Chromatin Remodeling in Addiction: BRG1-SMAD3 Interaction Contributes to Cued Reinstatement of Cocaine Seeking

Chronic cocaine use results in neuroplasticity in the nucleus accumbens that underlies vulnerability to relapse, even after protracted abstinence. It is well accepted that epigenetic changes in gene transcription contribute to the addicted phenotype, and three transcription factors in particular have received a great deal of attention: cyclic adenosine monophosphate response element binding protein, nuclear factor-κB, and ΔFosB. Targeting transcription factors is seen as a promising point of pharmacological intervention, in part because inhibition of a single transcription factor may pleiotropically alter multiple genes that contribute to the complex disorder. The most commonly studied epigenetic mechanisms of learning and memory are chromatin/histone and DNA modification (e.g., acetylation and methylation). Chromatin remodeling is a relatively understudied epigenetic mechanism in neuropsychiatric disease that is not to be confused with chromatin modification. Chromatin remodeling describes the process by which chromatin remodeling complexes (CRCs), composed of at least 15 subunits derived from 28 genes, hydrolyze adenosine triphosphate to disrupt histone-DNA interaction, reposition nucleosomes, and allow transcription factors to access specific portions of DNA (1).

The study by Wang et al. in this issue of Biological Psychiatry elucidates a novel mechanism by which chromatin remodeling contributes to cocaine-induced neuroplasticity. The major finding of this paper is that BRG1 binding to the SMAD3 transcription factor contributes to cocaine-induced changes in SMAD3 binding to target genes, and also to cue-induced reinstatement of cocaine seeking (2). BRG1 is a member of the switch/sucrose nonfermentable (SWI/SNF) neuron-specific CRC that contributes adenosine triphosphatase activity and allows transcription factor interaction with DNA. Specifically, Wang et al. showed that 10 days of restricted (2 hours/day; 1 mg/kg/infusion) access to cocaine self-administration and 7 days of subsequent abstinence upregulates expression of BRG1 and phosphorylation of SMAD3 in the accumbens, but not in the dorsal striatum. The authors went on to perform coimmunoprecipitation experiments in which they precipitated BRG1 and examined its association with SMAD3, and once again they found an increase specifically in the nucleus accumbens. They then performed quantitative chromatin immunoprecipitation in order to examine BRG1 binding to target genes, and reported a cocaine-induced upregulation of BRG1 binding to promoter regions for β-catenin (Ctnnb1), myocyte enhancement factor 2d (Mef2d), cyclic adenosine monophosphate–associated protein 2 (Cap2), and drebrin 1 (Dbn1). Finally, these cocaine-induced neuroadaptations were shown to be behaviorally relevant through site-specific inhibition of BRG1 and subsequent reduction in cue-induced reinstatement of cocaine seeking. Conversely, virally overexpressing BRG1 potentiated cued reinstatement, while neither of these manipulations had an effect on locomotor activity. Taken together, these data provide convincing preclinical evidence that BRG1-SMAD3 interaction is involved in cue-induced reinstatement. The authors adeptly acknowledge that determining whether PFI3, the BRG1 inhibitor used, can be administered systemically and penetrate the blood-brain barrier is paramount to determining the future therapeutic value of this research.

These experiments are related to a paper published in 2015 from the same laboratory that showed that activin receptor signaling is increased after chronic cocaine self-administration (3). Activin is a member of the transforming growth factor (TGF) superfamily of growth factors and is canonically upstream from the SMAD3 signaling pathway. Stimulation of the activin receptor and increased SMAD3 signaling contributes to cocaine-induced transcriptional and morphological plasticity (3). This previous paper examined contributions of SMAD3 to drug-primed rather than cue-induced reinstatement. Activin has also been implicated in neuroplasticity by inducing GluN2A phosphorylation and expression, an upward shift of the cocaine dose-response curve, and potentiation of cocaine-primed reinstatement (3,4). In addition to the effects of activin receptor stimulation on SMAD3 and cocaine seeking, this group has also shown that the TGF-β receptor 1 (TGFβ-R1), which is also upstream from SMAD3 signaling, is also increased after cocaine self-administration and abstinence (5). In an important series of control experiments, activin and SMAD3 signaling were each determined not to affect food reinforcement, indicating that these signaling pathways are specifically affected by drugs of abuse.

The current studies go on to further validate targets of the BRG1-SMAD3 complex that have previously been implicated in neuroplasticity or cocaine seeking. One in particular stands out as being previously implicated in cocaine addiction. Binding of BRG1-SMAD3 to the β-catenin promoter was increased after withdrawal from cocaine. β-Catenin is a multifunctional protein that both serves to anchor cell adhesion molecules called cadherins to the actin cytoskeleton and acts as a transcriptional coactivator that is a canonical member of the Wnt signaling pathway. It is increased in the striatum of both rodents and nonhuman primates after chronic cocaine exposure.

While chromatin remodeling and nucleosome repositioning complexes are becoming increasingly implicated in the fields of learning and memory, few studies have examined them specifically in cocaine addiction. The importance of this epigenetic mechanism was realized when genome-wide association studies determined that copy number variants of SWI/SNF (also called BAF, or BRG1-associated factor) proteins were related to intellectual disability (6). BAF53b is an accessory member of this CRC that has previously been implicated in neuroplasticity, namely long-term potentiation. When theta burst stimulation is applied to slices from BAF53b heterozygous knockouts or dominant negative expressing mice, both strains show a comparable induction of long-term potentiation to their wild-type littermates; however, this potentiation quickly decays back to baseline, indicating impairment of the maintenance phase of plasticity. This failure to maintain long-term potentiation was attributed to lack of phosphorylation of cofilin, an actin-shearing protein that, when phosphorylated, favors actin polymerization and dendritic spine head enlargement (7). To my knowledge, only one other study has examined CRCs in cocaine-related behaviors. Sun et al. found that overexpression of Baz1b, an accessory subunit of the imitation switch family of CRCs, drives potentiation of cocaine reward during self-administration in mice, and also confers resilience to chronic social defeat stress (8).

Wang et al. raise many interesting questions for future directions regarding chromatin remodeling in addiction and neuropsychiatric disease in general. Genome-wide association studies have implicated the SWI/SNF complex in intellectual disability, autism, and alcoholism (9,10), although its involvement in addiction to other drugs of abuse has not been reported in humans. A major hurdle to preclinical research focusing on specific subunits of chromatin remodeling complexes in addiction is the identification of specific binding partners between transcription factors and their coactivators within CRCs. Proteomic strategies more thoroughly examining the SWI/SNF complex and screening for transcription factors known to be relevant in addiction may be of use in identifying therapeutically relevant targets.