The Role of Histone and DNA Methylation Mechanisms in Alcohol Drinking and Anxiety-like Behaviors

Alcohol Use Disorder (AUD) is a multifactorial psychiatric illness and significant public health concern associated with millions of annual deaths worldwide and costing the United States hundreds of billions of dollars annually. Comorbid anxiety is common in AUD, exhibiting a reciprocal contributory relationship. This complex relationship is believed to be facilitated by the dominant anxiolytic effect of alcohol mediated by the amygdala, which has been extensively implicated in regulation of negative affective states associated with drug use, such as anxiety. To study this critical association, animal models of both intrinsic and acquired alcohol and anxiety-related phenotypes have been developed. One such model comprises the high-anxiety alcohol-preferring (P) versus non-preferring (NP) rats, which are selectively bred from Wistar rats for alcohol preference and non-preference, respectively. Additionally, because acquired tolerance to the anxiolytic effect of alcohol is believed to substantially influence alcohol intake, a model of rapid ethanol tolerance (RET) to the anxiolytic effects of alcohol is used and serves as an important index for chronic tolerance. Epigenetic mechanisms have recently emerged as strong candidates for treatment targets in psychiatric disorders, including AUD and anxiety disorders. Therefore, we investigated the role of histone and DNA methylation-based epigenetic mechanisms in anxiety and alcohol preference phenotypes, using the P/NP and RET models. The histone methyltransferase (HMT) known as G9a (KMT1C/EHMT2) and its downstream dimethylated histone H3 lysine 9 (H3K9me2) marker are both known to downregulate gene expression via inhibitory chromatin remodeling and have recently been implicated in regulating addiction and anxiety. Here, we determined that P rats exhibit innately higher mRNA and protein expression of G9a and H3K9me2, relative to NP rats in the central and medial amygdala (CeA and MeA) but not the basolateral amygdala (BLA). To determine gene-specific changes in G9a levels and function in P and NP rats, we utilized chromatin immunoprecipitation (ChIP) to evaluate the occupancy of G9a and H3K9me2 at the promoter regions of various genes of interest that were found to be differentially expressed between the amygdala of P and NP rats. Expression of the anxiolytic neuropeptide Y (Npy) is known to be lower in P rats relative to NP rats. The N-methyl-D-aspartic acid (NMDA) receptor subtypes 2a and 2b (Grin2a/2b), which are involved in synaptic plasticity, were respectively expressed higher and lower in P rats relative to NP rats. The mRNA expression of anxiogenic amygdalar pro-opiomelanocortin (Pomc) and melanocortin 4 receptor (Mc4r) were also higher in the P rat relative to NP rats. Additionally, higher protein expression of MC4r and the downstream cleavage product of POMC, α-MSH, was found in the CeA and MeA, but not the BLA of P rats. Despite globally higher G9a and H3K9me2 levels in the amygdala of P rats, ChIP results suggested that G9a influenced expression of both upregulated and downregulated genes, as the inhibitory G9a occupancy was higher at promoters of Npy and Grin2b and lower at the promoters of Pomc, Mc4r, and Grin2a in the P rat amygdala. Similarly, ChIP analysis revealed H3K9me2 occupancy echoed G9a occupancy at Pomc, Mc4r, and Grin2a promoters; however, H3K9me2 occupancy remained undifferentiated at the Grin2b and Npy promoters, where P rat G9a occupancy was higher. In recent years, the dynamic relationship between HMTs and DNA methyltransferases (DNMTs) has been increasingly elucidated, revealing reciprocal cooperation in downregulating gene expression. As such, we investigated DNMT status in the P and NP rat amygdala and found DNMT activity to be higher in P rats. Additionally, Dnmt1 and Dnmt3b, but not Dnmt3a, mRNA expression was higher in P rats, accompanied by higher DNMT3B, but not DNMT1, protein in P rat CeA and MeA. DNA methylation measurement using the Methylminer approach revealed higher methylation at the Grin2b and Npy gene promoter sites where G9a occupancy was higher in P rats and H3K9me2 was undifferentiated. We then investigated the effects of an intraperitoneal DNMT inhibitor treatment and reported decreased alcohol drinking behaviors in P rats and a concurrent increase in G9a occupancy at the promoters for Pomc, Mc4r, and Grin2b genes. G9a’s dynamic influence on biological pathways regulating P rat drinking and G9a’s documented involvement in anxiety inspired inquiry into its potential role in acquired behaviors. Utilizing the RET model, we determined that acute ethanol exposure reduces G9a and H3K9me2 in the CeA and MeA, but not the BLA, relative to saline-exposed controls. Two identical doses of acute ethanol 24 hours apart produced tolerance to ethanol’s anxiolytic effects, despite equivalent blood alcohol concentration to the acutely exposed group. Furthermore, two consecutive doses induced no changes in G9a or H3K9me2 protein levels in the CeA and MeA relative to controls, suggesting the tolerant phenotype is related to the hindrance of ethanol-induced attenuation of G9a and H3K9me2 expression and/or activity. Specifically, acute ethanol exposure reduced H3K9me2 and G9a occupancy at the Npy gene near a putative cAMP response element (CRE) site, whereas two consecutive doses 24 hours apart induced no such change. Notably, the transcription factor, CRE-binding protein (CREB), can potentially bind this site and has been heavily implicated in anxiety and addiction mechanisms via expression regulation of multiple genes, including Npy. This suggests that G9a-mediated chromatin remodeling may contribute to the expression regulation of the anxiolytic Npy in the amygdala. We then tried to determine if inhibition of G9a, before the second ethanol exposure, could recover sensitivity to ethanol-induced anxiolysis. Subacute IP administration of a G9a inhibitor not only reversed tolerance to anxiolysis but also reduced anxiety in controls. Treatment with the G9a inhibitor was also associated with reduced H3K9me2 and increased NPY protein levels in the CeA and MeA, further implicating G9a-mediated expression regulation of anxiolytic amygdalar NPY in this effect. Cumulatively, these findings suggest that G9a likely influences multiple anxiety and AUD-related phenotypes, including innate and alcohol-induced traits, via its complex chromatin remodeling mechanisms. Overall, G9a-mediated chromatin compression in the CeA and MeA is associated with both innate AUD behaviors and acquired tolerance to the anxiolytic effects of alcohol. We have shown that G9a modulates phenotypes through a diverse set of anxiolytic and anxiogenic pathways. G9a may be responsible for critical regulation of crucial genes in these pivotal pathways, thus ultimately contributing to the development of AUD. We believe this evidence suggests that targeting G9a-mediated mechanisms has significant merit as potential treatment for anxiety and AUD.