High times for cannabis: epigenetic imprint and its legacy on brain and behavior

Abstract

Extensive debates continue regarding marijuana (Cannabis spp), the most commonly used illicit substance in many countries worldwide. There has been an exponential increase of cannabis studies over the past two decades but the drug’s long-term effects still lack in-depth scientific data. The epigenome is a critical molecular machinery with the capacity to maintain persistent alterations of gene expression and behaviors induced by cannabinoids that have been observed across the individual’s lifespan and even into the subsequent generation. Though mechanistic investigations regarding the consequences of developmental cannabis exposure remain sparse, human and animal studies have begun to reveal specific epigenetic disruptions in the brain and the periphery. In this article, we focus attention on long-term disturbances in epigenetic regulation in relation to prenatal, adolescent and parental germline cannabinoid exposure. Expanding knowledge about the protracted molecular memory could help to identify novel targets to develop preventive strategies and treatments for behaviors relevant to neuropsychiatric risks associated with developmental cannabis exposure.

1. Introduction

The reduced perception regarding risks associated with marijuana (Cannabis sativa, Cannabis indica), as well as the growing industry evolving around recreational and medical cannabis, has lead to its increased use particularly among young people (SAMSHA, 2016). It is the first time in the United States’ history that adolescents smoke marijuana more than cigarettes, an increasing tendency since 2010 (Johnston et al., 2012; SAMSHA, 2016). Cannabis has low/moderate addictive properties (Gable, 2006) with only approximately 10–16% of users developing dependence (Anthony, 2006) yet, due to its prevalence today, millions of people in the United States and worldwide meet the clinical diagnosis for cannabis use disorder. This number far exceeds that of all other illicit drugs combined, even taking into consideration marijuana’s recent non-illicit status in several countries and states. While the great efforts taken to educate the public about the health risk of cigarettes have been successful, the pressure for marijuana legalization has contributed to teenagers being under the belief that marijuana is safe (SAMSHA, 2016).

In addition to recreational marijuana, ‘medical cannabis’ and cannabinoids are now being explored as potential therapies to treat various diseases and clinical symptoms. The health conditions studied thus far have been broad including chronic pain, spasticity due to multiple sclerosis, nausea and vomiting due to chemotherapy, depression, anxiety disorder, sleep disorder, psychosis and intraocular pressure associated with glaucoma to name a few. In a recent systematic study of randomized clinical trials (Whiting et al., 2015) only moderate to low quality evidence supported the beneficial clinical effects of medical cannabinoids. Another comprehensive review regarding the therapeutic and recreational use of cannabis and cannabinoids (National Academies of Sciences, 2017) suggest that the evidence is strong for the treatment of chronic pain and spasticity. Consistent conclusions in regard to mental health risk were the substantial evidence for a statistical association between cannabis and the development of schizophrenia or other psychoses (highest risk among the most frequent users) and the fact that initiating cannabis use at an earlier age is a risk factor for the development of cannabis abuse (National Academies of Sciences, 2017).

A growing body of literature has shown that the developing brain is especially sensitive to drugs compared with the adult brain (Anker and Carroll, 2010; Curran et al., 2016; Shahbazi et al., 2008; Zakharova et al., 2009) which is of important concern given that marijuana is the most commonly abused drug by two vulnerable populations — adolescents and pregnant women. Despite the perceived low health risk of cannabis use by the general public, there is now growing clinical awareness about the spectrum of behavioral and neurobiological disturbances associated with cannabis exposure such as psychosis, anxiety, depression, cognitive deficits, social impairments and subsequent drug addiction (Alegria et al., 2010; Bassir Nia et al., 2016; Charilaou et al., 2017; Crean et al., 2011; Feingold et al., 2017; Jutras-Aswad et al., 2009; Kedzior and Laeber, 2014; Leweke and Koethe, 2008; Malone et al., 2010; Morris et al., 2011; Sexton et al., 2016). These types of studies in recent years have begun to shift the perception of marijuana use being without any harm and emphasize the importance for more in-depth scientific investigations to address the potential long-term impact of cannabis use.

The impact of cannabinoid exposure on neurodevelopment is a central question since the brain undergoes rapid growth not only in the prenatal period but also during postnatal life, until early- to mid-adolescence. However, as with most human studies, data generated to date are clearly equivocal most likely due to multiple factors such as the dose and strain of cannabis and cannabinoids used, ratio of the cannabinoids and other entourage chemicals in the plant that is consumed, developmental time of exposure and genetic vulnerability that may modulate the risk for both the adverse effects and therapeutic potential. As such, scientific and medical questions are crucial to be asked about the long-term consequences of cannabis exposure on brain function and behavior. Controlled animal studies provide the potential to explore the behavioral and molecular consequences of cannabinoid exposure but also have evident challenges in recapitulating human usage of the drug such as dose range and route of administration. Nevertheless, information can be gleaned from existing human and animal studies that set the foundation for designing future studies to gain deeper neurobiological insights. In this article, we provide an overview of the current scientific data regarding vulnerabilities of the developing brain to cannabinoid exposure during particularly sensitive windows of development and its epigenetic legacy later in life.

2. Developmental “nature-nurture” interactions and cannabis use

Today, both scientists and clinicians recognize the importance of the prenatal and adolescent developmental periods in chronic and psychiatric disease. Elucidating associations between genotype, environment and phenotype has resulted in an impressive collection of data in relation to many substance use and neurodevelopmental disorders (Boivin et al., 2015; Brander et al., 2016; Enoch, 2012; Isles, 2015; Lv et al., 2013; Vrieze et al., 2012). Although the evidence is not absolute, an extensive body of literature has begun to specifically link marijuana use by pregnant women and during adolescence with adult mental health disturbances later in life (Barthelemy et al., 2016; Chadwick et al., 2013; Jutras-Aswad et al., 2009; Morris et al., 2011; Rubino and Parolaro, 2016). Cannabis exposure during these critical windows of development is thus expected to adversely affect neurobiological systems, resulting in long-lasting alterations in molecular mechanisms affecting neurocircuitry (Fig. 1). Determining how molecular mechanisms contribute to marijuana’s acute effects has been a central question in cannabis research during the last decade. However, much less is known about gene-environment interactions as they relate to the etiology of the complex neuropsychiatric phenotypes relevant to cannabis exposure.

Figure 1. Cannabis exposure during sensitive periods of development can impact epigenetic mechanisms, leading to persistent gene regulation and behavioral alterations.

The two most likely populations to use cannabis are pregnant women and adolescents (indicated by green arrow on top). These developmental phases also correspond to periods when the brain is most vulnerable to the influence of external cannabinoids via interfering with epigenetic mechanisms (shown below the schematic of the developmental cycle). Several epigenetic mechanisms that are relevant to the effects of cannabinoids can interfere with normal gene expression via interacting with DNA elements (e.g. promoters) and transcription factors (proteins that bind to the DNA) to regulate mRNA transcript levels from a gene. Specific regulatory mechanisms include DNA methylation (Me), positioning and post-translational modifications of nucleosomes (small blue balls), recruitment of the transcription complex (sequence-specific and basal transcription factors, RNA polymerase II), and non-coding RNAs. DNA methyltranserases (DNMT) generate 5-methylcytosine (pink stars) at CpG sites. Ten-eleven translocation (TET) proteins mediate the oxidation of 5-methylcytosine to 5-hydroxymethylcytosine (green stars), leading to demethylation of the DNA. Modifications of nucleosomal histone tails such as methylation (Me) and acetylation (Ac) are mediated by histone methyltransferases (HMT) and histone acetyltransferases (HAT), respectively. Small RNAs are produced from specific genes and either influence the transcription process or target protein-coding messenger RNAs for degradation. Germ cells (sperm, oocyte) are also sensitive to cannabinoids but the exact underlying epigenetic mechanisms remain to be determined.

What began as investigations focused on the intersection between genetics and developmental biology by scientists such as Conrad H. Waddington and Ernst Hadorn during the mid-twentieth century has evolved into the field we currently refer to as epigenetics. The term epigenetics, which was coined by Waddington in 1942, was derived from the Greek word “epigenesis” which originally described the influence of genetic processes on development (Van Speybroeck, 2002). This was subsequently expanded into broad studies on the molecular basis of Waddington’s observations regarding how environmental insults interact with the genetic material to cause certain phenotypic characteristics. Since then, a great number of research efforts have been focused on unraveling the epigenetic mechanisms related to gene-environment relationships in the context of substance use disorders. Below we provide an overview on behavioral and molecular brain alterations documented in adults exposed to cannabis during adolescence, in the offspring of women with cannabis use during pregnancy, and in subsequent generations conceived by individuals with cannabis-exposed germ cells. We propose and discuss a model for the development of cannabis-related abnormalities shown in Fig. 1 in the context of epigenetic molecular mechanisms.

3. Epigenetic regulatory mechanisms

According to the original definition, “an epigenetic trait is a stably heritable phenotype resulting from changes in a chromosome without alterations in the DNA sequence” (as proposed by Conrad Waddington); this view also implies heritability resulting in a phenotype (Baedke, 2013; Van Speybroeck et al., 2002). The epigenome provides the cellular context for environmental effects, including cannabis exposure during prenatal and early postnatal periods (Szutorisz and Hurd, 2016), therefore it is the most relevant biological target for the propagation of persistent abnormalities and aberrant neuronal processing (Fig. 1).

Generally, the interaction between genomic DNA elements (specific sequences with regulatory function), epigenetic modifiers, and transcription factors determines the expression state of genes. This network of processes is tightly coordinated in space and time, in the specification of different cell, tissue and organ types, and throughout the lifespan of the individual (Dambacher et al., 2013; Dillon, 2012) (Weake and Workman, 2010). In molecular biology, “epigenetic” typically has been used to refer to mechanisms that modulate gene expression without altering the genetic code. There are various epigenetic mechanisms including DNA methylation, nucleosomal structure and positioning, post-translational modifications of nucleosomal histones, histone replacement, and small RNA molecules that help to establish the molecular platform that maintains protracted effects on gene expression and ultimately behavior (Baubec and Schubeler, 2014; Dambacher et al., 2013; Dillon, 2012; Weake and Workman, 2010). In a biological mechanistic context, the complex interaction between genomic DNA elements (specific sequences with regulatory function), epigenetic modifiers and transcription factors determines the expression state of genes. This network of processes is tightly coordinated within cellular compartments, during the specification of different cell, tissue and organ types, and throughout the development and lifespan of the individual (Dambacher et al., 2013; Dillon, 2012; Weake and Workman, 2010).

Some of the most important ontogenetic regulatory decisions take place in early pre- and postnatal development and thus have critical implications in the influence of drug exposure during specifically sensitive periods. Epigenetic modifications that can regulate gene expression levels include DNA methylation, chromatin structure and remodeling, post-translational modifications of nucleosomal histones, histone replacement, and small RNA molecules that can influence protein production and transcription (Fig. 1). Mechanistic implications of several specific epigenetic processes that have thus far been linked to the effects of cannabis are described below.

3.1. DNA methylation

The role of DNA methylation in the regulation of gene expression is complex and highly dependent on genomic location, developmental stage, cell type, or disease state. Historically, CpG methylation (referring to cytosine and guanine in DNA sequence where “p” indicates that the “C” and “G” bases are connected by a phosphodiester bond) in promoter regions and transcriptional regulatory sequences has often been associated with gene repression, whereas methylation within the gene body is less understood (Baubec and Schubeler, 2014; Kato and Iwamoto, 2014). Accumulating evidence now also indicates that DNA methylation in brain is dynamic and its distribution changes throughout neuronal maturation and aging, in neurodevelopmental disorders, including addiction to drugs (Cheng et al., 2015; Feng et al., 2015). Mechanistically, DNA methylation (5-methylcytosine, 5mC) is generated by DNA methyltranserases (DNMTs). The oxidation of 5mC to 5-hydroxymethylcytosine (5hmC) by ten-eleven translocation (TET) proteins can prevent access to DNMTs and thereby can maintain an unmethylated state of the promoter, facilitating transcriptional activation (Branco et al., 2012). Interestingly, DNA methylation marks at specific gene loci have been shown to even persist during the maturation of germ cells (Szyf, 2013, 2015) and thus are interesting candidates for the transmission of cannabis effects from parent to child and possibly throughout multiple generations.

3.2. Histone modifications

On the protein level, the main epigenetic mechanism that has been implicated in neurobiological disturbances is posttranslational modifications of nucleosomal histones, which with the DNA that encircle them comprise the structural and regulatory unit of chromatin. Histones are subject to a variety of chemical modifications including but not limited to, lysine acetylation, lysine and arginine methylation, serine and threonine phosphorylation, and lysine ubiquitination and sumoylation (Bhaumik et al., 2007). These modifications occur primarily within the histone amino-terminal tails protruding from the surface of the nucleosome as well as on the globular core region, and have been shown to influence both the accessibility of genomic regions and the binding of transcription factors to the DNA (Cosgrove et al., 2004). Changes in acetylation and phosphorylation in response to drug exposure are often transient and associated with the quick activation of genes rather than the maintenance of an altered transcription state (Ciccarelli and Giustetto, 2014). However, histone lysine methylation is known to maintain stable gene expression alterations over long periods of time, and it is also the nucleosomal modification that has been associated with the persistent effects of marijuana and different cannabinoids in neurons and other cell types (Aguado et al., 2007; DiNieri et al., 2011; Tomasiewicz et al., 2012; Yang et al., 2014).

3.3. Non-coding (nc) RNAs

These functional RNA molecules are transcribed from DNA but are not translated into proteins. Many ncRNAs regulate gene expression at the transcriptional and post-transcriptional level. The ncRNAs that are known to be involved in epigenetic processes can be divided into two main groups — short ncRNAs (<30 nucleotides) and long ncRNAs (>200 nucleotides). The three major classes of short ncRNAs are microRNAs (miRNAs), short interfering RNAs (siRNAs), and piwi-interacting RNAs (piRNAs) (Chandra et al., 2015; Hegde et al., 2013; Jackson et al., 2014; Molina et al., 2011). While the exact genomic targets of specific cannabinoid-affected miRNAs remain to be characterized, these observations are mechanisticall intriguing given the variety of tissue-specific cellular and developmental processes that are influenced by miRNAs. Small RNAs have also received significant attention as regulators of multigenerational inheritance in a variety of organisms (Houri-Zeevi and Rechavi, 2017).

4. Cannabis and its neurobiological targets

4.1. Main chemicals in cannabis

The cannabis plant contains over 500 herbal compounds and cannabinoids constitute at least 100 of these (ElSohly et al., 2017). Cannabinoids interact with the endogenous systems of the body contributing to the user’s high (acute effect) and are also the molecular instigators of the long-term consequences of marijuana exposure. They are also expected to account for the ability of ‘medical cannabis’ to alleviate a variety of physiological and neuropsychiatric symptoms. Of the numerous cannabinoids, the two most extensively researched compounds are Δ⁹-tetrahydrocannabinol (THC) and cannabidiol (CBD). THC is known for its psychoactive properties (Martin-Santos et al., 2012) and CBD is a non-psychoactive cannabinoid shown to have anti-inflammatory effects, to protect neurons from injury or degeneration, to reduce anxiety, to attenuate drug craving in certain people, and to have antipsychotic properties (Hampson et al., 1998; Hurd et al., 2015; Leweke et al., 2012; Zuardi et al., 1982). Most strains of marijuana that has been cultivated and sold on the market over the last decade contain increasingly higher levels of THC and lower levels of CBD (Anker and Carroll, 2010; ElSohly et al., 2016; Guimaraes et al., 1994; Swift et al., 2013; Zuardi et al., 2006). Some studies indicate that, on average, variants of the Cannabis sativa species contain higher levels of THC to CBD and are commonly used for the characteristic ‘high’, whereas Cannabis indica has higher levels of CBD compared to THC and considered beneficial for its sedative, anxiolytic and analgesic properties (Hazekamp and Fischedick, 2012; Pearce et al., 2014).

4.2. Endocannabinoid (eCB) system

In the early 1990s, the neurobiological link between cannabis and its acute psychoactive effects were uncovered through identification of the eCB system. The brain creates its own set of cannabinoids that consist of lipid ligands and cannabinoid receptors, which mediate the actions of THC. The eCB system is responsible for the regulation of many important functions, such as appetite, sleep, emotion, memory and movement (Hillard, 2015; Kruk-Slomka et al., 2016; Moreira and Lutz, 2008; Prospero-Garcia et al., 2016; Tasker et al., 2015). The eCB system modulates synaptic function by on-demand synthesis and release of the ligands from the postsynaptic cell and the subsequent activation of cannabinoid receptors on the presynaptic neurons that attenuate excitatory and inhibitory neurotransmitter release within discrete neuronal circuits.

During development, CBRs play a central role in hardwiring the developing brain and contribute postnatally to the regulation of synaptic plasticity (Berghuis et al., 2007; Tortoriello et al., 2014). Two major types of cannabinoid receptors have been characterized in mammals: cannabinoid 1 receptors (CB1Rs) and cannabinoid 2 receptors (CB2Rs). CB1Rs are the most-abundant G protein-coupled receptors that are expressed in the adult brain, and they show particularly dense distribution in regions that are involved in reward processing and cognitive functions, such as the ventral pallidum, caudate putamen, nucleus accumbens (NAc), ventral tegmental area, amygdala, cingulate cortex, prefrontal cortex, and hypothalamus (Glass et al., 1997; Wang et al., 2003). CB1Rs directly inhibit the release of GABA, glutamate and acetylcholine, which produce widespread effects on neural signalling across many neurotransmitter systems (Lopez-Moreno et al., 2008). CB2Rs are expressed mainly in immune cells and the gut, although recent evidence suggests that they are also present in subsets of neurons, glia and endothelial cells of the brain (Atwood and Mackie, 2010).

The currently best-known eCB ligands are N-arachidonylethanolamide (anandamide (AEA)) and 2-arachidonoylglycerol (2-AG), which are synthesized upon induction by cleavage of membrane-bound precursors and immediately released through Ca2+-dependent mechanisms (Parsons and Hurd, 2015). AEA is derived from the phospholipid precursor N-arachidonoyl-phosphatidylethanolamine (NAPE) and, although the exact mechanisms for AEA formation are not known, the N-acyl-phosphatidylethanolamine-specific phospholipase D (NAPE-PLD) enzyme is likely to have a role in this process. 2-AG derives primarily from the hydrolytic metabolism of 1,2-diacylglycerol (DAG) by the sn-1-selective DAG lipases (DAGLs) DAGLα and DAGLβ. Once released into the extracellular space, eCBs are vulnerable to glial cell inactivation. AEA and 2-AG both exert agonist activity at CB1R and CB2R. AEA binds with slightly higher affinity to CB1R than to CB2R and, similar to THC, it exhibits low agonist activity at both receptors. 2-AG binds with essentially equal affinity at CB1R and CB2R and exhibits greater agonist efficacy than AEA.

The normal epigenetic control of the eCB system has recently been reviewed in (D’Addario et al., 2013). Various lines of evidence strongly suggest that the eCB signaling cascades mediated via CBRs regulate cellular functions in different tissues via epigenetic alterations in DNA methylation (Paradisi et al., 2008), histone methylation (Aguado et al., 2007), and miRNAs (Jackson et al., 2014). These data highlight the role of the eCB system in regulating cellular functions through epigenetic modifications and suggest that modulation of these mechanisms with cannabis use may have long-lasting neurobiological and functional impact.

5. Critical windows of development relevant to cannabinoid exposure

The study of epigenetics in relation to drugs of abuse has been a rapidly emerging field during the past several years, yielding important mechanistic revelations about different addictions and relevant neuropsychiatric disorders (Robison and Nestler, 2011; Sweatt, 2013). However, experimental data regarding potential epigenetic effects associated with cannabis exposure are still sparse in spite of the relatively easy accessibility and frequent use and abuse of this drug (Szutorisz and Hurd, 2016). Of the few published studies (most mainly focused on THC or synthetic cannabinoids), various neuropsychiatric phenotypes and epigenetic alterations that have been reported in association with developmental cannabinoid exposure as summarized in Table 1.

Table 1.

Cannabinoid exposure at sensitive periods of development associated with long-term behavioral and epigenetic disturbances.

Developmental period of exposure	Substance	Phenotypic disturbance later in life	Alteration observed in body	Epigenetic modification	Reference
In utero and adolescence	Cigarettes, alcohol (prenatal), Cannabis (adolescence)	Increased substance use in adolescence	Human gestational and adolescent blood	DNA methylation	(Cecil et al., 2016)
In utero	THC	Increased heroin seeking in adulthood	Adult rat brain (NAc)	Histone H3 methylation: H3K4me3, H3K9me2	(DiNieri et al., 2011; Spano et al., 2007)
Adolescence	THC	Increased heroin seeking in adulthood	Adult rat brain (NAc)	Histone H3 methylation: H3K9me2, H3K9me3	(Tomasie wicz et al., 2012)
Adolescence	Synthetic cannabinoid WIN55212.2	Memory impairment	Adult mouse brain (hippocampus)	DNA methylation	(Tomas-Roig et al., 2016)
Adulthood (at time of study)	Cannabis	Schizophrenia in adulthood	Human peripheral blood	DNA methylation	(Liu et al., 2014)
Adolescent parental germline	THC	Increased heroin seeking in adulthood	Adult rat brain (NAc)	DNA methylation	(Szutorisz et al., 2016; Watson et al., 2015)

5.1. Long-term effects of gestational cannabis exposure in the offspring

Various studies have evaluated the behavioral effects in the progeny of women who smoked cannabis when pregnant. Multiple review articles have previously addressed the phenotypic effects in humans (for example (Fried et al., 2003; Goldschmidt et al., 2008; Jutras-Aswad et al., 2009; Morris et al., 2011; Tomas-Roig et al., 2016). A number of preclinical animal studies have also demonstrated prenatal THC exposure on offspring behaviors and some suggested disturbances in gene expression (Campolongo et al., 2007; Rubio et al., 1998; Singh et al., 2006; Spano et al., 2007; Vela et al., 1998). Here we focus on findings linking cannabinoid exposure with epigenetic changes that are likely to cause dysregulation in the expression of genes functionally relevant to offspring neuropsychiatric phenotypes (Table 1).

Numerous investigations on the developmental effects of THC directly described molecular alterations highly relevant to addiction disorders. These studies focused in large part on the NAc, a critical neuroanatomical substrate underlying the pathophysiology of addiction (Everitt and Robbins, 2013; Girault, 2012; Koob and Volkow, 2010). Of the multiple epigenetic mechanisms, the regulation of histone modification is unique because they can have either positive or negative effects on gene transcription. Indeed, our previous studies revealed disturbances in the histone modification profile in the NAc of adult rats with prenatal THC exposure (DiNieri et al., 2011). This study identified decreased levels of the trimethylation of lysine 4 on histone H3 (H3K4me3), a transcriptionally permissive mark, increased levels of dimethylation of lysine 9 on histone H3 (H3K9me2), a repressive mark, as well as decreased RNA polymerase II association with the promoter and coding regions of the gene in the NAc. These THC-related chromatin modifications were linked to significant disturbances in the mRNA expression of the dopaminergic D2 receptor (Drd2) gene in both rats and humans that persisted into adulthood, emphasizing the enduring consequences of THC/cannabis exposure during gestational development.

More recently, Cecil et al. (Cecil et al., 2016) conducted the first genome-wide, relatively large human study to examine associations between DNA methylation alterations in blood from birth to early childhood and tobacco, cannabis and alcohol use later in adolescence in subjects from the Avon Longitudinal Study of Parents and Children (ALSPAC). They found that, at birth, variation in DNA methylation in gestational and adolescent blood across a tightly interconnected genetic network was associated with greater levels of substance use during adolescence, as well as an earlier age of onset amongst users. Affected genes included PACSIN1, NEUROD4 and NTRK2, which are implicated in neurodevelopmental processes. Although this study was not specific to the effects of cannabis exposure but also included cigarette and alcohol use, it provides valuable information on the relationship between gestational drug exposure, DNA methylation alterations detectable in the periphery, and substance abuse risk, including cannabis, later in life. Together, these findings highlight gestation as a sensitive window of biological vulnerability and provide evidence for abnormal epigenetic signatures of prenatal drug exposure and substance abuse vulnerability in postnatal life.

5.2. Consequences of adolescent cannabinoid exposure later in life

Marijuana use by teenagers and young adults often predates the abuse of harder drugs (known as the classic “gateway” concept), but the neurobiological underpinnings of such vulnerability are just beginning to be unraveled (Table 1). Converging evidence obtained from animal and human brain studies has shown that, similarly to Drd2, early cannabis exposure selectively alters the expression of opioid neuropeptide proenkephalin (Penk) in the mesocorticolimbic system, disturbances that persist into adulthood and modulate drug-seeking behavior later in life (Jutras-Aswad et al., 2012). Several years ago we reported a direct causal link between the upregulation of the Penk gene in the NAc due to THC exposure during adolescence and enhanced behavioral susceptibility to heroin seeking in adulthood (Tomasiewicz et al., 2012). On the chromatin level, persistent changes in repressive H3K9me2 and H3K9me3 were observed at the Penk locus in the NAc of adult rats following adolescent THC exposure, in line with enduring upregulation of Penk mRNA expression. This epigenetic effect represents a profound pathologic departure from the distinct developmental pattern of histone H3 methylation that normally occurs at Penk in the NAc across the transition from adolescence to adulthood. The chromatin landscape is highly complex, but trimethylation of H3K9 (a transcriptionally repressive mark) may account for the developmental transcriptional instability of NAc Penk due to adolescent THC exposure, allowing the Penk gene to be “primed” to respond to environmental cues later in life.

In the last decade not only marijuana but also synthetic cannabinoids have become increasingly popular among young people (Tournebize et al., 2016). While short-term effects on cognition and psychosis are normally observed in those individuals (Bassir Nia et al., 2016; Cohen et al., 2017; Murray et al., 2016; Spaderna et al., 2013), questions remain as to whether there may be enduring effects of adolescent synthetic cannabinoid use on adult brain and behavior. Preclinical studies have started to address this issue primarily with the use of the synthetic cannabinoid receptor agonist WIN55212.2, which acts similarly to THC but with much greater efficacy at the cannabinoid receptors. In a recent study, adult rats with a history of adolescent WIN55212.2 (self-administered or experimenter-administered) failed to show any long-term cognitive dysfunction as measured using working memory and spatial recognition tasks (Kirschmann et al., 2017). Although prefrontal GABAergic and glutamatergic signaling was altered in the prefrontal cortex of these animals, epigenetic alterations were not examined and remains to be investigated (Kirschmann et al., 2017). Another study (Tomas-Roig et al., 2016) that investigated the chronic administration of WIN55212.2 during adolescence in young adult mice did begin to address epigenetic consequences. Animals that received the drug during adolescence showed spatial memory disturbances in the Morris water maze, as well as a dose-dependent memory impairment in fear conditioning. Moreover, adolescent WIN55212.2 exposure increased adult hippocampal eCB levels and promoted DNA hypermethylation at the intragenic region of the intracellular signaling modulator Rgs7, which was accompanied by a lower rate of mRNA transcription of the gene, suggesting a potential causal relationship. RGS proteins are important regulators of striatal G protein-coupled receptors signaling (Ostrovskaya et al., 2014; Sjogren, 2011; Xie and Martemyanov, 2011). Although the concrete mechanisms underlying the behavioral observations remain to be elucidated, the study does demonstrate that exposure to a synthetic cannabinoid during adolescence could lead to epigenetic gene regulation abnormalities in adulthood. Nevertheless, the equivocal behavioral findings in the few papers published to date emphasize the need for more research focused on the effects of synthetic cannabinoids.

Of the different components of the eCB system, several investigations have focused on the epigenetic regulation of the Cnr1 gene, which encodes the CB1R. Specific genomic elements of the Cnr1 locus have been shown to interact with transcription factors, some of which are implicated in methylation of CpG sites in the DNA and histone posttranslational modifications (Lee et al., 2013; Mukhopadhyay et al., 2010; Nagre et al., 2015). CB1R expression has been reported to increase in peripheral blood lymphocytes of human schizophrenic patients with cannabis abuse and is inversely correlated to methylation of the CNR1 gene promoter (Liu et al., 2014). Interestingly, CNR1 mRNA expression levels and promoter DNA methylation detected in the blood was reported to relate to the intensity of cannabis craving as well as to the severity of nicotine, cannabis and alcohol consumption, suggesting a relevance of CNR1 epigenetic status to brain function and behavior.

In summary, these findings clearly show that cannabinoid exposure during adolescence and young adulthood can imprint on the epigenetic landscape of postnatal development and augment behavioral responses via the dysregulation of genes that have important neurobiological functions related to addiction risk.

5.3. Effects of germline cannabinoid exposure through multiple generations

A less obvious and still significantly unexplored question regarding the long-term consequences of prenatal cannabis exposure is whether there could be potential impact on subsequent generations. In recent years, findings in various disease states have demonstrated epigenetic aberrations that influence developmental risk and can be inherited through the germline from parent to child (Bohacek and Mansuy, 2013; Szyf, 2015). Several cases of parent-child transmission regarding drugs of abuse have been published, describing both behavioral phenotypes and molecular disturbances in the offspring of parents that were exposed to drugs before mating and conception (reviewed in (Vassoler and Sadri-Vakili, 2014)). Such studies have provided compelling evidence for stably heritable phenotypes resulting from epigenetic changes as originally described in the classic model of epigenetics (see section 3).

Though it is still a provocative concept, we have previously demonstrated that exposure of male and female adolescent rats before mating (“germline exposure”) leads to behavioral and molecular abnormalities in their unexposed offspring (Szutorisz et al., 2014). These studies revealed that adult progeny, themselves unexposed to THC, displayed increased work effort to self-administer heroin, demonstrating a cross-generational “gateway” most likely established in the parental germline before conception (Table 1). Importantly, acute effects of the drug are ruled out in these experiments since THC was no longer present in the body at the time of parental mating. Furthermore, all offspring were raised by surrogate dams never exposed to THC, ensuring that any cross-generational effects in the offspring were not due to drug-related abnormalities in maternal care early in life. Other investigators have also described cross-generational behavioral alterations using synthetic cannabinoid agonists. For example, adolescent female rats treated with WIN55212.2 before mating and pregnancy had progeny that exhibited increased morphine sensitivity (Byrnes et al., 2012; Vassoler et al., 2013). Neurobiologically, parental THC exposure has been associated with changes in the mRNA expression of cannabinoid, dopamine, and glutamatergic receptor genes in the striatum and altered synaptic plasticity in neurophysiological measures. Both sexes showed pronounced glutamatergic disturbances in the dorsal striatum in adulthood though stronger in females (Szutorisz et al., 2016).

Using the same paradigm as above, robust DNA methylation disturbances were detected in the NAc of adult rats with parental germline THC exposure in an epigenome-scale investigation (Watson et al., 2015). A key observation was the identification of DNA methylation alterations within an interaction network centered around the Dlg4 gene, encoding Psd-95, a membrane associated guanylate kinase scaffolding protein located in neural postsynaptic densities, involved in the regulation of dopamine-glutamate interactions (de Bartolomeis and Tomasetti, 2012). Previously, epigenetic dysregulation of Dlg4 has been linked to abnormal glutamatergic transmission involved in morphine conditioning (Wang et al., 2014), consistent with the earlier observations of increased heroin self-administration in adult offspring with germline THC exposure (Szutorisz et al., 2014). Many genes containing abnormal methylation are well-known regulators of neurotransmission and synaptic plasticity, including glutamate and kainite receptors, G-protein-coupled receptors, pre- and postsynaptic ion channels and scaffolding proteins, that have been speculated as susceptibility genes for psychiatric conditions such as schizophrenia, depression, autism and obsessive-compulsive disorder (Spiers et al., 2015; Wilson and Sengoku, 2013). Intriguingly, the dynamic control of DNA methylation and demethylation has recently been strongly implicated in the regulation of synaptic plasticity and drug addiction (Feng et al., 2015; Lunnon et al., 2016; Sweatt, 2013, 2016).

Multigenerational epigenetic effects occur when exposure to environmental stimuli triggers epigenetic alterations that are transmitted to the subsequent generation. Three different routes of multigenerational transmission have been described: fetal programming (e.g. maternal stress), behavioral/social transfer (e.g. interaction between parents and offspring), and germline transmission (Bohacek et al., 2013; Cowley and Oakey, 2012). In germline epigenetic inheritance, germ cells undergo meiosis to produce the gametes that could be vulnerable to alterations by parental THC exposure. In this context, it is important that the eCB system plays important roles not only in the development of a variety of somatic cells and physiological systems, but also in reproduction. It is known that both male and female reproductive tissues express cannabinoid receptors and eCBs and that in males, THC can disrupt the normal development of sperm cells (Banerjee et al., 2011; Bari et al., 2011). As an epigenetic correlate, studies on the impact of cannabinoids on male fertility have been conducted in Cnr1 null mutant mice that displayed higher histone retention in germ cells compared to wild type mice (Chioccarelli et al., 2010). In that study, CB1R expression was demonstrated to be necessary for spermiogenesis by controlling chromatin folding in sperm via the regulation of histone displacement. Marijuana-using women are also known to produce poor quality oocytes, associated with lower pregnancy rates (Klonoff-Cohen et al., 2006). Future studies are required to systematically address assess how possible epigenetic processes such as DNA methylation or chromatin regulation are disrupted by cannabinoids and are involved in the transmission of effects from parent to offspring and, potentially, throughout multiple generations.

6. Conclusions

The relationship between cannabis use and neuropsychiatric vulnerability is clearly complex, but the limited data accrued to date in this fast growing field already documents that early exposure during one’s lifetime leaves a long-term epigenetic memory mark which sets a legacy even onto future generations. As more studies are conducted, there are several aspects that must be considered in experimental design in order to gain greater in-depth insight and rigorous assessments to advance knowledge. These include fundamental knowledge regarding dose that better approximate human usage, timing or exposure, reversibility of behavior, and molecular events induced by developmental cannabis/cannabinoid use. Moreover, it will be important to investigate how different cannabinoids or other components of the cannabis plant and even their interactions (e.g. THC and CBD) influence behavioral, physiological and epigenetic effects in long-term users.

Due to technological advances, large genome-wide datasets are now available which makes it possible to perform complex computational analyses to help determine overlapping (e.g., transcriptome and epigenome) biological patterns that can guide the discovery of novel regulatory mechanisms. Another important variable that has not been well explored but acknowledged to be important in developmental questions regarding cannabis is potential sex-specific effects that have significant implications for disease vulnerability and treatment response.

Additionally, most information garnered to date about neurochemical and molecular mechanisms in brain are derived from homogenate approaches that are informative but limit cell-type specific knowledge. Direct insight about the epigenome and transcriptome within specific neural circuits such as the discrete striatal output pathways and prefrontal cortical circuits relevant to goal-directed behavior and decision-making would help to unravel the dynamic cellular mechanisms across development, linked to cannabis-related psychiatric disorders. One challenge of the cell-specific nature of epigenetic modifications is being able to track alterations across time in the brain within the same individual. As such, assessment of peripheral epigenetic marks in association with developmental cannabis exposure could provide an important opportunity to track the epigenetic trajectory across time in relationship to clinical outcomes. The ability to identify specific peripheral biomarkers in humans would also have significant translational value that could be integrated with animal models to allow mechanistic evaluation.

Overall, expanding knowledge about the protracted neurobiological signature of epigenetic memory associated with marijuana and various cannabinoids will identify novel targets to develop preventive strategies and treatments for behaviors relevant to neuropsychiatric risks due developmental cannabis exposure.

HIGHLIGHTS.

• Long-term developmental effects of cannabis largely lack in-depth scientific data.

• The epigenome underlies molecular and behavioral effects of cannabinoids.

• We discuss epigenetic dysregulation by prenatal, adolescent and germline cannabis.

• Expanding epigenetic knowledge will provide targets for treatment interventions.