Transcriptional and epigenetic regulation of microglia in substance use disorders

Abstract

Microglia are widely known for their role in immune surveillance and for their ability to refine neurocircuitry during development, but a growing body of evidence suggests that microglia may also play a complementary role to neurons in regulating the behavioral aspects of substance use disorders. While many of these efforts have focused on changes in microglial gene expression associated with drug-taking, epigenetic regulation of these changes has yet to be fully understood. This review provides recent evidence supporting the role of microglia in various aspects of substance use disorder, with particular focus on changes to the microglial transcriptome and the potential epigenetic mechanisms driving these changes. Further, this review discusses the latest technical advances in low-input chromatin profiling and highlights the current challenges for studying these novel molecular mechanisms in microglia.

Introduction

Substance use disorder (SUD), or addiction, is a chronic relapsing disease characterized by excessive substance use, despite negative consequences (1). SUD is not only a public health concern, but is also economic burden, as the economic impact of substance misuse and SUDs are ever increasing and is estimated to cost the United State over $400 billion annually in crime, lost productivity, and healthcare (US DHSS Surgeon General, NDTA DEA). Individuals that suffer from SUD will transition from an initial voluntary drug usage to compulsive drug seeking and use (2). Moreover, this compulsory drug use can be interrupted by periods of abstinence; however, these periods are typically short-lived, as individuals often relapse within 6 months of stopping use of the drug (3). Historically, studies have focused on how addictive drugs (i.e. cocaine, amphetamines, opioids) affect neuronal populations of the brain’s reward system, specifically focusing on the dopamine pathways, as these are central to the mechanisms of action of various drugs of abuse (4-6). Several therapeutics exist to treat SUD, consisting of medication-based treatment, various behavioral therapies (i.e. cognitive-behavioral therapy), or a combination of both (7). For example, naltrexone, an opioid receptor antagonist, is approved by the FDA to treat alcohol use disorder (AUD) and opioid use disorder (OUD), as it reduces alcohol and opioid cravings (8). However, these effects are short lived, and most patients treated for AUD show no significant effects relative to placebo after 1-3 months of naltrexone discontinuation (9, 10). In OUD patients, the adherence rate to naltrexone is particularly low since patients are typically required to be abstinent prior to starting treatment (11). Currently, medication-assisted treatment, or the use of partial opioid agonists such as methadone and buprenorphine, is most effective in reducing opioid use and increasing patient adherence to treatment (12, 13). Similarly, nicotine cessation therapeutics such as bupropion, and varenicline (14, 15), while effective, can require adherence of at least one month (16, 17). Therefore, there remains a critical need for treatments with high adherence rates that target specific pathways to induce cessation and prevent relapse.

Microglia, the resident immune cells of the central nervous system (CNS), have recently emerged as novel targets in SUD research (18-23), as they are present during critical periods of brain development and play an important role in shaping neural circuitry (24-36). Moreover, accumulating evidence suggests that microglia also play an active role in the development of SUDs and relapse (37, 38). In fact, the impact of microglia on drug-related behaviors can be seen most notably during adolescence, where they contribute to development of brain circuitry (30, 32, 34, 36, 39-41). This is further supported by evidence showing that early-life stress and/or adolescent exposure to drugs of abuse affects susceptibility to drug use later in life through microglial mechanisms (28, 38, 42-46). Specifically, microglia can respond to drugs of abuse through Toll-like receptor (TLR) signaling (47). Microglia express TLR1-9 (48), which recognize a variety of exogenous pathogen-associated molecular patterns (PAMPs) such as lipopolysaccharide (LPS) from gram-negative bacteria (49), as well as endogenous damage-associated molecular patterns (DAMPs) like high mobility group box 1 and heat shock proteins (50, 51). Interestingly, several drugs of abuse can bind the TLR4/myeloid differentiation factor (MD-2) complex, resulting in the release of numerous proinflammatory cytokines (52-54). Furthermore, microglia could offer insights into known sex differences in SUD (55-59), as these cells are sexually dimorphic and respond to stress in a sex-dependent manner (33, 60-66). For instance, microglia in adult male and female mice vary in density and size, as well as in their transcriptional profile across different brain regions (62), which can affect a variety of behavioral responses (67). Specifically, glucocorticoid signaling and synaptic remodeling via colony stimulating factor 1 (CSF1) impact behavioral response to chronic unpredictable stress in male mice more so than in females (68). Taken together, this evidence indicates that microglial activity in response to various ligands and conditions, including drugs of abuse and their misuse, must be tightly regulated. Therefore, understanding this regulation will yield further insight into the emerging role that microglia play in SUDs.

Several epigenetic mechanisms regulate gene expression (69). Specifically, histone post translational modifications (PTMs) can change the accessibility of chromatin to transcription factors and polymerases, which directly affects gene transcription (70, 71). Furthermore, substantial evidence supports that these mechanisms contribute to behavioral changes observed in SUDs (72). For example, histone acetylation is increased at the promoters of immediate early genes (cFos, FosB, BDNF, Cdk5) in the striatum following acute cocaine injection and cocaine self-administration in male mice (73). Similarly, inhibiting the repressive histone methylating enzyme G9a increases dendritic spine plasticity in the nucleus accumbens and increases cocaine preference in male mice (74). These examples highlight the importance of studying epigenetic modifications and chromatin remodeling in the context of SUD. However, much of this body of evidence, to date, has focused the role of these mechanisms in neuronal cells. Therefore, this review summarizes the current knowledge on the transcriptional and epigenetic changes that microglia undergo in SUDs, and the technical and conceptual challenges that face the field moving forward.

Alcohol

The innate immune system has long been implicated in AUD and dependence. Alcohol, or ethanol, crosses the blood brain barrier (BBB) and can elicit a neuroinflammatory response (75, 76). For instance, analysis of post-mortem brain tissue revealed increased ionized calcium-binding adaptor protein 1 (Iba1) immunoreactivity in the prefrontal cortex (PFC) of male chronic alcohol-dependent individuals, in addition to significant breakdown of BBB integrity (77). Iba1 is often used as a marker for microglia (78) in various inflammatory states (79), as membrane ruffling occurs during the transition from a ramified to ameboid morphology (80). Furthermore, recent single-nucleus RNA sequencing conducted on post-mortem brains from male individuals who were alcohol-dependent showed increased differentially expressed neuroinflammatory genes in oligodendrocytes, astrocytes, and microglia over neurons (81). Specifically, there were a large amount of differentially expressed microglial genes, most notably TLR2, as well as interferon signaling (81). Additionally, rodent studies have confirmed alcohol-induced neuroinflammation and the importance of microglia. For example, in male mice that underwent bidiurnal 2-bottle choice, microglial genes involved in the innate immune response, TLR signaling, and interferon signaling were upregulated with chronic ethanol consumption (82). Chronic intermittent ethanol (CIE) vapor exposure in male mice also resulted in enrichment of several gene expression networks in microglia relating to transforming growth factor-beta (TGF-β) signaling and impaired inflammatory response (83). Given that TGF-β signaling is neuroprotective and regulates homeostatic microglial function (84), the decrease in these modules following CIE indicates alcohol promotes neuroinflammation and disruption of normal microglial function. Altogether, these data suggest microglia are highly reactive to alcohol and may play a role in alcohol-induced neuroinflammation and dependence.

Alcohol induces lasting effects on microglial morphology and activity state (85). Additionally, changes to histone PTMs can drive altered microglial activity states following alcohol exposure. For instance, neonatal exposure to alcohol resulted in hyper-responsive adult hypothalamic microglia that showed increased histone 3 lysine 9 (H3K9) acetylation, a histone mark associated with active promoters (86), in the promoter regions of interleukin(IL)-6 and tumor necrosis factor-alpha (TNF-α) (38). Additionally, increased expression of microglial marker genes C-C Motif Chemokine Ligand 2 (CCL2) and translocator protein (TSPO) in the amygdala of male and female post-mortem samples from alcohol-dependent individuals could be attributed to increased histone 3 lysine 4 (H3K4) tri-methylation (87). Furthermore, dysregulation of inflammatory signaling by microglia was shown to be epigenetically regulated by histone lysine demethylase 6b (KDM6B) in both alcohol-dependent male rats that underwent alcohol vapor exposure and in post-mortem brains from males individuals who were alcohol-dependent (88).

Several studies have shown the therapeutic potential of microglia for treating AUD and alcohol dependence. Indeed, treatment with the anti-inflammatory drug minocycline has been shown to reduce alcohol intake in adult male and female mice undergoing the drinking in the dark paradigm of voluntary alcohol consumption (89). Most notably, microglial depletion in male mice using CSF1R inhibitor PLX5622 resulted in reduction of withdrawal-induced TNF-α and CCL2 expression (90). Moreover, microglial depletion in male mice reduced dependence-induced escalation of voluntary alcohol taking and alleviated anxiety-like behavior by decreasing neuroimmune signaling, preventing upregulation of glutamatergic-related genes in the medial PFC, and reducing glutamatergic and GABAergic signaling in the central amygdala (91). Taken together, these studies suggest microglia may be viable therapeutic targets for the treatment of alcohol dependence.

Opioids

Opioids bind to μ, κ and δ-opioid receptors (ORs) in the CNS (92, 93) and activate reward signaling primarily through μOR-mediated disinhibition of midbrain dopamine neurons (94, 95). Drugs that engage μORs are often the first line treatment for chronic, or severe post-operative pain (96). However, patients can develop tolerance, which could then lead to opioid misuse. Human post-mortem studies have recently shown evidence of increased neuroinflammation in the dorsolateral PFC and nucleus accumbens (NAc) of male and female individuals with OUD (97). Indeed, several rodent studies have suggested that TLR signaling underlies the development of tolerance to opioids and contributes to the addictive potential of these drugs (98-102). Interestingly, microglia express both ORs and TLRs (103, 104), which positions these immune cells as potential regulators of behavioral phenotypes associated with OUD. Furthermore, removal of MyD88, a critical co-adaptor protein for TLR signaling specifically in microglia, increased opioid addiction-like behaviors in male mice including prolonged extinction and enhanced reinstatement of morphine conditioned place preference (CPP) (105). Additionally, microglia-specific knockout of the μOR in both male and female conditional knockout (cKO) mice reduces analgesic tolerance as well as delays morphine tolerance, as measured by increased withdrawal latencies during hot plate tests over 12 days of treatment (106). The same study also showed attenuated opioid-induced hyperalgesia (OIH) in male cKO mice, and decreased withdrawal behaviors in female cKO mice (106). Lastly, microglia may contribute to neuroinflammation due to opioid exposure and misuse (107). For instance, microglia can respond to morphine in vitro resulting in increased IL-1β expression (108). On the other hand, opioids may also be neuroprotective and mitigate withdrawal by “switching off” microglia through inhibitory G protein-coupled receptor signaling (109).

Repeated opioid use results in numerous and persistent transcriptional and epigenetic changes to the CNS (97, 110). For instance, several RNA-sequencing studies of post-mortem tissue from opioid-dependent male and female individuals show upregulation of modules enriched for genes involved in immune and inflammatory processes, regulation of transcription, extracellular matrix, while also highlighting modules enriched for genes involved in synaptic transmission, neurotransmitter secretion, and nervous system development and differentiation (97, 111). Furthermore, the upregulation of these biological processes was the result of de-repression of gene transcription, exhibiting decreased repressive histone marks such as histone 3 lysine 27 (H3K27) tri-methylation (97). Interestingly, preclinical opioid studies in rodents confirm what is seen in OUD patient samples. For example, morphine self-administration in male mice results in upregulation of biological pathways related to immune response and oxidative stress, cell-to-cell signaling, and cell differentiation in the ventral striatum and midbrain (112). Additionally, morphine self-administration has been shown to increase numerous microRNAs, whose potential targets are involved in several immune-related pathways such as mitogen-activated protein kinase (MAPK) and TGF-β. Additionally, acetylation of H3K9 is implicated in OIH, tolerance, and physical dependence, and increases BDNF expression in the mouse spinal cord (113). However, increased histone deacetylase (HDAC) 1 expression can also restore morphine’s antinociceptive effect in a tail-flick assay and decrease TNF-α expression in morphine-tolerant male rats (114). Another mouse study also showed significant upregulation of inflammatory genes including IL-1β and interferon regulatory factor 1 in both the ventral and dorsal striatum following oxycodone self-administration (115). Moreover, exposure to opioids such as morphine during adolescence primes microglia and increases TLR4 signaling specifically in the NAc, predisposing male rats to drug-induced reinstatement of morphine CPP (44).

Reducing neuroinflammation has been shown to attenuate opioid related behaviors. For example in human clinical trials, ibudilast (AV-411), a phosphodiesterase inhibitor, reduced heroin craving and subjective ratings of pain in opioid dependent subjects (116). Additionally, ibudilast treatment during adolescent morphine exposure in male rats prevents morphine-induced reinstatement of CPP and prevents the increase of TLR4 protein levels in microglia compared to controls in adulthood (44). As well, inhibiting microglial activity using minocycline can decrease somatic symptoms of opioid withdrawal such as “wet dog” shakes and expression of several proteins of the mTOR and CREB signaling pathways, which are associated with morphine tolerance, dependence, and drug memory (117).

Opioid-induced neuroinflammation and drug-related behaviors can also be mitigated with early-life interventions. In fact, opioid-related behaviors are affected by early-life experience in a microglia-specific manner. For instance, neonatal handling of male rats results in demethylation of the IL-10 gene and reduces microglial activity, as well as attenuates reinstatement of morphine CPP following drug re-exposure (118). Additionally, inhibition of epigenetic enzymes can ameliorate opioid-induced neuroinflammation. Indeed, morphine has been shown to increase HDAC1, TNFR1, and TNF-α protein expression in the spinal cord of male rats (114). Intrathecal injection of the HDAC inhibitor resveratrol prevents these increases, as well as restores the analgesic properties of morphine, indicated by decreased tail-flick latency (114). Overall, these studies suggest microglia and their epigenome may be novel targets to treat OUD and dependence.

Stimulants

Although stimulants have not been the principal focus of many studies on neuroinflammation and SUD, there is a growing body of evidence suggesting these drugs do alter microglial activity states (23, 119). Overall, studying the effect of stimulants on the immune system in the human CNS has yielded variable outcomes, with much of the post-mortem data demonstrating inconsistent results when measuring glial numbers and expression of cell markers (18, 120). However, most results suggest a connection between stimulant abuse and neuroinflammation. For example, stimulants such as nicotine, cocaine, and amphetamines can increase expression of cytokine receptors, extracellular matrix proteins, and TLRs over the course of several hours in male mice (121). Stimulants act on the reward system by increasing dopamine through various means including inhibition of dopamine reuptake (cocaine, amphetamines, MDMA) (122-124), and through direct activation of midbrain dopamine neurons (nicotine) (125, 126). Interestingly, while nicotine can act as a cognitive stimulant, unlike cocaine and methamphetamine, nicotine decreases locomotion at higher doses (127-129). Nicotine and stimulant use are associated with production of reactive oxygen species (ROS), which can exacerbate neuroinflammatory states and disease (46, 130). Similar to alcohol and opioids, microglia can respond to stimulants such as cocaine and methamphetamine directly through the TLR4/MD-2 complex, which may increase activation of midbrain dopamine neurons and enhance dopamine release in the NAc (131, 132). To this point, human positron emission topography studies show nicotine smokers have lower TSPO expression as measured by radiotracer [11C]DAA1106 binding, suggesting suppressed microglial activity in these individuals (133). In fact, microglia express α7 nicotinic acetylcholine receptor, which may serve as a biological substrate for any direct actions of nicotine on microglial activation (134). Indeed, this could explain the functional deficits seen in microglia after treatment with the drug. Beyond nicotine, stimulants like cocaine and methamphetamine induce a proinflammatory microglial state, where these cells increase expression of cytokines including TNF-α, IL-1β, and IL-6 (131, 135-137), and increase phagocytosis and migration in vitro (138).

While microglia differentially respond to various stimulants, withdrawal from these drugs show microglia alter their gene expression in a similar manner (139-141). For instance, altered microglial morphological states and increased production of ROS can be seen during nicotine withdrawal in the NAc and striatum in male mice (139). These changes coincide with increases in anxiety-associated behaviors (139). Additionally, increased microglial reactivity correlates with impaired neurogenesis in the hippocampus of male mice (142), as well as increased hyperalgesia in male rats during nicotine withdrawal (143). This response is also seen in abstinent male rats following extended-access methamphetamine self-administration (144).

Stimulants have long lasting effects on the innate immune system and on microglia. Several microarray and RNA-sequencing studies demonstrate differential expression of immune-related genes. For example, there is significant overlap between upregulated genes associated with immune/inflammatory response, specifically those regulated by nuclear factor kappa B (NF-κB), in cocaine use disorder (CUD) or OUD (111). Additionally, methamphetamine administration results in delayed upregulation of immune-related genes in male mice, including NF-κB, interleukins and their receptors (145). Consistent with delayed effects following methamphetamine exposure, self-administered methamphetamine increased IL-6 production during abstinence and was associated with activation of nuclear factor of activated T-cell-mediated immune response, which is involved in regulating microglial response and cytokine expression (146).

Early-life stress and exposure to addictive drugs during adolescence increases susceptibility to misuse of stimulants later in life (23, 28). For example, prenatal exposure to nicotine results in downregulation of genes associated with TNF signaling and cytokine-cytokine receptor pathways, as well as increases microglial numbers and changes microglial morphology in the hippocampus of young male and female mice (postnatal day 20) (147). Another study showed adolescent exposure to nicotine increases cocaine self-administration in adult male and female mice in a microglia-specific manner through CXC3 motif chemokine receptor 1 - CXC3 chemokine ligand 1 signaling (43). Additionally, childhood trauma is correlated with increased plasma inflammatory markers in cocaine-dependent human subjects (46). The same study showed that exposing male and female mice to social stress during early life (postnatal day 14-22) sensitizes microglia and peripheral immune cells, and gene ontology analysis of microarray from blood RNA revealed enrichment of “inflammation mediated by chemokine and cytokine signaling” pathway during acute withdrawal from cocaine (46). Furthermore, neonatal exposure to an immune challenge increases methamphetamine-induced reinstated behavioral sensitization including locomotor activity, rearing, and stereotypy in adult male and female rats (148). Despite evidence that altered microglial dynamics early in life can influence later stimulant misuse, suggesting involvement of epigenetic regulation, most studies focus on the epigenetic landscape in neuronal populations. Therefore, there remains a significant need to characterize these changes in non-neuronal populations, particularly in microglia.

Nonetheless, numerous studies demonstrate therapeutic targeting of microglial function can normalize various behaviors related to stimulant use disorder. In human studies, ibudilast reduced reward-related subjective effects of methamphetamine including how “high” the subject felt and if the subject felt the “effect” of the drug (149). This was confirmed and further investigated in male rats using methamphetamine self-administration, where the authors showed that AV1013 (an analog of ibudilast) and minocycline significantly reduced drug intake (150). Minocycline also attenuates nicotine-induced increases in the number of microglia, as well as cocaine-seeking in male and female rats (43). Additionally, minocycline facilitates extinction of methamphetamine CPP and attenuates drug-primed reinstatement of CPP (151). Depletion of microglia is also therapeutic in stimulant-related addictive behaviors. For example, treatment with PLX5622 during nicotine treatment and withdrawal mitigated nicotine-induced increased NADPH oxidase 2 expression and ROS production and reduced anxiogenic behavior in male mice (139). Additionally, microglial depletion using PLX3397, another CSF1R inhibitor, blocked increased cocaine responding induced by nicotine (43). In support of the importance of TLR4 signaling in stimulant use disorders, pretreatment with TLR4 antagonist, (+)-Naloxone, prevented increased cocaine-induced dopamine release and expression of IL-1β in the NAc of male rats (131). Interestingly, in a separate study, the use of bromodomain and extra-terminal motif inhibitor JQ1, which can alter microglial activity (152), also reduced cocaine-induced increased expression of IL-1β in the NAc of male mice (153). Moreover, treatment with GSK-J4, a small molecule inhibitor of KDM6B known to increase macrophage activity (154), reduced reconsolidation of cocaine-conditioned memory and cocaine-primed reinstatement of CPP in male mice (155). Thus, in the context of stimulant use disorders, microglial function may be regulated by epigenetic mechanisms, and may therefore serve as a cellular target for curbing behavioral aspects of stimulant use disorders.

Cannabinoids

With the recent increased use and legalization of medical and recreational marijuana and cannabidiol (CBD) products (SAMHSA), there has been a renewed interest in studying the effects that exogenous cannabinoids have on the CNS. The primary mechanism of action for cannabinoids is modulated through cannabinoid receptor type 1 (CB1R) and cannabinoid receptor type 2 (CB2R) (156). While CB1R is primarily expressed on neurons, CB2R is expressed on immune cells (157), including microglia (158). Indeed, cannabinoids have anti-inflammatory properties (157). For example, cannabinoids can increase release of endogenous IL-1 receptor antagonist, which has neuroprotective effects against excitotoxicity (159). In fact, a recent study demonstrated CB2R is necessary for TLR signaling in mice, dampening microglial transcriptional and morphological response to TLR ligands such as LPS/Interferon-γ, polyinosinic-polycytidylic acid, and unmethylated CpG DNA, and that this occurs through phosphatidylinositol 3-kinase/protein kinase B and MAPK signaling pathways (160). Additionally, the endocannabinoid 2-arachidonylglycerol (2-AG) can increase microglial migration in vitro, and cannabinol and CBD, two nonpsychotropic chemicals in marijuana, inhibit cell migration by antagonizing CB2R (161). Furthermore, 2-AG and anandamide, another CB1R and CB2R ligand, contribute to endocannabinoid tone, which plays a role establishing sex differences in the developing brain by increasing the phagocytic activity of microglia (162).

Several studies have shown that Δ9-tetrahydrocannabinol (THC), the psychoactive component in marijuana, has long-lasting effects on the brain that are age- and sex-dependent (163, 164). Most notably, adolescent exposure to cannabinoids increases vulnerability to substance misuse later in life, suggesting underlying epigenetic regulation. For example, adolescent THC exposure increases heroin self-administration in male rats through decreased repressive trimethylation of H3K9, resulting in increased proenkephalin expression in the shell of the NAc (165). Interestingly, adolescent exposure to cannabinoid receptor agonist WIN55,212-2 increased nicotine self-administration in adult male mice, while nicotine self-administration was decreased in adult female mice (166). However, another study showed treatment with cannabinoid agonist CP 55,940 during adolescence increased acquisition of cocaine-self administration, as well as altered basal brain metabolic activity in adult female rats, but had no effect on males (167). Paradoxically, cannabinoids are also therapeutic in the treatment of SUD and can reduce drug-related behaviors and neuroinflammation. For instance, acute CBD reduced cue-induced drug-craving and anxiety in abstinent heroin users (168). Additionally, treatment with CB2R agonists reduced morphine-induced expression of inflammatory cytokines such as IL-1β, TNF-α, and IL-6 and altered microglial activity in vitro (169) and reduced hyperalgesia and allodynia in male rats (170). Furthermore, microinjection of CBD into the lateral cerebral ventricle prevented acquisition and expression of methamphetamine CPP in male rats (171). Finally, CBD attenuated cognitive deficits, altered microglial morphology, and expression of inflammatory cytokines TNF-α and IL-1β in male mice undergoing nicotine withdrawal (142). Since there is limited access to cannabis and THC due to its scheduling classification, there is a significant need for additional studies to better understand the epigenetic consequences of exogenous cannabinoid exposure, especially in microglia, that may underlie these therapeutic effects.

Current challenges and future directions

Microglia stand as a prime target for studying the behavioral consequences of SUDs. Through their dynamic regulation of neurodevelopment and role in neuroinflammation and psychiatric disease (Fig. 1), microglia stand as a key cell type to be further characterized. However, there are currently a number of theoretical and technical limitations when studying microglia, both as an individual cell type as well as in the context of SUDs.

Figure 1. Microglia regulate drug-related behaviors as a result of transcriptional changes throughout Substance Use Disorders.

Proposed mechanisms for microglial activation following exposure to various drugs of abuse. Microglia may be activated through both direct and indirect means. Following activation, microglia maintain an altered state, with possible changes to morphology, gene expression and epigenetic modifications that lead to functional deficits such as aberrant pruning of synapses and hyperexcitability. These changes are underscored by altered behavior across numerous paradigms. However, treatment with known anti-inflammatory drugs like AV-411 and the tetracycline antibiotic minocycline, as well as complete microglial depletion can mitigate these drug-induced changes both molecularly and behaviorally.

While there is a certain need to better characterize SUD-related gene expression and epigenetic changes in microglia, it is also necessary to further understand microglial activation dynamics from a broader perspective. Microglia have previously been classified through binary terms “M1” (pro-inflammatory) or “M2” (anti-inflammatory) as well as “resting” or “active”; however, recent advances in the field suggest a deeper complexity that extends beyond a binary nomenclature (172, 173). Developing a more comprehensive classification of microglia is essential to understanding the role they play in SUDs and other neurological disorders. In doing so, examining the transcriptional and epigenetic states of microglia can aid in these classifications and therefore broaden the pool of potential therapeutics to correct their altered state in SUDs that may contribute to adverse behavioral outcomes, such as increased drug craving and propensity for relapse.

From an in vitro perspective, the use of microglial cell lines offers a more cost-effective method to study and characterize microglia, as opposed to in vivo models, by directly measuring the effects of a drug. Additionally, in vitro studies have a faster turnaround, are very adaptable (e.g. transfection with siRNA, CRISPR, treatment with small molecules, etc.), and are useful for basic validation of molecular targets. Studies that utilize pharmacological interventions can be particularly modeled using established microglial cell lines such as the murine BV-2 (ATCC; EOC-20, CRL-2469) (174) or the human microglial clone 3 (ATCC; HMC3 CRL-3304) (175). However, while these cell lines display a similar cytokine profile of primary microglial cells (176, 177), they often have reduced expression of key microglial markers, and also display decreased functional activity (175). A recent advancement for studying microglia in vitro involves induced pluripotent stem cell-derived microglia, which are shown to be more transcriptionally and functionally similar to microglia in vivo (178-180); meaning they may offer additional insight into how microglia respond to drugs of abuse. Yet, a major limitation inherent to cell culture models in studying drug addiction is that these cannot account for changes that may be induced by the motivational or affective aspects of drug-seeking and craving. In fact, no in vitro models are currently available to replicate these behavioral phenotypes; therefore, the most reliable method is the use of in vivo behavioral paradigms. One such behavioral technique is intravenous self-administration (IVSA), which is considered the “gold standard” for preclinical studies, as it most closely reflects the motivational aspects and patterns of drug-taking and seeking in humans (181-183). Given how microglia rely on and communicate with other neural cells in both healthy and disease conditions (184, 185), in vivo models of SUDs are currently the most comprehensive for studying the contribution of microglia toward shaping the behavioral course of addiction. Additionally, as previous research has mainly focused on the role of neuronal populations in SUDs (186-188), microglia represent a novel cell population to examine within the context of these conditions.

Another major challenge to studying microglia is the ability to obtain a pure population of these cells from the brain for downstream analyses (Fig. 2). This is key to producing the transcriptomic and epigenetic profiles needed to better understand the role of microglia in SUDs. While the most common method of microglial purification utilizes Florescence Assisted Cellular Sorting (FACS), affinity purification with magnetic beads also yields highly pure populations, as it also targets microglia-specific markers with antibodies (189). Importantly, the multitude of receptors shared between microglia and macrophages (190) has been difficult to overcome, which inherently impacts purity in sorted populations. However, several of these issues have been resolved with the identification of microglia-specific markers such as TMEM119, which has resulted in the generation of several transgenic mouse lines (188). This recent advancement has also allowed for improved resolution when performing cell phenotyping studies using single-cell sequencing (191-193). Along similar lines, RiboTag-seq has been shown to reveal gene networks involved in morphine withdrawal (194), by analyzing the active translatome in microglia (195). While this method offers a number of advantages including compatibility with single-cell sequencing (196) and analyzing cell type specific profiles (197), care must be taken due to the inherent loss of RNA transcripts in this methodology. The discovery of new markers in combination with single-cell sequencing technologies has allowed for better characterization of microglial heterogeneity (198) as well as profiling chromatin accessibility (199). While recent studies have started to profile the epigenome using methods such as Assay for Transposase-Accessible Chromatin (ATAC) sequencing (200, 201) and chromatin immunoprecipitation (ChIP) sequencing (202, 203), the advent of lower input methodologies will enable microglia to be studied on a brain-region specific scale, rather than on a global scale, furthering our understanding of microglial transcriptomics and epigenetics in various disease models (204-206). For example, novel methods like CUT&RUN (Cleavage Under Targets & Release Using Nuclease) (207, 208) and CUT&Tag (Cleavage Under Targets & Tagmentation) (209, 210) allow for higher resolution using lower cell numbers than previous techniques. This new technology facilitates interrogation of transcription factors and histone modifications including various forms of histone methylation and acetylation to identify promoters and enhancers. Given the scarcity of data regarding the transcriptomic and epigenetic landscapes of microglia, these novel low-input, high-resolution methods hold promise to improve our understanding of not only SUDs, but also neurological disorders in general.

Figure 2. Methodologies for studying microglial genomics.

A number of novel technologies have enhanced the ability to sequence the genome and epigenome of microglia. As an initial barrier to profiling microglia, isolation of pure populations has recently been made easier with developments in both FACS and magnetic bead based affinity purification, with both methods achieving similar results. For RNA sequencing, the advent of low input and single cell, kits capable of creating complex libraries for sequencing has enabled investigators to reduce the number of microglia needed to understand their response to certain stimuli and delineate sub-populations within different brain regions. To study chromatin dynamics, methods such as CUT&RUN and CUT&Tag allow for regional microglial populations to be analyzed for various histone marks and transcription factors with very low cellular input, including single cells, at a fraction of the cost of ChIP-sequencing assays. These more recent advances can generate robust datasets comparable to those generated by ChIP-sequencing. ATAC-seq is a convenient method for profiling chromatin accessibility, however its utility is hampered by its limited ability to recognize gene enhancer loci. (FACS: Fluorescence Assisted Cellular Sorting, AFP: Affinity purification, scRNA: Single-cell RNA, CUT&RUN: Cleavage Under Targets & Release Using Nuclease, CUT&Tag: Cleavage Under Targets & Tagmentation, ATAC: Assay for Transposase-Accessible Chromatin) Created with BioRender.

From a translational standpoint, the therapeutic value of epigenetic targets remains promising. However, owing to the complex dynamics of epigenetic mechanisms during development and throughout the lifespan, development of treatments targeting specific epigenetic modifications has been challenging. While some studies show promise using HDAC inhibitors such as resveratrol for treatment of OUD (114), bromodomain-containing 4 inhibitor JQ1 for CUD (153), and GSK-J4 for AUD and CUD in vitro and in vivo, respectively (88, 155, 211), the pharmacokinetics, brain bioavailability, and lack of cell-type specificity of these drugs lend themselves to limited translational value. Furthermore, given the recent emergence of microglia in the study of SUDs, the importance of microglial chromatin architecture and the role that it may play in SUDs is not yet fully understood. Finally, epigenetic targets, while promising, are shared across numerous cell types, which may result in off-target effects. Since this review discusses several epigenetic changes noted in microglia for specific SUDs, repurposing of epigenetic drugs could provide a novel therapeutic avenue, but studies into novel epigenetic treatments should focus on cellular specificity to reduce off-target effects and increase efficacy.

While the fields of addiction and neuroscience continue to advance, recent evidence has positioned microglia as important mediators in the development and progression of SUDs. This review highlights recent advances in our understanding of microglia, from the shift in focus to glial involvement in SUD research as well as to setting our sights forward to crucial transcriptomic and epigenetic hypotheses regarding microglial regulation in SUDs. However, important gaps in this understanding still remain. While in vitro and clinically relevant in vivo behavioral models are essential to understanding the importance of microglia to addiction pathophysiology, genomics studies employing cell type-specific RNA or single cell sequencing applications in combination with low-input CUT&RUN/CUT&Tag chromatin profiling methods will provide the molecular depth to understanding the driving factors that may underlie drug-taking, craving and relapse. Further studies that better characterize these changes may shed light on the role of microglia in neuroinflammation, addiction, and may unlock a host of novel therapeutic targets for the treatment of SUDs.

Highlights.

• We explore the recent evidence supporting a role for microglia in regulating the behavioral aspects of substance use disorders.

• We focus on the transcriptional and proposed epigenetic mechanisms underlying microglial responses to drugs of abuse.

• We discuss the latest advances and challenges in low-input epigenetic/molecular profiling of microglia.