epigenetic regulation of rac1 induces synaptic remodeling in stress disorders and depression

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s

nature medicine advance online publication �

The treatment of major depressive disorder (MDD) is confounded by high rates of treatment resistance coupled with low probabilities of achieving lasting remission1. Such clinical realities, paired with an estimated depression-related economic burden in excess of $80 billion annually2,3, has led to a concerted effort to identify and develop new alternative therapeutic approaches on the basis of validated disease mechanisms to treat depression and related mood disorders. Although the monoamine hypothesis of depression4 has dominated depression treatment strategies, recent work suggests that the neuropathology of depression is stratified across a number of biological domains, such as inflammatory cytokines5, neurotrophic factors6 and glutamate7. One common mechanism linking these divergent systems is that they all robustly regulate excitatory synaptic structure8–10 and have been shown functionally to play a part in the etiology of depression and anxiety-like behavior6,11.

Although it is clear that synaptic loss occurs in cortical regions in depression12, a major question has been whether depression is caused by underlying synaptic alterations in subcortical reward cir-cuits. Evidence in rodent models shows that chronic stress regulates synaptic and structural plasticity in reward-related limbic regions13. Recent work from our laboratory showed that chronic social defeat stress induces structural and functional synaptic plasticity on

medium spiny neurons (MSNs) in the NAc13,14, a brain region known for controlling reward-related dysfunction such as social avoid-ance and anhedonia in depression15. Although the phenomenon of stress-induced structural plasticity has been observed across multiple species and stress models, the intracellular mechanisms underlying these neuronal alterations and their relevance to human depression are poorly understood.

Under nonpathological conditions, small Rho GTPases, such as RAC1, act as crucial modulators of synaptic structure. Here we report that RAC1 is transcriptionally downregulated through an epigenetic mechanism in the NAc of subjects with MDD, an effect that is reca-pitulated in the chronic social defeat stress model of depression-like behaviors. Functionally, Rac1 modulates depression-related behaviors such as social avoidance and anhedonia, as well as MSN structural plasticity. Overexpression of Rac1 by either viral gene therapy or epigenetic rescue has strong anti-stress efficacy in previously stress-susceptible mice.

RESULTSTranscriptional regulation of Rac1 in mice after social defeatPathway analyses of gene ontology groups from microarray data commonly identify cytoskeleton-related and small–Rho GTPase

1Fishberg Department of Neuroscience and Friedman Brain Institute, Graduate School of Biomedical Sciences at the Icahn School of Medicine at Mount Sinai, New York, New York, USA. 2Laboratory of Molecular Biology, The Rockefeller University, New York, New York, USA. 3Department of Brain and Cognitive Sciences, McGovern Institute, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA. 4Depressive Disorders Program, Douglas Mental Health University Institute and McGill University, Montréal, Québec, Canada. 5Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, Texas, USA. Correspondence should be addressed to S.J.R. ([email protected]).

Received 11 October 2012; accepted 14 January 2013; published online 17 February 2013; doi:10.1038/nm.3090

Epigenetic regulation of RAC1 induces synaptic remodeling in stress disorders and depressionSam A Golden1, Daniel J Christoffel1, Mitra Heshmati1, Georgia E Hodes1, Jane Magida1, Keithara Davis1, Michael E Cahill1, Caroline Dias1, Efrain Ribeiro1, Jessica L Ables2, Pamela J Kennedy1, Alfred J Robison1, Javier Gonzalez-Maeso1, Rachael L Neve3, Gustavo Turecki4, Subroto Ghose5, Carol A Tamminga5 & Scott J Russo1

Depression induces structural and functional synaptic plasticity in brain reward circuits, although the mechanisms promoting these changes and their relevance to behavioral outcomes are unknown. Transcriptional profiling of the nucleus accumbens (NAc) for Rho GTPase–related genes, which are known regulators of synaptic structure, revealed a sustained reduction in RAS-related C3 botulinum toxin substrate 1 (Rac1) expression after chronic social defeat stress. This was associated with a repressive chromatin state surrounding the proximal promoter of Rac1. Inhibition of class 1 histone deacetylases (HDACs) with MS-275 rescued both the decrease in Rac1 transcription after social defeat stress and depression-related behavior, such as social avoidance. We found a similar repressive chromatin state surrounding the RAC1 promoter in the NAc of subjects with depression, which corresponded with reduced RAC1 transcription. Viral-mediated reduction of Rac1 expression or inhibition of Rac1 activity in the NAc increases social defeat–induced social avoidance and anhedonia in mice. Chronic social defeat stress induces the formation of stubby excitatory spines through a Rac1-dependent mechanism involving the redistribution of synaptic cofilin, an actin-severing protein downstream of Rac1. Overexpression of constitutively active Rac1 in the NAc of mice after chronic social defeat stress reverses depression-related behaviors and prunes stubby spines. Taken together, our data identify epigenetic regulation of RAC1 in the NAc as a disease mechanism in depression and reveal a functional role for Rac1 in rodents in regulating stress-related behaviors.

http://www.nature.com/doifinder/10.1038/nm.3090

http://www.nature.com/naturemedicine/

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s

� advance online publication nature medicine

signaling pathways as being highly modulated by stress and depression-like behavioral paradigms, both in the NAc16,17 and other limbic regions18,19. This is consistent with a previously described phe-nomenological role of structural plasticity in stress and depression8,13. Small GTPases are responsible, to a large extent, for the regulation and reorganization of the actin cytoskeleton in a number of biologi-cal systems, including neuronal populations during synaptogenesis and dendritic spine maintenance20. They are modulated by upstream regulators such as guanine nucleotide exchange factors and GTPase-activating proteins, leading ultimately to the severing of the underly-ing actin cytoskeleton by cofilin20,21.

Although a number of preclinical rodent models of depression have been developed, models of prolonged social stress have been useful in revealing the molecular and cellular mechanisms underlying mood- or anxiety-related disorders22,23. In this model, male C57BL/6J mice are repeatedly subjected to bouts of social defeat by a larger and more aggressive CD-1 mouse, resulting in a majority of the smaller mice (termed susceptible) developing a syndrome highlighted by enduring social avoidance and anhedonia to previously pleasurable activities. Non-susceptible (resilient) mice, which do not show the social avoid-ance and anhedonic response, generally comprise one-third of the mice subjected to social defeat24. Therefore, the behavioral syndrome induced by social defeat allows for a detailed dissection of individual variations in the molecular mechanisms that underlie stress-related behavioral phenotypes. It is important to note that social defeat stress results in the development of anxiety-like behaviors in both the sus-ceptible and resilient populations, reinforcing the likelihood that

this model captures components of depression, post-traumatic stress disorder or even social anxiety16.

On the basis of the lines of evidence outlined above, we measured regulation of Rho GTPase–related genes in the NAc of susceptible and resilient mice after social defeat (Fig. 1; RT-PCR primer sequences are listed in Supplementary Table 1). Although we assessed 13 Rho GTPase–related genes, only Rac1 showed selective stress-induced transcriptional regulation in susceptible but not resilient mice (Fig. 1a). Forty-eight hours after the final defeat episode, Rac1 expression was strongly downregulated (Fig. 1b) and correlated with social avoidance behavior (Fig. 1c and Supplementary Fig. 1a–p). We did not generally observe the selective downregulation of Rac1 in susceptible mice in other brain structures. For example, in the prefrontal cortex (PFC), Rac1 mRNA was unaffected by social defeat stress, and in the dorsal striatum, Rac1 mRNA levels were decreased in both susceptible and resilient mice (Supplementary Table 2). Deletion of Rac1 in the PFC or dorsal striatum had no effect on social avoidance behavior (Supplementary Table 2).

Within the NAc, downregulation of Rac1 expression and social avoidance behavior were both long lasting (≥35 d) and partially reversible by chronic antidepressant treatment (20 mg per kg body weight imipramine intraperitoneally (i.p.) once daily for 35 d) (Fig. 1d and Supplementary Fig. 2a,b). A single acute injection of imipramine after chronic social defeat stress neither regulated Rac1 mRNA expression in the NAc (Fig. 1g) nor reversed social avoidance in susceptible mice (Supplementary Fig. 2c,d). Downregulation of Rac1 by chronic social defeat stress was associated with a decrease

e

Rac1

Gapdh

0

0.5

1.0

1.5

Rac

1 pr

otei

n fo

ld c

hang

e

5

*

8 7

Kalrn

Farp2

Tiam

1

Arhgef

6

Arhgef

7Vav

1Vav

3Rac

1Rac

2Rac

3Pak

1Lim

k Cfl

mR

NA

fold

cha

nge

a

Upstream Downstream

0

1

2

*

Control (5)Susceptible (8)Resilient (8)

4 24 480.5

1.0

1.5

2.0

*

Time (h)

Rac

1 m

RN

A fo

ld c

hang

e

b

Control (5–15)Susceptible (10–15)Resilient (5–16)

0.6 1.2 1.8 2.40

0.5

1.0

1.5

2.0

Social interaction ratio0

c


r = 0.289P = 0.027

0

0.5

1.0

1.5

0.6 1.2 1.8 2.4

Social interaction ratio

0


fr = 0.431P = 0.058

Control Susceptible0

0.5

1.0

1.5

2.0 Chronic vehicle Chronic imipramine

Rac

1 m

RN

A fo

ld c

hang

e

**

d

5 7 5 7

Rac

1 m

RN

A fo

ld c

hang

e

1 d 10 d

g

0

0.5

1.0

1.5 ControlDefeat

Social defeat stress

Acute vehAcute imi

15 15 5 5

Figure 1 Chronic social defeat stress decreases Rac1 mRNA expression in the NAc of susceptible but not resilient mice. (a) mRNA expression of the indicated small Rho GTPase–related genes in the NAc of C57BL/6J mice 48 h after their last social defeat exposure. Susceptible and resilient mice were subjected to chronic social defeat before social interaction testing, whereas control mice were housed identically but never exposed to social defeat sessions. (b) Temporal profile of Rac1 mRNA expression in the NAc 4, 24 and 48 h after chronic social defeat stress. (c) Correlation of Rac1 mRNA expression with social avoidance behavior at 48 h after chronic social defeat stress. (d) Rac1 mRNA expression after 35 d of imipramine or vehicle. (e) Total Rac1 protein expression in the NAc 48 h after chronic social defeat stress normalized to Gapdh (glyceraldehyde-3-phosphate dehydrogenase). (f) Correlation of total Rac1 protein with social avoidance behavior at 48 h after chronic social defeat. (g) Rac1 mRNA expression in the NAc 48 h after a single acute social defeat stress (left) and 48 h after acute injection of imipramine (imi; 20 mg per kg body weight i.p.) or vehicle (veh) in previously susceptible mice (right). All data are presented as the mean ± s.e.m.; group sizes are indicated in parentheses in each legend or within the graph bars. *P < 0.05 determined by Student’s t test (g), one-way analysis of variance (ANOVA) (a,b,e) or two-way ANOVA (d). The P values in c and f were determined by Pearson’s r.

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s


in the total amounts of Rac1 protein in the NAc of susceptible, but not resilient, mice 48 h after the final defeat session (Fig. 1e). There was a nonsignificant trend of Rac1 protein correlating with social avoidance behavior (Fig. 1f). The mRNA and protein levels of down-stream effectors of the Rac1 signaling cascade, such as p21-activated kinase 1 (Pak1) and LIM-domain containing protein kinase (Limk), were not affected by social defeat at similar time points (Fig. 1a and Supplementary Fig. 3a–c). Notably, a single social defeat episode did not alter Rac1 mRNA expression (Fig. 1g), indicating that only prolonged or severe stressors that induce marked depression-like behaviors are capable of inducing downregulation of Rac1.

Epigenetic regulation of Rac1 in mice after social defeatGenome-wide promoter analysis using chromatin immunoprecipi-tation (ChIP) paired to microarray analysis has recently identified increased repressive histone H3 Lys9 trimethylation (H3K9me3) and H3K27me3 in the NAc of mice along the promoter region of Rac1 after social isolation stress25. On the basis of our finding that Rac1 transcription is downregulated for extended periods after chronic social defeat stress, we next explored the possibility that this tran-scriptional decrease is controlled through an epigenetic mechanism. In our mouse brain tissue ChIP assay we sonicated genomic DNA, resulting in 350-bp to 750-bp fragments26. We designed primer pairs

(Supplementary Table 3) sequentially aligned from the promoter region and extending ~2,000 bp upstream of the transcription start site (TSS) of Rac1 (Fig. 2a). Using site-directed quantitative ChIP (qChIP), we examined permissive histone H3 acetylation (acH3) and repressive H3K27me3 in susceptible, resilient and control mice populations (behavioral data are shown in Supplementary Fig. 4). It is important to note that, unlike acH3, methylation at H3K27 has been shown to be either repressive or permissive under varying biological conditions. However, recent evidence suggests that in the adult mouse NAc, levels of H3K27me3 within the gene promoter region correlate with repressed gene expression27. Across both the promoter and the 2,000-bp upstream region, permissive acetylation was significantly reduced in the susceptible but not resilient population (Fig. 2b). Conversely, methylation was differentially modulated at the proxi-mal promoter and the immediate 1,000-bp upstream region; within the promoter region, resilient mice showed reduced methylation, and directly upstream of the promoter, susceptible mice had enhanced methylation (Fig. 2c).

As H3 acetylation was reduced in susceptible mice, we administered the class 1 HDAC inhibitor MS-275 locally into the NAc for 10 d by osmotic minipump (Fig. 2d). This resulted in a robust reversal of social avoidance in previously susceptible mice (Fig. 2e). Notably, Rac1 mRNA levels were normalized (Fig. 2f) and positively correlated

Figure 2 Epigenetic regulation of Rac1 after chronic social defeat stress. (a) A schematic of the mouse Rac1 gene promoter and upstream regulatory region. Purple uppercase text indicates the promoter region, and green lowercase text indicates the untranslated region upstream of the TSS. Blue and yellow boxes represent the promoter primer pair locations sequentially aligned throughout the sequence. (b,c) Profiles of permissive H3 acetylation (b) and repressive H3K27me3 (c) along the mouse Rac1 promoter and upstream gene sequence in the NAc 48 h after chronic social defeat stress. A pan-acH3 antibody recognizes all acetylation events along the H3 tail. Susceptible and resilient mice were subjected to chronic social defeat before social interaction testing, whereas control mice were housed identically but never exposed to social defeat sessions. (d) Schematic detailing the experimental protocol for chronic intra-NAc minipump infusion of the class 1 HDAC inhibitor (HDACi) MS-275. CSDS, chronic social defeat stress; Veh, vehicle. (e,f) Social interaction ratio (e) and NAc Rac1 mRNA expression (f) in susceptible mice treated with chronic MS-275 (100 µm) or vehicle. (g) Correlation of social avoidance behavior and NAc Rac1 mRNA expression 24 hours after social interaction testing. All mice were subjected to chronic social defeat stress, and susceptible mice were implanted with osmotic minipumps containing either MS-275 or vehicle. All data are presented as the mean ± s.e.m.; group sizes are indicated in parentheses in each legend or within the graph bars. *P < 0.05, **P < 0.01, ***P < 0.001 by one-way ANOVA (b,c) or Student’s t test (e,f). NC, no change. The P value shown in g was determined by Pearson’s r.

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s


with social interaction (Fig. 2g), suggesting that there is indeed a functional link between epigenetic regulation of Rac1 and social defeat behavior.

Regulation of RAC1 in subjects with MDDWe sought to corroborate the epigenetic mechanism found in our chronic social defeat model in postmortem NAc tissue derived from subjects with MDD (Fig. 3a). We used samples from two inde-pendently collected MDD cohorts, one from the University of Texas Southwestern Medical Center in Texas and a second from McGill University in Montreal. Although all subjects in both cohorts had been on antidepressant medication at some point in their life, roughly half of the subjects in each cohort were on medication at their time of death, as determined by toxicology reports (Supplementary Tables 4 and 5). As observed in our rodent model, RAC1 expression was strongly downregulated in the NAc in nonmedicated subjects with MDD from both cohorts (Fig. 3b). Results for the medicated subjects with MDD were somewhat mixed. In the Texas cohort, there were no significant differences in RAC1 mRNA levels between medicated and nonmedi-cated subjects; however, in the larger Montreal cohort, although the trend for reduced RAC1 mRNA levels was still evident, there was a bimodal distribution, with a subset of subjects having RAC1 mRNA levels similar to those of healthy, age-matched, nonmedicated control subjects. These findings are consistent with our mouse model, which showed that decreased Rac1 mRNA levels in susceptible mice were only normalized by chronic antidepressant treatment in ~50% of the mice (4 of 7 mice had social interactions ratios above 1.0; Supplementary Fig. 2a). In addition, NAc RAC1 mRNA levels were not reduced in a separate human cohort of individuals with a history of strong cocaine

abuse, suggesting that the effect is at least partly specific to MDD (Supplementary Fig. 5). These data are consistent with previous work28 showing that chronic cocaine administration in mice transiently decreases NAc Rac1 activity with no effect on Rac1 gene transcription. Thus, there is compelling evidence to suggest that RAC1 downregula-tion is not an artifact of addictive drugs or antidepressant medication, and is not a general marker of all psychiatric illnesses.

To examine chromatin structure around the RAC1 gene, we adapted our qChIP protocol for use in frozen human postmortem NAc tissue samples from subjects in the Texas cohort using a micro-coccal nuclease–based assay29, which reliably results in 146-bp to 165-bp fragments (Supplementary Fig. 6a,b). We used this method to create a high-resolution map of the human RAC1 promoter and the 1,000-bp upstream regulatory region by designing closely grouped sequentially aligned primer pairs (Supplementary Table 6) and site-directed qChIP (Fig. 3c). We selected this sequence length because of our observation that modifications in the Rac1 chromatin land-scape in mouse NAc tissue were observed mostly within the first 1,000 bp upstream of the TSS. We found decreased acetylation of the RAC1 gene sequences ~200 bp upstream and downstream of the TSS (Fig. 3d), whereas we found increased methylation only in the region ~200 bp upstream of the TSS (Fig. 3e). In each subject, we directly

Figure 3 Epigenetic regulation of RAC1 mRNA expression in subjects with MDD. (a) Coronal and sagittal schematic of the human brain highlighting regions of NAc dissection in subjects with MDD. Scale bar, 10 mm. (b) RAC1 mRNA expression in the NAc of subjects with MDD in two separate cohorts. The numbers of samples per cohort are shown in the table (for complete demographics, see Supplementary Tables 4 and 5). Control samples were obtained from healthy, age-matched patients with no previous documentation of depression or antidepressant medication use. (c) Schematic of the human RAC1 gene promoter region and its ~1,000-bp upstream sequence. Purple uppercase text indicates the promoter region, and green lowercase text indicates the untranslated region upstream of the TSS. Blue and yellow boxes represent the promoter primer pair locations sequentially aligned throughout the sequence. (d,e) Profile of permissive H3 acetylation (d) and repressive H3K27me3 (e) along the human RAC1 promoter and upstream sequence in the NAc. (f,g) Correlation of human NAc H3 acetylation (f) and human NAc H3K27me3 (g) levels with RAC1 mRNA expression. All data are presented as the mean ± s.e.m.; group sizes are indicated in parentheses in each legend. *P < 0.05, **P < 0.01 by one-way ANOVA (b) or Student’s t test (d,e). The P values shown in f and g were determined by Pearson’s r.

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s


correlated RAC1 mRNA expression with pro-moter acetylation (Fig. 3f) and methylation (Fig. 3g) status. acH3 enrichment ~200 bp upstream of the TSS significantly (P = 0.0004, r = 0.668) correlated with RAC1 mRNA levels. A reanalysis of the qChIP data using medica-tion status at time of death as a cofactor revealed no effect of medica-tion status on either acetylation or methylation along the upstream or downstream region of the RAC1 gene (Supplementary Fig. 7a,b).

Rac1 modulates depression-related behaviorNext we directly examined the role of Rac1 expression in the NAc in regulating depression-like behavior after either social defeat stress or a subthreshold microdefeat (Fig. 4). Unlike chronic social defeat, microdefeats do not result in the development of any social avoidance or anhedonia in microdefeated, but otherwise unmanipulated, control mice, which can help reveal pro-susceptibility factors.

To manipulate Rac1, we used herpes simplex virus (HSV)-mediated gene transfer of bicistronic constructs containing GFP and either dominant-negative Rac1 (RacDN), constitutively active Rac1 (RacCA) or Cre recombinase infused directly into the NAc (Fig. 4a,b). Viral-mediated overexpression of HSV-GFP-Cre in the NAc of floxed Rac1 (Rac1flox/flox) mice30 induced local knockdown of Rac1 expression (Supplementary Fig. 8a) and induced suscep-tibility to social avoidance behavior (Fig. 4c,d and Supplementary Fig. 9a,b) and anhedonia after a microdefeat, as measured by reduced sucrose preference (Fig. 4e). Similarly, HSV-mediated over-expression of RacDN recapitulated this pro-susceptibility phenotype (Fig. 4f,g and Supplementary Fig. 9c,d). Conversely, HSV-mediated overexpression of RacCA to increase Rac1 expression after chronic social defeat rescued the susceptible phenotype and restored social behavior to those observed in HSV-GFP–infected susceptible mice. Unstressed control mice were housed identically to defeated mice

but were never subjected to defeat before social interaction testing (Fig. 4h,i and Supplementary Fig. 9e,f).

Rac1 controls stress-induced formation of excitatory spinesChronic social defeat stress induces the formation of immature stubby dendritic spines on MSNs in the NAc13,14. These dendritic altera-tions are associated with a smaller postsynaptic density and a greatly increased frequency of excitatory drive that seems to be necessary and sufficient for the expression of depression-related phenotypes in mice13,14. To understand the mechanism leading to the induction of immature stubby dendritic spines on NAc MSNs, we first examined cofilin, a crucial downstream Rac1 target that is important for actin severing and has been shown to regulate immature spine structure31. We examined NAc MSN dendritic segments 48 h after chronic social defeat stress through a combinatorial approach using automated three-dimensional dendritic spine reconstruction from HSV-GFP–infected neurons overlaid with cofilin immunohistochemistry (Fig. 5a,b and Supplementary Fig. 10a). Chronic social defeat stress resulted in a selective increase in cofilin-positive immature stubby spines, but not with thin or mushroom spines or with total spine density (Fig. 5c). Within the tissue set used for immunohistochemical analysis, we also found a replication of previously published data13, where chronic social defeat stress resulted in an increase in total stubby spine den-sity on MSN dendrites (Supplementary Fig. 10b).

Next we examined whether overexpression of RacCA after chronic stress would be sufficient to prune away stubby spines (Fig. 5d). We injected HSV vectors after 10 d of chronic social defeat stress and

HSV-RacCA(n = 6/group)

HSV-Rac

DN

NAccNAcs

0246

4 2 0 –2 –4 –6 –8Bregma, A/P +1.10 mm

acaNAcs

aChronic social defeat stress

b

Microdefeat stress

HSV-GFP HSV-GFP-Cre0

50

100

150T

ime

in in

tera

ctio

n zo

ne (

s)

CSDS10 d

Social interactiontesting

Surgery

Recovery3 d

Microdefeat1 d

Socialinteraction

testingSurgery

Recovery3 d

NAcc

0

f

20

40

60

80

100

0

0.5

1.0

1.5

Soc

ial i

nter

actio

n ra

tio

HSV-GFP HSV-GFP-Cre0

20

40

60

80

Per

cent

age

sucr

ose

HSV-GFP HSV-GFP-Cre

c Rac1flox/flox

(n = 5–6/group)

HSV-RacDN(n = 12/group)

0

0.5

1.0

1.5

Soc

ial i

nter

actio

n ra

tio

Tim

e in

inte

ract

ion

zone

(s)

HSV-GFP

HSV-Rac

CA0

h

20

40

60

80

100

HSV-GFPT

ime

in in

tera

ctio

n zo

ne (

s)

d eTarget absentTarget present

**

****

**

g

0

0.5

1.0

1.5

Soc

ial i

nter

actio

n ra

tio

i

*

*

** *

*

HSV-Rac

DN

HSV-GFP

HSV-Rac

CA

HSV-GFP

Figure 4 Rac1 expression in the NAc modulates stress-related behavior in mice. (a) Representative schematic of the mouse NAc region and HSV expression. Left, GFP-labeled neurons in green show the spread of viral infection, which is limited to the NAc region. Scale bar, 50 µm. NAcs, NAc shell; NAcc, NAc core; aca, anterior commissure. Right, GFP-infected neurons express GFP in the distal dendrites of MSNs. Scale bar, 5 µm. (b) Schematic of chronic social defeat stress and subthreshold microdefeat protocols detailing the timing for viral infection and behavioral testing. CSDS, chronic social defeat stress. (c–e) Social avoidance behavior (c,d) and sucrose preference (e) in Rac1flox/flox mice subjected to subthreshold microdefeat after deletion of Rac1 by HSV-Cre or HSV-GFP in the NAc. (f,g) Social avoidance behavior in C57BL/6J mice subjected to subthreshold microdefeat after expression of a dominant negative Rac1 mutant (HSV-RacDN) or HSV-GFP in the NAc. (h,i) Social avoidance behavior of previously susceptible C57BL/6J mice injected with constitutively active Rac1 mutant (HSV-RacCA) or HSV-GFP in the NAc. All data are presented as the mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001 by two-way ANOVA (c,f,h) or Student’s t test (d,e,g,i).

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s


found that overexpression of RacCA reduced the density of stubby spines in susceptible mice (Fig. 5e). There was a nonsignificant trend toward a decrease in the number of thin spines after overexpression of RacCA (Fig. 5f), suggesting that RacCA may selectively target imma-ture spine structures. We found no effect of social defeat or Rac1 overexpression on the number of mushroom spines (Fig. 5g) or total dendritic spine density (Fig. 5h). Notably, social avoidance behavior was negatively cor-related with stubby spine density (Fig. 5i) but not with thin spine density (Fig. 5j). To deter-mine whether Rac1 is sufficient for immature dendritic spine formation on medium spiny neurons in the NAc, we performed dendritic

spine morphology analysis in conditional Rac1flox/flox mice after sub-threshold microdefeat stress (Fig. 6a). Deletion of NAc Rac1 by local-ized viral-mediated transfer of HSV-GFP-Cre increased both thin and stubby spine density in both nonmicrodefeated control mice and

e

HSV-GFP

HSV-Rac

CA0

0.1

0.2

0.3

0.4

0.5

Stu

bby

dens

ity µ

m–1 **

f

HSV-GFP

HSV-Rac

CA0

0.25

0.50

0.75

1.00

Thi

n de

nsity

µm

–1

0

0.2

0.4

0.6

Msu

hroo

m d

ensi

ty µ

m–1

HSV-GFP

HSV-Rac

CA

g

0

0.5

1.0

1.5

2.0

Tot

al d

ensi

ty µ

m–1

HSV-GFP

HSV-Rac

CA

hj

0 0.5 1.0 1.5 2.0 2.50

0.5

1.0

1.5


Thi

n de

nsity

µm

–1 r = –0.31P < 0.25

HSV-GFP, control

HSV-GFP, susceptible

HSV-RacCA, control

HSV-RacCA, susceptible

da

b

Control

Susceptible

GFP

GFP

Cofilin

Cofilin

Merge

Merge0

0.5

1.0

1.5

2.0

Per

cent

age

cofil

inpu

ncta

/con

trol

Stubb

yThin

Mus

hroo

mTot

al

*ControlSusceptible

c

i Susceptible, HSV-GFPControl, HSV-RacCASusceptible, HSV-RacCA

Control, HSV-GFP

0 0.5 1.0 1.5 2.0 2.50

0.2

0.4

0.6


Stu

bby

dens

ity µ

m–1

r = –0.72P < 0.01

Figure 5 Expression of constitutively active Rac1 (HSV-RacCA) in the NAc prevents chronic social defeat stress–induced stubby spine formation. (a,b) Representative ×63 confocal z-stack images of MSN dendritic segments from control and susceptible mice after chronic social defeat. Control mice were housed identically to susceptible mice before social interaction testing but were never subjected to social defeat. White arrows in the GFP images indicate dendritic stubby spines. White circles in the merge images highlight colocalization of GFP stubby spines and cofilin puncta, and yellow circles indicate non-colocalized stubby spines. Scale bar, 10 µm. (c) Percentage of cofilin+ stubby, thin, mushroom and total dendritic spines in the NAc. *P < 0.05 by Students t test. n = 48 neurons from 4–5 mice per group. (d) Representative three-dimensional reconstructions of HSV-infected MSN dendritic segments. White arrows indicate stubby spines, yellow arrows indicate mushroom spines, and blue arrows indicate thin spines. Scale bars, 5 µm. (e–i) Stubby (e), thin (f), mushroom (g) or total number of dendritic spines (h) in mice expressing HSV-RacCA or HSV-GFP in NAc MSNs after chronic social defeat stress. Correlation of social avoidance with stubby dendritic spine density (i) and thin dendritic spine density (j). *P < 0.05 by two-way ANOVA (e–h). n = 61 neurons from 3–5 mice per group. The P values in i and j were determined by Pearson’s r. All data are presented as the mean ± s.e.m.

HSV-GFP, control

HSV-GFP, microdefeat

Merge

HSV-GFP-Cre, control

HSV-GFP-Cre, microdefeat

a

HSV-GFP

0

0.2

0.4

0.6

Mus

hroo

m d

ensi

ty µ

m–1

Tot

al d

ensi

ty µ

m–1

d e

0 0.5 1.0 1.5 2.0 2.50

0.5

1.0

1.5


f

Stu

bby

dens

ity µ

m–1

r = –0.665P = 0.005

g

0 0.5 1.0 1.5 2.0 2.50

0.5

1.0

1.5


Thi

n de

nsity

µm

–1

r = –0.739P = 0.001

HSV-GFP-

Cre

HSV-GFP

0

1.0

2.0

3.0

HSV-GFP-

Cre

****

c

Thi

n de

nsity

µm

–1

HSV-GFP

HSV-GFP-

Cre

0

0.5

1.0

1.5 ***

HSV-GFP

HSV-GFP-

Cre

Stu

bby

dens

ity µ

m–1

b

0

0.5

1.0

1.5

******

MicrodefeatControl

Microdefeat, HSV-GFPControl, HSV-GFP-CreMicrodefeat, HSV-GFP-Cre

Control, HSV-GFP

Figure 6 Conditional deletion of Rac1 in NAc of Rac1flox/flox mice increases stubby and thin dendritic spines. (a) Representative three-dimensional reconstructions of HSV-infected MSN dendritic segments. White arrows indicate stubby spines, yellow arrows indicate mushroom spines, and blue arrows indicate thin spines. Scale bars, 5 µm. (b–e) Stubby (b), thin (c), mushroom (d) and total (e) spine density in mice after Rac1 gene deletion in the NAc. Control mice did not undergo subthreshold microdefeat stress, receiving only experimental HSV-GFP-Cre or HSV-GFP viral constructs. (f,g) Correlation of stubby (f) and thin (g) spine density with social avoidance behavior. All data are presented as the mean ± s.e.m. *P < 0.05, ***P < 0.001 by two-way ANOVA (b–e). n = 60 neurons from 4 mice per group. The P values in f and g were determined by Pearson’s r.

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s


microdefeated mice (Fig. 6b,c) but had no effect on mushroom spine density (Fig. 6d). We observed an additive effect on stubby spine density, where acute subthreshold microdefeat almost doubled stubby spine density in Rac1flox/flox mice injected with HSV-GFP-Cre (Fig. 6b). Overall, we observed an increase in total spine density (Fig. 6e). Both stubby and thin spine density correlated with social interaction ratio (Fig. 6f,g). Although future studies will need to examine these morphological changes in human postmortem tissue from subjects with MDD, new methodologies and techniques will need to be devel-oped for this purpose.

DISCUSSIONOur data suggest that chronic stress or depression reduces the expres-sion of RAC1 in the NAc of rodents and humans through an epige-netic mechanism. The expression of Rac1 is normalized by chronic imipramine treatment in mice after chronic social defeat and is posi-tively correlated with treatment response, suggesting that this adapta-tion may account for some of the effects of antidepressant treatment. However, in the majority of subjects with MDD on antidepressants at the time of death, RAC1 expression and chromatin modifications along the RAC1 promoter and upstream regions were not restored to those seen in control subjects, suggesting a need for more direct RAC1-targeting strategies to achieve therapeutic effects in this patient population. Functionally in mice, selective genetic deletion of Rac1 and viral-mediated gene transfer of Rac1 mutants or epigenetic modu-lation through class 1 HDAC inhibition revealed that Rac1 signaling in the NAc is both necessary and sufficient for social avoidance and anhe-donia behavioral responses and synaptic structural plasticity. Together these data suggest that small–Rho GTPase signaling pathways, and RAC1 specifically, are viable candidates for future development as antidepressant therapeutic agents when targeted to the NAc.

Indeed, small Rho GTPases have been implicated in the neu-ropathogenesis of several psychiatric disorders, including fragile X mental retardation32, Rett syndrome33, schizophrenia34 and sub-stance abuse28,35. Under nonpathological conditions, RAC1 has an important role in neuronal development36, axon guidance37 and learning and memory processes38. With regard to dendritic spines, as a general rule, increased synaptogenesis and functional synaptic plasticity are tightly correlated with the size and shape of a dendritic spine; as spines shift along a continuum from immature (stubby and thin) to mature (mushroom), they also shift toward greater synaptic strength and stability39,40. Within neuronal populations, there is a wealth of evidence supporting a role for RAC1 in dendritic spine morphogenesis and maintenance34,41–43 through modulation of cofilin spatiotemporal dynamics44,45.

This signaling cascade is positioned to regulate the delicate inter-face between extracellular stimuli and dynamic reorganization of the actin cytoskeleton to accommodate neuronal transduction. The transcriptional downregulation of this system consequently has an important role in setting the threshold for subsequent experience-dependent plasticity within the NAc. Our data suggest that stress-induced decreases in RAC1 expression result in concurrent increases in immature stubby spine formation and cofilin localization within these spines. Recent findings have confirmed that decreased p21-activated kinase 3 (PAK3) activity, an immediate downstream effecter of RAC1, leads to both an increase in the growth of new, immature spines and an impairment of plasticity-mediated spine stabiliza-tion that interferes with the formation of persistent stable spines46. A feature of chronic stress exposure is a shift toward synaptic instability8,47, and future treatments could be designed to increase the

synaptic stability of neurons within affected brain circuits. This ‘plas-ticity consolidation’ treatment strategy10 may act through a number of Rho GTPase–dependent mechanisms, and future studies will need to determine the optimal drug target.

Although there are clearly many acute biochemical events acting directly in the synapse that may mediate synaptic plasticity within the NAc, the data here point to a more sustained mechanism through epigenetic-mediated downregulation of RAC1, as a result of stressful experience, to cause social avoidance and anhedonia and synaptic restructuring events in the NAc. Although histone deactelyase inhibi-tors17 and histone demethylases48 have shown antidepressant efficacy, they more globally affect gene expression across the entire genome. Our data imply that the antidepressant efficacy of such epigenetic regulators is in part through their actions on RAC1. However, recent advances in gene-specific targeting of chromatin modifications using zinc finger artificial transcription factors49 allow epigenetic therapeu-tics to act on specific genes to reverse long-term chromatin disrup-tions. These may be used to clarify whether epigenetic modification at specific genes, such as Rac1, are necessary and sufficient to promote social avoidance and anhedonia. However, until these tools are more widely available for testing in humans, a more feasible therapeutic strategy may be to screen compounds that target either RAC1 activity or, more generally, excitatory synapse consolidation50.

We provide evidence that chronic stress induces long-term transcriptional downregulation of RAC1 through an epigenetic mechanism, robustly regulating NAc MSN synaptic structural and behavioral plasticity. Reversal of this endophenotype may present a new therapeutic venue for exploration and guide future therapeutics to target adaptive intracellular plasticity mechanisms.

METHODSMethods and any associated references are available in the online version of the paper.

Note: Supplementary information is available in the online version of the paper.

ACKNowLEDGMENTSWe thank Y. Zeng (University of Cincinnati) for providing Rac1flox/flox mice. This research was supported by US National Institute of Mental Health grant R01 MH090264, the Johnson and Johnson International Mental Health Research Organization Rising Star Award (S.J.R.), US National Institute of Mental Health grants P50 MH66172 and P50MH096890 (C.A.T.) and US National Institutes of Health grant T32 GM062754 (K.D.).

AUTHoR CoNTRIBUTIoNSS.A.G. and S.J.R. contributed to study design, data collection, analysis and writing. D.J.C., M.H., G.E.H., J.M., K.D., M.E.C., C.D., E.R., A.J.R., P.J.K. and J.L.A. contributed to data collection. J.G.-M. contributed to data analysis. R.L.N. contributed to viral design and packaging. G.T., S.G. and C.A.T. collected human brain samples and contributed to data analysis.

CoMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.

Published online at http://www.nature.com/doifinder/10.1038/nm.3090. Reprints and permissions information is available online at http://www.nature.com/reprints/index.html.

1. Krishnan, V. & Nestler, E.J. Linking molecules to mood: new insight into the biology of depression. Am. J. Psychiatry 167, 1305–1320 (2010).

2. Greenberg, P.E. et al. The economic burden of depression in the United States: how did it change between 1990 and 2000? J. Clin. Psychiatry 64, 1465–1475 (2003).

3. Kessler, R.C., Chiu, W.T., Demler, O., Merikangas, K.R. & Walters, E.E. Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Arch. Gen. Psychiatry 62, 617–627 (2005).





http://www.nature.com/reprints/index.html

http://www.nature.com/reprints/index.html

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

a r t i c l e s


4. Tissot, R. The common pathophysiology of monaminergic psychoses: a new hypothesis. Neuropsychobiology 1, 243–260 (1975).

5. Smith, R.S. The macrophage theory of depression. Med. Hypotheses 35, 298–306 (1991).

6. Duman, R.S., Heninger, G.R. & Nestler, E.J. A molecular and cellular theory of depression. Arch. Gen. Psychiatry 54, 597–606 (1997).

7. Palucha, A. & Pilc, A. The involvement of glutamate in the pathophysiology of depression. Drug News Perspect. 18, 262–268 (2005).

8. Christoffel, D.J., Golden, S.A. & Russo, S.J. Structural and synaptic plasticity in stress-related disorders. Rev. Neurosci. 22, 535–549 (2011).

9. Capuron, L. & Miller, A.H. Immune system to brain signaling: neuropsychopharma-cological implications. Pharmacol. Ther. 130, 226–238 (2011).

10. Manji, H.K. et al. Enhancing neuronal plasticity and cellular resilience to develop novel, improved therapeutics for difficult-to-treat depression. Biol. Psychiatry 53, 707–742 (2003).

11. Vidal, R. et al. New strategies in the development of antidepressants: towards the modulation of neuroplasticity pathways. Curr. Pharm. Des. 17, 521–533 (2011).

12. Kang, H.J. et al. Decreased expression of synapse-related genes and loss of synapses in major depressive disorder. Nat. Med. 18, 1413–1417 (2012).

13. Christoffel, D.J. et al. IκB kinase regulates social defeat stress-induced synaptic and behavioral plasticity. J. Neurosci. 31, 314–321 (2011).

14. Christoffel, D.J. et al. Effects of inhibitor of κB kinase activity in the nucleus accumbens on emotional behavior. Neuropsychopharmacology 37, 2615–2623 (2012).

15. Nestler, E.J. & Carlezon, W.A. Jr. The mesolimbic dopamine reward circuit in depression. Biol. Psychiatry 59, 1151–1159 (2006).

16. Krishnan, V. et al. Molecular adaptations underlying susceptibility and resistance to social defeat in brain reward regions. Cell 131, 391–404 (2007).

17. Covington, H.E. III et al. Antidepressant actions of histone deacetylase inhibitors. J. Neurosci. 29, 11451–11460 (2009).

18. Wei, Q. et al. Early-life forebrain glucocorticoid receptor overexpression increases anxiety behavior and cocaine sensitization. Biol. Psychiatry 71, 224–231 (2012).

19. Andrus, B.M. et al. Gene expression patterns in the hippocampus and amygdala of endogenous depression and chronic stress models. Mol. Psychiatry 17, 49–61 (2012).

20. Tolias, K.F., Duman, J.G. & Um, K. Control of synapse development and plasticity by Rho GTPase regulatory proteins. Prog. Neurobiol. 94, 133–148 (2011).

21. Kiraly, D.D., Eipper-Mains, J.E., Mains, R.E. & Eipper, B.A. Synaptic plasticity, a symphony in GEF. ACS Chem. Neurosci. 1, 348–365 (2010).

22. Berton, O. et al. Essential role of BDNF in the mesolimbic dopamine pathway in social defeat stress. Science 311, 864–868 (2006).

23. Nestler, E.J. & Hyman, S.E. Animal models of neuropsychiatric disorders. Nat. Neurosci. 13, 1161–1169 (2010).

24. Golden, S.A., Covington, H.E. III, Berton, O. & Russo, S.J. A standardized protocol for repeated social defeat stress in mice. Nat. Protoc. 6, 1183–1191 (2011).

25. Wilkinson, M.B. et al. Imipramine treatment and resiliency exhibit similar chromatin regulation in the mouse nucleus accumbens in depression models. J. Neurosci. 29, 7820–7832 (2009).

26. Kumar, A. et al. Chromatin remodeling is a key mechanism underlying cocaine-induced plasticity in striatum. Neuron 48, 303–314 (2005).

27. Feng, J. et al. Cocaine induced transcriptome and epigenome changes in mouse nucleus accumbens. Soc. Neurosci. Abstr. 458.12/T7 (2012).

28. Dietz, D.M. et al. Rac1 is essential in cocaine-induced structural plasticity of nucleus accumbens neurons. Nat. Neurosci. 15, 891–896 (2012).

29. Jiang, Y., Matevossian, A., Huang, H.-S., Straubhaar, J. & Akbarian, S. Isolation of neuronal chromatin from brain tissue. BMC Neurosci. 9, 42 (2008).

30. Chen, L., Melendez, J., Campbell, K., Kuan, C.Y. & Zheng, Y. Rac1 deficiency in the forebrain results in neural progenitor reduction and microcephaly. Dev. Biol. 325, 162–170 (2009).

31. Pontrello, C.G. et al. Cofilin under control of β-arrestin-2 in NMDA-dependent dendritic spine plasticity, long-term depression (LTD), and learning. Proc. Natl. Acad. Sci. USA 109, E442–E451 (2012).

32. Bongmba, O.Y., Martinez, L., Elhardt, M., Butler, K. & Tejada-Simon, M. Modulation of dendritic spines and synaptic function by Rac1: a possible link to Fragile X syndrome pathology. Brain Res. 1399, 79–95 (2011).

33. Chen, L.Y. et al. Physiological activation of synaptic Rac1 PAK (p-21 activated kinase) signaling is defective in a mouse model of fragile X syndrome. J. Neurosci. 30, 10977–10984 (2010).

34. Hayashi-Takagi, A. et al. Disrupted-in-Schizophrenia 1 (DISC1) regulates spines of the glutamate synapse via Rac1. Nat. Neurosci. 13, 327–332 (2010).

35. Golden, S.A. & Russo, S.J. Mechanisms of psychostimulant-induced structural plasticity. Cold Spring Harb. Perspect. Med. 2, a011957 (2012).

36. de Curtis, I. Functions of Rac GTPases during neuronal development. Dev. Neurosci. 30, 47–58 (2008).

37. Koh, C.-G. Rho GTPases and their regulators in neuronal functions and development. Neurosignals 15, 228–237 (2006).

38. Martinez, L.A. & Tejada-Simon, M. Pharmacological inactivation of the small GTPase Rac1 impairs long-term plasticity in the mouse hippocampus. Neuropharmacology 61, 305–312 (2011).

39. Holtmaat, A. & Svoboda, K. Experience-dependent structural synaptic plasticity in the mammalian brain. Nat. Rev. Neurosci. 10, 647–658 (2009).

40. Yoshihara, Y., De Roo, M. & Muller, D. Dendritic spine formation and stabilization. Curr. Opin. Neurobiol. 19, 146–153 (2009).

41. Luo, L. et al. Differential effects of the Rac GTPase on Purkinje cell axons and dendritic trunks and spines. Nature 379, 837–840 (1996).

42. Nakayama, A.Y., Harms, M.B. & Luo, L. Small GTPases Rac and Rho in the maintenance of dendritic spines and branches in hippocampal pyramidal neurons. J. Neurosci. 20, 5329–5338 (2000).

43. Tashiro, A. & Yuste, R. Regulation of dendritic spine motility and stability by Rac1 and Rho kinase: evidence for two forms of spine motility. Mol. Cell. Neurosci. 26, 429–440 (2004).

44. Kuhn, T.B. et al. Regulating actin dynamics in neuronal growth cones by ADF/cofilin and rho family GTPases. J. Neurobiol. 44, 126–144 (2000).

45. Rex, C.S. et al. Different Rho GTPase–dependent signaling pathways initiate sequential steps in the consolidation of long-term potentiation. J. Cell Biol. 186, 85–97 (2009).

46. Dubos, A. et al. Alteration of synaptic network dynamics by the intellectual disability protein PAK3. J. Neurosci. 32, 519–527 (2012).

47. Li, N. et al. mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists. Science 329, 959–964 (2010).

48. LaPlant, Q. et al. Dnmt3a regulates emotional behavior and spine plasticity in the nucleus accumbens. Nat. Neurosci. 13, 1137–1143 (2010).

49. Stolzenburg, S., Bilsland, A., Keith, W.N. & Rots, M.G. Modulation of gene expression using zinc finger–based artificial transcription factors. Methods Mol. Biol. 649, 117–132 (2010).

50. Russo, S.J., Murrough, J.W., Han, M.H., Charney, D.S. & Nestler, E.J. Neurobiology of resilience. Nat. Neurosci. 15, 1475–1484 (2012).

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicinedoi:10.1038/nm.3090

ONLINE METHODSAnimals. Male C57BL/6J mice (20–30 g; The Jackson Laboratory) were obtained at 7–8 weeks of age and allowed 1 week of acclimation to the hous-ing facilities before the start of experiments. Male Rac1flox/flox were obtained from Y. Zeng at the University of Cincinnati, Cincinnati, Ohio, USA and back-crossed onto a C57BL/6J background. Rac1flox/flox mice (20–30 g) were bred at the Mount Sinai School of Medicine and used at 7–8 weeks of age. Male CD-1 mice (35–45 g; Charles River Laboratories) used as aggressors were sexually experienced retired breeders at least 4 months of age. Aside from use in chronic social defeat experiments, all mice were single housed and maintained on a 12-h light, 12-h dark cycle with ad libitum access to food and water. Behavioral assessments and tissue collection were performed 1 h into the dark phase. Mouse procedures were performed in accordance with the Institutional Animal Care and Use Committee guidelines of the Mount Sinai School of Medicine.

Chronic social defeat stress. Chronic social defeat stress was performed as previously described16,22,24. CD-1 mice were screened for aggressive charac-teristics before the start of social defeat experiments on the basis of previously described criteria24 and housed within the social defeat cage (26.7 cm width × 48.3 cm depth × 15.2 cm height; Allentown Inc) 24 h before the start of defeats on one side of a clear, perforated Plexiglass divider (0.6 cm × 45.7 cm × 15.2 cm; Nationwide Plastics). Briefly, experimental C57BL/6J mice were subjected to a new CD-1 aggressor mouse for 10 min once per day over 10 consecutive days. After the 10 min of agonistic interaction, the experimental mice were removed to the opposite side of the social defeat cage and allowed sensory contact during the next 24 h period. Control mice were housed two mice per cage on opposite sides of the perforated divider and rotated daily in a manner similar to the defeat group but never exposed to aggressive CD-1 mice. Twenty-four hours after the final social defeat stress, the experimental C57BL/6J mice were singly housed. In a variation on this protocol, for tissue collected at more immediate time points (4 and 24 h) after social defeat, we exposed mice to nine consecutive defeats, performed behavioral testing for social avoidance and then performed the final tenth defeat before mice were euthanized and their brains removed at the necessary time point.

Microdefeat stress. To measure increased susceptibility to stress, we adapted a subthreshold variation on the chronic social defeat protocol, as previously described13,24. Using this protocol, experimental C57BL/6J mice were sub-jected to a new CD-1 aggressor for three consecutive 5-min defeat bouts, each separated by 15 min. Twenty-four hours later, the mice were assessed for social avoidance behavior. Under control conditions (when a naive C57BL/6J mouse is subjected to subthreshold microdefeat in the absence of other manipulations), this protocol does not result in social avoidance behavior but is sensitive to pro-susceptibility factors.

Social avoidance testing (social interaction test). Social interaction test-ing was performed as previously described22,24. All social interaction testing was performed under red-light conditions. Mice were placed in a new interac-tion open-field arena custom crafted from opaque Plexiglas (42 cm × 42 cm × 42 cm; Nationwide Plastics) with a small animal cage placed at one end. Their movements were then automatically monitored and recorded (Ethovision 3.0, Noldus Information Technology) for 2.5 min in the absence (target absent phase) of a new CD-1 mouse. This phase was used to determine baseline exploratory behavior. We then immediately measured 2.5 min of exploratory behavior in the presence of a caged CD-1 mouse (target present phase), again recording total distance traveled and duration of time spent in the interaction and corner zones. Social interaction behavior was then calculated as total time spent in each zone or as a ratio of the time spent in the interaction with target present divided by the time spent in the interaction zone with the target absent. All mice with a ratio above 1.0 were classified as resilient, and all mice below 1.0 were classified as susceptible.

Sucrose preference testing. To determine whether mice acquired anhedonic responses to any experimental manipulations, we performed a standard sucrose preference assay. Immediately after the final chronic social defeat stress session

or the final microdefeat bout, mice had their standard water bottle removed and replaced with two 50-ml conical tubes with sipper tops filled with water. After a 24-h habituation period, water from one 50-ml conical tube was replaced with 1% sucrose. All tubes were weighed, and mice were allowed 24 h to drink. Tubes were then reweighed, and their locations in the wire tops were switched before a second 24-h period of drinking. At the end of the second day of sucrose testing, sucrose preference was calculated as the total amount of sucrose consumption divided by the total amount of fluid consumed over the 2 d of sucrose availability.

Imipramine treatment. After social interaction testing, mice were sorted into either susceptible or resilient phenotypes on the basis of their interaction scores as described above. Control mice were housed identically to defeated mice for the duration of the experiment but were never exposed to social defeat sessions. Resilient mice were excluded from imipramine experiments, and susceptible and control mice were randomly divided into treatment groups. For each group (control or susceptible), mice received either daily i.p. injections of imipramine (20 mg per kg body weight) for 35 d or vehicle. All mice were killed 24 h after the last imipramine or vehicle injection, and NAc dissections were obtained by punch dissection as described previously25. Briefly, bilateral 14-guage NAc tissue punches were taken from 1-mm-thick coronal brain slices immediately after decapitation. Tissue punches were immediately placed on dry ice before being stored at −80 °C until being processed for quantitative PCR.

Transcriptional profiling. NAc samples were collected and processed as described previously25. Briefly, bilateral 14-gauge NAc, medial PFC or dorsal striatum punches were collected on ice after rapid decapitation and then imme-diately placed on dry ice and stored at −80 °C until use. We then isolated RNA using TRIzol (Invitrogen) homogenization and chloroform layer separation. The clear RNA layer was then processed (RNAeasy MicroKit, Qiagen) and analyzed with NanoDrop (Thermo Scientific). Five-hundred nanograms of RNA was then reverse transcribed to cDNA (qScript Kit, VWR). For quantitative PCR, cDNA was diluted to 500 µl, and 3 µl was used for each reaction. The reaction mixture consisted of Perfecta SYBR Green (VWR) (5 µl), forward and reverse primers (0.5 µl each), water (1 µl) and the cDNA template. Samples were then heated to 95 °C for 2 min followed by 40 cycles of 95 °C for 15 s, 60 °C for 33 s and 72 °C for 33 s. Analysis was done using the ∆∆C(t) method51. Samples were normal-ized to GAPDH. Primer pairs are shown in Supplementary Table 1 (Integrated DNA Technologies).

Western blotting. Bilateral NAc punches, taken as described above, were sonicated in a detergent-based lysis buffer (50 mM Tris-HCl, pH 7.5, 2 mM MgCl2, 150 mM NaCl, 1% Triton X-100, protease inhibitor cocktail (Roche), 1 µM sodium vanadate, 30 mM sodium pyrophosphate and 30 mM sodium fluoride) containing phosphatase and protease inhibitors (Sigma). Samples were then centrifuged for 30 min at 14,000 r.p.m. at 4 °C, and the superna-tant was collected and quantified with the Microplate BCA Protein Assay Kit (Thermo Scientific). Equal amounts of protein sample were separated by SDS-PAGE (Bio-Rad 4–20% polyacrylamide gels) and transferred to nitrocel-lulose membranes (Bio-Rad). The membranes were blocked in 5% milk and incubated overnight with primary antibodies in 5% BSA or milk. The primary antibodies used were to the following proteins: Rac1 (1:500, Cytoskeleton, Inc., ARC03), Pak1 (1:1,000, Cell Signaling Technologies, 2602) and Limk (1:1,000, Cell Signaling Technology, 3842). Antibody binding was detected using the Odyssey Fluorescence Detection System (LiCOR) with proprietary LiCOR secondary antibodies (1:50,000). All samples were normalized to GAPDH (1:50,000, Cell Signaling Technology, 2118). All quantification was performed with Odyssey Image Studio Software (Odyssey).

Mouse ChIP. Site-directed qChIP was performed for pan-acH3 and H3K27me3 as described previously26. Briefly, brains were rapidly removed, and bilateral 14-gauge NAc punches were taken from two adjacent 1-mm coronal slices con-taining the NAc. Punches were then immediately crosslinked in 1% formalde-hyde for 15 min at room temperature and quenched by the addition of glycine at a final concentration of 0.125 M. Punches were then washed five times in cold PBS containing proteinase inhibitors (1 mM PMSF, 1 µg/ml apoprotin

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicine doi:10.1038/nm.3090

and 1 µg/ml pepstatin A) and frozen on dry ice. For each sample, the total NAc punches from two mice were combined, balanced by social interaction ratio behavior ( that is, balanced so that all paired samples had similar social interac-tion ratios), resulting in eight NAc punches per sample.

Chromatin was solubilized and extracted by SDS detergent lysis buffer, which was followed by a series of sonications. First, samples were sonicated twice for 7 s with a desktop sonicator (40% power; Ultrasonic Processor, Cole Parmer) at low power in 500 µl SDS lysis buffer (1% SDS, 50 mM Tris-HCl, pH 8.1 and 10 mM EDTA) and then diluted with 1,100 µl ChIP dilution buffer (1.10% Triton X-100, 167 mM NaCl, 16.7 mM Tris-HCl, pH 8.1, 1.2 mM EDTA and 0.01% SDS) and sonicated (setting 4; Sonic Dismembranator 550, Fisher Scientific) at high power five times for 15 s per sonication. Four-hundred microliters of sheared chromatin was used for each immunoprecipitation brought to a final volume of 600 µl with ChIP dilution buffer. Five percent of each sample was put aside to be used as input for data normalization. This chromatin-shearing protocol reliably produces fragments 350–700 bp in length28. ChIP was then performed with 7 µg of each antibody to acetyl H3 and H3K27me3 (EMD Millipore, 06-599 and 07-449, respectively) per sample conjugated to magnetic Dynabeads (M-280 sheep antibody to rabbit IgG; Invitrogen). IgG control antibody–conjugated beads were also used and did not show enrichment (EMD Millipore, PP64B). Beaded antibodies were incubated with the immunoprecipitated chromatin overnight (~16 h) at 4 °C and then washed eight times in ChIP RIPA buffer (0.7% Na-deoxycholate, 500 mM LiCl, 50 mM 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES)-KOH, pH 7.6, 1% NP-40 and 1mM EDTA). Beads were removed by heating to 65 °C, shaken at 1,000 r.p.m. for 30 min on a Thermomixer and spun down before the supernatant was collected. Chromatin in the supernatant and input samples was then reverse crosslinked by heating to 65 °C overnight. DNA was then purified for RT-PCR analysis with a QIAquick PCR Purification Kit (Qiagen). Levels of specific histone modifi-cations at each gene promoter of interest were determined by measuring the amount of acetylated or methylated histone–associated DNA by quantitative real-time PCR, as detailed above. For the purposes of mapping out the chromatin landscape within, and upstream of, the Rac1 promoter, we used the primers listed in Supplementary Table 3 (Integrated DNA Technologies).

Human postmortem tissue. For the Texas cohort, whole-tissue NAc resections were collected and provided by the Dallas Brain Collection, where tissue is col-lected from the Dallas Medical Examiner’s Office and the University of Texas Southwestern’s Tissue Transplant Program under review of The University of Texas Southwestern Institutional Review Board52. For the Montreal cohort, whole-tissue NAc resections were collected by the Quebec Suicide Brain Bank at the Douglas Hospital Research Center under an approval of the Douglas Hospital Research Center’s Research Ethics Committee. Briefly, after obtaining next-of-kin permission, brain tissue was collected from cases at the local medical examiners offices. Subjects with known history of neurological disorders or head injury were excluded, and blood toxicology screened out any subject using illicit drugs or psychotropic medications. Clinical records and collateral information from telephone interviews with a primary caregiver were obtained for each case. Three to four mental health professionals carried out an extensive review of the clinical information and made independent diagnoses followed by a consensus diagnosis using the Diagnostic and Statistical Manual of Mental Disorders (DSM) IV cri-teria. Demographic characteristics associated with the tissue samples are listed in Supplementary Tables 4 and 5. The two groups were matched as closely as possible for race, gender, age, pH, postmortem interval (PMI) and RNA integrity number (RIN). Acceptable RIN values were within the range 6.1–9.5.

In each case, cerebral hemispheres were cut coronally into 1-cm to 2-cm blocks. The NAc was dissected from the appropriate coronal section (Fig. 3a) and immediately placed in a mixture of dry ice and isopentane (1:1 (vol:vol)). The frozen tissue was then pulverized on dry ice and stored at −80 °C.

pH and RIN determination was performed as previously described52. Briefly, tissue weighing approximately 150 mg was punched from the cerebellum, homogenized in 5 ml of ddH2O (pH adjusted to 7.00) and centrifuged for 3 min at 8,000g and 4 °C. The pH of the supernatant was measured in duplicate (Thermo-Electron Corporation). RIN determination was performed by isolat-ing total RNA using TRIzol (Invitrogen) followed by analysis with an Agilent 2100 Bioanalyzer.

Human ChIP. Site-directed qChIP was performed as previously described29 using a micrococcal nuclease–based assay, allowing for high-resolution mapping of the human gene promoter and upstream regions. Briefly, 50 mg of human NAc tissue was lightly homogenized with 550 µl douncing buffer (10 mM Tris-HCl, pH 8.0, 4 mM MgCl2 and 1 mM CaCl2) in a 1-ml loose-fit glass homog-enizer. The homogenate was then digested with micrococcal nuclease enzyme (2 units/ml) for 10 min in a 37 °C water bath. We have found that this reaction yields the optimal DNA basepair fragment size of ~150–160 bp (Supplementary Fig. 7). The digestion reaction was terminated by the addition of 10 mM EDTA, pH 8.0. The digested chromatin was then further incubated in SDS lysis buffer (1% SDS, 50 mM Tris-HCl, pH 8.1 and 10 mM EDTA) for 60 min on wet ice and lightly vortexed every 10 min. The lysed chromatin was then centrifuged at 3,000g for 20 min at 4 °C, and the supernatant was collected. Four-hundred microliters of the digested chromatin supernatant was then used for each ChIP brought to a final volume of 500 µl with incubation buffer (500 mM NaCl, 200 mM Tris-HCl, pH 8.0 and 50 mM EDTA, pH 8.0).

ChIP was then performed with 7 µg of each antibody to acetyl H3 and H3K27me3 (EMD Millipore, 06-599 and 07-449, respectively) per sample conjugated to magnetic Dynabeads (M-280 sheep antibody to rabbit IgG; Invitrogen). IgG control antibody–conjugated beads were also used and did not show enrichment (EMD Millipore, PP64B). Beaded antibodies were incubated with the immunoprecipitated chromatin overnight (~16 h) at 4 °C and then washed eight times in ChIP RIPA buffer (0.7% Na-deoxycholate, 500 mM LiCl, 50 mM HEPES-KOH, pH 7.6, 1% NP-40 and 1 mM EDTA). Beads were removed by heating to 65 °C and shaken at 1,000 r.p.m. for 30 min on a Thermomixer and then spun down to remove the supernatant. Chromatin in the supernatant and input samples was reverse crosslinked by heating to 65 °C overnight. DNA was then purified for RT-PCR analysis with a QIAquick PCR Purification Kit (Qiagen). Levels of specific histone modifications at each gene promoter of interest were determined by measuring the amount of acetylated or methylated histone–associated DNA by quantitative real-time PCR, as detailed above. For the purposes of mapping out the chromatin landscape within, and upstream of, the RAC1 promoter, the primers listed in Supplementary Table 6 were used (Integrated DNA Technologies).

Imaging and NeuronStudio. For spine analysis, dendritic segments 50–150 µm away from the soma were randomly chosen from HSV-infected cells that express XFP in a 4% paraformaldehyde (PFA)-fixed 200-µm coronal slice cover slipped in VectaShield setting medium. Images were acquired on a confocal LSM 710 (Carl Zeiss) for morphological analysis as described previously53. Neurons were selected from the NAc shell. To qualify for spine analysis, dendritic segments had to satisfy the following requirements: (i) the segment had to be completely filled (all endings were excluded), (ii) the segment must have been at least 50 µm from the soma and (iii) the segment could not be overlapping with other dendritic branches. Dendritic segments were imaged using a ×100 lens (numerical aperture 1.4; Carl Zeiss) and a zoom of 2.5. Pixel size was 0.03 µm in the x-y plane and 0.01 µm in the z plane. Images were taken with a resolution of 1024 × ~300–350 (the y dimension was adjusted to the particular dendritic segment to expedite imaging), the pixel dwell time was 1.58 µm/s and the line average was set to 4. An average of two dendrites per neuron on five neurons per mouse totaling ~2,500 dendritic spines per experimental group were analyzed. For quantitative analysis of spine size and shape, NeuronStudio was used with the rayburst algorithm described previously54. NeuronStudio classifies spines as thin, mushroom or stubby on the basis of the following values: (i) aspect ratio, (ii) head-to-neck ratio and (iii) head diameter. Spines with a neck can be classified as either thin or mushroom, and those without a neck are classified as stubby. Spines with a neck are labeled as thin or mushroom on the basis of head diameter. These parameters have been verified by comparison with trained human operators fully blinded across groups.

Immunohistochemistry. For cofilin immunohistochemisty, mice were perfused with 4% PFA as indicated in the above section. Forty-micrometer coronal slices were washed in PBST (0.4% Triton X-100 in PBS) and transferred to blocking solu-tion (0.2% PBST in 5% normal donkey serum) for 2 h before overnight incubation (0.2% PBST in 5% normal donkey serum) in primary antibody solution (Cofilin; 1:2,000, Abcam, 42824). Slices were then washed in 1× PBS and incubated for

©20

13 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicinedoi:10.1038/nm.3090

2 h with secondary antibody for fluorescence detection. Slices were cover slipped with DPX mounting media. For floxed Rac1 knockout validation, a monoclonal antibody to Rac1 (Millipore, 1:500, 05-389) was used with the same procedure.

Immunohistochemistry quantitative analysis. We acquired z-stack images of NAc dendritic segments and colocalized cofilin puncta on a LSM-710 (Carl Zeiss) confocal microscope using parameters identical to those used during the spine count experiments. This allowed us to determine dendritic spine morphol-ogy within the 40-µm tissue sections on each dendritic segment. To correct the photon point-spread function and improve resolution, we independently decon-volved the GFP and cofilin channels (Autodeblurr, AutoQuant) and remerged them into a single z stack. With the previously created NeuronStudio dendritic spine maps, we identified colocalized cofilin puncta manually on merged RGB images (ImageJ Software) and used orthogonal verification to ensure the puncta were located within discrete spine heads. All immunohistological images were taken with identical acquisition parameters using Zen software.

Stereotaxic surgery and viral gene transfer. We performed all surgeries under aseptic conditions using anesthetic as described previously13,55. Briefly, mice were anesthetized with a mixture of ketamine (100 mg per kg body weight) and xylazine (10 mg per kg body weight) and positioned in a small-animal stereotaxic instrument (David Kopf Instruments), and the skull surface was exposed. Thirty-three–gauge syringe needles (Hamilton Co.) were used to bilaterally infuse 0.5 µl of virus (1.5 × 108 infectious units/ml) expressing control GFP, Cre, RacDN or RacCA to activate or inhibit Rac1 signaling into the NAc (bregma coordinates: anteroposterior, +1.5 mm; mediolateral, +1.6 mm; dorsoventral, −4.4 mm) at a rate of 0.1 ml/min. All behavioral analyses were conducted between days 3–4 after infusion during maximal HSV expression.

Osmotic minipump surgery and infusion. Mice were anesthetized as detailed above. Mice were surgically implanted with two subcutaneous Alzet minipumps (model 1002, Durect) and bilateral guide cannulae (Plastics One) targeting the NAc. One day before surgery, two cannulae (28-gauge stainless steel) were filled with MS-275 (100 µm; provided by the Broad Institute) or 5% hydroxypropyl β-cyclodextrin vehicle (Trappsol, CTD), and each pedestal within the assembly was separately affixed by vinyl tubing to a minipump, each loaded with drug or vehicle. The minipumps were activated on the evening before surgery (by incubating them at 40 °C) to initiate a continuous delivery at 0.25 µl/h over 10 d. Briefly, the surgical procedure began with an incision over the skull, and the skin was spread apart under the scapulae to create an area for positioning the minipumps on the back. Bilateral cannulae were delivered into the NAc according to the bregma coordinates: anteroposte-rior, +1.5; mediolateral, +1.0; dorsoventral, −4.5. Cannulae were permanently fixed to the skull with Loctite skull adhesive (Henkel). Cannulae, tubing and minipumps were all secured under the skin using Vetbond tissue adhesive (3M) and two staples.

51. Tsankova, N. et al. Sustained hippocampal chromatin regulation in a mouse model of depression and antidepressant action. Nat. Neurosci. 9, 519–544 (2006).

52. Stan, A.D. et al. Human postmortem tissue: what quality markers matter? Brain Res. 1123, 1–11 (2006).

53. Radley, J.J. et al. Repeated stress induces dendritic spine loss in the rat medial prefrontal cortex. Cereb. Cortex 16, 313–320 (2006).

54. Rodriguez, A., Ehlenberger, D.B., Dickstein, D.L., Hof, P.R. & Wearne, S.L. Automated three-dimensional detection and shape classification of dendritic spines from fluorescence microscopy images. PLoS One 3, e1997 (2008).

55. Russo, S.J. et al. Nuclear factor κB signaling regulates neuronal morphology and cocaine reward. J. Neurosci. 29, 3529–3537 (2009).

epigenetic regulation of rac1 induces synaptic remodeling in stress disorders and depression

Documents