DeltaFosB in brain reward circuits mediates resilience to stress and antidepressant responses

Abstract

In contrast with the many studies of stress effects on the brain, relatively little is known about the molecular mechanisms of resilience, the ability of some individuals to escape the deleterious effects of stress. We found that the transcription factor DeltaFosB mediates an essential mechanism of resilience in mice. Induction of DeltaFosB in the nucleus accumbens, an important brain reward-associated region, in response to chronic social defeat stress was both necessary and sufficient for resilience. DeltaFosB induction was also required for the standard antidepressant fluoxetine to reverse behavioral pathology induced by social defeat. DeltaFosB produced these effects through induction of the GluR2 AMPA glutamate receptor subunit, which decreased the responsiveness of nucleus accumbens neurons to glutamate, and through other synaptic proteins. Together, these findings establish a previously unknown molecular pathway underlying both resilience and antidepressant action.

Figure 1. ΔFosB induction in NAc by social defeat mediates resilience

(a) Interaction zone times showing social avoidance in the susceptible mice only (n=4). [F(2,11)=34.91, P<0.001; Post-hoc test: ***P<0.001, versus “control”; ^##P<0.01 versus “susceptible”.] (b) Chronic social defeat induces ΔFosB in NAc as quantified on day 11. Resilient mice show greater ΔFosB induction in both core and shell of NAc versus control mice (n=3–4). [Core, F(2,11)=16.81, P<0.001; Shell, F(2,11)=39.9, P<0.001. Post-hoc test: ***P<0.001, **P<0.01 versus “control”; ^##P<0.01 versus “susceptible”]. (c) Representative photomicrographs of ΔFosB immunohistochemistry in NAc 24 hr after the last defeat. ac, anterior commissure. (d) Inducible bitransgenic mice overexpressing ΔFosB (Day 5: n=29–32; Day 11: n=6–15) don’t develop social aversion. [“Day 5”: Interaction F(1,118)=5.908, P<0.05; Post-hoc test: ***P<0.001 versus “no target”; “Day 11”: Significant effect of “target”, F(1,38)=13.20; A posteriori t-test, t=4.190, ***P<0.01 versus “no target”.] (e) Conversely, overexpression of ΔcJun increases susceptibility to social defeat with increased social aversion seen after 4 days of defeat (Day 5: n=15–23; Day 11: n=6–7). [“Day 5”: Interaction F(1,72)=4.198, P<0.05; Post-hoc test: *P<0.05 versus control “no target”; “Day 11”: Significant effect of “target” F(1,20)=13.16; A posteriori t-test, t=2.313, *P<0.05, t=3.801, **P<0.01 versus “no target”]. Whereas control littermate mice of the ΔFosB line show social aversion after 4 defeat episodes, control littermates of the ΔcJun line do not due to differences in genetic background. Such baseline differences between the lines are also seen in measures of anxiety- and depression-related behavior (see Online Methods and Supplementary Fig. 2).

Figure 2. Effect of social isolation on ΔFosB and on susceptibility to social defeat

(a) Long-term social isolation (n=4) decreases basal levels of ΔFosB in NAc shell [t=2.882, df=6, *P<0.05] and core [t=6.338, df=6, ***P<0.01]. (b) Representative brain sections showing ΔFosB levels in NAc of grouped-housed and isolated mice. (c) Social isolation triggers vulnerability to an acute social defeat (see Online Methods for details) based on the social avoidance measured on the following day (n=8–10), an effect rescued by viral-mediated overexpression of ΔFosB (n=12) in NAc of isolated mice. [Isolation: t=4.351, df=16, ***P<0.001, HSV-ΔFosB: t=3.030, df=22, **P<0.01.] Overexpression of ΔJunD (n=8–12) in NAc mimics social isolation by causing social avoidance after short-term social defeat [t=2.251, df=18, *P<0.05]. (d) Postmortem human NAc show decreased ΔFosB levels in depressed patients compared to matched controls (n=8) [t=3.416, df=14, **P<0.01]. ns, non-specific band unrelated to ΔFosB.

Figure 3. ΔFosB induction in NAc mediates the antidepressant effect of fluoxetine

(a) Chronic treatment with fluoxetine completely reverses the social avoidance, measured on day 11, induced by chronic (10 days, n=7) social defeat. [Interaction, F(1,24)=5.325, P<0.05, Post-hoc test: ***P<0.001 versus “defeat with vehicle”.] (b)ΔFosB levels in NAc shell measured by immunohistochemistry are increased after chronic fluoxetine treatment of control mice. Such levels are also increased in susceptible mice after chronic social defeat, with fluoxetine inducing a still further increase (n=4). [No interaction effect, F(1,12)=0.2122, significant effect of social defeat and antidepressant treatment; A posteriori t-test t=8.417, df=6 (“defeat fluoxetine”), t=4.516, df=6 (“control fluoxetine”), t=6.063, df=6 (“defeat vehicle”), ***P<0.001, **P<0.001 versus “control vehicle”.] Similar results were obtained in NAc core (See Supplementary Fig. 4). (c) Overexpression of ΔJunD in NAc blocks the antidepressant-like effect of chronic fluoxetine treatment (n=8). [Interaction, F(1,28)=6.121, P<0.05, Post-hoc test: ***P<0.001 versus “GFP vehicle”.]

Figure 4. Pro-resilience, antidepressant-like effect of GluR2 in NAc

(a) Resilient mice show increased GluR2 levels, and decreased GluR1 levels, versus control and susceptible mice (n=4). Conversely, susceptible mice show opposite changes. [GluR2: F(2,11)=69,89, P<0.001; Post-hoc test: ***P<0.001, *P<0.05 versus “control”, ^###P<0.001 versus “susceptible”; GluR1: F(2,11)=27.58, P<0.001; Post-hoc test: **P<0.01, versus “control”, ^###P<0.001 versus “susceptible”.] (b) Social defeat increases ΔFosB binding to the GluR2 promoter (n=5) [t=2.158, df=8, *P<0.05]. This effect was seen only for a region of the promoter that contains AP1 sites. (c) qPCR performed two or ten days after the last defeat revealed increased GluR2 mRNA levels in resilient mice (n=6–8) [Day 12, P>0.05, A posteriori t-test: Day 12, t=2.838, df=13, ^# P<0.05; Day 20, Group F(2,20)=8.739, P<0.05; Post-hoc test: **P<0.01 versus “control”, ^##P<0.01 versus defeat]. (d) Fluoxetine treatment increases GluR2 levels in NAc (n=4) [Core: t=3.778, df=6, **P<0.01; Shell: t=6.602, df=6, ***P<0.01] (e) Western-blotting revealed decreased GluR2 levels in NAc of depressed humans (n=8) [t=2.381, df=14, *P<0.05]. (f) Intra-NAc infusion of NBQX (n=8) had an immediate and persistent (one week) antidepressant effect in susceptible mice [Interaction, F(1,28)=6.128, P<0.05; Drug: ***P<0.001; Day 7: ***P<0.001 versus “vehicle”; A posteriori t-test: Day 1: t=2.156, df=15, *P<0.05 versus “vehicle”.] (g) Overexpression of GluR2 in NAc of susceptible mice reverses defeat-induced social avoidance (n=7–8), an effect that persists for at least 10 days, when viral expression has dissipated [Virus F(1,41)=9.553, P<0.01, Days F(2,41)=7.248, P<0.01, no effect on Interaction. A posteriori t-test: Day 15: t=2.702, df=12, *P<0.05; Day 22: t=2.008, df=12, *P<0.05.]

Figure 5. AMPA receptor composition is differentially regulated in susceptible and resilient mice

(a–b) Evoked EPSCs recorded after social defeat (between 2 and 28 days after the last defeat episode). (a) Sample traces of AMPAR EPSCs at two resting potentials. (b) A measure of rectification (EPSC_+40mV/EPSC_−80mV, n=6–10) from NAc neurons of control, susceptible, and resilient mice. [F(2,21)=8.773, P<0.01. Post-hoc test: **P<0.01, versus “control”; A posteriori t-test show a significant difference between susceptible and both control and resilient, t=3.482, df=11, ^##P<0.01 versus “controls” and t=3.146, df=14, ^##P<0.01 versus “resilient”, but not between resilient and control]. A decrease in the EPSC_+40mV/EPSC_−80mV ratio corresponds to an increase in inward rectification. (c) Current-voltage relationships for neurons recorded from control, susceptible, and resilient mice demonstrate changes in rectification at positive potentials only. (d) Linear regression analysis reveal significant correlation between rectification value and time interacting with the target (r=0.467, *P<0.05). (e–f) NASPM (200 μM, 5 min) decreased evoked AMPAR EPSC amplitude in susceptible mice. (e) Representative traces illustrating the effect of NASPM after 10 min of bath application. (f) Evoked EPSC amplitude with NASPM normalized to baseline (t=2.689, df=7, **P<0.01).

Figure 6. Pro-resilience, antidepressant-like effects of SC1 in NAc

(a) Changes in gene expression observed in NAc during resiliency overlap with those observed upon overexpression of ΔFosB (comparing datasets in refs and 24). The upper heatmap shows 106 genes significantly regulated (>1.5 fold; *P<0.05) in NAc by social defeat in resilient mice compared to control, and how those same genes are regulated in resilient mice vs. susceptible mice (lower heatmap) and by overexpression of ΔFosB in 11A mice (middle heatmap). Position of SC1 on the heatmaps is indicated. (b) Viral overexpression of SC1 in NAc reversed the social avoidance induced by chronic (10 days, n=17–24) social defeat. [Virus F(1,125)=7.002, P<0.01, Days F(2,125)=6.908, P<0.01, no effect on Interaction; A posteriori t-test: Day 15: t=1.875, df=43, *P<0.05, and on Day 22: t=2.138, df=39, *P<0.05.] Sus, susceptible. (c) Rats injected with HSV-GFP or HSV-SC1 into NAc were subjected to the forced swim test (n=8). SC1 overexpression had an antidepressant-like effect as measured by a decrease in time spent immobile [t=2.384, df=14 *P<0.05] and an increase latency to immobility [t=2.606, df=16, *P<0.05]. (d) Human NAc samples from depressed patients show a strong trend for decreased SC1 levels compared to matched controls (n=8) [t=1.922, df=14, ^#P=0.068].