Sex differences in stress-induced alcohol intake: a review of preclinical studies focused on amygdala and inflammatory pathways

Abstract

Clinical studies suggest that women are more likely than men to relapse to alcohol drinking in response to stress, however the mechanisms underlying this sex difference are not well understood. A number of preclinical behavioral models have been used to study stress-induced alcohol intake. Here we review paradigms used to study effects of stress on alcohol intake in rodents, focusing on findings relevant to sex differences. To date, studies of sex differences in stress-induced alcohol drinking have been somewhat limited, however, there is evidence that amygdala-centered circuits contribute to effects of stress on alcohol seeking. In addition, we present an overview of inflammatory pathways leading to microglial activation that may contribute to alcohol-dependent behaviors. We propose that sex differences in neuronal function and inflammatory signaling in circuits centered on the amygdala are involved in sex-dependent effects on stress-induced alcohol seeking and suggest that this is an important area for future studies.

Introduction

The consequences of chronic alcohol use represent a major personal, public health and financial burden. Historically, men have had higher rates of problematic alcohol use than women (Schulte et al. 2009). However, the trend for an increase in alcohol use disorders (AUD) among women is alarming, and recent analyses suggest an increase in problematic drinking in women in the United States of more than 80% over the past 10 years (Grant et al. 2017). While pharmacological treatments are available for AUD, they were developed exclusively or primarily with samples of men, (Anton et al. 2006) and none of the currently approved treatments are known to target the multiple factors that differentially maintain drinking in women. AUD is characterized by physical dependence and neuronal perturbations induced by repeated alcohol exposure. Withdrawal from alcohol leads to a number of negative effects, including changes in mood and induction of negative affect, but also has life-threatening consequences, including seizure and coma. Many AUD therapies focus on alcohol intake but other co-morbid conditions, such as depression, perpetuate the use of alcohol, likely as an attempt to cope with psychiatric symptoms. Not surprisingly, multiple studies have demonstrated strong co-morbidity between AUD and psychiatric disorders (https://pubs.niaaa.nih.gov/publications/arh26-2/81-89.htm), including generalized anxiety disorder, depression and post-traumatic stress disorder, all of which can contribute to harmful drug and alcohol use. Of particular concern is the fact that women are more prone to negative-reinforcement drinking (NRD), and thus, stress-related drinking relapse, compared to men (Nolen-Hoeksema and Hilt 2006). Because stress sensitivity and rates of anxiety disorders are twice as high in women than in men (Remes et al. 2016) and lifetime anxiety predicts poorer drinking outcomes in women (Farris et al. 2012), it is critical to identify the mechanisms by which neurobiological circuits that regulate these behaviors can contribute to NRD and alcohol consumption.

Clinical (Keren et al., 2014; Logrip et al., 2018; Peltier et al., 2019) and preclinical (Pelloux et al., 2005; Cozzoli et al., 2014; Peñasco et al., 2015; Bertholemey et al., 2016; Shaw et al., 2020) studies demonstrate heightened susceptibility to stress-induced drinking in females. Animal models can recapitulate the effects of sex and stress on alcohol intake. For example, exposure to a stressor can increase ethanol intake in mice, with female mice increasing their alcohol intake more rapidly than males (Cozzoli et al., 2014). These animal models provide the possibility of examining the underlying mechanisms of stress induced changes in alcohol-related behaviors and exploring sex differences in these mechanisms.

This review will provide an overview of the different animal models used to decipher the connection between stress and alcohol intake and how sex differences can modulate this interaction. Based on the established role of the amygdala in stress-relevant behaviors, we will then provide an overview of the role of neuronal networks centered on the amygdala in alcohol-related behaviors and explore how perturbation of amygdala activity by alcohol could alter these behaviors. Finally, given the increasing body of evidence that inflammatory pathways in the brain are recruited in response to both stress and alcohol exposure (Crews et al. 2017), we will provide an overview of the possible role of microglia, a key cellular component of the brain inflammatory response, in reshaping the neuronal networks that contribute to, and perpetuate, alcohol use. The goal of this review is to identify critical lines of research needed to gain a greater understanding of stress-induced alcohol use, and to evaluate sex differences in some of the critical mechanisms underlying these behaviors.

Rodent models of alcohol use and relevance for stress-induced drinking

Several studies have investigated the effects of stress on alcohol intake and preference in rodent models, and the results differ based on species, strain, type of stressor, and the timing of stress and alcohol exposure in the experimental design (Spanagel et al., 2014; Weera & Gilpin, 2019). However, only a limited number of rodent studies have considered potential sex differences in how stress can alter alcohol seeking behaviors. Preclinical studies of sex-dependent effects of stress on alcohol drinking could shed light on the biological mechanisms underlying female-specific increases in human alcohol drinking to cope with stress and negative affect (Peltier et al., 2019).

To model stress-induced increases in alcohol intake and preference in rodents, the first consideration is the route of administration and the duration of exposure to ethanol. Rodents typically do not self-administer ethanol in sufficient amounts to induce behavioral intoxication or to reach physiologically relevant blood alcohol concentrations measured in human drinkers. Several strategies have therefore been proposed to promote robust alcohol intake in rodents, including selective breeding of alcohol-preferring rodent strains and/or manipulation of schedule of access to alcohol (Becker, 2012; Becker & Ron, 2014).

Classical behavioral approaches to volitional administration of oral ethanol include access to an ethanol solution through an operant task (Heidbreder et al., 2007; Lopez & Becker, 2014; Sparta et al., 2009) or home cage drinking of an oral ethanol solution. Access to ethanol solutions can either be unlimited throughout the duration of the experiment (Crabbe et al., 2010; García-Pardo et al., 2017; Hwa et al., 2011), intermittent between days (Bloch et al., 2020; Hwa et al., 2011; Warnault et al., 2013), or limited to specific times of the day (Becker, 2012; Olney et al., 2018; Rhodes et al., 2005; Thiele & Navarro, 2014). These approaches differ in the extent of alcohol intake and blood alcohol levels achieved, with intermittent and limited access models allowing a much higher level of intoxication than unrestricted access (Hwa et al., 2011; Rhodes et al., 2005). Depending on the experimental design, however, reaching maximal alcohol intake in the absence of stress exposure could be counterproductive if the goal is to assess how stress can increase intake. In addition, the availability of choice between water and ethanol can provide stronger construct validity for the human condition, compared to models in which alcohol intoxication is reached because no other fluid is available (Cannella et al., 2019).

The nature and frequency of stress exposure is another factor that can be varied across experiments. Studies exploring stress-induced alcohol consumption have involved physical restraint (Farook et al., 2009; Marianno et al., 2017; Walker et al., 2015), social defeat (Newman et al., 2018; Norman et al., 2015), exposure to predator odors (Cozzoli et al., 2014; Hwa et al., 2011; Shaw et al., 2020), forced swim (Morais-Silva et al., 2015), footshocks (Breit & Chester, 2016; Cozzoli et al., 2014), pharmacologically-induced stress (Bertholomey et al., 2016; King and Becker, 2019; Ballas et al., 2021), or a combination of these stressors (Cozzoli et al., 2014). Interestingly, in male rats, forced swim and footshocks appear to elicit greater stress-induced alcohol drinking when compared to physical restraint, although this difference is not seen in mice (Noori et al., 2014). Importantly, stressors may differentially activate the hypothalamic-pituitary-adrenocortical (HPA) axis in male and female animals (Albrechet-Souza et al., 2020; Babb et al., 2013; Bland et al., 2005; Cozzoli et al., 2014), and thus the potential for sex-specific sensitivity to particular stressors and their effects on alcohol drinking should be an important point of consideration.

Human data strongly suggest that the interactions between stress and alcohol intake are bidirectional (Peltier et al., 2019): stress can prime individuals for subsequent alcohol seeking (Childs et al., 2011), while alcohol can also increase responsivity to stress (Becker & Koob, 2016; Bertholomey et al., 2016). Given the complex interactions between stress exposure and alcohol intake, complementary behavioral approaches have been developed to model different dimensions of the human condition. Several rodent models have focused on stress as a trigger for relapse in stress-induced alcohol seeking (Lê and Shaham, 2002). These studies have shown that a variety of stressors, notably footshocks (Lê et al., 1998, Lê et al., 2011), predator odor (King and Becker, 2019) or the α2-adrenoceptor antagonist yohimbine (Ballas et al., 2021; Bertholomey et al., 2016; Borruto et al., 2021; King and Becker, 2019; Lê et al., 2011), can robustly reinstate alcohol seeking in rodents previously trained in operant self-administration of alcohol that have subsequently undergone extinction. Importantly, these stress-induced reinstatement tests are done in the absence of an alcohol reinforcer. While these operant approaches can more accurately capture the effect of stress on alcohol seeking as a model for human relapse during abstinence, animal models that explore alcohol drinking in the home cage have the advantage of exploring how stress can alter alcohol drinking in the absence of cue- or action-triggered intake (Becker et al., 2011). In particular, home cage studies provide opportunities to study how drinking alcohol after stress exposure can ameliorate neuroadaptive imbalances arising from prior exposure to both alcohol and stress, and thus can be helpful in the mechanistic study of negative reinforcement drinking with relevance to relapse as well.

Only individuals who have learned that alcohol can reduce a negative affective state would be more likely to drink alcohol to alleviate stress (Heilig et al., 2010; Noori et al., 2014; Spanagel et al., 2014). Thus, animal models of stress-induced increases in alcohol intake require a history of prolonged ethanol exposure coupled with repeated exposures to stressors. For instance, male mice increase ethanol intake after repeated cycles of stress, but only if previously trained to consume high levels of alcohol, as can be achieved with chronic-intermittent exposure (CIE) (Anderson et al., 2016; Lopez et al., 2016) or a scheduled high-alcohol consumption paradigm (Finn et al., 2018). These two models incorporate cycles of binge alcohol intoxication followed by repeated withdrawal periods that dramatically increase alcohol intake compared to other paradigms (Holleran and Winder 2017). Withdrawal is thought to promote negative reinforcement drinking via alterations of the HPA axis (Blaine & Sinha, 2017; Koob, 2003; Rasmussen et al., 2000), thus having the potential to increase stress-induced drinking. Furthermore, a review by Becker and colleagues (Becker et al., 2011) noted that chronic exposure to stress is more likely to enhance alcohol drinking in rodents when compared to acute stressors. Interestingly, stress-induced effects on alcohol intake are only evident when the stress is removed in time from the availability of alcohol (Noori et al., 2014), potentially due to the time needed for stress-induced changes in neuroplasticity to alter alcohol seeking behavior (Spanagel et al., 2014). Taken together, these data suggest that the timing and frequency of both stress and alcohol exposure are likely to be critical parameters for modeling stress-induced alcohol drinking in rodents.

Not surprisingly, the limited number of preclinical studies that have explored stress-induced alcohol intake differ fundamentally in stress-alcohol timing and frequency, as is summarized in Table 1. Most of the studies that introduce the stressor prior to alcohol exposure focus on how stress occurring early in development determines future alcohol intake and preference during late adolescence and adulthood. Most of the studies done in adult rodents, in contrast, introduce alcohol exposure before repeated bouts of stress exposure. Nonetheless, only a small number of published studies have used both male and female rodents within the same experimental design (Table 2).

TABLE 1 -.

Stress Exposure prior to Alcohol Drinking

Study	Species, Strain	Sex	Stressor	Frequency and Length of Stress Exposure	Time between Stress and Alcohol	Alcohol Access	Frequency and Length of Alcohol Exposure	Stress-induced alcohol intake?
Pelloux et al. (2005)	Mice (CD1)	M/F	Tail suspension	Once, 6 mins	7 days	Continuous, TBC, increasing alcohol % every 8 days (3–20%)	Daily, 40 days	Yes
Cruz et al. (2008)	Mice (CFW)	M	Maternal separation	Daily, 14 days	45 days	Limited, 2 hr (DID) Operant responding, 30 min	Daily, 10 days Daily, 20 days	Yes Yes
Peñasco et al. (2015)	Rats	M/F	Maternal separation + Withdrawal + Restraint stress	Once, 24 hrs 2× 7 days, 1 wk apart Daily, 30 mins	18 days	Continuous, TBC Continuous, TBC Continuous, TBC	Daily, 22 days Daily, 4 days post stress Daily, 3 days post stress	No No Yes
Norman et al. (2015)	Mice (CFW)	M	Social defeat	Daily, 10 days	10 days	Continuous, TBC Intermittent, TBC + Operant responding (FR and PR) 30 min	Daily, 20 days 3x a week, 35 days. Daily, 35 days	Yes Yes Yes
Skelly et al. (2015)	Rats (Long-Evans)	M	Social isolation	Continuous, 6 weeks	8 weeks	Intermittent, TBC	3x a week, 6 weeks	Yes
Bertholomey et al. (2016)	Rats (Sprague Dawley)	M/F	Corticosterone in drinking water	Continuous, 20 days	10 days	Operant responding (PR), 1hr	Daily, 21 days	Yes
Newman et al. (2018)	Mice (C57BL/6J)	M	Social defeat	Daily, 10 days	10 days	Continuous, TBC Intermittent, TBC	Daily, 10 weeks 3x a week, 10 weeks	Yes Yes
Shaw et al. (2020)	Mice (C57BL/6J)	M/F	Predator odor	Daily, 15 days	12 days	Intermittent, TBC	3x a week, 4 weeks	No

TBC: two-bottle choice; DID: drinking in the dark.

TABLE 2 -.

Alcohol Drinking prior to Stress Exposure

Study	Species Strain	Sex	Alcohol Access	Frequency and Length of Initial Alcohol Exposure prior to Stress	Stressor	Frequency and Length of Stress Exposure	Stress-induced alcohol intake?
Farook et al. (2009)	Mice (C57BL/6J)	M	Continuous TBC	Daily, 7 days	Physical restraint	Daily for 5 days	Yes
Edwards et al. (2013)	Rats (Wistar)	M	Limited, 30 min, two-choice operant	Daily, 15 days	Predator odor	Once, 15 min	Yes
Cozzoli et al. (2014)	Mice (C57BL/6J)	M/F	Limited 2 hr (DID)	Daily, 15 days	One of each: Tail suspension Physical restraint Predator odor Foot shock Tail pinch	Each stressor applied at least 5 drinking sessions apart.	Yes (for predator odor and foot shock)
Walker et al. (2015)	Mice (C57BL/6J)	M	Continuous TBC	Daily, 3 weeks	Physical restraint + Forced swim	Daily, 7 days Daily, 2 days	No No
Anderson et al. (2016)	Mice (C57BL/6J)	M	Continuous TBC Intermittent TBC Limited, 2 hr (DID) Intermittent and limited (TBC, 2 hr) +alcohol vapor	Daily, 7 days 3x week, 1 week Daily, 1 week 3x week, 6 weeks Daily, 16hrs, 4 days	Forced swim Forced swim Forced swim Forced swim	3x week, for 4 weeks Same as above Daily, for 4 weeks, Same as above - cycle repeated 4 times on alternate weeks	No No No Yes
Lopez et al. (2016)	Mice (C57BL/6J)	M	Limited, TBC, 2 hr (DID) Limited, TBC, 2 hours (DID) + alcohol vapor exposure	Daily, 6 weeks Daily, until stable Daily, 16hrs, for 4 days	Physical restraint Social defeat Forced swim Social defeat Forced swim	Daily, 5 days Daily, 5 days Daily, 5 days Daily, 5 days, repeated 4x on alternate weeks Daily, 5 days, repeated 4 x on alternate weeks	No No No No Yes
Manjoch et al., 2016	Rats (Sprague-Dawley)	M	Continuous TBC	Daily, at least 7 days	Predator odor	Once, 15 min Re-exposed to context 2×15 min	Yes Yes
Finn et al., 2018	Mice (C57BL/6J)	M/F	Intermittent limited (single sipper, 30 min), then, continuous TBC	Every 3rd day, 7 sessions total Daily, 3 weeks	Predator odor	Every 2–3 days, 30 min, 4x.	Yes

TBC: two-bottle choice; DID: drinking in the dark.

A study done by Cozzoli and colleagues (Cozzoli et al., 2014) trained male and female mice using a restricted alcohol drinking schedule, in which mice could either drink alcohol or water in a daily 2-hour window. On selected days, mice were subjected to one of the following stressors: restraint, tail suspension, predator odor, footshocks or tail pinches, which were applied immediately prior to their drinking period. Of these stressors, only predator odor caused an increase in ethanol intake, with female mice showing a faster increase in intake compared to males (24 hours post stress vs 48 hours post stress; Cozzoli et al., 2014). Another study exposed male and female juvenile mice to predatory odor stress prior to introducing intermittent access to alcohol and water in their home cage several weeks later (Shaw et al., 2020). Although this study did not report any sex-specific changes in stress-induced alcohol drinking, male mice that had been stressed continued to drink alcohol longer compared to unstressed controls, even when alcohol reinforcement was devaluated by the addition of quinine (Shaw et al., 2020). Finally, a study done by Peñasco and colleagues (Peñasco et al., 2015) shows that periods of alcohol withdrawal and restraint stress in adult rats trained to drink alcohol result in a female-specific increase in alcohol intake, but only in those rats also exposed to maternal separation during adolescence. This study highlights that the timing and duration of stress and alcohol exposure are likely to be critical for identifying sex-specific increases in stress-induced alcohol drinking.

Overall, despite many advances in modeling stress-induced alcohol drinking in rodents, little attention has been placed on whether these behavioral models capture female-specific increases in alcohol intake after stress exposure, a phenomenon now well documented in humans (Peltier et al., 2019). It is therefore important to fine-tune existing behavioral approaches to capture this sex-specific dimension. A preclinical model that recapitulates the increased sensitivity to stress-induced drinking in females will be necessary for mechanistic explorations of the molecular, cellular and circuit-level basis for sex differences in alcohol intake.

Role of the amygdala in stress-induced alcohol intake

While behavioral models are beginning to show that female rodents may drink more in response to stress (Cozzoli et al., 2014), the neurocircuitry underlying sex differences in alcohol intake are mostly unknown (Becker and Koob 2016). The amygdala is likely to be involved in stress-induced alcohol intake, and potentially in sex-dependent differences in alcohol drinking, because it plays a pivotal role in the control of a wide range of behaviors related to stress, anxiety, fear, and alcohol intake. Importantly, the basolateral amygdala (BLA) underlies the complex control of behaviors that are related to both aversive (stress) and rewarding (acute alcohol intake) stimuli (Baxter and Murray 2002; Janak and Tye 2015; Crouse et al., 2020). Early investigations into the BLA suggested that the amygdala can rapidly detect negative emotional states and external stimuli to produce behavior that is adaptive to potential threats (Brown and Sharpey-Schafer 1888; Klüver and Bucy 1937; Weiskrantz 1956).

Neurons in the BLA and central amygdala (CeA) arise from distinct cell lineages. In the BLA, the majority of neuronal cells are excitatory projection neurons that are inhibited by a smaller number of local inhibitory interneurons (Janak and Tye 2015). BLA glutamatergic neurons project in part to the CeA (Roberto et al. 2012; Janak and Tye 2015), a striatal-like structure that is almost entirely composed of GABAergic inhibitory neurons, including both local interneurons and inhibitory projections to downstream regions such as the locus coeruleus (LC), bed nucleus of the stria terminalis (BNST), and periaqueductal grey (PAG) (Roberto et al. 2012; Spampanato et al., 2011; Janak and Tye 2015; Gilpin et al. 2015). Within the CeA, inhibitory neurons in the centrolateral (CeL) region act as a gate on activity of the centromedial (CeM) anxiety-promoting projection neurons (Ciocchi et al. 2010). Further, direct activation of the CeA by the BLA makes it the main output nucleus of the amygdala that drives neuroendocrine responses to stress (Sah et al. 2003).

The amygdala is a sexually dimorphic brain structure influenced by sex hormone signaling (Równiak et al. 2015; Price and McCool 2022). The balance of estrogen and androgen signaling can be disturbed by alcohol intake, which may contribute to maladaptive alcohol use (Morales et al. 2018;Dozier et al. 2019; Fulenwider et al. 2019; Lorrai et al. 2019; Scott et al. 2020; Priddy et al. 2017;Ford et al. 2004; Bertholomey and Torregrossa 2019). It should be noted that there are some discrepancies among animal models as to whether alcohol exposure alters the estrous cycle. In female rhesus monkeys, ethanol did not influence menstrual cycle length, including changes to the follicular or luteal phases, or progesterone levels (Dozier et al. 2019). In contrast, a study in female rats found that long durations of chronic intermittent ethanol intake disrupted the estrous cycle, and with longer exposure, there was an increased proportion of females in diestrus I and II compared to control females (Morales et al. 2018). With respect to effects of estrous cycle on alcohol intake, several studies show that non-human primates exhibit significantly higher alcohol intake during the luteal phase when compared to the follicular phase of the menstrual cycle (Dozier et al. 2019; Fulenwider et al. 2019; Lorrai et al. 2019; Scott et al. 2020; Priddy et al. 2017). Similarly, several lines of research show that estrogen levels are positively correlated with increased ethanol consumption (Bertholomey and Torregrossa 2019; Vandegrift et al. 2017; Molina-Martínez and Juárez 2020; Kerstetter et al. 2012; Larson and Carroll 2006; Hilderbrand and Lasek 2018; Juárez et al. 2002). For example, gonadectomized female rats show decreased binge drinking; however, when supplemented with 17beta-estradiol, ethanol consumption increased (Ford et al. 2004; Bertholomey and Torregrossa 2019). Conversely, gonadectomized males decreased alcohol self-administration when given replacement testosterone (Bertholomey and Torregrossa 2019).

A number of mechanisms may contribute to estrogen effects on alcohol intake. For example, ethanol-induced firing of VTA dopamine neurons is decreased when estrogen receptors are blocked in brain slices from female mice in diestrus (high estradiol), suggesting that estrogen heightens ethanol sensitivity of dopamine neurons (Vandegrift et al. 2017). Further, the response of dopamine neurons to ethanol was greater in ovariectomized mice following estradiol replacement (Vandegrift et al. 2017; Vandegrift et al. 2020). In addition to differences in activity of the dopamine system, estradiol has anxiolytic effects in female rodents (Koss et al. 2004; Tian et al. 2013) that are mediated through ERɑ and Erβ estrogen receptors (Österlund et al. 1998). Notably, there are regional differences in expression of these estrogen receptor subtypes, with high levels of ERɑ mRNA in BLA, while the CeA predominantly expresses ERβ (Österlund et al. 1998). Of note, ERβ is highly expressed in inhibitory, PV-expressing neurons in female rats in the amygdala, basal forebrain, and hippocampal regions (Blurton-Jones and Tuszynski 2002). Finally, amygdala ER expression levels are influenced by estradiol concentration (Österlund et al. 1998), likely contributing to differential responses across the estrous cycle.

In addition to gonadal hormones, allopregnanolone, a potent neurosteroid that increases GABA-A receptor signaling, can also increase alcohol consumption with differing effects across sex and species. Allopregnanolone can increase ethanol intake in male rodents and female monkeys (Sinnott et al. 2002; Rowlett et al. 1999; Grant et al. 2008; Grant et al. 1997; Dozier et al. 2019; Genazzani et al. 1998). Following chronic ethanol exposure in male monkeys, both tissue and circulating serum levels of allopregnanolone are significantly decreased in the amygdala, whereas, in a similar study of female monkeys subjected to chronic ethanol exposure, serum levels of allopregnanolone were unaffected (Beattie et al. 2017; Dozier et al. 2019). However, in a human clinical study of adolescent females, there was a significant increase in circulating allopregnanolone levels following alcohol intoxication (Torres and Ortega 2003). Taken together, these results emphasize the need for further research on the effects of steroidal hormones on the development and expression of AUD across sexes.

While steroid hormones can contribute to alcohol intake, numerous studies in males across animal species have found that alcohol acts primarily on GABA-A receptors (GABA_AR), potentiating receptor activity and enhancing inhibitory neurotransmission (Mihic 1999, Diaz et al. 2011, Floyd et al. 2004, McCool et al. 2003). Indeed, stress can increase ethanol self-administration via perturbation of the GABA system in male rats (Ostroumov et al., 2016). These interactions are particularly critical in the BLA. In male rats ethanol administration enhances GABA signaling onto BLA pyramidal cells, and can reduce anxiety-like and alcohol-seeking behaviors (Butler et al., 2014). However, physiological studies in rats have identified a decrease in inhibitory postsynaptic currents (IPSCs) after ethanol application to BLA slices (Zhu and Lovinger 2006; Ornelas and Keele 2018), suggesting alcohol decreases GABA signaling in the BLA. A single prolonged stress session alone did not result in significant sex differences in IPSCs recorded from male and female rat BLA slices; however, there were significant sex differences in neuronal excitability of BLA neurons when a single prolonged stress session was combined with ethanol exposure that was bath applied during recording (Ornelas and Keele 2018). Specifically, decreased neuronal spike firing was observed following ethanol application in BLA slices from female rats that were exposed to stress in-vivo (Ornelas and Keele 2018). Meanwhile, BLA slices from male rats that previously underwent stress showed a decreased hyperpolarization-activated, cyclic nucleotide-gated cation current (Ih) in response to acute ethanol application (Ornelas and Keele 2018). These results suggest that the neuronal network within the BLA is not only different in male and female animals, but that synergistic effects between stress and alcohol in this brain region could differ by sex.

Effects of alcohol in the amygdala are not limited to the BLA and are also observed in the central amygdala (CeA) and the bed nucleus of the stria terminalis (BNST or “extended amygdala”). Importantly, both the BLA and CeA send projections to the BNST (Miles and Maren 2019). There are striking similarities between the micro-circuitry of the BNST and the CeA and both are striatal-like in structure (Dong et al., 2000; Kash 2012). The BNST has been implicated in an increased drive to consume alcohol and also responds to stressful stimuli (Kash 2012). The Winder group has shown that there are extensive molecular adaptations and significant synaptic plasticity in the BNST following chronic alcohol exposure in male mice (Healey et al., 2008; Kash et al. 2009; McElligott and Winder 2009). The NMDA subclass of glutamate receptors (NMDAR) is activated by ethanol, and NMDARs in the ventral BNST become sensitized during acute withdrawal from chronic alcohol exposure in male mice in vivo (Kash et al. 2009). Chronic intermittent ethanol (CIE) exposure increased the probability of glutamate release in the stria terminalis of both male and female rats, with males starting to show a difference at 3 days of CIE, whereas it took 7 to 10 days of CIE to see the same effect in females (Morales et al. 2018); however, no sex difference was observed in withdrawal-induced anxiety in an elevated plus maze test after 3 days of CIE. These results indicate that a different synaptic mechanism drives expression of withdrawal-induced anxiety, possibly involving changes in activity of GABAergic interneurons (Morales et al. 2018).

Several studies have found that in brain regions where GABA_B receptors are expressed, such as the CeA, neurotransmission is potentiated by ethanol (see Fig. 1). Conversely, in the hippocampus, blocking GABA_BRs is required to observe alcohol-induced GABAergic transmission in male mice and rats (Roberto et al. 2003; Ariwodola and Weiner 2004; Nie et al. 2009), consistent with the idea that the ability of alcohol to facilitate GABA neurotransmission might be limited by GABA_BR-mediated presynaptic feedback (Ariwodola and Weiner 2004). Acute alcohol application to male rat BLA and CeA slices potentiates GABAergic transmission through pre- and post-synaptic mechanisms (Roberto et al., 2003), while decreasing glutamatergic activation (Roberto et al., 2012). Following chronic ethanol exposure in male ethanol-preferring rats, NMDARs are upregulated, leading to greater CeA excitability ex vivo (Obara et al. 2009). This is also true in other brain regions such as the hippocampus, where acute alcohol application inhibits glutamatergic transmission by decreasing transmission via NMDA and AMPA receptors, whereas chronic alcohol exposure up-regulates NMDA receptor-mediated transmission in male rats ex vivo and in vitro (Kalluri et al. 1998; Carpenter-Hyland et al. 2004; Carpenter-Hyland and Judson Chandler 2007). A compelling study that examined both male and female rats subjected to stress and then alcohol found that there were increased GABAergic miniature inhibitory postsynaptic currents (mIPSCs) and increased cytokine levels in CeA slices (Steinman et al. 2021). Interestingly, female rats that were exposed to stress in a familiar environment showed greater mIPSC frequency in the CeA, whereas males that were exposed to stress in either a novel or a familiar environment showed greater mIPSC amplitude (Steinman et al., 2021), suggesting that increases in GABAergic transmission may occur in CeA in both females and males, but through different mechanisms. These studies support the hypothesis that acute alcohol intake increases GABAergic transmission (Fig. 1), resulting in its ability to decrease behaviors relevant to anxiety; conversely, following chronic alcohol consumption glutamatergic transmission is enhanced and GABA transmission is decreased, leading to increased excitation/inhibition balance in a number of brain regions including the BLA and CeA, resulting in increased anxiety.

Figure 1. Hypothesized signaling in the BLA and CeA relevant to alcohol use disorder.

A) Following acute alcohol exposure there is a transient increase in inhibitory signaling onto excitatory pyramidal neurons in the BLA, leading to dampening of glutamatergic transmission. The CeA, in turn, receives decreased innervation from the BLA, leading to increased inhibitory output from the CeA. B) Following chronic ethanol exposure there is a decrease in GABAergic transmission from the interneurons in the BLA onto the pyramidal excitatory neurons of the BLA. There is also increased glutamatergic transmission within the BLA resulting from an increase in the AMPA/NMDAR ratio. Despite a decrease in CEL interneuron signaling, glutamatergic neurons in CeA receive greater input from PKCd+ interneurons from the CeM. The inhibitory output from the CeA is therefore dampened leading to an increase in alcohol seeking and anxiety.

Abbreviations: PFC: Prefrontal Cortex; BLA: Basolateral Amygdala; CEA: Central Amygdala; CEL: Centrolateral Amygdala; CEM: Centromedial Amygdala PV+: GABAergic parvalbumin-expressing interneurons; PKCd+: GABAergic protein kinase C delta-expressing interneurons; PN: Pyramidal Neurons; vHPC: Ventral hippocampus; BNST: Bed nucleus of stria terminalis; NAC: Nucleus accumbens; HYP: Hypothalamus; DVC: Dorsal vagal complex; VTA: Ventral tegmental area; PAG: Periaqueductal grey; PVT: Paraventricular nucleus of the thalamus

There are several sex differences in the types of interneurons found in amygdala subregions; for example, a higher density of calcium binding proteins (calbindin (CB)+ and parvalbumin (PV)+) has been observed in the BLA of female guinea pigs compared to males (Równiak et al. 2015). Furthermore, immunohistochemical studies in female rats identified a higher density of PV+ interneurons during diestrus and decreased density during proestrus (Blume et al. 2017). Because chronic ethanol consumption can increase the time spent in diestrus (Morales et al. 2018; Österlund et al. 1998; Blurton-Jones and Tuszynski 2002; Blume et al. 2017), future studies will be needed to determine whether a correlation exists between the total number of interneurons co-expressing PV and ERß and how this could underlie ethanol consumption.

GABAergic and glutamatergic mechanisms are necessary for the development and perpetuation of alcohol intake. There may be important differences in signaling mechanisms between sexes, although sex differences in glutamate and GABA signaling following alcohol use have not been studied as extensively. In one study, CeA neurons in male rats were shown to be sensitive to alcohol-induced inhibition of glutamatergic inputs to the structure, whereas female rats showed reduced sensitivity to alcohol-mediated inhibition of the CeA (Logrip et al., 2017). Thus, the interaction between neurons in different amygdala subregions could regulate alcohol consumption, with different interactions dominating in male and female animals.

In addition to changes in GABA signaling, epinephrine and norepinephrine (NE) levels are increased during withdrawal in individuals with AUD. A review of the literature has led to the hypothesis that noradrenergic signaling in the amygdala is modulated by both chronic alcohol use and by anxiogenic stimuli (Glavin 1985; Morilak et al., 2005). For example, several days of ethanol exposure can increase levels of NE and stress hormones such as cortisol in rats (Patterson-Buckendahl et al., 2005); however, a recent study found that alcohol does not stimulate the noradrenergic system directly, but instead, corticotropin releasing factor is required to increase norepinephrine release in the CeA of male rats following alcohol exposure (Hedges et al. 2020). Reducing noradrenergic tone may reduce stress-induced relapse to alcohol-seeking (Smith and Aston-Jones 2008). Clinically, guanfacine (an adrenergic receptor agonist) has shown efficacy in reducing smoking- and cocaine-induced relapse induced by stress, and can improve outcomes for patients with AUD (Fox et al., 2014; McKee et al., 2015). Guanfacine can decrease anxiety- and depression-related behaviors in mice, with similar behavioral effects in male and female animals, although sex differences in neuronal activation were observed in the BLA (Mineur et al., 2015). Reducing NE globally can also decrease ethanol preference in 2-bottle choice, ethanol conditioned place preference, and total ethanol consumption (Fitzgerald 2013). However, sex-dependent mechanisms through which the noradrenergic system mediates changes in alcohol intake following stress have not been studied systematically. Table 3 summarizes recent studies on the effects of the combination of alcohol and stress on synaptic and molecular processes. The studies summarized in Table 3 were mostly carried out using male rodents, highlighting the need to include both female and male animals to identify potential sex-specific differences in the effects of stress and alcohol on the underlying neurocircuitry in brain areas relevant to AUD.

Table 3 -.

Synaptic and molecular effects due to alcohol and stress

Study	Species Strain	Sex	Alcohol Access	Stress	Brain Region	Alcohol/Stress Affect	Manipulation and effect
Sillaber et al. 2002	Crhr1 KO mice	M	TBC, continous, 40+ days	FST, SD	Hippo +, Amy −, NaC+	↑ NMDA- NR2B	NA
Edwards et al. 2013	Wistar rats	M	TBC, 30 mins, 7 days	PO	mPFC+, dmPFC+, CeA+, BLA+	↑ pERK	NA
Delis et al. 2013	Drd2 KO mice	M	TBC	CMS	Global expression of Drd2 +/− and Drd2−/−	Drd2+/− and −/− ↑consumption when exposed to stress. Drd2 +/+ ↑etoh when not stressed	NA
Walker et al. 2015	Rxfp3 KO mice	M	TBC	CMS/RS/FST	global	Rxfp3 KO mice reduced ethanol preference after stress	NA
Ostroumov et al. 2016	Rats (Long-Evans)	M	OESA	RS	VTA	In Vivo: ↓ DA Neruon firing In Slice: shift to GABAA Signaling	GABA and DA pharmacological manipulation
Ornelas and Keele 2018	Rats (Sprague Dawley)	M/F	in slice bath solution etoh	SPS	BLA+	In Slice: ↓ spike firing in BLA in F hyperpolarization activated current ↓ M	NA
Morales et al. 2018	Rat (Sprague Dawley)	M/F	CIE (Vapor inhalation)	Withdr-wal	MT − ST + BLA +	In Slice: changes in presynaptic glutamate release in vivo: change in estrus cycle and anxiety	Electrophysiology paired pulse in ST and BLA
Padula et al. 2020	Mouse (C57BL/6J)	M	CIE, TBC	FST	BLA+		systemic - KCa2.1–2.3 channel activator decreased drinking
Domi et al. 2021	Rats (Wistar)	M	Punishment-resistant self-administration, TBC	FS	CeA(PKCd+)	↑ CeA (PKCd+) cell expression in rats drinking despite stress	hm4Di in CeA PKCd+ cells decreased drinking
Steinman et al. 2021	Rats (Wistar)	M/F	TBC	FS in FAM and NOV	CeA+	↑ CeA GABAergic mIPSC ↑ cytokine levels	NA

Abbreviations: CIE: chronic Intermittent ethanol, TBC: two bottle choice, OESA: operant ethanol self admin, PO: predator odor, RS: restrain stress, SPS: single prolonged stress, CMS: chronic mild stress, FST: Forced swim test, FS: Foot shock, FAM: familiar, NOV: novel, F: female, M: male, BLA: basolateral amygdala, CeA: central lateral amygdala, Rxfp3: relaxin family peptide receptor 3, Drd2: dopamine receptor 2, PKCd: protein kinase C delta, KCa2.1–2.3: calcium activated potassium type 2, mPFC: medial prefrontal cortex, dmPFC: dorsal medial prefrontal cortex, hm4Di: human muscarinic 4 receptor designer receptor CNOactivated inhibitor, mIPSC: miniature inhibitory post synaptic current

+ indicates tested and found effects, - indicates tested but found no effects. Arrows indicated increasing or decreasing respectively

Contribution of microglia to stress-induced alcohol intake

Extensive human clinical studies have highlighted the role of inflammation in the etiology of stress and AUD. Several studies have demonstrated an increase in peripheral cytokine levels in subjects with AUD, particularly interleukin-6 (IL-6) and tumor necrosis factor α (TNFα), both of which are associated with alcohol craving and other affective changes, suggesting a potential link between inflammation and AUD-related behaviors (Laso et al., 2007; Gonzalez-Quintela et al., 2008; Heberlein et al., 2014). Research has shown alterations of immune-related genes in the brains of individuals with AUD (Lewohl et al., 2000; Mayfield et al., 2002; Liu et al., 2006; Crews et al., 2013; Vetreno et al., 2021); of particular note are increases in expression of microglial markers (He & Crews, 2008), as well as genetic and epigenetic alterations in microglia (Ponomarev et al., 2012; Brenner et al., 2020). Similarly, depression is associated with increased markers of microglial activation (Torres-Plata et al., 2014), and positron emission tomography (PET) studies have revealed alterations in brain inflammation in vivo in subjects with major depressive disorder (MDD: Holmes et al., 2018; Li et al., 2018; Richards et al., 2018) and AUD (Hillmer et al., 2017; Kalk et al., 2017; Kim et al., 2018).

Several studies have identified sex-specific effects of stress and alcohol consumption on microglial number and function. While some studies have shown heightened microglial responses to stress in females (Gildawie et al., 2020; Bekhbat et al., 2021), others have shown increased susceptibility in males (Woodburn et al., 2021); these differences may be due to differences in timing and type of stressor. As shown in Table 4, several preclinical studies have identified alcohol-induced increases in microglial number and activation in male rodents, as demonstrated through expression of Iba1, a marker upregulated in activated microglia (Sasaki et al., 2001), phagocytic markers CD68 (Kurishima et al., 2000) and CD11b (Ehlers, 2000), and the chemokine receptor Cx3Cr1 (Jurga et al., 2020), as well as measurements of cell morphology and binding of PET ligands to TSPO, a mitochondrial protein associated with neuroinflammation (Notter et al., 2018). Less work has included females, but there are data to suggest that alcohol consumption has heightened inflammatory effects in females, including upregulation of microglia-related genes, cytokines, and chemokines (Pascual et al., 2017), as well as increases in microglial number and activation (Barton et al., 2017). Thus, sex differences in immune function may underlie sex differences in AUD and the heightened susceptibility of women to stress-induced drinking. Women have higher levels of IL-6 (O’Connor et al., 2007; Chapman et al., 2009) and binge-drinking-induced endotoxin (Bala et al., 2014), which are associated with social disconnection and depressed mood (Moieni et al., 2015). Furthermore, autoimmune diseases are more prevalent in women (Whitacre, 2001); these data point to the possibility of increased immune activity in women that may prime heightened reactivity to challenges such as psychosocial stress and alcohol consumption.

TABLE 4 –

Effects of Alcohol & Stress on Microglia in Rodents

Study	Species, Strain	Sex	Alcohol Access	Stressor	Effects on Microglia
Fernandez-Lizarbe et al., 2009	Mice (C57BL/6J)	F	3d 4 g/kg IP	none	Increased microglial activation (CD11b IR)
Alfonso-Loeches et al., 2010	Mice (C57BL/6J)	F	5mo continuous TBC	none	Increased microglial activation (CD11b IR)
McClain et al., 2011	Rats (Sprague-Dawley)	M	4d 5 g/kg IG every 8h	none	Increased microglial activation (morphology)
Ehrlich et al., 2012	Rats (Sprague-Dawley)	M	12mo continuous drinking	none	Increased microglial activation (Iba1 and CD11b IR)
Qin & Crews, 2012b	Mice (C57BL/6J)	M	10d 5 g/kg IG	none	Increased microglial activation (Iba1 IR)
Marshall et al., 2013	Rats (Sprague-Dawley)	M	4d 5 g/kg IG every 8h	none	Increased microglial number (Iba1+ cells) and activation (TSPO ARG; CD11b IR)
Zhao et al., 2013	Rats (Sprague-Dawley)	M	25d intermittent IG	none	Increased microglial activation (CD11b IR)
Marshall et al., 2016	Rats (Sprague-Dawley)	M	4d 5 g/kg IG every 8h Second 4d binge 7d later	none	Increased microglial activation (CD11b IR); decreased microglial number (Iba1+ cells) Further increase in microglial activation; increased microglial number
Avila et al., 2017	Mice (C57BL/6J)	M	3w continuous drinking	none	Increased microglial activation (Iba1 IR)
Barton et al., 2017	Rats (Long-Evans)	M/F	4d 5 g/kg IG	none	Increased microglial number (Iba1+ cells) and activation (morphology) in females only
Walter & Crews, 2017	Mice (C57BL/6J)	M	3, 4.5, or 6 g/kg IG	none	4.5 or 6 g/kg increased microglial gene expression (Iba1 and CD68 mRNA)
Walter et al., 2017	Rats (Wistar)	M	5 g/kg IG 1h before stress 25d intermittent 5 g/kg IG 42d before stress	2h restraint + partial water immersion	Increased microglial activation (CD11b IR) Increased microglial activation (CD11b IR)
Lowe et al., 2020	Mice (C57BL/6J)	F	42d continuous drinking	none	Increased microglial activation (morphology) and decreased phagocytic activity (Iba1/CD68 colocalization)
Marshall et al., 2020	Rats (Sprague-Dawley)	M	2d 5 g/kg oral gavage 4d 5 g/kg oral gavage	none	Decreased microglial number (Iba1+ cells), increased microglial dystrophy (morphology) Decreased microglial number (Iba1+ cells), increased microglial dystrophy (morphology)
Socodato et al., 2020	Mice, C57BL/6J	M	10d 1.5 g/kg oral gavage	none	Increased microglial number (Cx3CR1+ cells, Iba1+ cells) and activation (CD11b IR, CD45 IR, morphology, Iba1 PE)
Tournier et al., 2020	Rats (Wistar)	M	14d intermittent 3 g/kg IP	none	Increased microglial activation (TSPO VT)
Warden et al., 2020	Mice (C57BL/6J)	M	4w intermittent TBC	none	Increased microglial number (Iba1+ cells)
West et al., 2020	Rats (Long-Evans)	M/F	3w 4 g/kg IG every 7d 8w 4 g/kg IG every 7d	none	No changes Increased microglial number (Iba1+ cells) and activation (morphology)
Aranda et al., 2021	Rats (Wistar)	M	9w intermittent self-administration, 2w abstinence, 3w reinstatement	none	Increased microglial activation (Iba1 IR and morphology)
Lee et al., 2021	Mice (C57BL/6J)	M	28d continuous TBC	28d social isolation	Alcohol potentiated stress-induced increases in microglial number (Iba1+ cells) and activation (morphology)

Abbreviations: ARG: autoradiography; IG: intragastric; IR: immunoreactivity; PE: protein expression; TBC: two-bottle choice; V_T: total volume of distribution

The preclinical literature highlights microglia as key mediators of the brain’s response to stress and alcohol consumption (see Fig. 2). Microglia are the brain’s resident macrophages; in their resting state, they display a ramified morphology and monitor the brain environment. Detection of a toxin or stressor triggers classical activation, in which microglia transform into a more ameboid morphology, upregulate expression of various pro-inflammatory factors, and work to phagocytose debris and dead cells (Fig. 2). Once the threat has been addressed, microglia transition into an anti-inflammatory, alternative activation state (Block et al., 2007; Colton et al., 2009).

Figure 2. Alcohol-mediated changes in microglial markers and activity.

A) At rest, GABA_A receptors, TLRs, and ERs are expressed on microglia. Resting microglia express several markers, including phagocytic markers such as CD11b, CD45, and CD68, and the pruning marker CX3CR1. B) Alcohol activates microglia via GABA_A receptors and TLRs; stress also results in microglial activation. The increase in activated microglia results in phagocytosis and synaptic pruning, which drives neurodegeneration. This coincides with release of ROS and pro-inflammatory cytokines such as TNFα, which also contribute to neurodegeneration. Enhanced neurodegeneration results in long-term changes in excitation/inhibition balance in brain areas such as the amygdala, contributing to emotional dysregulation and increased stress responses, leading to potentiation of alcohol consumption and further driving microglial activation. Estrogens acting through ERs inhibit microglial activation and the associated signaling cascades.

Abbreviations: ER: Estrogen Receptor; TLR: Toll-like Receptor; GABA_A: GABA-A Receptor; ROS: Reactive Oxygen Species.

In addition to playing a role in phagocytosis, microglia can also alter synaptic structure and function in the CNS (Tremblay and Majewska 2011; Tremblay et al., 2011). At baseline, microglia are physically associated with neuronal synapses, and react dynamically to changes in the microenvironment (Nimmerjahn et al. 2005). Microglia play a critical role in synaptic pruning via chemokine (C-X3-C motif) ligand 1 (Paolicelli et al. 2011; Paolicelli and Gross 2011). The mechanisms by which synapse number is regulated in vivo remain to be elucidated, but in vitro experiments suggest that microglia control synaptic activity by regulating synapse number (Schafer et al., 2012). Thus, microglial activation could contribute to reorganization of neuronal networks via synaptic pruning.

While acute microglial activation is likely adaptive, and the classic activation phenotype is necessary for maintenance of a healthy brain, prolonged activation of microglia results in oxidative stress and ultimately, neurotoxicity (Block et al., 2007; Colton, 2009; Franco & Fernández-Suárez, 2015). Alcohol and stress both activate microglia, and chronic exposure to either can drive persistent microglial activation, causing hypersensitivity of the neuroimmune system and dramatic neurodegeneration (Crews et al., 2017). Microglia tend to be more responsive in females compared to males, both at baseline (Schwarz et al., 2012) and in response to binge alcohol consumption (Pascual et al., 2017; Barton et al., 2017), suggesting that the transition to maladaptive microglial signaling could be more pronounced in females.

Alcohol induces activation of microglia via toll-like receptor 4 (TLR4; Fernandez-Lizarbe et al., 2009; Alfonso-Loeches et al., 2010; Crews et al., 2011; Fernandez-Lizarbe et al., 2013). Subsequent activation of nuclear factor κB (NF-κB) stimulates release of TNFα and other pro-inflammatory cytokines, which drive apoptosis in surrounding neurons (Crews et al., 2006; Boyadjieva & Sarkar, 2010; McClain et al., 2011; Walter & Crews, 2017). Sustained TLR4 activation due to chronic alcohol consumption prolongs microglial activation, shifting the brain into a state of maladaptive microglial signaling (Alfonso-Loeches et al., 2010; Vetreno & Crews, 2012). Another mechanism of alcohol-induced neurodegeneration occurs through the release of reactive oxygen species (ROS) from activated microglia (Boyadjieva & Sarkar, 2013; Qin & Crews, 2012a). The increase in neuronal death due to inflammation and oxidative stress disrupts cortico-limbic circuitry and may contribute to further alcohol consumption driven by heightened anxiety and deficits in executive function (Crews et al., 2011). In fact, pharmacological inhibition of microglial function can reduce alcohol consumption (Agrawal et al., 2014; Israel et al., 2021), while ethanol drinking induces upregulation of immune regulatory pathways in males (females were not investigated). These signaling cascades should therefore be investigated as potential mechanisms underlying sex-related differences observed in the neuroimmune response to alcohol (Finn et al., 2018).

Gonadal hormones likely contribute to sex differences in neuroimmune signaling. While the expression of estrogen receptors (ERs) on microglia and the anti-inflammatory effects of estrogen signaling have been well-documented (Johann & Beyer, 2013; Acosta-Martínez, 2020), the role of androgen receptors (ARs) remains unclear. Immunocytochemical analyses found ER and AR expression in microglia of male rats, but only after brain injury (García-Ovejero et al., 2002); however, an ex vivo study measuring receptor expression by PCR only identified ER transcripts (but not AR) in male and female microglia at baseline and showed that ER expression was downregulated after an immune challenge (Sierra et al., 2008). This discrepancy could be due to technical differences (Sierra et al., 2008), as other studies have shown ER expression in microglia using PCR or immunocytochemical measurement in cell culture (Baker et al., 2004; Liu et al., 2005; Bruce-Keller et al., 2008).

Estrogens and androgens exert anti-inflammatory effects by regulating microglia (Baker et al., 2004; Barreto et al., 2007; Yang et al., 2020) which may also contribute to the neuroprotective effect of estrogen (Liu et al., 2005). Furthermore, estrogen can inhibit microglial reactive oxygen species (ROS) production, phagocytic activity, and release of TNFα (Bruce-Keller et al., 2000; Liu et al., 2005; Acosta-Martínez, 2020). One study found that simvastatin, a lipophilic statin with estrogenic activity, reduces depressive-like behavior, upregulates ER expression, and inhibits microglial activation in ovariectomized rats (Menze et al., 2021). Another study found that both estrogens and androgens reduce microglial complexity at baseline in males, and that estrogens are necessary for stress-induced microglial remodeling in females (Bollinger et al., 2019). Although preclinical research suggests a facilitatory effect of estrogens on alcohol drinking in females, and an inhibitory effect of androgens in males (Finn, 2020), further studies are needed to understand the relationship between gonadal hormones, microglia, and alcohol-related behavior.

Microglial activation by ethanol relies, at least in part, on GABA signaling (Domercq et al. 2013), and human microglia express GABA_A receptors (Domercq et al. 2013). Furthermore, rodent studies suggest that accumulation of microglia in the hippocampus correlates with decreased GABA transmission and greater neuronal excitability, as measured by induction of long-term potentiation (LTP; Nistico et al. 2013). Microglial motility is also modulated by NE and inflammation increases α2AR expression on microglia, shifting microglial responses to NE release (Gyoneva and Traynelis 2013).

Research in male rodents has demonstrated that, like alcohol exposure, stress leads to activation of microglia, increasing release of pro-inflammatory cytokines and ROS and leading to neuronal death (Lu et al., 2014; Cheng et al., 2019). However, it remains unclear whether this also occurs in females (Bollinger, 2021). In males, microglial inhibition can reverse the depressive effects of stress and promote neurogenesis (Han et al., 2019). Stress (Frank et al., 2007) and alcohol (Qin & Crews, 2012b) can both activate microglia in males, making them hypersensitive to subsequent inflammatory stimuli. Furthermore, studies of male rodents have shown that alcohol and stress interact to enhance microglial activation (Walter et al., 2017), whereas inhibiting microglial activity can reverse the escalation in drinking and in anxiety-like behavior associated with alcohol dependence (Warden et al., 2020). Thus, chronic alcohol consumption can make the male brain more susceptible to stress-induced inflammation and vice versa, potentiating subsequent neurodegeneration, which in turn drives further emotional dysregulation and alcohol consumption (see Fig. 2). However, more research is needed to understand the neuroimmune effects of stress and alcohol in females.

Conclusions

Sex differences in stress-induced alcohol intake contribute to increased relapse to alcohol drinking in women. Several preclinical behavioral studies have demonstrated sex differences in stress-alcohol interactions, suggesting that rodent models can be useful in identifying mechanisms underlying sex-specific contributions to alcohol drinking. Limited access drinking paradigms coupled with repeated stress exposure appear to be most useful in studying stress effects on alcohol intake in rodents. Female rodents appear to be more sensitive to stressors in these drinking paradigms than males, although not many studies have used both sexes. Physiological studies have demonstrated some differences in ethanol effects on GABA and glutamate signaling in amygdala that could contribute to sex-dependent effects of alcohol. In addition, female animals are more likely to mount a neuroimmune response to stress and show microglial activation in response to alcohol. These observations suggest potential mechanisms for sex differences in stress-induced microglial perturbations, alcohol use and stress-induced alcohol intake. One hypothesis is that microglia are activated by ethanol exposure, reshaping neuronal dendritic arborization in several brain areas including the amygdala, leading to greater sensitivity to stress and increasing subsequent alcohol intake. Chronic activation of neuroinflammatory networks and microglia leading to neurodegeneration could lead to permanent deficits in the balance between GABA and glutamate signaling in these networks, leading to even greater sensitivity to alcohol-related behaviors. Future work should focus on identifying activity in brain systems that is most divergent across sexes in response to alcohol intake, whether sex differences in microglial activation contribute to stress-induced drinking behavior, and whether treatments that target the immune system may be more efficacious for women with AUD. This will involve additional model development to identify the patterns and timing of exposure to stress and alcohol that reveal sex differences in physiology and behavior.