The Effects of Cannabis sativa on the Sperm Epigenome

We have a rudimentary understanding of the consequences of preconception exposure to cannabis. As the most commonly used illicit psychoactive drug, cannabis prevalence is rapidly increasing across the United States (US), and consumers are increasingly perceiving it as safe. Recreational cannabis use is especially common among American men, rendering the paternal preconception environment potentially vulnerable to deleterious effects. Parental cannabis use has been associated with adverse developmental outcomes in offspring, but little is known about how such phenotypes are transmitted. Gametic epigenetic changes – the collection of molecular modifications made to DNA and histone proteins that play a role in regulating gene activity – provide one potential explanation. In a pilot study, our group recently demonstrated that cannabis use in humans, and exposure to delta-9-tetrahydrocannabinol (THC, the main psychoactive component of cannabis) in rats, is associated with significantly altered levels of DNA methylation in sperm. Epidemiological studies have further associated prenatal cannabis use with an increased risk of numerous teratologies including neurodevelopmental disorders and cardiovascular defects. While these studies illuminate the risks associated with cannabis and the prenatal environment, little attention has been paid to the effects of paternal preconception exposures alone on such congenital anomalies. There remains an urgent need to investigate the effects of cannabis on the sperm epigenome as use increases across the globe. Further, it is critical to investigate these effects on genes that are uniquely positioned to contribute to early development. The hypothesis of this research was that cannabis exposure is associated with heritable, but reversible, changes in DNA methylation in sperm at genes important for early life development. Broadly, the main objectives of this thesis research were to generate meaningful data to contribute to the gaps in knowledge of how cannabis can impact sperm DNA methylation. The results of this thesis research are described beginning in chapter 2. An initial pilot study from our group used reduced representation bisulfite sequencing (RRBS) to analyze methylation changes in sperm between cannabis users and controls. One gene identified as being significantly differentially methylated was an autism candidate gene, Discs-Large Associated Protein 2 (DLGAP2). Quantitative bisulfite pyrosequencing confirmed that an intronic region of DLGAP2 was significantly hypomethylated in the sperm of cannabis exposed men compared to controls. Use of human fetal brain tissues demonstrated that there is a significant, sex-specific, inverse relationship between DNA methylation and gene expression at this locus. A paternal rat model of THC exposure was used to determine whether these effects at Dlgap2 were heritable. Bisulfite pyrosequencing identified significant changes in rat sperm DNA methylation at Dlgap2, as well as significant losses of methylation at the same locus in F1 nucleus accumbens (NAc) tissues. Having demonstrated that autism candidate gene DLGAP2 shows functional changes in DNA methylation, the next question addressed was whether different routes of THC exposure, and exposure to different drugs, could similarly impact DNA methylation at a select group of neuroactive genes. Sperm DNA from rats exposed to THC or vehicle control via oral gavage underwent RRBS. Bisulfite pyrosequencing of sperm DNA from rats exposed to injected THC or vehicle control was performed to examine methylation at regions identified by RRBS. Sperm DNA from rats exposed to nicotine or vehicle control underwent pyrosequencing at the same regions. Lastly, two publicly available datasets were investigated to determine significant overlap between a known list of autism candidate genes and a list of genes with bivalent chromatin, a unique epigenetic feature. In the sperm of rats injected with THC and those nicotine-exposed, significant differential methylation at five of seven neurodevelopmentally active genes that were initially identified as significantly altered by oral gavage was identified. It was further discovered that autism candidate genes are significantly enriched for genes containing bivalent chromatin. Enrichment of both autism candidate genes and genes possessing bivalent chromatin was identified in the human RRBS dataset of genes significantly differentially methylated in sperm of cannabis users. These studies demonstrated THC and nicotine exposure in rats can impact DNA methylation in sperm at neuroactive genes. Further, this work provides initial evidence that genes with bivalent chromatin may be particularly vulnerable to DNA methylation changes resulting from environmental exposures. The fourth chapter of this thesis research employed a novel in vitro model of human spermatogenesis to identify the impact of exposure to a cannabis smoke extract (CSE) on DNA methylation at two groups of genes important for early life development. Human embryonic stem cells (hECS) exposed to CSE or vehicle control were differentiated into a mixed population of spermatogonial stem-like cells (SSCs), primary spermatocyte-like cells, secondary spermatocyte-like cells, and round haploid spermatid-like cells over a ten-day period. Following differentiation, flow cytometry was performed to isolate SSC-like cells and haploid spermatid-like cells for DNA methylation analyses. Methylation was first analyzed at a group of imprinted genes. Significant effects of exposure were identified in SSC-like cells at Sarcoglycan Epsilon (SGCE) and in haploid spermatid-like cells at Paternally Expressed 3 (PEG3) and Growth Factor Receptor-Bound Protein 10 (GRB10). Next, methylation was assessed at a group of genes randomly chosen genes from the Simons Foundation Autism Research Initiative (SFARI) autism candidate gene list, half of which possessed bivalent chromatin at the specific CpG sites analyzed and half of which did not. Significant methylation changes were identified in SSC-like cells at genes from the SFARI list possessing bivalent chromatin, but not at genes from the SFARI list without this epigenetic modification. These results support the hypothesis that bivalent chromatin may make genes more vulnerable to environmental exposures. Chapter five of this thesis research addressed the potential heritability of the impacts of exposure to CSE. Changes in F0 sperm DNA methylation were initially identified via whole genome bisulfite sequencing (WGBS) and methylation changes were validated at select genes via bisulfite pyrosequencing. Methylation changes validated in F0 sperm at the gene 2-Phosphoxylose Phosphatase 1 (Pxylp1) were similarly present in F1 sperm, while changes validated in F0 sperm at Gamma-Aminobutyric Acid Type A Receptor Subunit Beta2 (Gabrb2) and Metastasis Suppressor 1-Like Protein (Mtss1l) were similarly present in F1 hippocampal, and NAc and hippocampal tissues, respectively. Further, for Mtss1l a significant, sex-specific relationship between DNA methylation and gene expression in offspring NAc was demonstrated. Phenotypically, rats born to CSE-exposed fathers exhibited significant cardiomegaly relative to those whose fathers were CSE-naïve. Finally, chapter six of this thesis research addressed whether or not the effects of cannabis on human sperm DNA methylation were reversible. Men were recruited to participate in the study as cannabis users and non-user controls. Semen samples were collected at baseline, and then again following an 11-week cannabis-abstinence period. WGBS was performed on all sperm samples. There were no significant differences between users and controls based on demographic information or measured semen parameters. WGBS quantified DNA methylation changes in sperm. Importantly, a reduction in the magnitude of methylation difference between users and controls after the abstinence period relative to the methylation difference present before abstinence was observed. However, select genes retained their altered methylation patterns after abstinence, suggesting not all cannabis-induced effects were ameliorated. Bioinformatic analysis of genes associated with significantly differentially methylated CpG sites revealed terms associated with nervous system development, cardiovascular system development, and embryonic development. Together, this suggests that the abstinence period is at least partially effective at resolving the methylation changes observed following cannabis use at genes important for early development. This research adds to the emergent literature that cannabis is able to impact DNA methylation in sperm at genes important for early life development. It demonstrates in rodents the ability of this exposure to induce heritable epigenetic and phenotypic effects in offspring. Further, it provides the first evidence that abstinence from cannabis use might help resolve the methylation changes that arise in sperm following this exposure. Future work should assess the ability of this exposure to impact offspring methylation and neurodevelopmental outcomes in children and should define how long abstinence from cannabis use must last to produce the most robust amelioration effects.