Ventral striatal regulation of CREM mediates impulsive action and drug addiction vulnerability.

Increased heroin addiction risk in impulsive animals(a–c) Impulsive SHRs exhibit a greater preference for the immediate reward as a function of delay (a), commit more errors of commission during the delay period (b), and show increased novelty seeking in an open-field arena (c) (for ITD, n = 15 SHR and 16 WKY; for open-field, n = 8 per group). (d) After an acute, single-dose injection of heroin, SHRs exhibit elevated extracellular dopamine in the AcbC (upper panel) and increased locomotor activity (lower panel) (for microdialysis, n = 4 SHR and 6 WKY and shown relative baseline; for open-field, n = 7 per group and shown relative saline-treated animals from panel c). (e–h) Impulsive SHRs also exhibit elevated heroin self-administration (e) and cumulative intake (f), attenuated extinction (g), and enhanced reinstatement (h) (n = 5 SHR and 6 WKY). All data are expressed as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001 relative WKY; †P < 0.05, †††P < 0.001 relative extinction day; independent or paired t-tests with Holm-Šidák correction, see results text for ANOVA statistics.

Crem in the AcbC alters impulsive action and associates with genetic and epigenetic alterations in impulsive animals(a) Crem was reduced in the Acb of drug-naïve SHRs, specifically in the AcbC but not AcbSh (inset) (for array, n = 6 per group; for single-assay validation, n = 6 per group for AcbC, 5 SHR and 6 WKY for AcbSh). (b) Acute heroin induced Crem in the AcbC of SHRs but not WKYs, while in the AcbSh, heroin produced similar increases in both strains (n = 6 SHR and 5 WKY and shown relative saline-treated animals in panel a). (c) To study the causal contribution of Crem to impulsivity, HSV-Crem-Gfp was infused into the AcbC of SHRs resulting in over-expression (n = 4 Gfp and 5 Crem-Gfp, 2 independent experiments). (d–f) While over-expression did not impact preference for the large-reward (d), it decreased errors of commission during ITD (e) and inactive lever pressing during heroin self-administration (f) (for ITD, n = 5 Gfp and 7 Crem-Gfp; for SA, n = Gfp and 6 Crem-Gfp). (g–i) In the AcbC of SHRs, the permissive histone marks AcH3 (g) and H3K4me3 (h) were depleted near the Crem TSS, while DNA methylation was elevated upstream within the promoter (i) (n = 8 per group for ChIP, 8–10 per group for methylation). (j) Schematic of the rat Crem illustrating exons (numbered black bars), conserved regions (gray bars), and PCR primers used for ChIP-qPCR (top), bisfulfite sequencing (middle) and Sanger sequencing (bottom), all numbered relative the transcriptional start site (TSS). Single-nucleotide polymorphism within the first non-coding exon of rat Crem (indicated by •). Sequencing results from outbred Long Evans (LE) strain shown for reference. All data are expressed as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001 relative WKY or Gfp alone; †††P < 0.001 relative vehicle treatment (shown as horizontal line); independent t-tests with Holm-Šidák correction (for behavior) and Fisher’s LSD (for qCPR), see results text for ANOVA statistics.

Genetic variation in CREM associates with substance use(a) In a post-mortem population composed of individuals with opiate abuse, the rs12765063 variant associated with increased risk of opiate abuse even after stratifying on the basis of ethnicity. (b) In our well-curated postmortem brain collection composed of ethnically homogenous heroin abusers, this variant associated with shorter duration of drug use, which, given the fact that brains were collected at death due to opiate overdose, suggests that individuals with the A-allele more rapidly progressed to lethal intoxication. (c) This is consistent with the finding that A-allele carriers were also more likely to experience a previous, albeit non-lethal overdose. (d) The additive genotypic effect of the A-allele on higher total number of symptoms for any drug dependence can clearly be seen. Homozygotes for the G/G genotype had an average 2.38±0.08 symptoms while homozygotes for the A/A genotype had an average of 3.36±0.61 symptoms. All data are expressed as mean ± s.e.m. unless otherwise indicated on y axis. For sample sizes, see Supplementary Table 4. *P < 0.05, ****P < 0.0001 relative control participants; †P < 0.05, relative G/G genotype; independent t-tests with Holm-Šidák correction (versus control participants) or Fisher’s LSD (versus G/G), see results text for linear and non-linear regression.

Genetic variation in CREM associates with impulsivity in the context of substance abuse(a) In a mixed cohort of preschool-aged children (3–5 years-old) with and without ADHD, CREM rs12765063 genetic variation associated with the proportion of individuals engaging in dangerous activities (Ab = absent, M = mild, S = severe). (b–c) This polymorphism associated with ankle movement (b) and self-reported DSM hyperactivity/impulsivity score (c). (d) Stratifying this population based on ADHD status at follow-up, the genetic association between CREM and reported use was almost exclusively driven by participants with persistent ADHD, not remitters. (e) In an adult population composed of cannabis dependent and non-dependent individuals, this genetic variant strongly associated with impulsivity and persistence, as well as negative affect. (f) Variation in impulsivity significantly mediated the association between CREM genotype and substance dependence severity as measured by cannabis consumption (values adjacent to arrows reflect regression coefficient ± standard error). All data are expressed as mean ± s.e.m. unless otherwise indicated. For sample sizes, see Supplementary Table 4. *P < 0.05, ***P < 0.001, ****P < 0.0001 relative control participants; †P < 0.05, ††P < 0.01 relative G/G genotype; independent t-tests with Holm-Šidák correction (versus control participants) or Fisher’s LSD (versus G/G), see results text for linear and non-linear regression.

Impulsivity and heroin abuse share a common ventral striatal gene expression network that relates to CREM and neuroplasticity(a,b) To focus on networks related to both impulsivity and heroin abuse, we compared differentially expressed genes derived from two datasets. While there was only a modest overlap between heroin and SHR altered genes (a), the shared genes are disproportionately regulators of neuroplasticity (b, upper panel) and targets of Crem (b, lower panel). (c) Of these common genes, Adgrl3—a gene previously associated with ADHD—is differentially regulated in the AcbC of SHRs at baseline and after acute heroin exposure (n = 10 per group for baseline study, n = 6 SHR and 5–6 WKY for acute heroin study). (d) Primary striatal neurons (co-cultured with cortical neurons) over-expressing Crem for two days in vitro exhibited significantly less dendritic spines at DIV14 (St = Stubby, Th = Thin, Mu = Mushroom) (n = 11 cells per group from 2–3 independent experiments). All data are expressed as mean ± s.e.m. unless otherwise indicated on x axis. ***P < 0.001 relative WKY or Gfp alone; †P < 0.05 relative vehicle; independent t-tests with Holm-Šidák correction, see results text for ANOVA statistics.