You are on page 1of 14

Please cite this article in press as: Zannas AS, West AE.

Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003

Neuroscience xxx (2014) xxx–xxx

REVIEW
EPIGENETICS AND THE REGULATION OF STRESS VULNERABILITY
AND RESILIENCE
A. S. ZANNAS a1 AND A. E. WEST b* Key words: epigenetics, neural plasticity, psychological
a
Department of Psychiatry and Behavioral Sciences, Duke stress, resilience, vulnerability, DNA methylation.
University Medical Center, Durham, NC 27710, USA
b
Department of Neurobiology, Duke University Medical Contents
Center, Durham, NC 27710, USA Introduction 00
Stress, adaptations, resilience, and vulnerability 00
Adaptive and maladaptive responses to stressors 00
Abstract—The human brain has a remarkable capacity to Comparing stress in rodents and humans 00
adapt to and learn from a wide range of variations in the Stressor-induced changes in neuronal DNA methylation 00
environment. However, environmental challenges can also Environmental stressors regulate the writers and erasers
precipitate psychiatric disorders in susceptible individuals. of cytosine methylation 00
Why any given experience should induce one brain to adapt Stressor-dependent regulation of the methyl-cytosine
while another is edged toward psychopathology remains reader MeCP2 00
poorly understood. Like all aspects of psychological func- Epigenetic regulation of the HPA axis 00
tion, both nature (genetics) and nurture (life experience) Epigenetic homeostasis of the HPA axis 00
sculpt the brain’s response to stressful stimuli. Here we Epigenetics of the HPA axis in humans 00
review how these two influences intersect at the epigenetic Conclusions, directions and challenges for future research 00
regulation of neuronal gene transcription, and we discuss Emerging methodologies for the study of the epigenome 00
how the regulation of genomic DNA methylation near key Mechanisms of stress-dependent epigenetic regulation 00
stress-response genes may influence psychological sus- Understanding the epigenetics of stress in humans 00
ceptibility or resilience to environmental stressors. Our goal References 00
is to offer a perspective on the epigenetics of stress
responses that works to bridge the gap between the study
of this molecular process in animal models and its potential
usefulness for understanding stress vulnerabilities in
humans.
This article is part of a Special Issue entitled: Epigenetics INTRODUCTION
in Brain Function. Ó 2013 IBRO. Published by Elsevier Ltd.
Psychological stress is a major contributor to morbidity,
All rights reserved.
mortality, and health-care costs. Psychiatric disorders
directly linked etiologically to stress, including unipolar
major depressive disorder (MDD) and posttraumatic
stress disorder (PTSD), are among the top causes of
*Corresponding author. Address: Duke University Medical Center, disability and disease burden. MDD is ranked as the
Box 3209, Bryan Research Building, Room 301D, Durham, NC number one worldwide cause of years lived with
27710, USA. Tel: +1-919-681-1909. disability and is projected to be the second leading
E-mail addresses: anthony.zannas@duke.edu (A. S. Zannas), cause of global disease burden (combined morbidity
west@neuro.duke.edu (A. E. West).
1 and mortality) by the year 2020 (Lopez and Murray,
Current address: Max Planck Institute of Psychiatry, Munich 80804,
Germany. 1998). Less evidence exists on the epidemiology of
Abbreviations: 5hmC, 5-hydroxymethyl cytosine; 5mC, 5-methyl PTSD, but it is estimated that PTSD may affect as
cytosine; ACTH, adrenocorticotropin hormone; AVP, arginine much as 10% of the general population and is among
vasopressin; BDNF, Brain-Derived Neurotrophic Factor; BER, base
excision repair; Crf, Corticotrophin-Releasing Factor; CUMS, chronic the 20 leading causes of disease burden (Galea et al.,
ultra mild stress; DMRs, differentially methylated regions; DSM, 2005; Freed et al., 2010). Furthermore, psychological
Diagnostic and Statistical Manual for Mental Disorders; ELS, early stress can precipitate or perpetuate other psychiatric
life stress; FKBPs, FK506 binding proteins; GDNF, Glial-Derived
Neurotrophic Factor; GR, glucocorticoid receptor; GREs, glucocorticoid
disorders, such as schizophrenia, dementia, and
response elements; HPA, hypothalamic–pituitary–adrenal; MBD, addiction (van Winkel et al., 2008; Tsolaki et al., 2009;
methyl-DNA binding domain; MDD, major depressive disorder; Sinha et al., 2011), and can negatively affect the course
MeCP2, methyl-CpG binding protein 2; NGFI-A, nerve growth factor- of several non-psychiatric conditions, including
inducible protein A; PBCs, peripheral blood cells; PET, positron
emission tomography; PTG, posttraumatic growth; PTSD, cardiovascular disease, cancer, and HIV infection
posttraumatic stress disorder; PVN, paraventricular nucleus; SNPs, (Cohen et al., 2007; Kaltsas et al., 2012). Many of these
single nucleotide polymorphisms.

0306-4522/13 $36.00 Ó 2013 IBRO. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
1
2 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

conditions are potentially preventable and/or treatable. accessibility and/or activation state of gene regulatory
Improving the ability to predict which individuals are elements, which are the sequences encoded within
susceptible to stressors would offer the opportunity for genomic DNA that serve as docking sites for
early intervention and could have enormous implications transcription factors. The consequences of chromatin
for reducing stress-related disability and health-care regulation for gene transcription are context dependent
costs. and determined by both the location and the nature of
Risk for both MDD and PTSD is determined at least in the epigenetic modifications induced. For example,
part by genetics (Kendler and Prescott, 1999; Sartor increasing the accessibility of a gene regulatory element
et al., 2012). Concordance rates for monozygotic and could either increase or decrease transcription of nearby
dizygotic twins estimate heritability at about 37% for genes depending on whether an activator or a repressor
MDD (Sullivan et al., 2000) and 30–35% for PTSD binds at that site, and structural changes that open a
(Skelton et al., 2012). A few genetic polymorphisms gene regulatory element may have no effect on gene
have been associated with vulnerability to stressors. For expression whatsoever if the regulatory binding protein is
example, single nucleotide polymorphisms (SNPs) in not expressed.
FKBP5, a gene coding for a glucocorticoid receptor When established over the course of cellular
chaperone, has been shown to interact with childhood differentiation, a primary function of chromatin regulation
abuse to predict adult PTSD symptoms (Binder et al., is to underlie the ability of a single genome to be read
2008), and the 5-HTTLPR polymorphism in the out into a multiplicity of distinct cell-type specific gene
promoter region of the serotonin transporter gene expression programs. However, many of the same
(SLC6A4) has been reported to moderate the effects of chromatin regulatory factors that operate during
stress on depressive phenotypes in some but not all development remain expressed in terminally
populations tested (Caspi et al., 2003; Risch et al., differentiated neurons of the mature brain, and it is
2009; Karg et al., 2011). The heritable component of becoming increasingly clear that the expression and/or
stress responses can be replicated and studied in function of many of these factors in neurons is subject
laboratory animal models, as distinct mouse strains to regulation by environmental stimuli including
have been shown to display different behavioral stressors. In postmitotic neurons, chromatin regulation
responses to presentation of the same stressors thus provides a mechanism to translate environmental
(Uchida et al., 2011; Russo et al., 2012; Scharf and exposures into plasticity of gene expression.
Schmidt, 2012). Nonetheless, large-scale genome-wide Importantly, because structural regulation of chromatin
association studies have failed to show consistent can be very persistent (e.g. the lifelong inactivation of
genome-wide significant loci for MDD (Ripke et al., one X chromosome in female cells), the epigenome has
2013), and given that twin studies have revealed a the potential to encode a molecular memory of past
substantial role for unique environmental risk in the events that can influence gene expression, neuronal
etiology of stress disorders (Sullivan et al., 2000), function, and future behaviors. At the level of the
interest has increasingly turned toward identifying behaviors we are discussing here, this molecular
gene  environment interactions as a more effective mechanism would be relevant if some experience-
way to predict individual stress vulnerabilities. induced chromatin states were to change patterns of
One important point of gene  environment gene expression that alter adaptability to stressors, thus
interactions occurs at the level of the epigenetic changing the likelihood that an individual will display
regulation of gene transcription. The term epigenetics is vulnerability or resilience to future stressors.
derived from the Greek prefix ‘‘epi,’’ which means upon In this article, we discuss the evidence that epigenetic
or over, and genetics, which refers to the sequence of regulation of chromatin plays a role in determining the
an individual’s DNA. In common parlance, ‘‘epigenetic’’ adaptive or maladaptive nature of neural and behavioral
is often used in reference to a wide variety of phenotypic responses to environmental stressors. Our purpose is
differences between individuals that are independent of not to provide an exhaustive review of all the literature
the underlying genetic code, such as when genetically linking chromatin and stress, but rather to offer a
identical twins have dissimilar personalities. By contrast, perspective on this topic by focusing on selected stress-
at the molecular level the word has come to signify a inducible epigenetic regulatory mechanisms and specific
specific set of gene regulatory mechanisms that are gene targets that contribute directly to the behavioral
mediated by biochemical modifications of genomic DNA responses induced by environmental stressors. Given
along with its associated histone proteins and non- the diverse usage of the terms ‘‘stress,’’ ‘‘adaptation,’’
coding RNAs, which together comprise chromatin ‘‘vulnerability,’’ and ‘‘resilience,’’ we first define these
(Bonasio et al., 2010). Epigenetic transcriptional terms in the context of animal models and human
regulatory mechanisms include DNA methylation, studies. We then offer an example of how the epigenome
posttranslational modifications of histones, and changes influences stress responses by focusing on the evidence
in the position of the secondary structures formed by for DNA methylation-dependent regulation of genes in
DNA and histones called nucleosomes, all of which can the hypothalamic–pituitary–adrenal (HPA) axis, a key
be collectively referred to as the epigenome. What these regulatory pathway for coordinating organismal stress
forms of chromatin regulation have in common is that responses. Finally, we discuss the implications of these
they impact gene transcription by altering the data for and the challenges of applying this knowledge to

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx 3

the identification, prevention, and treatment of stress- event has been traditionally considered capable of
related psychiatric disorders. causing PTSD only if it involves the ‘‘actual or
threatened death or serious injury, or a threat to the
STRESS, ADAPTATIONS, RESILIENCE, AND physical integrity of self or others’’ and induces ‘‘intense
fear, helplessness, or horror’’ (American Psychiatric
VULNERABILITY
Association, 2000). Although this criterion was recently
The term stress can be used to refer to either a stressful revised in the 5th edition of DSM, intensity remains
stimulus, the organism’s response to a stressful stimulus, an important characteristic of traumatic events,
or the consequences of this response (Levine, 2005). By which should involve ‘‘death, threatened death, actual
contrast, the term stressor has been more consistently or threatened serious injury, or actual or threatened
used to denote the physical and emotional challenges sexual violence.’’ (American Psychiatric Association,
that threaten organismal homeostasis, and stress 2013).
response denotes the organism’s behavior in the face of Yet even when encountering a stressor of high
such challenges (Chrousos and Gold, 1992; intensity, only a small minority of individuals will develop
Karatsoreos and McEwen, 2011). Because stressors stress disorders. Whereas more than two thirds of the
are physiologically coupled to stress responses, some general population experiences an event meeting the
researchers have advocated for use of the more holistic definition of ‘‘traumatic’’ in their lives, only 5–10% of the
term allostasis to describe the process by which population develops PTSD (Galea et al., 2005).
variations in internal function allow an organism to Similarly, only 20–25% of the individuals exposed to
achieve stability in the face of constantly changing intense stressors develop MDD (Cohen et al., 2007).
environments (Karatsoreos and McEwen, 2011). These tendencies of individuals to exhibit adaptive
However for the purposes of this article, to differentiate versus maladaptive stress responses are widely
environmental effects on the epigenome from the described in terms of resilience and vulnerability (or
biochemical and behavioral consequences of chromatin susceptibility) respectively. Traditionally, research
regulation, we will use the term stressor to denote a studies tend to define vulnerability as the induction of
challenge posed by the environment and the term stress pathological processes upon stressor exposure (e.g.
response for the molecular, hormonal, neural, and/or onset of MDD following a divorce), whereas the
behavioral response evoked by such a stressor. absence of psychiatric symptoms despite stressor
exposure has been viewed as resilience. However,
Adaptive and maladaptive responses to stressors these descriptions risk implying that resilience is a
passive process, whereas experimental evidence
Although the terms stressor and stress response indicates that resilience can be actively invoked by
commonly carry negative connotations, stress molecular, hormonal, neural, and/or behavioral
responses can be either adaptive or maladaptive and mechanisms that either counteract already established
even similar responses can have different adaptive vulnerabilities or prevent the expression of vulnerable
value in distinct environmental contexts. For example, phenotypes (Russo et al., 2012). Such active
combat veterans exhibit vigilance and fear responses mechanisms not only prevent the development of
that are adaptive in the combat zone because they vulnerability states, but may also lead to a state of
decrease the likelihood of injury and death. However, enhanced functioning following stress. In this sense,
these same responses become maladaptive if they resilience may reflect a state of heightened adaptability
generalize to other settings and persist after return to induced by exposure to previous stressors. This model
civilian life. Other determinants of the adaptive value of is supported by studies in rodents and primates showing
the stress response include the intensity, type, and that, compared with controls, animals exposed to brief
timing of the stressor. Stressors that occur early in intermittent stressors during critical developmental
development have greater and more enduring effects windows – an exposure termed ‘‘stress inoculation’’ –
than stressors that occur in adulthood, effects that vary exhibit subsequently enhanced novelty seeking
depending on stressor intensity, duration and frequency behavior, decreased anxiety, and greater stress
(Hoffmann and Spengler, 2012; Russo et al., 2012). tolerance in adulthood (Parker et al., 2004; Russo et al.,
Chronic excessive stress carries a high risk of 2012). In parallel, studies in humans have described the
detrimental consequences, whereas acute intermittent phenomenon of posttraumatic growth (PTG), which is
stress may be essential for successful adaptation to characterized by positive psychological change following
changing natural and social environments (McEwen, exposure to challenging experiences, and which in
2008). Mild to moderate stress may enhance cognitive some cases has been reported to be a more common
and emotional learning via the induction of neural phenotype than PTSD following trauma exposure
plasticity and the adaptive remodeling of neural circuits, (Levine et al., 2008).
and may thus be a necessary component of learning
and successful psychotherapy (Cozolino, 2010). The
Comparing stress in rodents and humans
importance of stressor intensity for determining stress
response has been further highlighted over the years by Rodent models are commonly employed to study the
the definition of Post-Traumatic Stress Disorder (PTSD) molecular basis of stress responses. We refer the
in the Diagnostic and Statistical Manual for Mental interested reader to comprehensive reviews of these
Disorders (DSM). In the 4th edition of DSM, a stressful behavioral paradigms published elsewhere (Krishnan

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
4 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

and Nestler, 2011; Scharf and Schmidt, 2012). Despite these limitations, rodent behavioral models
Vulnerability and resilience are commonly defined in continue to represent the most efficient approach to
rodents based on the degree to which individual animals elucidating the molecular and cellular mechanisms that
with a common genotype or two groups of animals with underlie the etiopathogenesis of psychiatric disorders.
different genotypes show differential outcomes, such as Most importantly, rodent models offer access to brain
social avoidance, anhedonia, or metabolic syndrome, tissue, which is essential for elucidating the circuit basis
following a common exposure to any one of a number of these disorders. In the next section, we review how
of experimentally delivered stressors, such as chronic research from rodent models has helped to provide
social defeat stress, forced swim, or inescapable foot evidence for a brain circuit plasticity-based theory of the
shock (Russo et al., 2012). These paradigms offer a epigenetics of stress responses. We then consider how
strategy both for elucidating molecular hypotheses of this knowledge can be used to inform human studies of
the mechanisms linking stressor exposure with specific HPA axis genes in peripheral tissues to improve the
stress responses, as well as for testing the efficacy of identification and treatment of patients with stress
potential therapeutic drugs. disorders.
Nonetheless, there exist inherent limitations in our
ability to interpret the emotional connotations of rodent
behaviors. Thus it is important to critically evaluate the STRESSOR-INDUCED CHANGES IN NEURONAL
use of ‘‘susceptible’’ (implying maladaptive) and
DNA METHYLATION
‘‘resilient’’ (implying adaptive) to describe behaviors in
experimental rodent models with respect to how findings Methylation of cytosines in the genomic DNA of mammals
from these studies may be applied to psychiatric is a functionally important epigenetic mechanism of
disorders. Misinterpretation of the adaptive value of transcriptional regulation that is essential for cellular
rodent behaviors is particularly problematic when using differentiation and organismal development. In most cell
data from a rodent model to predict stress vulnerabilities types, cytosine methylation occurs predominantly at
in humans. For example, immobility time in the forced CpG dinucleotides, although recent studies have
swim and tail suspension tests is often considered to be revealed that neurons contain a uniquely large
a depressive-like behavior because the ability of drugs complement of methylated cytosines occurring in the
to reduce rodent immobility in these paradigms is a CpH (e.g. CpA, CpC, or CpT dinucleotides) context
strong predictor of their antidepressant efficacy in (Lister et al., 2013; Varley et al., 2013). At a
humans (Cryan et al., 2005). However, immobility in mechanistic level, DNA methylation is a complex
these tests may be also viewed as an active adaptive epigenetic mark with multiple effects on transcription.
response to conserve energy until escape becomes Genome-wide sequencing studies have suggested that
possible. Finally, since the animal has never been on average >70% of all CpG dinucleotides across the
placed in this environment before and cannot anticipate genome are methylated, meaning there are on the order
that it will be removed shortly, a third interpretation is of at least 35 million individual sites of DNA methylation
that floating is a passive coping response to a helpless in a single human cell (Lister et al., 2009). The
situation. As this example highlights, the risk of functional consequences of DNA methylation at any
anthropomorphizing rodent behavior can be limited by given cytosine depend both on the genomic context of
focusing description of rodent behavioral paradigms and the site and whether methylation recruits or blocks the
their outcomes in terms of the face validity (similar recruitment of transcriptional regulatory proteins. When
symptoms), construct validity (similar pathophysiological found in differentially methylated regions (DMRs),
causes), and pharmacological validity (similar drug regions of imprinted genes, or in areas of high CpG
responses) of each test with respect to the psychiatric density, such as CpG islands in gene promoters,
disorder of interest. methylation is usually associated with gene silencing.
A more intractable limitation of conventional rodent This repression can occur either by recruiting the
models is their inability to recapitulate the intricate association of methyl-DNA binding proteins that scaffold
relationship between humans and the stressors they chromatin repressors, or by inhibiting the association of
encounter throughout their lives in complex social transcriptional activators or architectural proteins such
environments. Both genetic and epigenetic factors as CTCF with their regulatory elements (Wang et al.,
influence behaviors and thus may also influence in 2012). The functional consequences of low-density
many cases a person’s choice of social environment cytosine methylation remain less well understood,
and his/her likelihood of encountering stressors, as well though it is likely that DNA methylation within promoters
as impact the person’s response to those environmental and enhancers contributes to developmental regulation
stressors. Furthermore, whereas consistent behavioral of gene expression (Meissner et al., 2008; Mohn et al.,
responses across a population in multiple stress 2008), whereas DNA methylation at sites distant from
paradigms is the usual standard required to transcription start sites has been found in at least some
experimentally implicate a molecular pathway in stress cases to correlate with gene activation (Varley et al.,
responses, human behavior is far more variable, with 2013). Finally, high levels of DNA methylation are found
distinct resilient and vulnerable phenotypes (defined by over actively transcribed gene bodies, where its function
symptoms, social functioning and behavior) often co- remains unknown (Hellman and Chess, 2007; Varley
presenting within the same disorder. et al., 2013).

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx 5

The persistence of cytosine methylation is often cited the hippocampus but not the prefrontal cortex (Roth
as evidence that this chromatin mark could serve as a et al., 2011). The fact that these stressor-induced
form of molecular memory on a behavioral timescale. changes in Bdnf DNA methylation are selective for
DNA methylation is required for the extremely persistent specific brain regions that are physiologically relevant to
(e.g. lifelong) repression of gene transcription that the environmental exposure suggests they could
characterizes transposon silencing, monoallelic contribute to the BDNF-dependent brain plasticities
expression of imprinted genes, and X chromosome underlying behavioral responses to these stressors.
inactivation (Goll and Bestor, 2005). Moreover, cytosine
methylation has an established mechanism of Environmental stressors regulate the writers and
inheritance across cycles of cell division via the activity erasers of cytosine methylation
of the maintenance methyltransferase Dnmt1, which
targets hemimethylated cytosines on newly replicated Despite these correlations, it is not trivial to determine
DNA (Goll and Bestor, 2005). However, if DNA whether stimulus-dependent changes in DNA
methylation were irreversible in postmitotic cells, it methylation are regulatory (i.e. causative for changes in
would have little relevance for mediating behavioral gene transcription and subsequent behavioral plasticity)
adaptations to environmental stimuli. Thus, it was the or merely an epiphenomenon reflecting changes in gene
experimental demonstration that DNA methylation activation. One clue that stressors may act to directly
patterns in the adult brain are dynamic and subject to regulate DNA methylation comes from the analysis of
regulation by environmental stimuli that spurred the the enzymes that regulate the methylation and
most intense interest in neuronal epigenetics (Meaney demethylation of cytosines. In mammals, methyl groups
and Szyf, 2005; Roth and Sweatt, 2009). are added to cytosines by a small group of three
Evidence for rapid, stimulus-dependent regulation of enzymatically active DNA methyltransferases – Dnmt1,
DNA methylation in the adult brain was provided in Miller Dnmt3a, and Dnmt3b (Goll and Bestor, 2005). All three
and Sweatt (2007), who showed that fear conditioning methyltransferases remain expressed in post-mitotic,
increased or decreased DNA methylation in the fully differentiated neurons of the mature brain where
hippocampus at single CpG sites in the plasticity- they could contribute to stimulus-dependent changes in
associated Ppp1ca and Reln genes respectively. These DNA methylation patterns (Feng et al., 2010). Active
DNA methylation changes were detected within 1 h of demethylation of DNA can occur by at least two
fear conditioning and returned to baseline 24 h later, potentially non-exclusive enzymatic processes – base
revealing an unanticipated liability of this chromatin mark excision repair (BER) and oxidation (Niehrs and
(Miller and Sweatt, 2007). Due to the technical Schafer, 2012; Pastor et al., 2013). BER is mediated by
challenges of measuring DNA methylation, most studies, a complex of proteins including the Gadd45 family of
like this one, assay methylation at only a limited number scaffolds, the Aid/APOBEC family of cytosine
of sites. However, it is likely that stimulus-dependent deamidases and the methyl-DNA binding protein Mbd4,
DNA methylation changes simultaneously at a large whereas oxidation of 5-methyl cytosine (5mC) to 5-
number of sites across the genome. For example, using hydroxymethyl cytosine (5hmC) is mediated by the Tet
a sampling method to assess DNA methylation across family of proteins (Tet1–3).
the entire genome, Guo et al. (2011) discovered that Several of these DNA methylation regulatory proteins
1.4% of >200,000 CpGs measured in hippocampal are subject to stressor-dependent changes in expression
extracts showed rapid loss or gain of methylation in the (Fig. 1). For example, LaPlant et al. (2010) showed that
24 h after a period of synchronous neuronal activation 10 days of chronic social defeat stress in mice drove a
induced by electroconvulsive seizure (Guo et al., 2011). small (10–15%) but significant increase in expression of
Determining which, if any, of these changes in DNA Dnmt3a mRNA in the NAc that persisted for at least
methylation have biological consequences for neuronal 10 days following the final defeat. Viral-mediated
plasticity remains challenging. overexpression of Dnmt3a attenuated the social
Stressors are among the environmental stimuli that interaction response in a submaximal defeat stress
can change DNA methylation patterns in the brain, and, paradigm and reduced the latency to immobility in a
as we will review below, various stress paradigms have forced swim test, both of which represent pro-
been shown to decrease methylation of multiple genes depressive-like behaviors; thus, these data suggest that
encoding intermediaries in the HPA axis. Neural the induced expression of Dnmt3a in the NAc after
plasticity genes, such as the gene coding for Brain- chronic social defeat stress contributes to the
Derived Neurotrophic Factor (BDNF), are also targets of development of depressive-like behaviors in response to
regulation by DNA methylation. For example, exposure this stressor (LaPlant et al., 2010). Potential gene
of rat pups to a rodent model of abuse with daily targets of social defeat stress-induced Dnmt3a in the
disruptive caregiving during the first postnatal week is NAc have not yet been identified, however a study from
associated with increased methylation near Bdnf exons an independent group linked chronic social defeat
IV and IX and reduced Bdnf mRNA expression in the stress-induced decreases in mRNA expression of a
prefrontal cortex but not the hippocampus (Roth et al., different methyltransferase, Dnmt3b, with reduced
2009), whereas exposure of adult rats to predator stress methylation of the promoter of the gene coding for
and social instability leads to increased DNA Corticotrophin-Releasing Factor (Crf) in the
methylation and decreased mRNA expression of Bdnf in paraventricular nucleus (PVN) of the hypothalamus

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
6 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

Fig. 1. Molecular mechanisms of stressor-induced regulation and their effects on stress responses. Cytosine methylation is mediated by members
of the DNA methyl transferase (DNMT) family (‘‘writers’’), whereas demethylation is facilitated by proteins including Gadd45b and the Tet family
(‘‘erasers’’). Proteins like MeCP2 that bind methylated cytosines and recruit transcriptional coactivators are the ‘‘readers’’ of this mark. The boxes
show DNA methylation relevant changes that occur following specific stress paradigms. Red boxes represent epigenetic modifications associated
with stress vulnerability, whereas green boxes represent modifications associated with stress resilience. To the right of the boxes are transcriptional
consequences of these chromatin regulatory events. The lack of known transcriptional consequences is shown with question marks. White circles,
unmethylated cytosines, black circles, methylated cytosines. Circled P represents a stimulus-regulated phosphorylation site on MeCP2. NAc,
nucleus accumbens; PVN, paraventricular nucleus of the hypothalamus; LHb, lateral habenula.

(Elliott et al., 2010). This stressor simultaneously induced CUMS (Uchida et al., 2011). Interestingly, both
expression of Gadd45b in the PVN, raising the possibility susceptible and stress-resilient mouse strains showed
that active DNA demethylation along with depressed increased methylation of the Gdnf promoter after
methyltransferase activity could contribute to stressor- CUMS, but in the susceptible strain this increased
induced changes in promoter methylation status. CRF methylation was associated with decreased Gdnf
secreted from neurons of the PVN is a key regulator of expression whereas in the resistant strain it was
the HPA axis and exposure to chronic social defeat associated with increased expression of Gdnf. This
stress in this study was associated with increased Crf difference between the two strains appears to arise from
mRNA expression only in the PVN of mice that showed the distinct complexes of proteins recruited to the sites
stressor-induced social avoidance, again suggesting of promoter methylation. In the susceptible strain the
that the stressor-induced regulation of DNA methylation transcriptional co-repressor Hdac2 was recruited in
in this context was pro-depressive. response to increased promoter methylation, whereas in
Changes in DNA methylation have also been the resistant strain the transcriptional activator CREB
associated with resilience in a different experimental became bound.
stressor paradigm called chronic ultra mild stress
(CUMS). The CUMS paradigm utilizes chronic exposure
Stressor-dependent regulation of the methyl-cytosine
to a series of mild environmental and social stressors
reader MeCP2
(e.g. small cage, paired housing, light exposure, cage
tilt, etc.) that over time induce depressive-like behaviors How could the same DNA methylation event in two mouse
(Uchida et al., 2011). Distinct genetic strains of mice strains induce opposite effects on gene regulation? Unlike
differ in the degree to which they show CUMS-induced the enzymes that serve as ‘‘writers’’ and ‘‘erasers’’ of DNA
depressive-like behaviors, suggesting that this paradigm methylation, which regulate the amount and distribution of
can be used to identify molecular mechanisms of this mark, the methyl-DNA binding ‘‘reader’’ proteins
resilience. Exposure to CUMS for six weeks was found transduce DNA methylation into the local recruitment of
to induce mRNA expression of both Dnmt1 and Dnmt3a transcriptional regulatory complexes. Thus these
but not Dnmt3b in the ventral striatum of a susceptible proteins provide a compelling substrate for encoding
mouse strain. One target of DNA methylation-dependent functional specificity upon methylated DNA (Fatemi and
regulation that correlated with behavior in this study is Wade, 2006). Among the small family of methyl-DNA
the promoter of the gene coding for Glial-Derived binding domain (MBD) proteins, methyl-CpG binding
Neurotrophic Factor (GDNF). Expression of GDNF in protein 2 (MeCP2) is of particular importance in the
the ventral striatum of mice was shown to inversely brain, because loss-of-function mutations in MECP2
correlate with depressive-like behaviors, and cause the severe human neurodevelopmental disorder
overexpression of GDNF in the ventral striatum of the Rett Syndrome (Chahrour and Zoghbi, 2007). MeCP2 is
susceptible strain improved social interaction after widely bound across the neuronal genome in a pattern

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx 7

that closely mirrors the distribution of methylated came in a study from our lab, which examined
cytosines (Skene et al., 2010). Once bound to depressive-like behaviors and antidepressant responses
methylated DNA, MeCP2 acts as a scaffold to recruit in MeCP2 Ser421Ala knock-in mice (Cohen et al., 2011;
the binding of multiple chromatin regulators including Hutchinson et al., 2012b). We found that treating mice
histone deacetylases, histone methyltransferases, the with the antidepressants citalopram and imipramine
NCoR co-repressor complex, and the SWI/SNF family induced Ser421 phosphorylation of MeCP2 in specific
helicase Atrx (Nan et al., 1998; Fuks et al., 2003; Nan parts of the mesolimbic system (Hutchinson et al.,
et al., 2007; Baker et al., 2013; Lyst et al., 2013). If 2012a; Hutchinson et al., 2012b). We hypothesized that
MeCP2 were able to differentially nucleate the formation the slow, but persistent, ability of antidepressants to
of distinct protein complexes under different sets of reverse stress-induced depressive-like behaviors might
cellular conditions, this could provide a mechanism to require the activation of phosphoSer421-MeCP2-
confer differential transcriptional specificities upon dependent gene transcription. To address this question,
common DNA methylation events. Consistent with this we subjected MeCP2 Ser421Ala knock-in mice and their
possibility, recent studies have shown that wild-type littermates to chronic social defeat stress. In
phosphorylation of MeCP2 can regulate its protein– this paradigm, 10d of repeated social defeat induces a
protein interactions (Gonzales et al., 2012; Ebert et al., persistent social avoidance that can be ameliorated by
2013). MeCP2 is a target of stimulus-regulated chronic but not acute imipramine treatment (Berton
phosphorylation at multiple sites, including Ser80, et al., 2006). Both MeCP2 Ser421Ala knock-in mice and
Ser86, Ser274, Thr308, and Ser421 (Zhou et al., 2006; their wild-type littermates showed social avoidance after
Tao et al., 2009; Ebert et al., 2013). The biochemical 10d of defeat. However, whereas the MeCP2 wild-type
consequences of most of these phosphorylation events mice increased their social interaction times after
remain to be fully understood, however phosphorylation chronic imipramine, the MeCP2 Ser421Ala knock-in
of MeCP2 at Thr308 induced by the activation of mice did not, suggesting that MeCP2 phosphorylation is
neuronal intracellular calcium signaling pathways has required to mediate the behavioral plasticities induced
been shown to disrupt the association of MeCP2 with by this antidepressant therapy (Hutchinson et al.,
the NCoR repressor complex, providing a mechanism to 2012b). Interestingly, only chronic but not acute
reduce DNA methylation-dependent repression of a imipramine induced MeCP2 Ser421 phosphorylation in
target gene independent of changes in DNA methylation neurons of the lateral habenula, which is a brain region
itself (Ebert et al., 2013). strongly implicated in the regulation of stress-induced
Evidence that stressors can act via MeCP2 to regulate depressive-like behaviors via its ability to directly
stress responses comes from two studies that have linked modulate the firing of dopaminergic and serotonergic
inducible MeCP2 phosphorylation with stressor-induced neurons in brainstem nuclei (Hikosaka, 2010; Li et al.,
neural adaptations (Fig. 1). In the first study, periodic 2011). This context- and circuit-specific regulation of
infant–mother separation was used to induce early life MeCP2 phosphorylation provides an important piece of
stress (ELS) in postnatal mice (Murgatroyd et al., 2009). evidence linking phospho-Ser421 MeCP2 with
10 days of ELS was sufficient to induce increased behavioral responses to stressors. We propose that
expression of the mRNA encoding the HPA axis factor future studies of MeCP2-dependent transcriptional
arginine vasopressin (AVP) in the PVN of the regulation in the lateral habenula may elucidate new
hypothalamus for up to 1 year. This increase in Avp molecular mechanisms of the neural adaptations that
mRNA expression was associated with decreased CpG underlie antidepressant action.
methylation of an intragenic Avp enhancer and
decreased association of MeCP2 with this enhancer,
EPIGENETIC REGULATION OF THE HPA AXIS
raising the possibility that the persistent change in Avp
expression arose as a result of an MeCP2- and DNA The context- and circuit-specificity of stressor-induced
methylation-dependent process. Interestingly, DNA methylation-dependent regulation of gene
phosphorylation of MeCP2 at Ser421 (called Ser439 in transcription is a key piece of evidence supporting the
this paper with respect to its position in the mouse potential relevance of these changes for the alterations
MeCP2 protein) was substantially elevated in the PVN in brain function that underlie stress-induced behaviors.
of 10-d-old mice that were exposed to ELS compared However, these findings present a major challenge to
with age-matched controls. This study further showed the study of DNA methylation and stress in humans. If
that CaMKII-mediated phosphorylation of MeCP2 at postmortem brain tissue is available, it is unlikely to
Ser421 reduces its association with the Avp enhancer have been obtained in temporal correlation with stressor
and that CaMKII relieves MeCP2-dependent repression exposures or stress responses, limiting the study of how
of Avp in a manner that depends on the phosphorylation chromatin regulation in these tissues relates to the
of MeCP2 at Ser421. On the basis of these data the behavioral context. Furthermore, human epigenetic
authors propose that environmental stressors may act studies often rely on biochemical analysis of peripheral
via the regulation of MeCP2 to effect long-lasting blood cells (PBCs). PBCs can be easily and repeatedly
changes in gene regulation in response to ELS. obtained from human subjects and are excellent source
Evidence that Ser421 phosphorylation of MeCP2 is materials both for genetics and for studying the
functionally required for neural adaptations to stressors epigenetic regulation of immune cell function. However,

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
8 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

these cells have a very limited ability to report the cell 2005; Meaney et al., 2007; Hoffmann and Spengler,
type- and brain region-specific epigenetic regulations of 2012). In humans, both HPA hyper- and hyporeactivity
gene transcription that underlie the adaptive versus have been associated with psychiatric disorders. For
maladaptive nature of behavioral responses to example, HPA hyperreactivity has been associated with
environmental stressors. Nonetheless, given the history of childhood abuse and has been linked to
limitations of modeling many behavioral aspects of vulnerability to stress and heightened risk for the
human stress responses in laboratory animals, there is development of depressive episodes (Heim et al.,
a significant need to find ways in which the study of 2008). On the other hand, excessive suppression of the
epigenetic regulation of gene expression in human cells HPA axis, an effect thought to result from GR
and tissues can enhance understanding of stress hypersensitivity, has been associated with traumatic
vulnerability and resilience. Here we highlight one events and PTSD (Binder, 2009; Morris et al., 2012).
promising area – the epigenetic regulation of HPA axis Taken together, these findings show that homeostasis
genes – both because the regulation of these genes is of the HPA axis is essential for health maintenance and
physiologically relevant in peripheral tissues, and stress resilience, and these data also suggest that the
because their peripheral regulation has the potential to HPA axis reactivity is programed by ELS experiences.
provide a read-out of central neural adaptations in HPA Accumulating evidence suggests that the homeostatic
axis regulation integrated over time. set-point of the HPA axis is influenced by DNA
methylation-dependent regulation of multiple genes
Epigenetic homeostasis of the HPA axis along the axis (Fig. 2). The most studied HPA target of
epigenetic regulation is the gene encoding the GR
The HPA axis plays a central role in coordinating
(Nr3c1) in the hippocampus, and several studies in both
physiological responses to stress in organs throughout
rodents and humans suggest that DNA methylation of
the body. Detection of environmental stressors by the
the Nr3c1 promoter is tightly linked to stress
limbic system induces neurons in the PVN of the
vulnerability and resilience. Hippocampal Nr3c1 is
hypothalamus to produce CRF and AVP. CRF and AVP
upregulated in rats that are either handled or exposed to
act centrally in the brain to trigger secretion of
high maternal care during the first postnatal days, which
adrenocorticotropin hormone (ACTH), which in turn
then mediates enhanced glucocorticoid feedback, long-
leads to peripheral secretion of glucocorticoids from the
term decreased HPA axis responsivity, and stress-
adrenal glands into the blood stream (Chrousos and
resilient phenotypes during adulthood (Meaney et al.,
Gold, 1992). Circulating glucocorticoids reach and exert
1985; Liu et al., 1997). By contrast, ELS or low maternal
effects on essentially every organ, including the brain,
care decreases hippocampal Nr3c1 expression,
via activation of the glucocorticoid receptor (GR)
increases HPA responsivity, and predisposes to stress-
(Nicolaides et al., 2010). The GR is a ligand-dependent
vulnerable phenotypes in adults (Meaney et al., 2007;
transcription factor. In the absence of its ligand, the GR
Mueller and Bale, 2008). In both cases, expression of
forms a cytoplasmic multiprotein complex with protein
the mRNA encoding the GR is inversely correlated with
folding chaperones, including heat shock proteins and
DNA methylation of CpG residues in the Nr3c1
FK506 binding proteins (FKBPs), which regulate GR
promoter that serves as a binding site for the
activity. Following ligation by glucocorticoids, the GR
transcription factor nerve growth factor-inducible protein
dissociates from this complex and translocates to the
A (NGFI-A). Thus, by driving demethylation of the Nr3c1
nucleus, where it regulates transcription of
promoter, which then permits NGFI-A-dependent GR
glucocorticoid-responsive genes. The specific pattern of
expression, high maternal care during a critical period in
transcriptional regulation is distinct for each body organ
the first week of life is thought to be able to determine
and cell type depending on the cell-type specific
the set-point of HPA axis responsivity in adulthood
accessibility of GR binding sites in the genome (Lu and
(Weaver et al., 2004; Mueller and Bale, 2008).
Cidlowski, 2005). In the brain, a key consequence of
Perhaps the most exciting data from these studies are
glucocorticoid-dependent GR activation is the rapid
inhibition of genes encoding intermediates in the HPA the evidence that pharmacological manipulation of the
axis itself (Russell et al., 2010), thus creating a negative epigenome in adults appears to be sufficient to reverse
feedback loop to restrain stressor-induced HPA activity. the behavioral effects of these early life exposures. For
HPA axis homeostasis is important for stress example, administration to adult rats of a histone
resilience and health maintenance, whereas HPA deacetylase inhibitor, which increased GR expression,
dysregulation has been associated with stress was sufficient to reduce the elevated stress responses
vulnerability and disease phenotypes. In rodents, HPA found in those animals that had received low maternal
reactivity is reduced following either high maternal care care as pups. Conversely, administration to adult rats
or brief maternal separation during early life, but in both of L-methionine, which increases Nr3c1 promoter
cases it is associated with phenotypes of decreased methylation and decreases GR expression, increased
fearfulness and anxiety in adulthood (Levine, 2005; stress responses in rats that had received high maternal
Meaney et al., 2007; Hoffmann and Spengler, 2012). By care (Weaver et al., 2004, 2005). This evidence that
contrast, HPA hyperreactivity is observed in rodents pharmacological treatment of adults can reverse the
exposed to prenatal stress or prolonged maternal early epigenetic programming effects of low maternal
separation during early development, and is related to care both at the level of gene expression and at the
heightened fearfulness and anxiety in adults (Levine, level of behavior has profound implications for the

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx 9

Fig. 2. Schematic representation of epigenetic regulation of the hypothalamic–pituitary–adrenal (HPA) axis. The left side of the figure depicts the
epigenetic modifications thought to confer vulnerability and the right side the ones thought to confer resilience. Stimulatory loops between sites of
the axis are shown in blue, whereas inhibitory loops are shown in red. Potential directions for future research are suggested with question marks.
GR, glucocorticoid receptor; HIP, hippocampus; PVN, paraventricular nucleus.

potential of applying our understanding of epigenetic study found that a history of childhood sexual abuse
mechanisms of gene transcription to the treatment of and its severity positively correlated with the methylation
stress disorders in humans. levels of the NR3C1 promoter as measured in PBCs of
adult patients with MDD and borderline personality
disorder (Perroud et al., 2011). Another study examined
Epigenetics of the HPA axis in humans
the effects of maternal depression and anxiety in the
Translatability of these findings from rodents to humans is third trimester of pregnancy on the methylation status of
supported by a number of recent studies. NR3C1 is one of the NR3C1 promoter and HPA responsivity of infants.
the genes that have been studied in postmortem human This study found that NR3C1 promoter methylation in
brains allowing for a direct comparison of neural neonatal leucocytes was directly correlated with prenatal
regulation between rodent experimental models of exposure to greater levels of maternal depression or
environmental stressors and humans. For example, one anxiety and furthermore predicted higher salivary
postmortem study examined the relationship between cortisol levels in response to stressors at 3 months of
childhood abuse and methylation of the human NR3C1 age (Oberlander et al., 2008).
promoter in postmortem brain tissue from suicide In addition to the NR3C1 gene, experimental evidence
victims (McGowan et al., 2009). As compared with non- indicates that additional targets of epigenetic regulatory
abused suicide victims or non-suicide controls, suicide pathways relevant to HPA axis homeostasis are the
victims with history of childhood abuse were found to chaperone molecules involved in intracellular
have both decreased hippocampal GR mRNA glucocorticoid signaling. The most studied member of
expression and increased NR3C1 promoter methylation. this family is the FK506 binding protein 51 (FKBP51/
The GR is expressed in peripheral tissues in addition to FKBP5), which is encoded by the FKBP5 gene.
the brain, and the regulated release of circulating HPA Glucocorticoids induce expression of FKBP5, which in
axis regulatory factors from the hypothalamus provides turn modulates GR sensitivity via an ultra-short negative
a mechanism for epigenetic plasticity of gene regulation feedback loop. As a result, increased FKBP5 activity is
in the brain to be communicated to these peripheral thought to lead to GR resistance and slower recovery of
tissues. If similar epigenetic mechanisms control NR3C1 stress-induced increases in HPA activity (Binder, 2009).
in both the brain and peripheral tissues, then the Four SNPs in the FKBP5 gene were found in one study
NR3C1 gene may be a promising candidate for to interact with the severity of child abuse as a predictor
identifying epigenetic regulatory events relevant to of adult PTSD, suggesting a role for FKBP5 in the
stress responses. Several studies have suggested that gene  environment interactions relevant to stress
similar to regulation of NR3C1 in the brain, increased vulnerabilities (Binder et al., 2008). GR induces the
methylation and decreased expression of this gene in expression of FKPB5 via its association with and
PBCs is also a marker of past environmental stressors regulation of glucocorticoid response elements (GREs)
and maladaptive stress responses. For example, one in distal enhancers that associate with the FKBP5

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
10 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

promoter through a chromatin looping mechanism. One of can differentially affect stress responses via the
the disease-associated FKBP5 SNPs impairs the epigenome and determining to what degree this gene
association of intron 7 GREs with the FKBP5 promoter, regulatory mechanism actually underlies stress
implicating this chromatin interaction in stress vulnerability. We see three major challenges to bridging
vulnerabilities (Klengel et al., 2013). Furthermore, a this gap: (1) developing improved methods to
history of childhood trauma in subjects with this allele experimentally demonstrate the functional importance of
was found to correlate with reduced methylation, as specific epigenetic marks for both gene transcription
measured in PBCs, of CpG dinucleotides in and around and stress responses, (2) characterizing the
the GREs in intron 7 of the FKBP5 gene (Klengel et al., mechanisms of epigenetic regulation in vivo, and (3)
2013). Decreased methylation of this intronic enhancer translating concepts, experimental paradigms, and
is associated with higher induction of FKBP5 expression results between model animal and human studies.
upon GR activation, leading through the FKBP5-
dependent feedback regulation of GR to increased GR Emerging methodologies for the study of the
receptor resistance and a higher HPA axis set-point. epigenome
These results highlight the fact that stress-induced
epigenetic modifications of genes in the HPA axis may The fact that environmental stressors can induce changes
be moderated by genetic polymorphisms and by the in DNA methylation that are both closely correlated with
transcriptional state of stress-related molecules during changes in expression of stress-responsive genes and
critical developmental periods. Notably, a study in mice restricted to physiologically relevant brain regions argue
showed that the DNA methylation levels of Fkbp5 as in favor of their functional importance for behavior. Still,
measured in peripheral blood correlate with the 4-week in most cases whether identified changes in DNA
total glucocorticoid burden independent of any disease- methylation or other chromatin marks mediate changes
associated alleles (Lee et al., 2011). Although the in gene expression or whether they merely reflect those
biological relevance of this study for human psychiatric changes remains unknown. In the case of DNA
patients is limited by use of corticosterone methylation, one of the most significant limitations to
administration as opposed to chronic stress exposure in understanding its functional importance has simply been
the study, these findings suggest that Fkbp5 methylation the challenge of measuring this chromatin mark. The
levels in peripheral blood may be a promising biomarker very first studies characterizing DNA methylation in
for measurement of stress burden. Future studies will neurons genome-wide are only now beginning to appear
need to determine whether Fkbp5 methylation also in the literature (Lister et al., 2013; Varley et al., 2013).
correlates broadly with levels of chronic stress and As understanding of the distribution and regulation of
whether this finding holds true in humans. this mark grows it is likely that new functionally
meaningful models for DNA methylation will emerge.
For example, already the evidence that DNA
CONCLUSIONS, DIRECTIONS AND methylation is highly concentrated over both silenced
gene promoters and active gene bodies suggests that
CHALLENGES FOR FUTURE RESEARCH
the consequences of methylation are context-specific. In
In this review, we have discussed how exposure to addition the recent observation that neurons alone
psychosocial stressors can induce lasting epigenetic among the differentiated mammalian cell types studied
modifications that carry the potential to shape stress to date possess significant methylation of cytosines
responses. Over the past decade, a wealth of studies outside the CpG context raises the possibility that
have revealed that stressors are just one of many neurons use unique mechanisms of chromatin
environmental exposures that can impact the extent, regulation that cannot be discovered in other cell types.
type, and genome-wide distribution of epigenetic marks With affordable next-generation sequencing now
in the brain. The concept that stress vulnerability and/or available to a broad range of researchers, it can be
resilience may result from cumulative, environmentally expected that the speed of chromatin data in the
induced epigenetic marks that persist over time has literature will continue to accelerate, providing a strong
important implications for our understanding of the foundation for the development of new hypotheses of
pathophysiology of psychiatric disorders. For example, epigenetic functions.
studies in humans suggest a ‘‘kindling hypothesis’’ for Beyond mapping DNA methylation and other
the relationship between stressors and MDD, wherein chromatin marks, the expansion of the molecular toolkit
the requirement for potent environmental stressors for experimentally manipulating the epigenome foretells
progressively diminishes with repeated depressive an upcoming period of rapid advances in functional
episodes (Kendler et al., 2000). Although the genomics. One of the most exciting among these
mechanisms of this kindling-like effect have not yet developments is the CRISPR methodology, which co-
been elucidated, it is plausible that stressor-induced opts bacterial proteins that recognize DNA sequences in
epigenetic marks may underlie changes in stress order to experimentally recruit chromatin regulatory
vulnerability over the course of the disorder. If this were enzymes to specific gene regulatory elements (Gaj
the case, then pharmacological treatments directed at et al., 2013; Mali et al., 2013a). The CRISPR system
manipulating the epigenome might become useful tools utilizes the Streptococcus thermophilus protein Cas9,
to impact the course of stress disorders. However, a which is part of an innate immunity system against
vast gulf remains between establishing that stressors attack by bacteriophage. Cas9 has two functional

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx 11

domains, one of which recruits the protein to a specific how epigenetic marks at multiple genes in distinct
DNA sequence using an RNA guide, while the other neural circuits interact, evolve, and accumulate over the
possesses nuclease activity. In early 2013 multiple timecourse of stress disorders in model animals may
research groups published methods harnessing the use cast light on the mechanisms shaping stress
of Cas9 for the site-specific editing of the mammalian vulnerability and guide the development of stage
genome (Cho et al., 2012; Cong et al., 2013; Hwang specific, personalized therapeutic interventions.
et al., 2013; Jiang et al., 2013; Mali et al., 2013b). As we have discussed, epigenetic control
Shortly thereafter several research groups deleted the mechanisms are involved in multiple regulatory events
nuclease domain of Cas9 or replaced it with artificial that maintain homeostasis of the HPA axis. These data
transcription factors to artificially regulate the site- suggest a model in which these epigenetic marks may
specific activation of gene expression from specific be acquired cumulatively over the course of multiple
genomic sites (Cheng et al., 2013; Gilbert et al., 2013; exposures to environmental stressors and act
Perez-Pinera et al., 2013; Qi et al., 2013). Given the synergistically to determine the homeostatic set-point of
ease, speed, and low cost of developing guide RNAs to the HPA axis. Consequently, changes in this set-point
direct Cas9 to any genomic site, and the potential to may confer stress vulnerability or resilience to future
use the modular structure of Cas9 to experimentally stressors. As we note in Fig. 2, more evidence exists on
recruit epigenetic modifiers to gene regulatory elements, epigenetic modifications that confer vulnerability and
CRISPR is very likely to lead to rapid new much less on mechanisms of resilience. By contrast, at
understanding of how epigenetic regulation of the the behavioral level we have discussed evidence
genome impacts gene expression, cellular function, and showing that resilience can result from active processes
ultimately, organismal behaviors. (Russo et al., 2012), which not only prevent the
development of disorders but may also lead to states of
Mechanisms of stress-dependent epigenetic enhanced functioning following stress. This state of
regulation knowledge likely reflects the predilection of research
efforts to examine biological correlations with constructs
Animal models of the epigenetics of stress responses defined by clusters of psychiatric symptoms, rather than
have been most valuable for establishing the molecular studying positive outcomes as endpoints of interest.
and cellular mechanisms by which stressor-induced Furthermore, although symptom-based definitions of
changes in chromatin regulation impact behavioral psychiatric disorders form the basis of current
stress responses (Fig. 1). For example we now know psychiatric diagnostic manuals and guide utilization of
that stressor-inducible plasticity of DNA methylation is health care, it should be borne in mind that psychiatric
mediated at least in part through the regulated disorders are defined phenomenologically and do not
expression, and likely the regulated function and/or necessarily reflect known underlying pathophysiology.
recruitment, of writers and erasers of this chromatin Thus, studies examining dichotomous outcomes, such
mark. The functional effects of stressor-induced as the presence or absence of specific behaviors, might
changes in DNA methylation are context-dependent, not be the most informative when studying a process as
and they can lead to either increases or decreases in nuanced as stress vulnerability and/or resilience. We
gene expression depending on the position of the propose that future studies linking epigenetics and
chromatin modification with respect to gene regulatory stress reponses should focus on active mechanisms
elements and the recruitment or blockade of protein that confer resilience and use continuous behavioral
interactions with the underlying DNA sequence. A stress response measures as outcomes of interest.
further layer of functional specificity is conferred by the
stressor-dependent recruitment and/or post-translational
Understanding the epigenetics of stress in humans
modification of the readers of DNA methylation, as
exemplified by Ser421 phosphorylation of MeCP2. Most Ultimately, both researchers and psychiatrists hope that
importantly, studies in animal models have revealed that epigenetics findings from animal models will be
stressor-induced changes in brain chromatin are highly translatable to the understanding of stress responses
selective for the neural circuits that underlie behavioral and the treatment of stress disorders in humans.
stress responses. These data provide the strongest Sufficient data have been accumulated to demonstrate
causal evidence to date suggesting their direct linkage that exposure to stressors can change epigenetic marks
to the pathophysiology of stress disorders. Using the in humans as well as model animals. However to be
power of mouse genetics to selectively measure and useful for treatment or diagnostic purposes, chromatin
disrupt epigenetic regulation of gene transcription in modifications of specific genes would need to serve as
these functional circuits will be key in revealing the biomarkers – meaning that the epigenetic processes
functional meaning of these marks. One outcome of measured could be said to directly reflect the severity of
these studies could well be the discovery that different the disease process and/or measure the effect of
epigenetic mechanisms in distinct brain regions are at treatment. A major hurdle to achieving this goal is the
play at different developmental periods and stages of fact that gene expression changes underlying mental
any stress disorder. If this were the case, then the illnesses occur within the brain, where biochemical
appropriate therapeutic interventions targeting these analyses of epigenetic marks are not possible in living
mechanisms would need to be varied over the course of humans. In the future the development of molecular
a long-term disorder like MDD. Examining systematically imaging technologies may overcome this limitation by

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
12 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

using methods like positron emission tomography to Chahrour M, Zoghbi HY (2007) The story of Rett syndrome: from
monitor the activation of chromatin regulatory enzymes clinic to neurobiology. Neuron 56:422–437.
Cheng AW, Wang H, Yang H, Shi L, Katz Y, Theunissen TW,
in the intact brain (Wang et al., 2013). However for now,
Rangarajan S, Shivalila CS, Dadon DB, Jaenisch R (2013)
biomarkers detectable in peripheral tissues like blood Multiplexed activation of endogenous genes by CRISPR-on, an
remain the most tractable targets for study. RNA-guided transcriptional activator system. Cell Res.
Genes of the HPA axis currently offer one of the more Cho SW, Kim S, Kim JM, Kim JS (2012) Targeted genome
promising biomarkers for stress responses in humans. engineering in human cells with the Cas9 RNA-guided
The evidence that similar epigenetic mechanisms endonuclease. Nat Biotechnol 31:230–232.
regulate the glucocorticoid receptor gene NR3C1 in both Chrousos GP, Gold PW (1992) The concepts of stress and stress
system disorders. Overview of physical and behavioral
PBCs and the central nervous system suggests that homeostasis. JAMA 267:1244–1252.
blood samples could provide a biologically meaningful Cohen S, Gabel HW, Hemberg M, Hutchinson AN, Sadacca LA,
assessment of processes that contribute to stress Ebert DH, Harmin DA, Greenberg RS, Verdine VK, Zhou Z,
responses in the brain. Because animal studies have Wetsel WC, West AE, Greenberg ME (2011) Genome-wide
shown that Nr3c1 methylation predicts behavioral activity-dependent MeCP2 phosphorylation regulates nervous
responses to future stress, putatively measuring system development and function. Neuron 72:72–85.
Cohen S, Janicki-Deverts D, Miller GE (2007) Psychological stress
epigenetic modifications of this or other HPA axis genes
and disease. JAMA 298:1685–1687.
in blood samples may represent a useful strategy for Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, Hsu PD, Wu X,
identifying individuals at risk of maladaptive stress Jiang W, Marraffini LA, Zhang F (2013) Multiplex genome
responses to environmental stressors. In addition, by engineering using CRISPR/Cas systems. Science 339:819–823.
serving as a read-out of overall HPA axis activity, Cozolino L (2010) The neuroscience of psychotherapy. New
analysis of epigenetic marks on HPA axis genes in the York: W.W. Norton & Co. pp. 20–22.
Cryan JF, Mombereau C, Vassout A (2005) The tail suspension test
periphery could be a useful indicator of the efficacy of a
as a model for assessing antidepressant activity: review of
wide variety of therapies for stress disorders, whether pharmacological and genetic studies in mice. Neurosci
these drugs directly target chromatin regulatory Biobehav Rev 29:571–625.
enzymes or not. Finally, the evidence that at least some Ebert DH, Gabel HW, Robinson ND, Kastan NR, Hu LS, Cohen S,
of the chromatin marks laid down on HPA axis genes by Navarro AJ, Lyst MJ, Ekiert R, Bird AP, Greenberg ME (2013)
ELS can be pharmacologically modified in adult animals Activity-dependent phosphorylation of MECP2 threonine 308
raises the intriguing possibility that epigenetic modifiers regulates interaction with NcoR. Nature [epub ahead of print].
Elliott E, Ezra-Nevo G, Regev L, Neufeld-Cohen A, Chen A (2010)
might be useful modalities for the treatment of stress- Resilience to social stress coincides with functional DNA
related psychiatric disorders in humans. Nonetheless, methylation of the Crf gene in adult mice. Nat Neurosci
the safety and potential usefulness of these emerging 13:1351–1353.
modalities remain to be determined. Fatemi M, Wade PA (2006) MBD family proteins: reading the
epigenetic code. J Cell Sci 119:3033–3037.
Feng J, Zhou Y, Campbell SL, Le T, Li E, Sweatt JD, Silva AJ, Fan G
(2010) Dnmt1 and Dnmt3a maintain DNA methylation and
REFERENCES regulate synaptic function in adult forebrain neurons. Nat
Neurosci 13:423–430.
American Psychiatric Association (2000) Diagnostic and statistical Freed MC, Goldberg RK, Gore KL, Engel CC (2010) Estimating the
manual of mental disorders. 4th ed. Washington, D.C.: American disease burden of combat-related posttraumatic stress disorder in
Psychiatric Association. United States veterans and military service members. In: Preedy
American Psychiatric Association (2013) Diagnostic and statistical VR, Watson RR, editors. Handbook of disease burdens and
manual of mental disorders. 5th ed. Washington, D.C.: American quality of life measures. London, UK: Springer. p. 1527–1548.
Psychiatric Association. Fuks F, Hurd PJ, Wolf D, Nan X, Bird AP, Kourzarides T (2003) The
Baker SA, Chen L, Wilkins AD, Yu P, Lichtarge O, Zoghbi HY (2013) Methyl-CpG-binding protein MeCP2 links DNA methylation to
An AT-hook domain in MeCP2 determines the clinical course of histone methylation. J Biol Chem 278:4035–4040.
Rett syndrome and related disorders. Cell 152:984–996. Gaj T, Gersbach CA, Barbas 3rd CF (2013) ZFN, TALEN, and
Berton O, McClung CA, Dileone RJ, Krishnan V, Renthal W, Russo CRISPR/Cas-based methods for genome engineering. Trends
SJ, Graham D, Tsankova NM, Bolanos CA, Rios M, Monteggia Biotechnol 31:397–405.
LM, Self DW, Nestler EJ (2006) Essential role of BDNF in the Galea S, Nandi A, Vlahov D (2005) The epidemiology of post-
mesolimbic dopamine pathway in social defeat stress. Science traumatic stress disorder after disasters. Epidemiol Rev
311:864–868. 27:78–91.
Binder EB (2009) The role of FKBP5, a co-chaperone of the Gilbert LA, Larson MH, Morsut L, Liu Z, Brar GA, Torres SE, Stern-
glucocorticoid receptor in the pathogenesis and therapy of Ginossar N, Brandman O, Whitehead EH, Doudna JA, Lim WA,
affective and anxiety disorders. Psychoneuroendocrinology Weissman JS, Qi LS (2013) CRISPR-mediated modular RNA-
34(Suppl 1):S186–S195. guided regulation of transcription in eukaryotes. Cell
Binder EB, Bradley RG, Liu W, Epstein MP, Deveau TC, Mercer KB, 154:442–451.
Tang Y, Gillespie CF, Heim CM, Nemeroff CB, Schwartz AC, Goll MG, Bestor TH (2005) Eukaryotic cytosine methyltransferases.
Cubells JF, Ressler KJ (2008) Association of FKBP5 Annu Rev Biochem 74:481–514.
polymorphisms and childhood abuse with risk of posttraumatic Gonzales ML, Adams S, Dunaway KW, LaSalle JM (2012)
stress disorder symptoms in adults. JAMA 299:1291–1305. Phosphorylation of distinct sites in MeCP2 modifies cofactor
Bonasio R, Tu S, Reinberg D (2010) Molecular signals of epigenetic associations and the dynamics of transcriptional regulation. Mol
states. Science 330:612–616. Cell Biol 32:2894–2903.
Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, Guo JU, Ma DK, Mo H, Ball MP, Jang MH, Bonaguidi MA, Balazer JA,
McClay J, Mill J, Martin J, Braithwaite A, Poulton R (2003) Eaves HL, Xie B, Ford E, Zhang K, Ming GL, Gao Y, Song H
Influence of life stress on depression: moderation by a (2011) Neuronal activity modifies the DNA methylation landscape
polymorphism in the 5-HTT gene. Science 301:386–389. in the adult brain. Nat Neurosci.

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx 13

Heim C, Newport DJ, Mletzko T, Miller AH, Nemeroff CB (2008) The Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D, Freedman A,
link between childhood trauma and depression: insights from HPA Sharma S, Pearson D, Plotsky PM, Meaney MJ (1997) Maternal
axis studies in humans. Psychoneuroendocrinology care, hippocampal glucocorticoid receptors, and hypothalamic–
33(6):693–710. pituitary–adrenal responses to stress. Science
Hellman A, Chess A (2007) Gene body-specific methylation on the 277(5332):1659–1662.
active X chromosome. Science 315:1141–1143. Lu NZ, Cidlowski JA (2005) Translational regulatory mechanisms
Hikosaka O (2010) The habenula: from stress evasion to value-based generate N-terminal glucocorticoid receptor isoforms with unique
decision-making. Nat Rev Neurosci 11:503–513. transcriptional target genes. Mol Cell 18(3):331–342.
Hoffmann A, Spengler D (2012) DNA memories of early social life. Lyst MJ, Ekiert R, Ebert DH, Merusi C, Nowak J, Selfridge J, Guy J,
Neuroscience. Kastan NR, Robinson ND, de Lima AlvesF, Rappsilber J,
Hutchinson AN, Deng JV, Aryal DK, Wetsel WC, West AE (2012a) Greenberg ME, Bird A (2013) Rett syndrome mutations abolish
Differential regulation of MeCP2 phosphorylation in the CNS by the interaction of MeCP2 with the NCoR/SMRT co-repressor. Nat
dopamine and serotonin. Neuropsychopharmacology Neurosci 16:898–902.
37:321–337. Mali P, Esvelt KM, Church GM (2013a) Cas9 as a versatile tool for
Hutchinson AN, Deng JV, Cohen S, West AE (2012b) engineering biology. Nat Methods 10:957–963.
Phosphorylation of MeCP2 at Ser421 contributes to chronic Mali P, Yang L, Esvelt KM, Aach J, Guell M, DiCarlo JE, Norville JE,
antidepressant action. J Neurosci 32:14355–14363. Church GM (2013b) RNA-guided human genome engineering via
Hwang WY, Fu Y, Reyon D, Maeder ML, Tsai SQ, Sander JD, Cas9. Science 339:823–826.
Peterson RT, Yeh JR, Joung JK (2013) Efficient genome editing McEwen BS (2008) Central effects of stress hormones in health and
in zebrafish using a CRISPR-Cas system. Nat Biotechnol disease: understanding the protective and damaging effects of
31:227–229. stress and stress mediators. Eur J Pharmacol 583:174–185.
Jiang W, Bikard D, Cox D, Zhang F, Marraffini LA (2013) RNA-guided McGowan PO, Sasaki A, D’Alessio AC, Dymov S, Labonte B, Szyf M,
editing of bacterial genomes using CRISPR-Cas systems. Nat Turecki G, Meaney MJ (2009) Epigenetic regulation of the
Biotechnol 31:233–239. glucocorticoid receptor in human brain associates with
Kaltsas G, Zannas AS, Chrousos GP (2012) Hypothalamic–pituitary– childhood abuse. Nat Neurosci 12:342–348.
adrenal axis and cardiovascular disease. In: Hjemdahl P et al., Meaney MJ, Aitken DH, Bodnoff SR, Iny LJ, Tatarewicz JE, Sapolsky
editors. Stress and cardiovascular disease. London, RM (1985) Early postnatal handling alters glucocorticoid receptor
UK: Springer. p. 71–88. concentrations in selected brain regions. Behav Neurosci
Karatsoreos IN, McEwen BS (2011) Psychobiological allostasis: 99(4):765–770.
resistance, resilience and vulnerability. Trends Cogn Sci Meaney MJ, Szyf M, Seckl JR (2007) Epigenetic mechanisms of
15:576–584. perinatal programming of hypothalamic–pituitary–adrenal function
Karg K, Burmeister M, Shedden K, Sen S (2011) The serotonin and health. Trends Mol Med 13(7):269–277.
transporter promoter variant (5-HTTLPR), stress, and depression Meaney MJ, Szyf M (2005) Maternal care as a model for experience-
meta-analysis revisited: evidence of genetic moderation. Arch dependent chromatin plasticity? Trends Neurosci 28:456–463.
Gen Psychiatry 68:444–454. Meissner A, Mikkelsen TS, Gu H, Wernig M, Hanna J, Sivachenko A,
Kendler KS, Prescott CA (1999) A population-based twin study of Zhang X, Bernstein BE, Nusbaum C, Jaffe DB, Gnirke A,
lifetime major depression in men and women. Arch Gen Jaenisch R, Lander ES (2008) Genome-scale DNA methylation
Psychiatry 56:39–44. maps of pluripotent and differentiated cells. Nature 454:
Kendler KS, Thornton LM, Gardner CO (2000) Stressful life events 766–770.
and previous episodes in the etiology of major depression in Miller CA, Sweatt JD (2007) Covalent modification of DNA regulates
women: an evaluation of the ‘‘kindling’’ hypothesis. Am J memory formation. Neuron 53:857–869.
Psychiatry 157:1243–1251. Mohn F, Weber M, Rebhan M, Roloff TC, Richter J, Stadler MB, Bibel
Klengel T, Mehta D, Anacker C, Rex-Haffner M, Pruessner JC, et al M, Schubeler D (2008) Lineage-specific polycomb targets and de
(2013) Allele-specific FKBP5 DNA demethylation mediates gene- novo DNA methylation define restriction and potential of neuronal
childhood trauma interactions. Nat Neurosci 16:33–41. progenitors. Mol Cell 30:755–766.
Krishnan V, Nestler EJ (2011) Animal models of depression: Morris MC, Compas BE, Garber J (2012) Relations among
molecular perspectives. Curr Top Behav Neurosci 7:121–147. posttraumatic stress disorder, comorbid major depression, and
LaPlant Q, Vialou V, Covington 3rd HE, Dumitriu D, Feng J, et al HPA function: a systematic review and meta-analysis. Clin
(2010) Dnmt3a regulates emotional behavior and spine plasticity Psychol Rev 32(4):301–315.
in the nucleus accumbens. Nat Neurosci 13:1137–1143. Mueller BR, Bale TL (2008) Sex-specific programming of offspring
Lee RS, Tamashiro KL, Yang X, Purcell RH, Huo Y, Rongione M, emotionality after stress early in pregnancy. J Neurosci
Potash JB, Wand GS (2011) A measure of glucocorticoid load 28(36):9055–9065.
provided by DNA methylation of Fkbp5 in mice. Murgatroyd C, Patchev AV, Wu Y, Micale V, Bockmuhl Y, Fischer D,
Psychopharmacology (Berl) 218:303–312. Holsboer F, Wotjak CT, Almeida OF, Spengler D (2009) Dynamic
Levine S (2005) Developmental determinants of sensitivity and DNA methylation programs persistent adverse effects of early-life
resistance to stress. Psychoneuroendocrinology 30:939–946. stress. Nat Neurosci 12:1559–1566.
Levine SZ, Laufer A, Hamama-Raz Y, Stein E, Solomon Z (2008) Nan X, Hou J, Maclean A, Nasir J, Lafuente MJ, Shu X, Kriaucionis S,
Posttraumatic growth in adolescence: examining its components Bird A (2007) Interaction between chromatin proteins MECP2 and
and relationship with PTSD. J Trauma Stress 21:492–496. ATRX is disrupted by mutations that cause inherited mental
Li B, Piriz J, Mirrione M, Chung C, Proulx CD, Schulz D, Henn F, retardation. Proc Natl Acad Sci USA 104:2709–2714.
Malinow R (2011) Synaptic potentiation onto habenula neurons in Nan X, Ng HH, Johnson CA, Laherty CD, Turner BM, Eisenman RN,
the learned helplessness model of depression. Nature Bird A (1998) Transcriptional repression by the methyl-CpG-
470:535–539. binding protein MeCP2 involves a histone deacetylase complex.
Lister R, Mukamel EA, Nery JR, Urich M, Puddifoot CA, et al (2013) Nature 393:386–389.
Global epigenomic reconfiguration during mammalian brain Nicolaides NC, Galata Z, Kino T, Chrousos GP, Charmandari E
development. Science 341:1237905. (2010) The human glucocorticoid receptor: molecular basis of
Lister R, Pelizzola M, Dowen RH, Hawkins RD, Hon G, et al (2009) biologic function. Steroids 75(1):1–12.
Human DNA methylomes at base resolution show widespread Niehrs C, Schafer A (2012) Active DNA demethylation by Gadd45
epigenomic differences. Nature 462:315–322. and DNA repair. Trends Cell Biol 22:220–227.
Lopez AD, Murray CC (1998) The global burden of disease, 1990– Oberlander TF, Weinberg J, Papsdorf M, Grunau R, Misri S, Devlin
2020. Nat Med 4:1241–1243. AM (2008) Prenatal exposure to maternal depression, neonatal

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003
14 A. S. Zannas, A. E. West / Neuroscience xxx (2014) xxx–xxx

methylation of human glucocorticoid receptor gene (NR3C1) and anxiety on subsequent alcohol relapse and treatment outcomes.
infant cortisol stress responses. Epigenetics 3:97–106. Arch Gen Psychiatry 68:942–952.
Parker KJ, Buckmaster CL, Schatzberg AF, Lyons DM (2004) Skelton K, Ressler KJ, Norrholm SD, Jovanovic T, Bradley-Davino B
Prospective investigation of stress inoculation in young (2012) PTSD and gene variants: new pathways and new thinking.
monkeys. Arch Gen Psychiatry 61:933–941. Neuropharmacology 62:628–637.
Pastor WA, Aravind L, Rao A (2013) TETonic shift: biological roles of Skene PJ, Illingworth RS, Webb S, Kerr AR, James KD, Turner DJ,
TET proteins in DNA demethylation and transcription. Nat Rev Andrews R, Bird AP (2010) Neuronal MeCP2 is expressed at near
Mol Cell Biol 14:341–356. histone-octamer levels and globally alters the chromatin state.
Perez-Pinera P, Kocak DD, Vockley CM, Adler AF, Kabadi AM, Mol Cell 37:457–468.
Polstein LR, Thakore PI, Glass KA, Ousterout DG, Leong KW, Sullivan PF, Neale MC, Kendler KS (2000) Genetic epidemiology of
Guilak F, Crawford GE, Reddy TE, Gersbach CA (2013) RNA- major depression: review and meta-analysis. Am J Psychiatry
guided gene activation by CRISPR-Cas9-based transcription 157:1552–1562.
factors. Nat Methods 10:973–976. Tao J, Hu K, Chang Q, Wu H, Sherman NE, Martinowich K, Klose RJ,
Perroud N, Paoloni-Giacobino A, Prada P, Olie E, Salzmann A, Schanen C, Jaenisch R, Wang W, Sun YE (2009)
Nicastro R, Guillaume S, Mouthon D, Stouder C, Dieben K, Phosphorylation of MeCP2 at Serine 80 regulates its chromatin
Huguelet P, Courtet P, Malafosse A (2011) Increased methylation association and neurological function. Proc Natl Acad Sci USA.
of glucocorticoid receptor gene (NR3C1) in adults with a history of Tsolaki M, Kounti F, Karamavrou S (2009) Severe psychological
childhood maltreatment: a link with the severity and type of stress in elderly individuals: a proposed model of
trauma. Transl Psychiatry 1:e59. neurodegeneration and its implications. Am J Alzheimers Dis
Qi LS, Larson MH, Gilbert LA, Doudna JA, Weissman JS, Arkin AP, Other Demen 24:85–94.
Lim WA (2013) Repurposing CRISPR as an RNA-guided platform Uchida S, Hara K, Kobayashi A, Otsuki K, Yamagata H, Hobara T,
for sequence-specific control of gene expression. Cell Suzuki T, Miyata N, Watanabe Y (2011) Epigenetic status of Gdnf
152:1173–1183. in the ventral striatum determines susceptibility and adaptation to
Ripke S, Wray NR, Lewis CM, Hamilton SP, Weissman MM, et al daily stressful events. Neuron 69:359–372.
(2013) A mega-analysis of genome-wide association studies for van Winkel R, Stefanis NC, Myin-Germeys I (2008) Psychosocial
major depressive disorder. Mol Psychiatry 18:497–511. stress and psychosis. A review of the neurobiological
Risch N, Herrell R, Lehner T, Liang KY, Eaves L, Hoh J, Griem A, mechanisms and the evidence for gene–stress interaction.
Kovacs M, Ott J, Merikangas KR (2009) Interaction between the Schizophr Bull 34:1095–1105.
serotonin transporter gene (5-HTTLPR), stressful life events, and Varley KE, Gertz J, Bowling KM, Parker SL, Reddy TE, Pauli-Behn F,
risk of depression: a meta-analysis. JAMA 301:2462–2471. Cross MK, Williams BA, Stamatoyannopoulos JA, Crawford GE,
Roth TL, Lubin FD, Funk AJ, Sweatt JD (2009) Lasting epigenetic Absher DM, Wold BJ, Myers RM (2013) Dynamic DNA
influence of early-life adversity on the BDNF gene. Biol Psychiatry methylation across diverse human cell lines and tissues.
65:760–769. Genome Res 23:555–567.
Roth TL, Sweatt JD (2009) Regulation of chromatin structure in Wang C, Schroeder FA, Hooker JM (2013) Visualizing epigenetics:
memory formation. Curr Opin Neurobiol 19:336–342. current advances and advantages in HDAC PET imaging
Roth TL, Zoladz PR, Sweatt JD, Diamond DM (2011) Epigenetic techniques. Neuroscience.
modification of hippocampal Bdnf DNA in adult rats in an animal Wang H, Maurano MT, Qu H, Varley KE, Gertz J, Pauli F, Lee K,
model of post-traumatic stress disorder. J Psychiatr Res Canfield T, Weaver M, Sandstrom R, Thurman RE, Kaul R, Myers
45:919–926. RM, Stamatoyannopoulos JA (2012) Widespread plasticity in
Russell GM, Henley DE, Leendertz J, Douthwaite JA, Wood SA, CTCF occupancy linked to DNA methylation. Genome Res
Stevens A, Woltersdorf WW, Peeters BW, Ruigt GS, White A, 22:1680–1688.
Veldhuis JD, Lightman SL (2010) Rapid glucocorticoid receptor- Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S,
mediated inhibition of hypothalamic–pituitary–adrenal ultradian Seckl JR, Dymov S, Szyf M, Meaney MJ (2004) Epigenetic
activity in healthy males. J Neurosci 30(17):6106–6115. programming by maternal behavior. Nat Neurosci 7(8):
Russo SJ, Murrough JW, Han MH, Charney DS, Nestler EJ (2012) 847–854.
Neurobiology of resilience. Nat Neurosci 15:1475–1484. Weaver IC, Champagne FA, Brown SE, Dymov S, Sharma S,
Sartor CE, Grant JD, Lynskey MT, McCutcheon VV, Waldron M, Meaney MJ, Szyf M (2005) Reversal of maternal programming of
Statham DJ, Bucholz KK, Madden PA, Heath AC, Martin NG, stress responses in adult offspring through methyl
Nelson EC (2012) Common heritable contributions to low-risk supplementation: altering epigenetic marking later in life. J
trauma, high-risk trauma, posttraumatic stress disorder, and Neurosci 25(47):11045–11054.
major depression. Arch Gen Psychiatry 69:293–299. Zhou Z, Hong EJ, Cohen S, Zhao WN, Ho HY, Schmidt L, Chen WG,
Scharf SH, Schmidt MV (2012) Animal models of stress vulnerability Lin Y, Savner E, Griffith EC, Hu L, Steen JA, Weitz CJ, Greenberg
and resilience in translational research. Curr Psychiatry Rep ME (2006) Brain-specific phosphorylation of MeCP2 regulates
14:159–165. activity-dependent Bdnf transcription, dendritic growth, and spine
Sinha R, Fox HC, Hong KI, Hansen J, Tuit K, Kreek MJ (2011) Effects maturation. Neuron 52:255–269.
of adrenal sensitivity, stress- and cue-induced craving, and

(Accepted 3 December 2013)


(Available online xxxx)

Please cite this article in press as: Zannas AS, West AE. Epigenetics and the regulation of stress vulnerability and resilience. Neuroscience (2014),
http://dx.doi.org/10.1016/j.neuroscience.2013.12.003

You might also like