Serum response factor and cAMP response element binding protein are both required for cocaine induction of ΔFosB

Abstract

The molecular mechanism underlying induction by cocaine of ΔFosB, a transcription factor important for addiction, remains unknown. Here, we demonstrate a necessary role for two transcription factors, cAMP response element binding protein (CREB) and serum response factor (SRF), in mediating this induction within the mouse nucleus accumbens (NAc), a key brain reward region. CREB and SRF are both activated in NAc by cocaine and bind to the fosB gene promoter. Using viral-mediated Cre recombinase expression in the NAc of single- or double-floxed mice, we show that deletion of both transcription factors from this brain region completely blocks cocaine induction of ΔFosB in NAc, whereas deletion of either factor alone has no effect. Furthermore, deletion of both SRF and CREB from NAc renders animals less sensitive to the rewarding effects of moderate doses of cocaine when tested in the conditioned place preference (CPP) procedure and also blocks locomotor sensitization to higher doses of cocaine. Deletion of CREB alone has the opposite effect and enhances both cocaine CPP and locomotor sensitization. In contrast to ΔFosB induction by cocaine, ΔFosB induction in NAc by chronic social stress, which we have shown previously requires activation of SRF, is unaffected by the deletion of CREB alone. These surprising findings demonstrate the involvement of distinct transcriptional mechanisms in mediating ΔFosB induction within this same brain region by cocaine versus stress. Our results also establish a complex mode of regulation of ΔFosB induction in response to cocaine, which requires the concerted activities of both SRF and CREB.

Figure 1.

CREB and SRF mediate ΔFosB induction by chronic cocaine. A, ΔFosB accumulation, measured immunocytochemically, after chronic cocaine is unaffected by NAc-specific knock-out of CREB. B1, Injection of AAV–Cre–GFP into NAc of Creb^fl/fl–Srf^fl/fl mice results in the knock-out of CREB and SRF protein in Cre-expressing neurons. Injection of AAV–GFP has no discernable effect. B2, B3, Such selective knock-out of CREB and SRF from NAc completely blocks ΔFosB induction in NAc by chronic cocaine (n = 4 per group). *p < 0.05; **p < 0.01 versus saline.

Figure 2.

Cocaine-induced increases in phospho-SRF and phospho-CREB correlate with increased ΔFosB expression in NAc. Chronic cocaine increases phospho-SRF (A) and phospho-CREB (B) protein levels in NAc 1 h after the last injection detected by Western blots. Acute cocaine induces similar changes, although they are not significant. C, D, No differences were observed in total levels of SRF or CREB (n = 6–8 per group). *p < 0.05 versus saline.

Figure 3.

Chronic cocaine increases the binding of phospho-SRF and phospho-CREB to the fosB gene promoter. A, ChIP demonstrates enrichment of phospho-SRF and phospho-CREB at the fosB gene promoter after chronic cocaine. No significant change was observed for total CREB or NFκB/p65 binding. B, C, Phospho-SRF and phospho-CREB display increased binding at the promoter of their respective target genes, egr2 and glur1 (n = 6–12 per group). *p < 0.05 versus saline.

Figure 4.

Loss of CREB and SRF decreases behavioral responses to cocaine. A, Local deletion of CREB from NAc increases CPP to moderate doses of cocaine (n = 8–10 per group). No differences were seen at low and high dose of cocaine. B, In contrast, local knock-out of SRF and CREB from NAc reduces place conditioning to cocaine at low and moderate dose but not high dose of cocaine. Note that AAV–GFP-injected double-floxed mice show stronger cocaine place conditioning at every dose tested compared with AAV–GFP-injected Creb^fl/fl mice because of differences in genetic background. All experiments were done in parallel on the same day in independent batches of animals. *p < 0.05, ^#p = 0.06 versus AAV–GFP. C, Loss of CREB resulted in an increase cocaine sensitization compared with controls (n = 10 per group). D, Consistent with the decrease preference for cocaine, SRF and CREB local knock-out reduced the locomotor sensitization to cocaine (n = 10 per group). Note that both selective manipulation did not affect acute response to cocaine at 12 mg/kg. *p < 0.05, ***p < 0.001 versus cocaine day 1; ^#p < 0.05, ^###p < 0.001 versus AAV–GFP cocaine day 3; ^§p < 0.05 versus AAV–GFP cocaine day 1.

Figure 5.

Effect of NAc-specific CREB knock-out on stress-induced ΔFosB accumulation and anxiety-like behavior. A, NAc-specific CREB knock-out has no effect on ΔFosB accumulation after exposure to chronic social defeat stress. *p < 0.05 versus control. B, Local deletion of CREB from NAc decreases locomotor activity and time spent in the center of a novel open field, as well as time spent in the open arms in the O-maze. *p < 0.05, **p < 0.01 versus AAV–GFP.