Endocannabinoid signaling and epigenetics modifications in the neurobiology of stress-related disorders

Abstract Stress exposure is associated with psychiatric conditions, such as depression, anxiety, and post-traumatic stress disorder (PTSD). It is also a vulnerability factor to developing or reinstating substance use disorder. Stress causes several changes in the neuro-immune-endocrine axis, potentially resulting in prolonged dysfunction and diseases. Changes in several transmitters, including serotonin, dopamine, glutamate, gamma-aminobutyric acid (GABA), glucocorticoids, and cytokines, are associated with psychiatric disorders or behavioral alterations in preclinical studies. Complex and interacting mechanisms make it very difficult to understand the physiopathology of psychiatry conditions; therefore, studying regulatory mechanisms that impact these alterations is a good approach. In the last decades, the impact of stress on biology through epigenetic markers, which directly impact gene expression, is under intense investigation; these mechanisms are associated with behavioral alterations in animal models after stress or drug exposure, for example. The endocannabinoid (eCB) system modulates stress response, reward circuits, and other physiological functions, including hypothalamus–pituitary–adrenal axis activation and immune response. eCBs, for example, act retrogradely at presynaptic neurons, limiting the release of neurotransmitters, a mechanism implicated in the antidepressant and anxiolytic effects after stress. Epigenetic mechanisms can impact the expression of eCB system molecules, which in turn can regulate epigenetic mechanisms. This review will present evidence of how the eCB system and epigenetic mechanisms interact and the consequences of this interaction in modulating behavioral changes after stress exposure in preclinical studies or psychiatric conditions. Moreover, evidence that correlates the involvement of the eCB system and epigenetic mechanisms in drug abuse contexts will be discussed.


Key epigenetic mechanisms in stress response
Epigenetics focuses on the interaction between the environment and genome, whereby gene expression is modulated in response to new stimuli. In the short or long term after stress experiences, several epigenetic marks can be modified to adapt the organism to the environment; these marks can even be transmitted between generations. Because of a maladaptive response, these epigenetic changes can predispose the organism to diseases. There are three major epigenetic mechanisms for the regulation of gene expression: DNA modifications, histone modifications, and interference RNAs. Below, we will briefly describe each one of them, associating them with modifications in important stress-related systems.

DNA methylation
The DNA is passive to chemical modifications that do not change the nucleotide sequence but regulate gene transcription. Among many modifications described in the literature, cytosine methylation at the 5 position (5mC) is the most investigated. The DNA-methyltransferases (DNMTs) family of enzymes in mammals, comprising DNMT1, DNMT3a, and DNMT3b, catalyzes the reaction of a methyl group addition to a cytosine [95]. The reaction occurs mainly in CpG dinucleotides, distributed throughout the genome but more concentrated in CpG islands in promoter regions of the genes. 5mC is recognized by methyl-CpG binding proteins (MBPs), such as methyl-CpG binding protein 2 (MeCP2), commonly related to psychiatry diseases [96,97]. 5mC is usually associated with the repression of gene expression, but it also can promote gene expression activation depending on the location of the CpG island in the gene body [98]. Despite some stability, DNA methylation is a dynamic process constantly subject to reversal (demethylation) by active and passive forms. The passive form occurs through DNA damage or replication, while the active form is a process based on methyl-cytosine modifications through the ten-eleven translocation (TET) enzyme family and the activation-induced cytidine deaminase/apolipoprotein B mRNA-editing enzyme complex (AID/APOBEC) [98].
Both DNA hyper-and hypomethylation of stress-related genes are found in various neuropsychiatric and neurological diseases [99][100][101][102]. Most clinical studies focus on genes related to glucocorticoid response and serotonin neurotransmission [100,103]. The promoter region of the glucocorticoid receptor gene (NR3C1) appears to be extremely sensitive to DNA methylation. Various stressor events, including child abuse, war and genocide-related trauma, maternal depression, or violence during pregnancy, correlate with increased [104][105][106][107][108][109][110] and decreased [105,111,112] methylation levels in NR3C1. Methylation of NR3C1 is also altered in the post-mortem brains of suicide victims and is associated with childhood traumatic experiences [113,114]. In these works, methylation levels frequently are inversely proportional to NR3C1 expression, and low levels of GR can directly impact the feedback of glucocorticoid release and HPA activity, contributing to altered responses to stress. Regarding the serotoninergic system, the serotonin transporter gene (SLC6A4) methylation levels are related to traumatic events, child abuse, work stress, and depression [115][116][117][118][119][120][121], and although it does not correlate with mRNA expression in the blood of patients, the hypomethylation of SLC6A4 is proposed to be a biomarker of diagnosis and drug response to major depression [118,122].

Histones modifications
Histones are proteins that, together with DNA, make the chromatin and organize its packaging state. There are four types of histones: H2A, H2B, H3, and H4. The addition of chemical groups to amino acid residues of the histones alters their binding to DNA modulating the access of transcription factors. The addition of group acetyl, or acetylation, is a modification primarily associated with gene expression. It occurs due to the action of histone acetyltransferases (HATs) and is erased by histone deacetylases (HDACs). Histone methylation is another modification related to gene expression or repression depending on the location of the methyl group. It is catalyzed by histone methyltransferases and reversed by histone demethylases, also occurring on lysine residues [123,124].
In humans, there are only a few works exploring histone mark changes in the context of stressful experiences and neuropsychiatric diseases. In these studies, they found differences in tri-methylation of H3 at lysine 27 (H3K27me3) and at lysine 4 (H3K4me3) levels in the brains of suicide victims when it was compared between control and depressive groups [125][126][127]. On the other hand, in animal models, several types of stressors, such as maternal separation, social stress, restraint stress, and chronic mild stress, induce alterations in HDACs and histone acetylation/methylation levels in a global or gene-specific manner [128][129][130]. Interestingly, antidepressant drugs with different mechanisms of action, such as ketamine, imipramine and fluoxetine, not only ameliorate behavior alterations after the stress but also alters HDACs activity and expression, or impact levels of acetylation at specific histone residues associated with important genes related to synaptic plasticity, such as Nr2b and Bdnf genes [131][132][133][134][135][136][137][138].

Possible crosstalk between the eCB system and epigenetic mechanisms in stress response and psychiatric disorders (Endo)cannabinoid control of epigenetic mechanisms
Cannabis use induces epigenetic alterations, supporting a relationship between the eCB system and epigenetic mechanisms. For instance, Cannabis-dependent patients have reduced methylation of the CB 1 receptor gene promoter (Cnr1) and increased CB 1 expression in the blood [150]. The same profile of methylation and CB 1 expression was observed in peripheral blood lymphocytes of patients with schizophrenia reporting the use of Cannabis [151]. Moreover, prenatal Cannabis exposure decreased D 2 receptors mRNA in Nucleus Accumbens (NAc) and amygdala of aborted fetuses, which was replicated in an animal model. In this model, there was increased di-methylation at lysine 9 of H3 (H3K9me2), a repressive mark, and decreased H3K4me3, mentioned earlier, an enhancer mark, and RNA polymerase II at the Drd2 gene locus [152,153], supporting the role of an epigenetic mechanism induced by Cannabis in decreasing D 2 expression. These changes have implications for drug addiction, which will be discussed later, and other psychiatric conditions. For instance, the chronic administration of a CB 1 agonist to adolescent male rats has been implicated in greater susceptibility to stress and anxiety-like behavior, in addition to increase DNMT and global methylation levels in the PFC of their adolescent offspring [154]. Also, paternal activation of CB 2 receptors was implicated in impaired offspring growth via reduced expression of TET enzymes and altered DNA methylation in several genes [155].
Regarding treatment with CBD, it was demonstrated that acute CBD treatment decreased immobility in mice in the forced swimming test, similar to what was observed with DNMT inhibitors (5-AzaD and RG108). Interestingly, the combination of ineffective doses of CBD and DNMT inhibitors induced similar antidepressant effects, suggesting CBD effects could be directly modulating DNMTs. In fact, all drugs prevented the swimming stress-induced reduction of the DNA methylation in the PFC and the increase in the hippocampus. However, whereas the DNMT activity was decreased by swimming stress in the PFC and increased in the hippocampus, CBD could only counteract the first in this work [156]. In contrast, the hippocampal neurodegeneration induced by iron administration in neonatal rats, which induces mitochondrial DNA methylation alterations, was reverted by treatment with chronic CBD during adulthood [157]. Finally, subacute treatment with CBD induced hypomethylation of DNMT3a in the mouse hippocampus [158], a mechanism already shown to induce gene expression related to neurogenesis [159,160]. These pieces of evidence suggest a role for DNA methylation in CBD effects in animal stress models. As described before, many of these makers are also involved in stress-related disorders, and CBD has anxiolytic/antidepressant effects in psychiatric patients [161]. Therefore, it is possible to suggest that these CBD effects may involve DNA methylation; however, there are no studies in humans with this analysis, which would be very useful for better conclusions.
Histone modifications may also be involved in effects mediated by cannabinoids. Repeated co-administration of THC and CBD increased the acetylation in lysine 9 (H3K9ac) and 14 (H3K14ac) of H3 in the ventral tegmental area of adult mice [162]. In another study, acute CBD treatment increased levels of methylation and acetylation markers H3K4me3, H3K27me3, and H3K9ac in the cerebral cortex. In contrast, it decreased H3K9ac levels in the hypothalamus and H3K4me3 in the pons in rats, demonstrating that its effects are brain area-specific [163].
Chronic unpredictable stress (CUS) increased the nuclear expression and activity of HDAC2 and decreased the expression of CB1 levels, mainly in glutamatergic neurons, in the mouse cingulated cortex. Moreover, CUS reduced the expression of H3K9ac associated with CB 1 and Neuropeptide Y (NpY) genes. They also showed that URB597, a FAAH inhibitor, which is expected to increase anandamide levels, reverted stress effects in the Npy gene, but not in Cnr1, and anxious behavior [164]. Like stress, a TRPV 1 agonist (capsaicin), which increased immobility in the forced swimming test, increased HDAC2 expression in the mouse DG of the hippocampus [165] and enriched HDAC2 expression at Dlg4, Syp, Gria1, and Gria2 gene promoters, all related to neuroplasticity [166]. In contrast, genetic deletion of TRPV 1 receptors, which induced an antidepressant-like phenotype, reduced HDAC2 levels in the same brain region and consequently increased levels of H3 and H4 global acetylation. In addition, TRPV 1 knockout mice, contrary to what was observed with capsaicin injection, showed increased levels of plasticity and neurogenesis-related genes in the hippocampus, in addition to being resilient to stress [166]. Therefore, considering anandamide activates CB 1 and TRPV 1 receptors, we suggest that the bell-shaped profile of anandamide effect on behavior may be mediated by differential effects on CB 1 and TRPV 1 receptors, among other mechanisms, regulating the expression and activity of HDAC2, histone acetylation levels, gene expression, and neuroplasticity.
As described before, miRNAs play an essential role in gene expression regulation, being implied in several diseases. eCB system activity, in turn, seems to regulate miRNA expression and, therefore, could impact the pathogenesis and treatment of stress-related diseases. In mice, chronic mild stress (CMS) increased expression of some miRNAs in the PFC (miR-9-5p, miR-128-1-5p, and miR-382-5p) [167] that target Drd2, Clock, Map2k, Mapk1, and Bdnf genes [168][169][170], all related to the physiopathology of depression. Moreover, CMS induced decreased expression of others miRNAs (miR-16-5p, miR-129-5p, and miR-219a-5p) [167], which target Slc6a4, Htr2a, Bdnf , Grm7, Camk2a, and Camk2g genes, which are also related to depression physiopathology and antidepressant response [168,[171][172][173][174]. In the same study, stressed animals treated with anandamide showed increased expression of all these miRNAs compared with the vehicle group; the depressive-like effect of stress in the forced swimming test was reverted [167]. Moreover, early life stress-induced depressive-like behavior in rats and downregulated miR-16 in males and miR-135a in females in the mPFC. These changes were reversed when rats received a FAAH inhibitor [175].
Additionally, lower expression levels of let-7d miRNA were observed in the cortex and hippocampus of CB 1 knockout mice or after CB 1 knockdown in zebrafish embryos. Conversely, the knockdown of let-7d in zebrafish embryos increased the expression of CB 1 receptors, suggesting negative feedback in this regulation [176,177]. Moreover, let-7d overexpression in adult mouse hippocampus induced anxiolytic-and antidepressant-like effects [178]. Thus, it is arguable that anxiolytic and antidepressant effects induced by CB 1 activation are promoted by let-7d expression and that this mechanism may be impaired in psychiatric conditions, such as depression. Furthermore, the anxiolytic and antidepressant effects induced by let-7d increased expression may occur, between other mechanisms, by the negative regulation of dopamine D3 receptors, mu-opioid receptors, TLX, an orphan nuclear receptor, and upregulation of miR-9, regulating neuroplasticity, cellular proliferation, neuronal differentiation, and migration [178][179][180].
In summary, the eCB system activity regulates the expression and activity of epigenetic enzymes, such as TETs, DNMTs, and HDACs, which result in differential global and specific-site levels of DNA and histone modifications. Moreover, the eCB system is also involved in miRNA expression regulation. All these alterations change gene expression related to neurotransmission, neurogenesis, and neuroplasticity. These mechanisms, also altered by stress, may be involved in the development of psychiatric conditions; therefore, more studies are needed to better understand how they work in physiological and pathological conditions to determine if they could be targets for treating these conditions.
So far, only a few studies combine stress protocols, modulation of the eCB system, and evaluation of epigenetic output, highlighting the need for more studies addressing that combination. These studies are summarized in Table  1. How eCB system molecules modulate epigenetic factors in stress-related contexts are outlined in Figure 1A.

Epigenetic control of the endocannabinoid system
Although the relevance of the epigenetic mechanisms to the activity of the endocannabinoid system is well known [181][182][183][184], this regulation in the context of stress is less explored. The investigation of epigenetic control of the eCB system relies mainly on the regulation of Cnr1 and Faah genes. This specificity can be explained by the attention these two genes receive due to their pharmacological importance in physiology and disease and their susceptibility to being regulated by epigenetic marks.
DNA methylation levels of Cnr1, for example, is reported to be inversely proportional to mRNA and protein expression of the gene [185][186][187][188][189][190][191]. Cnr1 methylation pattern is recurrently found altered in a variety of situations, such as diet [190,192,193], patients with schizophrenia [189], and THC consumption [150]; Cnr1 is also susceptible to demethylation by the agent 5-aza-2-deoxycytidine [194]. On the other hand, Faah hypermethylation is associated with alcohol consumption [195], while hypomethylation, along with an increase in mRNA and protein expression, is related to Alzheimer's disease patients [196]. Nonetheless, the methylation of Cnr1 and Faah in stressful conditions  are less explored. Chronic stress induces depressive-like behavior and results in hypermethylation in Cnr1 [188], including in several CpG islands of Cnr1 gene in sperm of the stressed rats and in their offspring's brains [197]. In PTSD patients, the Cnr1 gene was one of several uniquely methylated genes found in patient's PBMC [198], suggesting variations in CB 1 expression could be involved in the pathology of this disease, as well as other psychiatric diseases, as discussed in previous reviews [67,199]. Histone modifications are also found in Cnr1 and Faah genes. Ethanol treatment in mice is correlated with an increase in histone acetylation marker H4K8a and Cnr1 expression and a decrease in histone methylation marker H3K9me2 in the neocortex and hippocampus [200,201]. On the other hand, histone methylation marker H3K9me2 and mRNA expression of the Cnr1 gene are induced at dorsal root ganglion in a model of nerve injury in mice [202]. Although no change in DNA methylation was observed in a model of binge-eating behavior, histone acetylation H3K9ac associated with the Faah gene and its mRNA expression decreased after frustration stress [203]. Moreover, after exposure to CUMS, histone acetylation H3K9ac decreased in the Cnr1 gene, although its mRNA expression remained unchanged [164].
Several miRNAS are reported as modulators of genes of the eCB system, including genes of Cnr1, Cnr2, and Faah (Table 2). Except for the miR-let-7d, which inhibits but does not have Cnr1 as a direct target [176], all described miRNAs have predicted pairing to their targets so far [176,192,204,205,206,207,208,209,210,211]. All these miRNAs directly bind to the transcript, inhibiting eCB-related gene expression.
Similar to what is seen in studies evaluating DNA methylation and histone modifications related to the eCB system, most works describe miRNAs regulating Cnr1 expression in the context of stressful or psychiatric conditions. For instance, miR-128 is down-regulated in the blood of PTSD patients [224] but is up-regulated in the brain of depressed subjects [225]; however, it is also reported to increase in blood after 12 weeks of antidepressant treatment [146]. Overall, these data could suggest that miR-128 participates in disease and treatment response, but this pattern can be different depending on the psychiatric condition and the evaluated tissue. In animal models, miR-128 up-regulation is found in the amygdala, PFC, and hippocampus of stressed mice [220,225,226], which indicates it can participate in brain functions. Another miRNA, miR-301a, is also up-regulated in the brain of depressed suicide victims [227] and chronically stressed rats [228]. miR-494 findings in the blood and brain are contrasting: it was upregulated in blood of major depression patients [235] and after antidepressant treatment [146], in depression episodes [232], in a PTSD rat model [233]; however, it was downregulated in the brain of depressed suicide victims [227] or in the brain of acutely stressed rats [219]. miR-494 also had an anxiolytic effect in ethanol-exposed rats [231]. Moreover, miR-29a is up-regulated in the blood of stressed students but down-regulated in treatment-resistant depression patients [214,215]. After restraint stress, rats subjected or not to maternal separation have up-regulation of miR-29a in the amygdala and PFC [216], which is increased in the cerebral spinal fluid of MDD patients [217]; it is also increased in the frontal cortex of mice exposed to acute restraint stress [219] but decreased in the frontal cortex after chronic stress [218] and in the hippocampus 1 h after footshock stress in a fear conditioning paradigm [220]. Similarly, 1 h after fear conditioning, there was also a reduction in the miR-30b expression in the hippocampus [220], as seen after acute restraint stress [219], whereas chronic stress increases the same miR-30b in the hippocampus [223]. Another miRNA, miR-let-7d, appears to be important in various stress processes. It was reduced in the blood of MDD patients, and increased after antidepressant treatment [146,235]; its levels changed in the PFC, hypothalamus, hippocampus, and amygdala of animal models after different types of stressors [220,223,234,236]. Meanwhile, overexpression of miR-let-7d in the hippocampus has anxiolytic and antidepressant effects in mice, corroborating its function in behavior and potential impact in neuropsychiatry diseases [178]. Although there is no evidence of alterations in humans, miR-338-5p and miR-23a are altered after protocols of social stress and chronic unpredictable stress [212,223,229].  ↑: miRNA up-regulation; ↓: miRNA down-regulation; ARS, acute restraint stress; CRS, chronic restraint stress; CSD, chronic social defeat; CUMS, chronic unpredictable mild stress; MDD, major depressive disease; mPFC, medial prefrontal cortex; MS, maternal separation; PFC, prefrontal cortex, PTSD, post-traumatic stress disorder.
These works evidence a complex control by different miRNAs, including similarities or differences among patients with different psychiatric conditions and differences in animal models involving stress exposure. There are two miRNAs reported to modulate Cnr2 in animal models of stress: miR-187-3p and miR-665. miR-187-3p is regulated in several stress protocols, including chronic mild stress, acute restraint stress, psychological stress, and after fear conditioning, indicating that miRNA as having an important role in stress responses in general [216,223,229,237,238]. Both acute and chronic stressors can decrease miR-187-3p expression in the amygdala, while it was up-regulated after the evaluation of extinction of conditioned fear memory [238]. Additionally, miR-665 is altered in the amygdala, being up-regulated after chronic mild stress [223]. miR-411, for the best of our knowledge, is the only FAAH miRNA regulated by stress, and it is only found regulated in animal models; it was increased in the hypothalamus, PFC, and hippocampus after maternal separation or chronic unpredictable mild stress [239][240][241].
eCB system genes are considerably sensitive to epigenetic control, particularly under stressful experiences, although the mechanisms are not completely elucidated. The discussed evidence highlights the importance of epigenetic mechanisms in the eCB system response to stress and in its dysfunction. For instance, histone modifications are fundamental to memory consolidation and extinction [242,243], and intervention in these processes could be key to treating trauma-related disorders. In fact, some stressors can induce histone modifications in the Faah and Cnr1 genes, which can reverberate or not in mRNA expression; moreover, the histone modifications, as acetylation, is one of the proposed mechanisms for the action of antidepressant drugs [244]. Even when the gene or the protein expression is not altered, epigenetic markers can influence the gene expression pattern in response to the environment. DNA methylation and miRNA expression are already suggested as biomarkers of disorders, prognosis, treatment prediction, and response. Considering the findings with the CB 1 receptor in human and animal models, epigenetic modifications in the Cnr1 gene are promising biomarkers in neuropsychiatry conditions [67].

Possible cross-talk between eCB system and epigenetics in drug abuse and stress
The eCB system is critical to the reward-related effects of dopamine, which is involved in the neurobiological mechanism underlying drug addiction [245]. Indeed, the modulation of the eCB system regulates molecular and behavioral responses promoted by distinct addictive drugs, including psychostimulants and alcohol [246,247]. Stress is an important risk factor in the neurobiology of drug addiction [248,249]. Previous stress exposure is correlated to the vulnerability to developing the disorder and the reinstatement of drug seeking. Interestingly, behavioral and molecular evidence indicates that the eCB system is a required element in the ability of stress to modulate drug responses [250,251]. This convergence is consistent with the fact that exposure to addiction drugs promotes changes in important brain structures also involved in stress biology, such as the PFC, nucleus accumbens, hippocampus, and amygdala, which are also important targets for cannabinoids [252]. In this way, similarly to what was described for stress events, exposure to addiction drugs also modulated the eCB system, which involves epigenetic mechanisms.
For instance, cocaine self-administration (SA) promotes histone modifications and chromatin looping in the eCB system-associated genes [253]. Animals exposed to cocaine demonstrated increased H3K4me3 enrichment on the hippocampus's promotor regions of FAAH and DAGLα coding genes. Moreover, using a 4C-seq approach targeting the Cnr1 promoter, authors also demonstrated that cocaine SA induces remodeling of chromatin loops in the hippocampus and the NAc, suggesting that 3D chromatin architecture at the Cnr1 locus was substantially changed

Figure 2. Mechanisms involved in the cross-talk between the eCB system and epigenetic mechanisms in stress-and drug abuse-related contexts
(A) In stressful contexts, interference with the eCB system can modulate a wide range of epigenetic factors. CB 1 receptors, for example, modulate the expression of DNMTs and the microRNA let-7d. Moreover, the inhibition of FAAH and consequent increase in AEA levels, which may act at CB 1 receptors, increases the expression of several miRNAs (miR-9-5p, miR-128-1-5p, miR-382-5p, miR-16-5p, miR-129-5p, miR-219a-5p, miR-16, and miR-135a) and H3K9ac levels at npy gene, and decreases the expression and activity of HDAC2. Moreover, TRPV 1 receptors activation increases the expression of HDAC2 and reduces global H3/H4 acetylation levels. Finally, CB 2 receptors have been shown to reduce TET enzyme levels. (B) Stress can regulate eCB genes through epigenetics tools. CB1 expression is reported to be sensitive to hypermethylation and increased levels of H3K9ac of the gene and affected by miRNAs (miR-23a, miR-29a, miR-29b-3, miR-30b, miR-128, miR-301a, miR-338-5p, miR-494, and miR-let-7d). FAAH expression may also be altered by H3K9ac and the miR-411. Furthermore, CB2 is one target of miR-187-3p and miR-665 expression. (C) The cross-talk between the systems in the context of drug abuse is very diversified since drugs with different mechanisms of action promote different alterations. For example, alcohol increases H4K8ac in the CB1 gene, and its protein expression is related to the down-regulation of MeCP2. Alcohol also down-regulates DNMT1 and DNMT3a and upregulates HDAC1, HDAC2, and HDAC3, and all these effects are blocked by CB1 antagonism. Moreover, DNA methylation of the FAAH gene is affected by alcohol exposure. Cocaine consumption is reported to increase H4K9me3 in FAAH and DAGLα genes. Regarding exposure to cannabinoids, THC induces global levels of H3K4me3 and H3K9me2 and can increase or decrease H3K9me3 depending on the exposure. More details about these mechanisms can be found in the main text. Dashed arrows indicate inhibition or reduction. Continuous arrow indicate induction or increase. Question mark indicates that the CB1 involvement after FAAH inhibition was not directed tested.
following cocaine exposure [253]. Pieces of evidence also have demonstrated that the eCB system undergoes epigenetic modulation by alcohol, as briefly mentioned before. A blind epigenome-wide analysis of datasets that explored hazardous drinkers and binge drinkers versus controls evidenced that Faah hypermethylation is associated with alcohol consumption [195]. Accordingly, an elevation in the expression of CB 1 associated with increased H4K8ac at the Cnr1 promoter was observed in adult mice exposed to alcohol on postnatal day 7 (PD7) [254]. These results provide evidence that epigenetic mechanisms contribute to altered regulation of the eCB system in response to specific abuse drugs.
Interestingly, epigenetic changes promoted by exposure to alcohol also appear to be modulated by the eCB system. Nagre and colleagues observed that treatment with ethanol in PD7 mice impaired DNA methylation through reduced DNA methyltransferases (DNMT1 and DNMT3A) levels; these effects were reversed by the blockade of CB 1 before ethanol treatment [255]. Similarly, alcohol exposure at the PD7 was associated with enhanced HDAC1, HDAC2, and HDAC3 expression, which was also prevented by administering a CB 1 receptor antagonist before alcohol exposure [256]. Moreover, using a similar protocol, another study demonstrated that exposure to ethanol activates the apoptotic caspase-3 enzyme via CB 1 in neonatal mice and causes a reduction in MeCP2 levels [257]. Regarding miRNA processes, a reduction in the expression of brain CB 1 was coupled with an increased complementary miR-26b in a mouse model of fetal alcohol spectrum disorders [258].
Corroborating the role of the eCB system in the regulation of drug response during development, converging pieces of evidence support that treatment with THC in early phases of development promotes epigenetic changes [259]. For instance, prenatal THC exposure significantly modifies the histone methylation profile in the NAc. Subjects exposed to THC during prenatal stage showed a decreased level of the H3K4me3 [152]. Similarly, persistent changes in H3K9, increased dimethylation and reduced trimethylation, were observed in the NAc of adult rats following adolescent THC exposure [260]. Another study observed a significant increase of H3K9me2 in the hippocampus and the amygdala of female rats exposed to THC during adolescence [261]. Moreover, using the same adolescent THC exposure, there was an enhancement in H3K9me3 in the nucleus accumbens, hippocampus, and PFC [261,262]. Preconception THC exposure also disrupts DNA methylation in the NAc, with cross-generational effects. In a study comparing rats with or without parental THC exposure, 1027 differentially methylated regions (406 hypermethylated and 621 hypomethylated) associated with parental THC exposure were found in the subsequent generation, even though they were not directly exposed to the drug [263].
Confirming the correlation between the inheritance of paternal epigenetic changes and cannabinoid exposure, developmental changes in the offspring were associated with premating paternal THC exposure [264]. Exposure to cannabinoids has also been associated with changes in sperm DNA methylation. The analysis of sperm DNA from adult rats exposed to 2 mg/kg of THC for 12 days identified 627 genes whose methylation status was altered [265]. Similarly, significant differential methylation of genes related to neurodevelopment was observed in the sperm of rats exposed to THC via oral gavage [266]. Similarly to the preclinical reports, substantial changes in both hypoand hyper-DNA methylation, with the latter predominating, were determined in the sperm methylome of marijuana smokers [265]. The impact of cannabis exposure on DNA methylation status also was investigated directly in human spermatogenesis in vitro. The results revealed alterations in DNA methylation levels of genes related to autism, HCN1, and NR4A2 [267]. These studies provide compelling evidence that preconception exposure to cannabinoids can impact reproduction and paternal epigenetic inheritance, potentially leading to altered DNA methylation patterns that have an impact on gene expression and developmental outcomes in offspring.
Altogether these findings support the idea that the eCB system is involved in regulating epigenetic mechanisms and has an essential role in the effects of addictive drugs. Since this response profile also was observed with stress exposure and considering the role of stress in the neurobiology of the substance use disorder, future studies might evaluate the involvement of the eCB system in the modulation of drug addiction by stress.

Final remarks
In the last two decades, much attention was directed toward understanding how exposure to different stressors could result in long-term changes in the organism that could result in psychiatric disorders. In this context, a boom of studies evaluating epigenetic changes in animal models and a run to find epigenetic markers related to psychiatric conditions arose, bringing many new understandings in the neurobiology of psychiatric conditions.
Among several physiological systems affected by epigenetic modulation, one has, in particular, been in the spotlight of scientists for more than 20 years: the endocannabinoid system. As overviewed in this review, the eCB system has a fundamental role in controlling many functions, including the fine control of stress response and circuits involved in drug abuse. Although not fully explored, eCB system genes are sensitive to epigenetic control [183,184,268,269]. The discussed evidence highlights the importance of epigenetic mechanisms in the eCB system response to stress, drugs of abuse, and the dysfunctions caused by them. As epigenetic marks can persist, the long-term alteration in the expression of cannabinoid-related proteins may be part of triggering diseases, particularly after stressors or substance use disorder. More recently, as discussed, many studies have investigated if the behavioral consequences of exposure to stressors in animals' models could result in epigenetic regulation of the eCB system. Changes in miRNAs that regulate eCB system molecules, for example, are observed after acute protocols of stress in animal models but also in postmortem brains of depressive subjects. Epigenetic changes can persist through generations, indicating how stress and drug exposure, for example, can modify the neurobiology along the generations.
As also discussed, in animal models, cannabinoids, including THC and CBD, promote several behavioral changes related to psychiatric disorders and induce epigenetic modifications, mainly related to DNA methylation and histone modifications. Besides, Cannabis use in humans appears to induce epigenetic changes not only in the eCB system but also in the dopaminergic system and others, indicating a potential mechanism by which it could lead to psychiatric disorders, including substance use disorder. Finally, exposure to cannabinoids during critical periods of brain development can induce persistent brain and behavioral changes in adulthood.
As summarized in this review, therefore, there appears to have a close relationship between modulation of the eCB system and evaluation of epigenetic changes (DNA methylation, histones modifications, and miRNAs) under stress conditions (Figure 2A) and how epigenetic markers under stress conditions, mainly miRNAs, influence the expression of eCB-related molecules ( Figure 2B). Furthermore, common drugs of abuse, including alcohol, cocaine, and cannabis (THC), could promote their long-term effects by promoting epigenetic changes that impact the eCB system ( Figure 2C). The elucidation of epigenetic mechanisms controlling, or being controlled by, the eCB system in stress-related disorders is essential to better understand the neurobiology of those disorders and to provide new treatment approaches. Finally, understanding the cross-talk between those systems can potentially lead to the identification of biomarkers, such as miRNAs, which could help to predict the course of the disease and treatment response.

Data Availability
Data sharing is not applicable.

Competing Interests
The authors declare that there are no competing interests associated with the manuscript.