause of GR-mediated reression [15].

The nGRE in the hCRH promoter is separated by as few as 25

bp with the canonical CRE, a distance that clearly permits

functional interactions at the promoter [16]. Since, in vivo the order


tbrunton

Cross-Out

tbrunton

Inserted Text

h and L

of activation of the cAMP and glucocorticoid signaling pathways

may vary considerably, and this is known to affect responses at the

level of neuroendocrine secretion [17], we tested the hypothesis

that effective repression of the cAMP-induced hCRH proximal

promoter depends on the relative timing of GR activation in the

At-T20 cell-line.

Results

Dexamethasone pre- or simultaneous co-treatment withFSK

FSK treatment led to a robust and progressive stimulation of the

CRH-promoter activity that was evident for luciferase induction

from 1 hour to at least 5 hours (fig. 2A). In line with previous

reports [14,18], simultaneous DEX co-treatment strongly sup-

pressed the FSK-induced stimulation of the hCRH-promoter

activity (fig. 2A). DEX co-treatment resulted in up to 75%

repression of the FSK-induced promoter activity after 3 hours

treatment (fig. 2B). To test our hypothesis that the order of

activation of both signaling cascades affects the level of GR-

mediated repression, we initiated the DEX treatment at different

time points prior to or after initiation of the 3-hours FSK

treatment (fig. 2C). We compared the DEX-induced repression in

the different groups to the simultaneous co-treatment group,

which was set at 100% repression. Two hours of DEX pre-

treatment resulted in a significantly increased repression, suggest-

ing that a slower mechanism requiring de novo protein synthesis is

responsible for the additional repression (data not shown).

Activation of the GR up to one hour prior to FSK treatment

did not affect the levels of repression (fig 2C, first three time

points). Of note, DEX treatment alone (0.1 mM) did not suppress

the basal activity of the CRH-promoter, indicating that basal

CRH drive is not governed by CREB/CRE dependent mecha-

nisms (data not shown).

Dexamethasone treatment applied after FSKWhen DEX was applied after forskolin stimulation of the CRH

promoter, the time-window separating both treatments was of

great consequence for the level of repression (fig. 2C). A

10 minutes delay in the onset of DEX treatment relative to the

FSK treatment resulted in a loss of 20% repression. Strikingly, a

30 minutes delay resulted in a 50% loss of GR-mediated

repression, indicating the importance of the relative time of

treatment. Clearly, the reduced time of DEX exposure is not

proportional to the loss of GR-mediated repression pointing to a

‘GR resistance’. Because FSK treatment induces a progressive

increase of the CRH-luc promoter activity over a period of at least

5 hours (fig. 2A) we assume that FSK-induces binding of CREB to

the promoter over that period. However, the first hour following

FSK treatment is critical for the GR to mediate effective

repression.

To assess whether FSK treatment alters the translocation

properties of the GR to the nucleus, we quantified GR-

immunoreactivity in the different conditions. The data show that

DEX treatment induces maximal nuclear GR-immunoreactivity

(GR-ir) as early as 10 minutes after treatment (figure 3). No

difference in nuclear GR-ir was observed between the 10 and

30 minutes DEX treatment groups, suggesting that the ‘GR

resistance’ is not due to delayed translocation to the nucleus

(fig. 3A). In addition, FSK treatment did not influence transloca-

tion dynamics of the GR although it is known that PKA activation

can modulate the steroid sensitivity by enhancing DNA binding

properties of GR [19]. In sum, the translocation data provide

evidence that GR should be capable of modulating gene

transcription as early as 10 minutes after treatment and that

FSK treatment does not interfere with translocation related

mechanisms.

Promoter specificityPosttranslational modification such as phosphorylation is known

to affect DNA binding properties, transcriptional activation and

stability of numerous nuclear receptors including GR [19].

Although translocation to the nucleus was not affected by FSK

treatment, we tested whether FSK influenced the transcriptional

activity of the GR in these cells. We measured the effect of FSK

and DEX co-treatment on a positively regulated promoter (a

Figure 1. Simplified representation of the hCRH-luc promoter and known response elements. Schematic representation of thecomposite hCRH proximal promoter, as present in the reporter construct. Although only the known nGRE and CRE have been indicated, manyresponse elements have been identified within the used reporter construct, such as a functional estrogen response element half site [28], and severalputative AP1 sites [14,25]. In addition, some of the listed factors act on sequences that are not present in reporter construct [29].doi:10.1371/journal.pone.0004327.g001

Timing of CRH Repression


synthetic GRE-containing promoter; TAT3-luc; [20]). FSK co-

treatment synergistically induced transcription on an exclusively

GRE-containing promoter compared to DEX treatment alone

(fig. 4). FSK treatment prior DEX treatment resulted in an

increased transcriptional activity of the GR. Likewise, the longer

the time of FSK co-treatment the higher the transcriptional

activity of the GR.

Discussion

The current data demonstrate that time-dependent interactions

between GR and cAMP/CREB can occur at the level of the CRH

gene, where these factors seem to functionally compete for the

same promoter. While similar interactions have been described for

the secretion of ACTH from the pituitary [17], we now show that

cAMP induced ‘GR-resistance’ occurs at the level of a single

promoter, and that it is not a global cellular phenomenon, but

gene-specific.

Using the nGRE that was reported to be functional in these cells

as a working model [14], the observed ‘primacy’ effect for

transcription factor action at the CRH promoter may be due to

the close proximity of the two response elements involved. The

spacing of the elements is such that it is unlikely that both GR and

CREB may bind simultaneously in an independent manner [16].

Sterical hindrance at the promoter due to the formation of larger

protein complexes may be responsible for the importance of

timing of stimuli. Alternatively, in view of the dynamic nature of

transcription factor-DNA interactions, CREB-mediated chroma-

tin remodeling events that disfavor GR-binding may account for

the decreased GR efficacy observed after prior cAMP elevations.

Interestingly, the analogous dependence of timing of both

cAMP and GR that exists for ACTH secretion [17], which

obviously is not linked to the activity of the exogenous reporter

plasmid, suggests that the phenomenon of time-dependence occurs

at multiple genes. Any gene regulated in a parallel manner will

allow better hypotheses as to the mechanism that is responsible for

the time dependent effects. POMC and CRH seem to depend on

the same coregulator molecule, namely SRC-1 [18,21]. In this

respect it would be of great interest to also study negative

regulation of the endogenous POMC gene in these cells under

similar conditions as were used for the CRH reporter construct.

Although numerous studies were devoted to understanding the

regulation of CRH gene expression in the paraventricular nucleus

of the hypothalamus, it is still a topic of debate whether the

activated-GR directly acts on the promoter region of the gene or

that different mechanisms are responsible for the repression of

CRH gene after stress. Bali et al. convincingly demonstrated in

organotypic slice cultures that the GR directly acts on the

paraventricular neurons to repress FSK-induced activity. Howev-

er, the molecular mechanisms underlying this GR-mediated

repression are still unknown. Guardiola-Diaz et al. suggested in

1996 that glucocorticoid repression occurs via interactions

between the GR and the cAMP-responsive element-binding

proteins [15], rather than via direct DNA binding of GR. In

contrast, Dorin et al. provided evidence, also in the same cell line as

used in present study, that the nGRE in the promoter is essential

for repression by glucocorticoids [13]. It would certainly be of

interest to study whether CREB phosphorylation status changes as

a consequence of GR activation at different time points, and test

the hypothesis that it is inversely related with the extent of GR

repression. However, while CREB-driven transcription is re-

pressed by glucocorticoids on the composite hCRH promoter, it is

unaffected on a 56CRE-containing promoter [18]. With the

possible caveat that the 56CRE may be not allow detection of

subtle changes in CREB function, these data point to gene/

promoter specificity of any direct CREB-GR interactions.

On the other hand, FSK-induced PKA can modulate glucocor-

ticoid signaling both on the composite hCRH and the exclusively

36GRE-containing promoters. Therefore, PKA activation can

determine the transcriptional outcome at glucocorticoid target genes,

Figure 2. Relative timing of DEX and FSK treatment determines efficacy of GR-dependent repression of CRH-promoter activity. (2A)FSK-stimulation progressively induces the CRH-promoter activity in the Att20 cells over time. Co-treatment with DEX resulted in a repressed CRH-activity. 2B) CRH-promoter activity expressed as percentage of maximal induction after 3 hours forskolin (FSK) treatment (filled bar). Simultaneous co-treatment with DEX (open bar) resulted in a strong repression of the CRH-promoter activity. (2C) The repression induced by DEX in the co-treatmentgroup was set at 100%. All groups were treated for three hours with FSK. Different time of onset of the DEX treatment relative to the FSK treatmentresults in a significant loss of repression when DEX treatment is started 10 minutes after FSK treatment (*). FSK treatment leads to a progressiveincrease in CRH-luc promoter activity over a period of at least 5 hours (inset). Values represent group averages 6 SD.doi:10.1371/journal.pone.0004327.g002



independent of the presence of CREs in the promoter. We postulate

that there is no cross-talk between the GR and CREB off the DNA

but that PKA activation modulates GR-mediated transcription by

changing e.g the phosphorylation status of coregulator proteins.

Chromatin immunoprecipitation assays should be used to test the

interactions between GR and (phospho-)CREB at the CRH

promoter, and demonstrate lack of direct GR-CREB interactions

at non-composite GRE-containing promoters.

An unexplained phenomenon is that, in contrast to the situation

in PVN, glucocorticoids induce, rather than repress, CRH gene

expression in the placenta and amygadala [2,22–24]. The opposite

effect of GR in these cells may rather relate to differential presence

of transcription factors or coactivators such as SRC1a [18,25]. One

principle difference in cellular context between CRH containing

cells in PVN and other tissues is that activation of the CRH gene in

the paraventricular cells often will be accompanied by increased

activation of the HPA axis, causing a quick rise in glucorticoid levels

and GR activation. However, current data should be interpreted in

the context of regulation of the CRH-promoter in the PVN, and do

not give insights in the mechanisms governing the cell-specific

effects of glucocorticoids on CRH expression.

It is well known that acute exogenous steroid treatment

effectively suppresses stress-induced expression of CRH mRNA

in rats [26]. However, the current study using a model system

shows that repression is markedly attenuated if GR activation is

initiated with as little as a 10 minutes delay. The critical time-

window for effective repression by glucocorticoids may have

interesting implications in the control of CRH expression in vivo.

The order of activation of both signaling pathways is variable, and

depends on the history of stress and glucocorticoid exposure, as

well as the circadian and ultradian pulsatility of glucocorticoid

levels [27]. Therefore, it is likely that effective GR-mediated

repression of the stress-induced CRH mRNA expression will only

occur in specific situations. We conclude that the differences in

timing of stimulatory and repression signals are of consequence for

adaptation of the organism to stress.


Reporter assays0.16106 cells were transiently transfected in 24-wells plate using

Lipofectamine 2000 (Invitrogen, Breda, The Netherlands) accord-

ing to the manufacturer’s instructions. Per well, 200 ng of the

hCRH-luc reporter plasmid [18] or the 36GRE containing

TAT3-luc reporter were transfected. The day after transfection,

the cells were treated with 10 mM forskolin (Calbiochem,

Darmstadt, Germany) and/or 0.1 mM of the synthetic glucocor-

ticoid dexamethasone (DEX) and assayed for luciferase activity.

Figure 3. Translocation of the GR occurs within 10 minutes after treatment. (3A) Time course of GR-ir in different treatment groups. DEXalone and FSK + DEX co-treatment, but not FSK alone show nuclear GR staining after 10 minutes treatment. (3B) Control IgG staining show specificityof the GR-specific antibody. (3C) Nuclear quantification of GR-ir after 10 minutes treatment. Values represent average 6SEM.doi:10.1371/journal.pone.0004327.g003



Experiments were performed at 4 wells per condition, and were

repeated at least three times. Statistical analysis was performed

using one way analysis of variance (ANOVA) and statistical

significance (*) was determined with Tukey’s multiple comparison

tests with p,0.05.

Immunofluorescent staining of the GRA day prior stimulation, 306103 AtT-20 cells were grown in

chamber slides. Following stimulation, cells were fixed in 4%

paraformaldehyde, permeabilized with Triton X-100 and blocked

with 5% normal goat serum. Cells were incubated with a GR-

specific antibody (M20; dilution 1:500; Santa Cruz biotechnolo-

gies) during 60 minutes, washed and subsequently incubated for

60 minutes with a secondary goat anti-rabbit Alexa Fluor 488

antibody (dilution 1:750; Invitrogen, Breda, The Netherlands).

After incubation, cells were washed and counterstained for 10 min

with Hoechst 33528. All sections were embedded in polyaqua-

mount (Polysciences, Inc.) and visualized with an immunofluores-

cence microscope (Leica DM6000). Control cells were incubated

with equal amounts of non-immune rabbit serum (Santa Cruz),

which was used as substitute for the primary antibodies. Nuclear

immunoreactivity was measured using ImageJ 1.32j software

(NIH, USA).


Conceived and designed the experiments: SvdL OCM. Performed the

experiments: SvdL. Analyzed the data: SvdL OCM. Wrote the paper:

SvdL ERdk OCM.

References

1. Liu Y, Kamitakahara A, Kim AJ, Aguilera G (2008) Cyclic adenosine 39,59-monophosphate responsive element binding protein phosphorylation is required

but not sufficient for activation of corticotropin-releasing hormone transcription.Endocrinology 149: 3512–3520.

2. Swanson LW, Simmons DM (1989) Differential steroid hormone and neuralinfluences on peptide mRNA levels in CRH cells of the paraventricular nucleus:

a hybridization histochemical study in the rat. J Comp Neurol 285: 413–435.

3. Zoeller RT, Tan SW, Tyl RW (2007) General background on the hypothalamic-

pituitary-thyroid (HPT) axis. Crit Rev Toxicol 37: 11–53.



5. Van LP, Spengler DH, Holsboer F (1990) Glucocorticoid repression of 39,59-

cyclic-adenosine monophosphate-dependent human corticotropin-releasing-

hormone gene promoter activity in a transfected mouse anterior pituitary cellline. Endocrinology 127: 1412–1418.

6. Spengler D, Rupprecht R, Van LP, Holsboer F (1992) Identification andcharacterization of a 39,59-cyclic adenosine monophosphate-responsive element

in the human corticotropin-releasing hormone gene promoter. Mol Endocrinol6: 1931–1941.

7. Kraus J, Hollt V (1995) Identification of a cAMP-reponsive element on thehuman proopiomelanocortin gene upstream promoter. DNA and Cell Biology

14: 103–110.

8. Drouin J, Trifiro MA, Plante RK, Nemer M, Eriksson P, et al. (1989)

Glucocorticoid receptor binding to a specific DNA sequence is required for

hormone-dependent repression of pro-opiomelanocortin gene transcription. MolCell Biol 9: 5305–5314.

9. Hagiwara M, Brindle P, Harootunian A, Armstrong R, Rivier J, et al. (1993)Coupling of hormonal stimulation and transcription via the cyclic AMP-

responsive factor CREB is rate limited by nuclear entry of protein kinase A. MolCell Biol 13: 4852–4859.

10. Delghandi MP, Johannessen M, Moens U (2005) The cAMP signalling pathwayactivates CREB through PKA, p38 and MSK1 in NIH 3T3 cells. Cell Signal 17:

1343–1351.

11. Wolfl S, Martinez C, Majzoub JA (1999) Inducible binding of cyclic adenosine

39,59-monophosphate (cAMP)-responsive element binding protein (CREB) to acAMP-responsive promoter in vivo. Mol Endocrinol 13: 659–669.

12. Kovacs KJ, Sawchenko PE (1996) Sequence of stress-induced alterations inindices of synaptic and transcriptional activation in parvocellular neurosecretory

neurons. J Neurosci 16: 262–273.

13. Malkoski SP, Handanos CM, Dorin RI (1997) Localization of a negative

glucocorticoid response element of the human corticotropin releasing hormonegene. MolCell Endocrinol 127: 189–199.

14. Malkoski SP, Dorin RI (1999) Composite glucocorticoid regulation at afunctionally defined negative glucocorticoid response element of the human

corticotropin-releasing hormone gene. Mol Endocrinol 13: 1629–1644.

15. Guardiola-Diaz HM, Kolinske JS, Gates LH, Seasholtz AF (1996) Negative

glucorticoid regulation of cyclic adenosine 39, 59-monophosphate-stimulatedcorticotropin-releasing hormone-reporter expression in AtT-20 cells. Mol

Endocrinol 10: 317–329.

16. Pearce D, Matsui W, Miner JN, Yamamoto KR (1998) Glucocorticoid receptor

transcriptional activity determined by spacing of receptor and nonreceptor DNAsites. J Biol Chem 273: 30081–30085.

17. Shipston MJ, Antoni FA (1992) Inactivation of early glucocorticoid feedback bycorticotropin-releasing factor in vitro. Endocrinology 130: 2213–2218.

18. van der Laan S, Lachize SB, Vreugdenhil E, de Kloet ER, Meijer OC (2008)Nuclear receptor coregulators differentially modulate induction and glucocor-

ticoid receptor-mediated repression of the corticotropin-releasing hormone gene.

Endocrinology 149: 725–732.

19. Weigel NL, Moore NL (2007) Steroid receptor phosphorylation: a keymodulator of multiple receptor functions. Mol Endocrinol 21: 2311–2319.

20. Liu W, Wang J, Sauter NK, Pearce D (1995) Steroid receptor heterodimeriza-tion demonstrated in vitro and in vivo. ProcNatlAcadSciUSA 92: 12480–12484.

21. Winnay JN, Xu J, O’Malley BW, Hammer GD (2006) Steroid receptorcoactivator-1-deficient mice exhibit altered hypothalamic-pituitary-adrenal axis

function. Endocrinology 147: 1322–1332.

22. Makino S, Gold PW, Schulkin J (1994) Effects of corticosterone on CRH mRNA

and content in the bed nucleus of the stria terminalis; comparison with the effectsin the central nucleus of the amygdala and the paraventricular nucleus of the

hypothalamus. Brain Res 657: 141–149.

23. Watts AG (2005) Glucocorticoid regulation of peptide genes in neuroendocrine

CRH neurons: a complexity beyond negative feedback. Front Neuroendocrinol

26: 109–130.

Figure 4. The FSK induced GR resistance is specific for the CRHpromoter. TAT3-luc (GRE-containing promoter) activity is expressed aspercentage of maximal induction after 4 hours DEX treatment (filledbar; t = 0). All groups (hatched bars) were treated for 4 hours with DEXand only the time of onset of FSK treatment was different. Forskolintreatment strongly enhanced the transcriptional rate of GR at all timepoints (# indicates significantly different from DEX group with p,0.05).Pre-treatment with FSK resulted in the highest potentiation of the GRtranscriptional rate.doi:10.1371/journal.pone.0004327.g004



24. King BR, Nicholson RC (2007) Advances in understanding corticotrophin-

releasing hormone gene expression. Front Biosci 12: 581–590.25. King BR, Smith R, Nicholson RC (2002) Novel glucocorticoid and cAMP

interactions on the CRH gene promoter. Mol Cell Endocrinol 194: 19–28.

26. Ginsberg AB, Campeau S, Day HE, Spencer RL (2003) Acute glucocorticoidpretreatment suppresses stress-induced hypothalamic-pituitary-adrenal axis

hormone secretion and expression of corticotropin-releasing hormone hnRNAbut does not affect c-fos mRNA or fos protein expression in the paraventricular

nucleus of the hypothalamus. J Neuroendocrinol 15: 1075–1083.

27. Windle RJ, Wood SA, Lightman SL, Ingram CD (1998) The pulsatilecharacteristics of hypothalamo-pituitary-adrenal activity in female Lewis and

Fischer 344 rats and its relationship to differential stress responses. Endocrinol-

ogy 139: 4044–4052.

28. Chen XN, Zhu H, Meng QY, Zhou JN (2008) Estrogen receptor-alpha and -

beta regulate the human corticotropin-releasing hormone gene through similar

pathways. Brain Res 1223: 1–10.

29. Lalmansingh AS, Uht RM (2008) Estradiol regulates corticotropin-releasing

hormone gene (crh) expression in a rapid and phasic manner that parallels

estrogen receptor-alpha and -beta recruitment to a 39,59-cyclic adenosine 59-

monophosphate regulatory region of the proximal crh promoter. Endocrinology

149: 346–357.



Stress-Induced Depression and Comorbidities: …...Stress-Induced Depression and Comorbidities: From...

Documents

Transcript of Stress-Induced Depression and Comorbidities: …...Stress-Induced Depression and Comorbidities: From...