Stress-Induced Depression and Comorbidities: …...Stress-Induced Depression and Comorbidities: From...
Transcript of 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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4265
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
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4265
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
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4265
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
Psychiatry 64: 293–301.
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
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4265
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4266
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4266
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
Editor: Bernhard Baune, James Cook University, Australia
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e3590
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e3590
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
Stress Biomarkers in Women
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e3590
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.
Stress Biomarkers in Women
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e3590
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.
Stress Biomarkers in Women
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e3590
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
Editor: Bernhard Baune, James Cook University, Australia
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4324
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?
Materials and Methods
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4324
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4324
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4324
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4324
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4324
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
COR peak week 5 NIm (n = 16) 7/9 11/5
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 1 NIm (n = 7) 3/4 4/3
COR peak week 5 Im (n = 14) 3/11 5/9
Phi = 0.435 p = 0.052 Phi = 0.436 p = 0.051
COR peak week 5 NIm (n = 6) 4/2 5/1
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 1 NIm (n = 3) 3/0 3/0
COR peak week 5 Im (n = 10) 4/6 6/4
Phi = 0.105 p = 0.639 Phi = 0.000 p = 1.000
COR peak week 5 NIm (n = 10) 3/7 6/4
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4324
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
COR basal week 5 Im (n = 24) 6/18 11/13
Phi = 0.257 p = 0.104 Phi = 0.226 p = 0.154
COR basal week 5 NIm (n = 16) 8/8 11/5
reboxetine group (n = 20) 7/13 10/10
COR basal week 1 Im (n = 13) 2/11 5/8
Phi = 0.560 p = 0.012 Phi = 0.314 p = 0.160
COR basal week 1 NIm (n = 7) 5/2 5/2
COR basal week 5 Im (n = 14) 2/12 5/9
Phi = 0.663 p = 0.003 Phi = 0.436 p = 0.051
COR basal week 5 NIm (n = 6) 5/1 5/1
mirtazapine group (n = 20) 7/13 12/8
COR basal week 1 Im (n = 15) 3/12 7/8
Phi = 0.545 p = 0.015 Phi = 0.471 p = 0.035
COR basal week 1 NIm (n = 5) 4/1 5/0
COR basal week 5 Im (n = 10) 4/6 6/4
Phi = 0.105 p = 0.639 Phi = 0.000 p = 1.000
COR basal week 5 NIm (n = 10) 3/7 6/4
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e4324
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e4324
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.
Author Contributions
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.
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e4324
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
Psychiatry 40: 493–500.
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
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e4324
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.
DEX/CRH Test and Prediction
PLoS ONE | www.plosone.org 12 January 2009 | Volume 4 | Issue 1 | e4324
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
Editor: Bernhard Baune, James Cook University, Australia
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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.
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4328
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4328
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
KAR, Stress & Steroids
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4328
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
KAR, Stress & Steroids
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4328
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
KAR, Stress & Steroids
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4328
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.
Author Contributions
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.
KAR, Stress & Steroids
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4328
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.
KAR, Stress & Steroids
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4328
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
Editor: Bernhard Baune, James Cook University, Australia
Received September 11, 2008; Accepted October 9, 2008; Published January 30, 2009
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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],
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4329
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e4329
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
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e4329
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.
Author Contributions
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.
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e4329
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.
72. Holsboer F (2000) The corticosteroid receptor hypothesis of depression.
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.
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e4329
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.
Reversing Juvenile Stress
PLoS ONE | www.plosone.org 12 January 2009 | Volume 4 | Issue 1 | e4329
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
Editor: Bernhard Baune, James Cook University, Australia
Received September 16, 2008; Accepted October 28, 2008; Published January 30, 2009
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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)-
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4330
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.
Materials and Methods
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e4330
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
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e4330
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.
Author Contributions
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.
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e4330
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.
Exercise-Epigenetics-Behavior
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e4330
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
Editor: Bernhard Baune, James Cook University, Australia
Received September 10, 2008; Accepted October 21, 2008; Published January 30, 2009
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.
* E-mail: [email protected]
. 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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4331
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e4331
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
Body Weight under Stress
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e4331
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%
Body Weight under Stress
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e4331
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.
Author Contributions
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.
25. American Psychiatric Association (2000) Diagnostic and Statistical Manual of
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.
Body Weight under Stress
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e4331
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.
Body Weight under Stress
PLoS ONE | www.plosone.org 12 January 2009 | Volume 4 | Issue 1 | e4331
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
Editor: Bernhard Baune, James Cook University, Australia
Received September 10, 2008; Accepted October 17, 2008; Published January 29, 2009
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
¤ 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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e3659
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.
Materials and Methods
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e3659
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
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e3659
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].
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e3659
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
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e3659
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
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e3659
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
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e3659
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
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e3659
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].
Author Contributions
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.
J Neuroendocrinol 14: 499–505.
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.
Stress Regulates M6a in Axons
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e3659
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
Editor: Bernhard Baune, James Cook University, Australia
Received September 15, 2008; Accepted October 14, 2008; Published January 30, 2009
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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].
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e3675
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.
Materials and Methods
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e3675
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 =
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e3675
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
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e3675
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.
Author Contributions
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.
Biol Psychiatry 52: 386–392.
3. 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: 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.
Biol Psychiatry 49: 1023–1039.
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.
6. Holsboer F (2000) The corticosteroid receptor hypothesis of depression.
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.
Biol Psychiatry 59: 1087–1096.
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.
J Neuroendocrinol 18: 629–631.
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.
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e3675
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.
J Neuroendocrinol 12: 5–12.
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.
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 12 January 2009 | Volume 4 | Issue 1 | e3675
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.
Early Stress and Neurogenesis
PLoS ONE | www.plosone.org 13 January 2009 | Volume 4 | Issue 1 | e3675
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
Editor: Bernhard Baune, James Cook University, Australia
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
. 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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4325
[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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4325
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.
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4325
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
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4325
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
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4325
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
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4325
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
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4325
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
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e4325
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
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e4325
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.
Author Contributions
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.
15. Gold PW, Chrousos GP (2002) Organization of the stress system and its
dysregulation in melancholic and atypical depression: high vs low CRH/NEstates. Mol Psychiatry 7: 254–275.
16. de Kloet ER, Joels M, Holsboer F (2005) Stress and the brain: from adaptation
to disease. Nat Rev Neurosci 6: 463–475.
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.
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e4325
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.
Rhythmicity & HPA Reactivity
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e4325
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
Editor: Bernhard Baune, James Cook University, Australia
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4326
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.
Materials and Methods
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 7 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 8 January 2009 | Volume 4 | Issue 1 | e4326
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
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 9 January 2009 | Volume 4 | Issue 1 | e4326
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.
Author Contributions
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
1. American Psychiatric Association (1994) Diagnostic and Statistical Manual of
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.
10. de Kloet ER, Joels M, Holsboer F (2005) Stress and the brain: from adaptation
to disease. Nat Rev Neurosci 6: 463–475.
11. Holsboer F (2000) The corticosteroid receptor hypothesis of depression.
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
Psychiatry 57: 261–267.
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 10 January 2009 | Volume 4 | Issue 1 | e4326
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
Psychiatry 9: 326–357.
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
Psychiatry 27: 625–631.
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.
CMS-Anhedonia-Activity
PLoS ONE | www.plosone.org 11 January 2009 | Volume 4 | Issue 1 | e4326
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
Editor: Bernhard Baune, James Cook University, Australia
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.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
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
PLoS ONE | www.plosone.org 1 January 2009 | Volume 4 | Issue 1 | e4327
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
PLoS ONE | www.plosone.org 2 January 2009 | Volume 4 | Issue 1 | e4327
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
Timing of CRH Repression
PLoS ONE | www.plosone.org 3 January 2009 | Volume 4 | Issue 1 | e4327
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.
Materials and Methods
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
Timing of CRH Repression
PLoS ONE | www.plosone.org 4 January 2009 | Volume 4 | Issue 1 | e4327
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).
Author Contributions
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.
4. de Kloet ER, Joels M, Holsboer F (2005) Stress and the brain: from adaptation
to disease. Nat Rev Neurosci 6: 463–475.
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
Timing of CRH Repression
PLoS ONE | www.plosone.org 5 January 2009 | Volume 4 | Issue 1 | e4327
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.
Timing of CRH Repression
PLoS ONE | www.plosone.org 6 January 2009 | Volume 4 | Issue 1 | e4327