Chronic alcohol use is a leading cause of preventable death in the United States (Center for Disease Control, 2025). It causes or exacerbates health and social problems, increases mortality rate, and costs society nearly a quarter trillion dollars each year (c.f., National Institute of Alcohol Abuse and Alcoholism [NIAAA], 2023; Rehm and Imtiaz, 2016). There are only a limited number of FDA-approved pharmacotherapies for alcohol use disorder (AUD) (e.g., Koob and Mason, 2016). Therefore, it is critical that we identify novel, efficacious pharmacotherapeutics for AUD. Genetic studies indicate that individuals with a family history positive (FHP) for alcoholism are more susceptible to developing AUD than family history negative (FHN) subjects (e.g., Lai et al., 2022). However, not every FHP individual develops an AUD, and FHN individuals can often develop AUD. Environmental risk factors such as other drugs of abuse, stress, history of child abuse, poverty, etc., can also contribute to the development of AUD (Deeken et al., 2020). Thus, the interaction of genes and environmental risk factors combine to enhance the risk for developing an AUD (Nieratschker, Batra, and Fallgatter, 2013). In addition, environmental risk factors can alter epigenetics, resulting in changes in phenotypes by modifying the chromatin structure or controlling gene expression without affecting the underlying DNA sequence (e.g., Peedicayil, 2023). Therefore, interactions between genetics, epigenetics, and environmental risk factors, all play a role in the development of AUD (e.g., Nieratschker, Batra, and Fallgatter, 2013). A commonly studied epigenetic modification is induced by histone (de)acetylation processes, which generally increase genetic transcription or accelerate rates of protein production (e.g., Moore, 2017; Pandey and Bohnsack, 2020). Clinical and pre-clinical studies have shown that alcohol exposure can alter histone acetylation and deacetylation via histone acetyltransferases (HATS), histone deacetylase (HDAC) enzymes, and other epigenetic processes (López-Moreno et al., 2015; Pandey et al., 2017; De Carvalho et al., 2021).

Epigenetic modifications are stable but can be reversed. For example, administrating HDAC inhibitors (HDACi) can reverse some of these epigenetic modifications, such as (de)acetylation. HDACs are repressors of gene transcription by deacetylating histone proteins, leading to a transcriptionally repressed chromatin structure. HDACs are divided into zinc-dependent and nicotinamide-adenine-dinucleotide (NAD)-dependent sirtuins (Hull et al., 2016). The HDAC class I family (HDACs 1, 2, 3, and 8), class IIa and IIb families (HDACs 4, 5, 6, 7, 9, and 10), as well as class IV family (HDAC 11) are zinc-dependent, and widely expressed in the brain (Hull et al., 2016; Palmisano and Pandey, 2017). In particular, isoforms of class I, II, and IV HDACs are expressed in brain neurons (Broide et al., 2007). There are low levels of HDACs found in astrocytes and HDACS 2, 3, 4, 5, and 11 are expressed in oligodendrocytes (Graff and Tsai, 2013). Class I HDACs, except for HDCAC six (only cytoplasm), are located in both the nucleus and cytoplasm (Gibson and Murphy, 2010). As noted above, HDACs are implicated in the development of alcohol dependence, ethanol-seeking, and other ethanol/drug-associated behaviors (Werner et al., 2021; Palmisano and Pandey, 2017; Pandey et al., 2017; Kamat et al., 2016). In vivo ethanol exposure and ethanol withdrawal can increase the expression of Class I HDACs, particularly HDAC2, in rodents (Arora et al., 2013; Pandey et al., 2015; Bohnsack et al., 2018; Chen et al., 2019), while in vitro exposure to alcohol can upregulate gene and protein expression levels of HDAC2 in human cells (Agudelo et al., 2011; Agudelo et al., 2016). Agudelo and colleagues (2016) found that all class I HDACs (1, 2, 3, and 8) when examined ex vivo in human primary monocyte-derived dendritic cells, from alcohol users, displayed significantly higher HDAC-associated RNA and protein levels compared to their controls (Agudelo et al., 2016).

The first generation pan HDACi (i.e., an inhibitor of Class I and Class II HDACs), trichostatin A (TSA), is a common HDACi used to examine the involvement of HDACs in excessive ethanol intake by rodents. TSA reduces voluntary ethanol drinking in alcohol-preferring (P) rats, to a greater extent than that observed in alcohol-nonpreferring (NP) rats (Sakharkar et al., 2014). It also reduced ethanol-drinking by outbred Wistar rats using an intermittent 2-bottle choice protocol (Bohnsack et al., 2018). These findings suggest that TSA attenuates ethanol consumption in both genetically selected and non-selected rodents. Clinically, HDAC inhibitors are used as anticancer treatments. However, pan-HDAC inhibitors lead to more adverse side effects such as diarrhea, nausea, vomiting, fatigue, thrombopenia, and neutropenia (c.f., Hontecillas-Prieto et al., 2020). These adverse side effects are reduced when treated with more selective HDACi’s (c.f., Hontecillas-Prieto et al., 2020). To date, there have been no published research examining the effects of entinostat (predominantly an HDAC1 and HADC3 inhibitor), quisinostat (second generation pan HADCi), and tubastatin-A (selective HDAC6 inhibitor) on ethanol consumption in rodents.

The risk for developing an AUD can be modeled through rodent bidirectional selective breeding for high ethanol consumption and preference vs. low ethanol consumption and nonpreference. The P and high alcohol drinking (HAD1) rat lines rats satisfy criteria put forth for an animal model of alcoholism (c.f., Bell et al., 2012; 2017; McBride and Li, 1998; McBride et al., 2014; Murphy et al., 2002). Research on genetic factors contributing to a vulnerability for developing high ethanol-consuming behavior have paralleled findings with FHP, versus control, individuals (e.g., Li et al., 1993; alpha-synuclein; Liang et al., 2003). Both P and HAD rats achieve pharmacologically, often exceeding that associated with binge-drinking (80 mg%), relevant blood alcohol concentrations under limited access, continuous access, and relapse-like conditions (Bell et al., 2006; 2011; 2012; 2014; Dhaher et al., 2012; Murphy et al., 1986; 2002). Therefore, the objectives of the present study were to examine the effects of entinostat, quisinostat, and tubastatin-A on ethanol drinking by adult male HAD1 and P rats.

Ethanol and water intakes were determined as the difference in the weight of glass bottle, and its contents, prior to and after each time-point (2-, 4-, and 24-h). Ethanol intake was then converted to g of absolute (i.e., corrected for ethanol’s specific gravity) ethanol consumed/kg of body weight/unit of time measurement (g/kg/time-period), water-intake was measured as ml/time-period, food-intake was measured as g/kg body weight, and body weight measured in grams. IBM SPSS version 29 was used to analyze data, and GraphPad version 10.3.1 was used to graph the data (mean ± SEM). Data were analyzed with mixed ANOVAs, with repeated measures for time-point and test day, and appropriate simple main effects with one-way ANOVAs, followed by Dunnett’s t-test for planned comparisons (i.e., dose vs. control). P-values <0.05 were considered statistically significant.

The major findings of the present study were that the HDAC inhibitors (HDACi) entinostat, with selectivity for HDAC1 and HDAC3 (Figure 3; Table 1), and quisinostat, a second generation pan-HDACi (Figure 5; Table 1), reduced ethanol intake in HAD1, while only entinostat reduced ethanol intake in P rats (Figure 2; Table 1). The HDACi tubstatin-A, with selectivity for HDAC6, did not alter ethanol intake in HAD1 or P rats (Figures 6, 7; Table 1). In the current study, all doses of entinostat reduced ethanol intake in HAD1 rats both acutely (2-h and 4-h post-injection) and chronically (24-h). In contrast, the highest dose of entinostat was the most effective in reducing 24-h ethanol intake in P rats. These findings suggest that entinostat has a more robust effect in reducing ethanol drinking by HAD1 rats than by P rats. One possible explanation for the differences is the different genetic backgrounds of the rats, although both rat lines were selectively bred using the same criteria (>5 g of ethanol/kg bodyweight/day; Bell et al., 2016; Bell et al., 2017; McBride et al., 2014). P rats were selected from a closed colony of Wistar rats at Walter Reed Army Hospital and transferred to the Indiana University School of Medicine in Indianapolis, Indiana, United States (cf., McBride and Li, 1998).

In contrast, HAD1 rats were derived from N/NIH heterogeneous stock rats (derived from eight inbred strains [ACI, BN, BUF, F344, M520, MR, WKY, and WN]) at Indiana University School of Medicine in Indianapolis, Indiana, United States (cf., Bell et al., 2012). Another possibility is entinostat’s higher binding specificity for HDAC1 (IC ₅₀ values = 190 nM) and HDAC3 (IC ₅₀ values = 420 nM), over other HDACs (Marks et al., 2011). This may have limited its effect on P rat drinking. It has been reported that P rats do not have innate differences in HDAC 1, 3, 4, 5, or six protein levels within the amygdala compared to their control non-alcohol preferring (NP) rats, but P rats have an innately higher level of HDAC2 compared to NP rats (Moonat et al., 2013). NP rats are the low alcohol-consuming counterpart from the bidirectional selection that yielded P rats (Bell et al., 2012; McBride and Li, 1998). Therefore, entinostat effects on P rats may be via inhibition of HDAC2 (binding specificity IC₅₀ value = 1900 nM; Marks et al., 2011). No published work to date has examined HDAC levels in HAD1 rats, compared with their low alcohol consuming low-alcohol-drinking (LAD1) counterparts (c.f., Bell et al., 2012), or HDAC levels in other brain areas of P rats.

Our findings are in line with previous studies, which demonstrated that entinostat can reduce ethanol intake in rodents. For example, systemic administration of entinostat (also known as MS-275) can robustly reduce excessive ethanol intake in ethanol-dependent rats using an oral operant ethanol self-administration procedure while not having an effect in non-dependent rats (Simon-O'Brien et al., 2015). Jeanblanc et al. (2015) reported that intracerebroventricular administration of entinostat reduced operant ethanol self-administration, decreased motivation to consume ethanol, and blocked relapse. Even after it was discontinued, Entinostat effectively reduced reacquisition to ethanol self-administration in rats (Jeanblanc et al., 2015). This may be due to entinostat’s long half-life (i.e., 9.8 h in rats and >50 h in humans) (Yang et al., 2014; Kurmasheva et al., 2019).

Entinostat has also been shown to reduce binge-like ethanol drinking in mice without altering saccharin consumption at a dose as high as 20 mg/kg (Warnault et al., 2013). This indicates that entinostat can selectively decrease ethanol drinking without altering the intake of natural rewards. However, the current study demonstrated that the highest dose (5 mg/kg) reduced food intake in both HAD1 and P rats during the last 2 or 3 test days without altering body weight. Thus, repeated administration of entinostat may result in possible cumulative effects on food intake. In addition, there may be species-dependent effects that differ between mice and rats. Entinostat tended to increase water intake, which may be a compensatory effect on liquid intake.

Ethanol has a strong modulating effect on the neuroimmune system (Crews et al., 2015; Cui et al., 2014; Erickson et al., 2019). The overexpression of neuroinflammatory factors observed following excessive ethanol consumption has been postulated to be a critical biological link between excessive ethanol consumption and the development of AUD (Crews et al., 2015; Cui et al., 2014; Erickson et al., 2019). Moreover, previous work from our laboratories has demonstrated that treatments with anti-neuroinflammatory agents can attenuate ethanol intake in HAD1 and P rats (ibudilast: Bell et al., 2015; PDE4 modulators; Franklin et al., 2015). Inhibition of HDACs can also reduce neuroinflammatory responses (Kannan et al., 2013; Patnala et al., 2017; Suliman et al., 2012; Xia et al., 2017). HDAC1 and HDAC3, like other Class I HDACs, are involved in innate immunity and cytokine production (Daskalaki et al., 2018; Kouzarides, 2007; Shakespear et al., 2011). Durham et al. (2017) reported that treatment with entinostat or siRNA knockdown of HDAC1 can reduce cytokine expression [i.e., tumor necrosis factor-α (TNF)-α and interleukin 6 (IL-6)] in microglia activated with lipopolysaccharide (LPS). Inhibition of HDAC3 has also been reported to reduce LPS-stimulated toll-like receptors in microglia, attenuate TNF-α and IL-6 production, and partially inhibit signal transducer and activator of transcription 3 (STAT3) and STAT5 as well (Xia et al., 2017). The latter are activated by cytokines and growth factors (e.g., Xia et al., 2017). Taken together, this suggests that entinostat’s ability to reduce ethanol intake in HAD1 and P rats may partly be due to its ability to inhibit inflammatory responses.

Quisinostat is a second-generation Class I and Class II HDAC inhibitor (Gatla et al., 2019), a good brain penetrant (Lo Cascio et al., 2023) with a half-life of about an hour in mice (Lo Cascio et al., 2023), and 8.8-h half-life in humans (Venugopal et al., 2013). At subnanomolar concentrations, it is specific for HDAC1 (IC ₅₀ values = 0.1 nM) and HDAC2 (IC ₅₀ values = 0.3 nM) (Lo Cascio et al., 2023), with ≥14-fold selectivity against the other Class I HDACs (i.e., HDAC3 [IC ₅₀ values = 4.8 nM] and HDAC 8 [IC ₅₀ values = 4.26 nM] as well the Class II HDACs; Arts et al., 2009). Compared with entinostat, quisinostat was not as effective in reducing HAD1 rats’ ethanol intake at the lowest and middle doses. Only the highest dose of quisinostat effectively decreased HAD1 rats' 24-h ethanol intake on all test days, with the greatest decrease on the third and fourth days. For the 2-h and 4-h drinking time points, quisinostat only effectively reduced ethanol intake on the third and fourth test days.

Interestingly, quisinostat did not significantly alter the P rats’ ethanol intake at any time point. Like entinostat, the highest dose of quisinostat tended to increase water intake suggesting a possible compensatory effect on liquid intake. However, HAD1 and P rats’ food intake was significantly reduced during the last 2 to 3 test days. To our knowledge, this is the first study to examine quisinostat’s effects on ethanol intake.

As previously mentioned, chronic ethanol exposure is associated with an overexpression of class I HDACs, particularly HDAC2, in limbic brain regions (i.e., hippocampus and amygdala). Therefore, it is possible that quisinostat’s attenuating effects on ethanol intake may, in part, be via inhibition of HDAC2 activity. However, Quisinostat was not effective in reducing P rats' ethanol drinking even though it has been reported that P rats have innately higher nuclear HDAC2 activity in the central and medial amygdala compared with their NP counterparts (Moonat et al., 2013). In addition, these authors reported that small interfering RNA (siRNA) knock-down of HDAC2 expression in the central nucleus of the amygdala (CeA) of P rats corrected ethanol-induced deficits by attenuating anxiety-like behaviors and inhibited ethanol-drinking behaviors in P rats (Moonat et al., 2013). A subsequent study found that trichostatin-A (TSA) can decrease HDAC2 levels as well as correct deficits in brain-derived neurotrophic factor, activity-regulated cytoskeleton-associated protein, and neuropeptide Y expression levels within the CeA of P rats (Sakharkar et al., 2014). It is noteworthy that this same trio of genes/proteins has been implicated in the development of AUD (Pandey et al., 2017).

Similar to pre-clinical research, HDAC2 expression in the amygdala was reported to be higher in the post-mortem brains of AUD subjects compared with control subjects (De Carvalho et al., 2021). An in vitro study revealed that ethanol exposure of a human neuronal cell line (SK-N-MC) upregulates gene and protein expression levels of HDAC2 through mechanisms involving oxidative stress, which often leads to inflammation (Agudelo et al., 2011). Moreover, TSA inhibition of HDAC2 gene and protein expression decreased reactive oxygen species production (Agudelo et al., 2011). Furthermore, TSA inhibited the effects following ethanol exposure of all class I HDACs (Agudelo et al., 2016). Collectively, these findings indicate that ethanol modulates class I HDAC activity and inhibition of class I HDACs may be useful for the treatment of AUD.

Quisinostat’s ineffectiveness in reducing P rat’s ethanol intake may be due to the sample size of five. However, the significant findings for water and food intake in P rats suggest against this conjecture. Another possibility is that a higher dose of quisinostat may be needed to reduce ethanol intake in P rats.

Tubastatin-A is highly specific for HDAC6 as a Class IIb HDAC with an IC50 of 15 nM for HDAC 6 (Butler et al., 2010). This was the only epigenetic compound tested in the current study that did not alter ethanol intake in HAD1 or P rats. This suggests HDAC6 may not regulate ethanol intake, at least in HAD1 and P rats. HDAC6 substrates, such as tubulin, heat shock proteins, and cortactin (Shen et al., 2020), are known to be affected by alcohol consumption (Azizov et al., 2023; Labisso et al., 2018; Mandrekar et al., 2008). The potential ineffectiveness of tubastatin-A could be due to alcohol altering different classes of the HDAC6 substrates than those targeted by tubastatin-A. Despite its low brain penetrance (Kozikowski et al., 2019) and a plasma half-life between 1 and 2 h in mice (Jochems et al., 2014; Shen et al., 2020), higher doses of tubastatin-A have shown promise in other animal disease models (e.g., Alzheimer’s disease, Parkinson’s disease, cardiac and pulmonary diseases; Shen et al., 2020). Therefore, another possible explanation is that a higher dose of tubastatin-A may be needed to reduce ethanol intake in rats.

Interestingly, López-Moreno et al. (2015) using RT qPCR reported that ethanol operant self-administration did not alter HDAC6 gene expression in rats, but acute ethanol intoxication in humans increased HDAC6 gene expression in peripheral blood samples. However, Hdac6 mRNA levels were found to be increased in the central and basolateral amygdala in ethanol-dependent male rats compared to non-dependent male rats (De Carvalho et al., 2021; Repunte-Canonigo et al., 2015). Moreover, a recent clinical study reported an increase in HDAC6 levels in the amygdala of AUD individual’s post-mortem brain tissue (De Carvalho et al., 2021). Inhibition of HADC6 has also been shown to attenuate LPS-induced inflammation markers in macrophages such as TNF-α, IL-1β, and IL-6 (Zhang et al., 2019). Therefore, further research is warranted to examine HDAC6 involvement in AUD as well as examine the effects of other HDAC6 inhibitors.

The current study does have some limitations. One limitation is that we did not examine female P and HAD1 rats. It is possible that female rats may respond differently to the HDAC inhibitors compared to male rats. The current study also did not examine whether the attenuating effects of the HDAC inhibitors were specific to ethanol drinking or whether they would reduce other reward drinking solutions, such as sucrose. Another limitation is that we did not investigate the effects of alcohol on HDAC genes and proteins across the different alcohol models, nor did we explore how the various HDAC inhibitor treatments might modify the effects of alcohol on HDAC gene and protein levels. Therefore, future studies should focus on assessing potential sex differences with different HDAC inhibitors and other rewards, as well as measuring gene and protein levels before and after HDAC inhibitor administration.

In conclusion, the current study indicated that entinostat effectively reduced ethanol intake in two genetic rat models selected for high ethanol consumption, while quisinostat had a robust effect, it was effective in only one of these genetic rat models. Thus, genetic background may influence the effectiveness of certain HDAC inhibitors in reducing ethanol consumption. Finally, this study provides further evidence that targeting Class I HDACs may be an effective treatment for AUD.