Alcohol use disorder (AUD) is complex psychiatric disorder that is characterized by excessive alcohol drinking, alcohol dependence and relapses even after long periods of abstinence (Siomek-Gorecka et al., 2021). It is a devastating public health problem and according to a WHO Global status report on alcohol and health in 2018, harmful use of alcohol resulted in about three million deaths worldwide in 2016 (World Health Organization, 2018). Alcohol and other drugs of abuse have been shown to change the synaptic plasticity and function of neurons in specific brain regions, resulting in long term changes at molecular, cellular and behavioural level (Berkel and Pandey, 2017). It has been shown that during the development of AUD, both genetic and environmental factors play an important role (Siomek-Gorecka et al., 2021). The fact that alcoholism tends to run in families has long been known and almost all twin studies have shown that heritability in AUD ranges from 40 to 60% (Ducci and Goldman, 2008; Verhulst et al., 2015).

In this study, we focused on alcohol-induced aberrant DNA methylation. DNA methylation is considered to be the most stable epigenetic modification with critical roles in brain development (Lister et al., 2013; Kinde et al., 2015). DNA methylation has been an important area of research also in the study of molecular mechanism to psychiatric disorders (Liu et al., 2018). Accumulated evidence has suggested that abnormalities in global methylation and methylation of genes could play a role in the pathogenesis of many forms of mental illness (Kuehner et al., 2019). In mammals, DNA methylation occurs predominantly at cytosine-guanine (CpG) dinucleotide sites, of which approximately 60–80% are methylated (Wu and Zhang, 2017). Furthermore, these CpG-rich regions are often methylated near gene promoters, which typically causes gene silencing (Deaton and Bird, 2011; Li and Zhang, 2014). DNA methylation is catalysed by DNA methyltransferases (DNMTs), including DNMT1 (maintenance DNMT), DNMT3a and DNMT3b (de novo DNMT) (Okano et al., 1999; Bestor, 2000; Goll and Bestor, 2005). To establish DNA methylation, DNMTs relocate the methyl group taken from S-adenosylmethionine (SAM) and covalently attach it to the fifth carbon of the pyrimidine ring of cytosine, forming 5-methylcytosine (5mC) (Jaenisch and Bird, 2003; Moore et al., 2012).

DNA demethylation caused by removal of the 5mC modification, can be performed in several ways. First, the modification can be removed by passive demethylation, where methylation maintenance process is impaired and no new 5mC marks are copied during DNA replication. Second, ten-eleven translocation enzymes (TET1–3) actively demethylate DNA. TET enzymes can mediate the oxidation of 5mC in order to generate hydroxymethylcytosine (5hmC) that can be further modified (Wu and Zhang, 2017). Previously the 5hmC was considered only as a transient marker of demethylation, but now it is treated as a stable epigenetic mark itself (Hahn et al., 2014). Several studies have shown that 5hmC accounts for approximately 40% of epigenetic cytosine modifications in the brain, which is a much higher density than in any other tissue found in the body (Li and Liu, 2011; Loh et al., 2017). 5hmC is found to be highly enriched within active gene promoters and in gene bodies of highly expressed genes (Wu et al., 2011). Several studies indicate that both hypomethylation and hypermethylation can be observed in postmortem human alcoholic brain (Manzardo et al., 2012; Wang et al., 2016). Although previous reports suggest a role of DNA methylation in alcohol-related behaviours, the exact molecular mechanisms by which DNA methylation contributes to long-term neuroadaptations in alcohol dependence are far more complex and remain to be clarified (Tulisiak et al., 2017; Pucci et al., 2019).

Alcohol affects many parts of the brain and exerts its effects through different neurotransmitter systems. One of the systems affected by ethanol is the endogenous opioid system (EOS), that has been shown to play critical role in development of alcohol and other drug addictions (Gianoulakis, 2001; Sirohi et al., 2012). In alcohol addiction, ethanol consumption abnormally stimulates the release of endogenous opioid peptides and activates their main receptors in the brain (Le Merrer et al., 2009; Erikson et al., 2018). The three main endogenous ligands and their receptors are the beta-endorphins (βEND), which binds to mu (µ) opioid receptor (MOR), the enkephalins (ENK), which binds to delta (δ) opioid receptors (DOR) and the dynorphins (DYN), which binds to kappa (κ) opioid receptor (KOR) (Sirohi et al., 2012). These endogenous ligands (βEND, ENK, and DYN) are produced from larger precursor peptides known as proopiomelanocortin (POMC), preproenkephalin (PENK) and preprodynorphin (PDYN) by proteolytic cleavage.

The drugs of abuse, including alcohol, induce an increase in the extracellular concentration of dopamine in the Nucleus Accumbens (NAc), and it was suggested that may mediate their rewarding and reinforcing properties (Di Chiara and Imperato, 1985, 1988; Le Merrer et al., 2009). Additionally, activation of MOR and DOR in the NAc enhances the extracellular concentration of dopamine in the NAc (Latimer et al., 1987; Di Chiara and Imperato, 1988; Longoni et al., 1991). Given the role of NAc in different affective and addiction disorders, the presence of KOR/DYN system in the NAc and the changed regulation of KOR/DYN in the brain during AUD, it is feasible that KOR/DYN changes in the brain are altered by epigenetic modification and neuroadaptation caused by long term alcohol use (Wee and Koob, 2010). Therefore, our aim for this study was to replicate and extend previous findings on the aberrant DNA methylation and expression of selected genes in the NAc, a critical region of the reward pathway, following intermittent voluntary ethanol intake. We hypothesized that long-term gene expression changes in the brain, via epigenetic modifications, underlie molecular mechanisms of AUD and relapses after abstinence. To test this hypothesis, we used two different rat strains, the standard laboratory Wistar rat and the alcohol preferring Alko Alcohol (AA) rat. The AA rat line is produced by selective breeding based on high voluntary alcohol consumption, and therefore, these rats are naturally prone to high alcohol consumption, which is hypothesized to occur due to abnormal function of opioidergic mechanisms (Eriksson, 1968; Hyytiä and Sinclair, 1989; Sommer et al., 2006; Oinio et al., 2021). Alterations in Kor/Pdyn and Dnmt3a mRNA, promoter methylation and hydroxymethylation levels were determined. Furthermore, we analysed global DNA methylation and hydroxymethylation, DNMT/TET activity and mRNA levels of other AUD-relevant genes from EOS and dopamine systems.

In this work, we investigated whether intermittent ethanol exposure affects the expression and enzymatic activity of the epigenetic DNA modifiers DNMT and TET using the alcohol-preferring rat model. Previously, we have shown that addictive substances and environmental factors, may affect the transcription of epigenetic DNA modifiers in different brain regions (Anier et al., 2018; Urb et al., 2020). Thus, it can be assumed that ethanol causes similar epigenetic changes in different regions of the brain. As the NAc plays a central role in the mechanisms of drug addiction, we focused our study on this area.

We report three key findings. First, we show that intermittent ethanol exposure increased significantly the mRNA levels of DNMTs and these changes were associated with significant increases in enzymatic activity in AA-ethanol rats compared with control rats in the NAc. No significant changes were observed in TET mRNA levels or enzymatic activity. Second, we demonstrate a significant increase in the relative percentage of global methylation and hydroxymethylation in the AA-ethanol group compared with the control groups. Finally, our selected gene expression and promoter methylation status analysis revealed a significant increase in the mRNA levels of Pdyn, Kor, and Dnmt3a; however, we failed to convincingly demonstrate that the aberrant DNA methylation in promoter areas are responsible for the changed gene expression of these genes.

Despite a decade of research, there is no consensus on alcohol-induced brain DNA methylation in experimental animal models. Previous studies in rodents have been suggested that chronic ethanol consumption induces DNA hypomethylation with histone acetylation (Mahadev and Vemuri, 1998; Berkel and Pandey, 2017). Subsequent studies in experimental animal models suggests that chronic ethanol administration tends to increase DNA methylation in various parts of the brain (Warnault et al., 2013; Barbier et al., 2015). In this study, we found that intermittent alcohol drinking in AA-rats increased significantly Dnmt3a and to a lesser extent both Dnmt1 and Dnmt3b mRNA levels, and the increase in expression of these genes was associated with increased DNMT enzymatic activity in NAc. Our data are consistent with previously published results, for instance, excessive alcohol drinking increased Dnmt1 expression and reduced histone H4 acetylation in NAc of rodents and reducing DNA methylation by inhibiting the activity of DNMT with systemic administration of 5-azacitidine prevented excessive alcohol use in mice (Warnault et al., 2013). Similarly, another study of NAc and medial prefrontal cortex (mPFC) from rats following weeks of abstinence from ethanol showed a significant increase in Dnmt1 expression and in global DNA methylation that was associated with a downregulation of a cluster of synaptic genes specifically within mPFC neurons (Barbier et al., 2015).

Only a few reports have been described the aberrant DNA methylation in the alcohol-preferred rat model. Qiao et al. (2017) demonstrated that, both of the mRNA and protein levels of DNMT3A and -3B in mPFC were upregulated after 35 days of alcohol exposure and this upregulation could be reversed by 5-Aza-2′-deoxycytidine treatment. Recently, Cui et al. (2020) showed, using the reduced representative bisulfite sequencing technology, that the methylation level of promoter region in the mPFC of rats exposed to chronic alcohol was significantly increased. In addition, chronic alcohol exposure increased the mRNA and protein levels of DNMT3B and methyl CpG binding protein 2. However, short term alcohol exposure did not affect their expression. Therefore, based on previous reports and our results, accumulating data indicate that repeated ethanol exposure increases DNMTs expression in various brain regions in different animal models of AUD.

There are fewer data on alcohol-induced DNA demethylation in brain and other tissues. Finegersh et al. (2015) demonstrated that repeated intermittent alcohol exposure increased Tet1 mRNA expression in the NAc but not in cortex in a mouse AUD model. However, in a recently published study, Ji et al. (2019) found that ethanol exposure to liver cells reduces Tet1 and 5hmC formation is involved in hepatocyte apoptosis in alcoholic liver disease progression. We found no changes in Tet1-3 mRNA levels or DNMT activity in NAc in the rat AA model. The effect of repeated ethanol administration on TET gene expression and enzymatic activity in different brain regions has not been previously studied in a rat model. Thus, results suggest that the effect of repeated ethanol on TET gene expression and enzymatic activity is more variable than DNMTs expression and may depend on the species of animal and the tissue under examination.

The data collected suggest that both DNMTs and TETs may function as epigenetic editors in the mechanisms of drug addiction induced by substance of abuse (Saha et al., 2021). The exact mechanism underlying ethanol-mediated changes in DNMT and TET gene expressions is poorly understood. It has been suggested that ethanol alters receptors (glutamate, dopamine, GABA) and ion channels which subsequently alter intracellular signalling cascades leading to the activation or inhibition of transcription factors and other nuclear proteins, eventually changing the expression of DNA modifying enzymes (Nestler, 2014). Taken together, our results from AA-ethanol rats suggest that repeated intermittent exposure to ethanol increases DNMT activity and it may be one of the molecular mechanisms for aberrant DNA methylation.

Changes in DNMT enzymatic activity implies possible downstream changes in percent global 5mC. To gain additional information on aberrant DNA methylation induced by repeated alcohol drinking, the global levels of 5mC and 5hmC in the NAc were measured. Our results demonstrated an increase in percent 5mC in AA-ethanol rats compared with control groups. These changes concur with our data showing increased DNMT enzymatic activity in NAc in AA-ethanol rats.

Previous studies in rodent models of AUD have not studied the global level of 5mC and 5hmC in the NAc after intermittent administration of ethanol. Surprisingly, we also found a statistically significant increase in global 5hmC levels. We speculate that an increase in 5mC (a precursor of 5hmC) leads to increases in 5hmC even if TET enzymatic activity is not changed. Thus, an increase in 5hmC levels may be a compensatory process to ensure optimal DNA demethylation in the cell. Changes in the global levels of 5mC and 5hmC also suggest that a new DNA methylation and demethylation balance develops following repeated ethanol exposure in the NAc. We have found similar phenomena also after repeated psychostimulant exposure (Anier et al., 2018). Therefore, our results support the findings that prolonged ethanol drinking increases the extent of DNA methylation at the whole-genome level and that these changes may underlie altered gene expression changes in the NAc.

In addition to the changes in ethanol-induced global methylation, we were interested in whether repeated ethanol drinking also affects the methylation state of promoters of selected genes in the AA rat model. There is currently no consensus on genes that could serve as marker genes in AUD models. Therefore, we selected candidate genes that have been studied in various alcohol models.

We examined the transcription of several genes from dopaminergic and opioidergic systems (Figure 4, Supplementary Figures S1, S2) previously associated with the AUD. In addition, we also analyzed DNA methyltransferases. There were no changes in Dat gene, but we found significant increases in Drd1, Drd2, Kor, Pdyn, and DNMT3a mRNA levels in AA-ethanol rats. The results of the methylation status in the promoter areas of selected genes were more complex. The Pdyn promoter region under study was significantly hypomethylated, while in the Kor and Dnmt3a promoter regions we saw a trend towards hypomethylation, but it did not reach statistical significance. Previous studies of promoter methylation status had revealed several genes that are affected by the interplay between alcohol and epigenetic regulation and that may play a role in alcohol addiction (Basavarajappa and Subbanna, 2016). Some of these genes exhibited hypermethylated promoters with others showing an opposite trend, suggesting that alcohol’s effects on DNA methylation are diverse and may be affected by numerous factors, including developmental stage, functional state of the cell, and specific gene targets in specific cell types. Using prenatal ethanol exposure model, Wille-Bille and colleagues demonstrated (Wille-Bille et al., 2018) upregulation of Pdyn and Kor mRNA levels in the ventral tegmental area (VTA) in infant and adolescent rats and Kor mRNA levels in the prefrontal cortex in infant rats. The changes in gene expression in the VTA were accompanied by a reduction of DNA methylation at the Pdyn gene promoter, and by a reduction of DNA methylation at the Kor gene promoter. These results suggest that different patterns of alcohol consumption in experimental animals may similarly affect the methylation and transcription of Pdyn and Kor gene promoters.

However, our results demonstrate a discrepancy between increased DNMT expression, increased global methylation and increased transcription of the selected genes in the NAc. The results of our selected genes (Kor, Dnmt3a) showed that the increase in transcription after repeated alcohol consumption in AA rats was only partially explained by a 5mC decrease in promoter, suggesting that other regulatory mechanisms may be involved. Accumulating data suggest that the DNA methylation and histone modification interact and influence each other and together may fine-tune gene expression (Cedar and Bergman, 2009; Krishnan et al., 2014). Alternatively, promoter regions selected for methylation status analysis were not representative. The possible weakness of 5hmC and 5mC detection methods at the specific CCGG is that it evaluates the change in methylation patterns in a very short region of the promoter, which makes it not necessarily representative.

Interestingly, the changes in the 5hmC level of the selected gene promoters were variable. This may indicate that after repeated alcohol administration a new balance between 5mC and 5hmC levels may be formed in the promoter region of some genes (e.g., Kor, Pdyn). However, no definitive conclusions about 5hmC changes can be drawn from the analysis of the three genes.

Finally, the results of our study indicate that the epigenetic changes also contribute to the AA-rat model susceptibility to ethanol drinking rather than being a pure response to alcohol exposure. This speculation is supported by the differences in DNMT enzymatic activity and the marker genes promoter methylation in AA-water rats compared to Wistar-water rats. The underlying mechanism for these epigenetic changes is unclear, but we hypotheses that the stress level of an individual rat in the early stages of development may affect epigenetic mechanisms of basal level and some of these epigenetic changes may persist into adulthood (Anier et al., 2014) and cause the higher-level ethanol consumption in adulthood. Further studies should elucidate possible mechanisms of ethanol sensitivity in AA-rats.

Our results, using alcohol preferring AA-rat model, support previous findings that repeated alcohol exposure tends to increase DNMT activity and may induce the aberrant DNA methylation in the NAc. The aberrant DNA methylation together with other epigenetic mechanism, such as histone modification can be the specific molecular mechanisms mediating dependence-induced neuroadaptations induced by alcohol consumption.