Studies of ethanol
The number of studies focused on DNA methylation changes in ethanol addiction has grown in recent years (Krishnan, Sakharkar, Teppen, Berkel, & Pandey, 2014; Kyzar & Pandey, 2015; Ponomarev, 2013). Many such studies come from human subjects, particularly from peripheral blood samples. This work has provided valuable information on ethanol action.
One unique feature of ethanol, with respect to DNA methylation, is that ethanol metabolism directly affects methyl donors throughout the body. For example, chronic ethanol exposure can induce a folate deficiency, resulting in decreased levels of S-adeno- sylmethionine (SAM) and increased levels of its precursor homocysteine (Blasco et al., 2005; Hamid, Wani, & Kaur, 2009). As SAM is the primary methyl donor for most biological reactions, this effect of ethanol would be expected to decrease DNA methylation. In addition, the ethanol metabolite, acetaldehyde, was shown to inhibit DNMT activity in vitro (Garro, McBeth, Lima, & Lieber, 1991). Decreased expression of Dnmt3a and Dnmt3b, but not Dnmtl, was observed in whole blood of alcoholics, with a significant negative correlation between Dnmt3b expression and blood ethanol concentrations (Bonsch et al., 2006). Although the effect of ethanol on DNMT expression in adult brain is still unclear, a human study using microarray technology identified a 20-30% downregu- lation of Dnmtl across the superior frontal cortex and amygdala of alcoholics (Ponomarev, Wang, Zhang, Harris, & Mayfield, 2012). In parallel, in adolescent rat brain, it was recently reported that acute ethanol exposure significantly inhibited DNMT activity in the amygdala and bed nucleus of stria terminalis (BNST). However, this same study also demonstrated contradicting increases in Dnmtl and Dnmt3a mRNAs in the BNST together with no mRNA alterations in amygdala (Sakharkar et al., 2014). These findings indicate that DNMT enzymatic activity may not associate with transcriptional alterations. They also recommend further analyses, together with protein levels of DNMTs, in a defined brain region, cell type, and developmental stage. In fact, the same group carried out such an approach on cultured astrocytes and found that DNMT activity and DNMT3a protein expression were both attenuated by ethanol (X. Zhang et al., 2014), which may be involved in astrocyte-mediated inhibition of neuronal plasticity by ethanol exposure. Consistent with the aforementioned reduction of SAM after chronic ethanol administration, global DNA demethylation was observed in liver and colon of alcoholics (Choi et al., 1999; Lu et al., 2000). However, a significant increase (10%) of genomic DNA methylation in blood of patients with alcoholism was reported previously (Bonsch, Lenz, Reulbach, Kornhuber, & Bleich, 2004). This not only suggests ethanol-induced alteration in global DNA meth- ylation is tissue specific and cell-type specific but also warrants future replication studies with the addition of mapping DNA methylation genome-wide.
Altered methylation of numerous genes, most derived from candidate gene approaches, has been associated with alcohol use disorders (Table 8.1). For example, the gene encoding the N-methyl-D-aspartate (NMDA) receptor NR2B underwent promoter demethylation after chronic ethanol treatment in primary mouse cortical neuronal cultures. This was associated with increased transcription of the gene (Marutha Ravindran & Ticku, 2005). Furthermore, a clinical study demonstrated that the degree of methyla- tion of the NR2B gene promoter in peripheral blood during withdrawal negatively correlated with the severity of ethanol consumption (Biermann et al., 2009). Similarly, elevated DNA methylation of the HERP and SNCA genes associated with repression of gene transcription in peripheral blood of alcoholics (Bleich et al., 2006; Bonsch, Lenz, Kornhuber, & Bleich, 2005). Barbier et al. (2015) reported that in rats after 3 weeks of abstinence from chronic ethanol drinking, there was increased DNA methylation, and reduced expression, of several genes encoding synaptic proteins involved in neurotransmitter release in the medial prefrontal cortex (mPFC). They further showed that intra- mPFC infusion of RG108 blocked the increased ethanol consumption observed under these conditions along blockade of the downregulation of four of the seven regulated mRNA transcripts. Finally, viral-mediated suppression of one of these transcripts (Syt2) increased alcohol drinking, thus directly linking DNA methylation—regulated changes in gene transcription to alcohol addiction (Barbier et al., 2015).
However, this negative correlation between DNA methylation and gene expression is not always seen. By analyzing the PDYN gene in the dorsolateral PFC of human alcoholics postmortem, DNA methylation status was examined on CpG sites overlapping with single-nucleotide polymorphisms (SNPs) known to be associated with alcohol dependence. Interestingly, a methylation increase at such CpG-SNP sites within the 3' untranslated region (3' UTR) was observed and positively correlated with Pdyn transcription and with increased vulnerability to alcohol dependence (Taqi et al., 2011). This observation further emphasizes that the effects of DNA methylation on gene transcription are complex and region specific.
The differential methylation of certain genes observed in peripheral blood suggests that such epigenetic profiles might serve as a valid biomarker for alcohol use disorders. Although further work is needed to determine how DNA methylation patterns in blood compare with those in several relevant brain regions, the same genes need not be influenced in blood and brain for peripheral DNA methylation measures to serve an important biomarker function. To better account for DNA methylation changes associated with alcohol use disorders, a recent consortium carried out a genome-wide analysis of DNA methylation in peripheral blood of 18 monozygotic twin pairs discordant for alcohol use disorders and identified 77 differentially methylated regions most of which (68%) were hypermethylated (Ruggeri et al., 2015). The 3'-PPM1G (protein phosphatase 1G) gene locus displayed the most significant association with alcohol use disorders in the genome-wide analysis and in a replication cohort (Ruggeri et al., 2015). Moreover, PPM1G hypermeth- ylation was associated with increased activation of the subthalamic nucleus as measured by functional brain imaging, early escalation of alcohol use, and increased impulsiveness. This study illustrates the power of genome-wide approaches, as discussed in more detail later.