Genomic and Epigenomic Studies Performed Using Next-Generation Sequencing: In collaboration with the Section of Human Neurogenetics (David Goldman), we have performed ChIP- and RNA-SEQ using archived brain from rhesus macaques reared under both normal and stressful conditions. While fewer than 1,000 rhesus SNPs were known, we identified more than 500,000 SNPs using this gene-centric approach. Non-synonymous SNPs that were similar to those in humans were identified (32 in number). Many of these are present in or around genes at which variation has been shown to predict differential sensitivity to various neurodevelopmental and/or psychiatric disorders. Most studies examining epigenetic effects of early stress in the brain have been performed in rodents. This is quite limiting, since many of the systems critical to stress responding differ between rodents and catarrhine primates (Old World monkeys, apes and humans). We have performed genome-wide examination of epigenetic effects and gene expression changes resulting from stress exposure. We performed Chromatin immunoprecipitation (ChIP) using an antibody directed toward H3K4me3, which marks active promoters, and identified gene regulatory regions at which H3K4me3 binding was altered as a function of early life stress exposure (266 in number). Among these were genes involved in neurodevelopment, and epigenetic/transcription regulation. Stress-induced epigenetic changes observed in early stress-exposed monkeys may represent predictive adaptive responses. Such responses are proposed to occur when an individual is exposed pre- or post-natally to challenge, as a means of increasing adaptability to an unpredictable environment. In this sense, the altered H3K4me3 binding that we observe among early stress-exposed monkeys could be considered epigenetically programmed environmental readiness, which could promote environmental reactivity. In the face of prolonged stress exposure or as part of the normative aging process, however, these changes could also potentially contribute to stress-induced pathology and allostatic load. Consistent with this hypothesis, when we examined effects of early stress on mRNA expression levels using RNA-SEQ, we observed differences in expression for genes involved in neuroplasticity, energy balance, and neuroprotection. The mRNA expression differences that we observe in early stress exposed macaques are those that, therefore, would be considered to be reflective of chronic stress-induced pathology, or allostatic load. Some of these changes overlap with those observed with chronic alcohol exposure as well. Cross-Species Genetic Variation and Stress-Induced Epigenetic Regulation of the BDNF Gene: Adaptations to unpredictable or stressful environmental conditions can occur at both the species and the individual levels. Identification of genes that are critical to stress adaptation may be achieved by searching for presence of functionally similar variants that occur in multiple species. Searching for genes at which there is both shared genetic variation and epigenetic regulation may be a particularly powerful approach for identifying genes that play critical roles in stress adaptation. There were several genes in which both conserved genetic variation and epigenetic regulation were present. Notable among these was the BDNF gene, which encodes a neurotrophin that is critical for neuronal survival, development and plasticity. We found a non-synonymous SNP in the pro-BDNF domain (Val46Met), and H3K4me3 promoter binding was increased in hippocampus of PR animals. We identified a non-synonymous BDNF SNP in rhesus that is similar in location and function to the human Val66Met SNP, which predicts individual differences in cognition and susceptibility to psychiatric disorders. We find that in hippocampus, a brain region important to environmental assessment, there is increased H3K4me3 binding in response to early stress. This could be considered a predictive adaptive response, as it would translate into increased readiness to express BDNF in response to environmental challenge. It has been proposed that the human loss-of-function 66Met allele has been maintained by balancing selection, and some studies indicate there to be recent positive selection for this and other BDNF variants. It may also be that stress-induced epigentic regulation of BDNF permits loss-of-function variation to be maintained in selected human populations. Because of the role of this growth factor in neuronal plasticity, we wanted to determine whether there were effects of rearing condition and BDNF genotype on early infant developmental trajectories. We confirmed the existence of this polymorphism (Val46Met) by dideoxy sequencing and performed analyses to determine whether genotype predicted results of Brazelton Neonatal Assessment, which we were able to do via our collaboration with Steve Suomi in the Laboratory of Comparative Ethology at NICHD. We found that early stress exposure produced effects Orientation, Motor Maturity, Activity and State Control and found both main and moderating effects of Val46Met genotype, with allelic effects being more pronounced among PR monkeys. Whether there are persistent effects of this variant on stress reactivity or other behavioral parameters that are predictors of alcohol problems is currently being tested. Cross-Species Analysis of the Oxytocin Receptor (OXTR) Gene: We have also been performing more refined, locus based searches for candidates that might be important to behavior, reward and stress reactivity. One system of potential relevance to the addictions via its effects on anxiety and reward is the oxytocin system. Using an informatics approach, we determined that there is evidence for selection at the Oxytocin receptor gene (OXTR). We found there to be cross-species conservation and, in mice and canids, evidence of purifying selection. These were largely restricted to the first intron and second exon, indicating that this is may be a particularly significant region. We also found that there is a selective sweep across the 10 kB region encoding the OXTR gene in humans and, among these SNPs were multiple non-synonymous SNPS in the second OXTR intron. Also indicating that this might be an interesting region is the fact that that there was a coding variant in the second exon that defines the Homo lineage (ie, is common to human and Neandertal). We identified 4 SNPs at the rhesus Oxytocin receptor gene (OXYR), all of which were located in the second exon. We are currently performing deep re-sequencing to determine if there are any other rare variants within this region. Oxytocin is a neuropeptide that produces affiliative, amnesic and anxiolytic affects. Given its roles in some of these processes, it has been proposed as a potential therapeutic agent for the treatment of anxiety and stress-related disorders, including PTSD. Functional variation that influences oxytocin system function may, therefore, be a particularly good candidate for performing gene x stress interaction studies. We found that within the 2.5 kB region encoding the first intron and second exon of the human OXTR gene, that there is an abundance of CpG islands and H3K4me3 and HeK4me1 binding peaks (both of which mark active promoters), indicating that this region is not only evolutionarily dynamic, but has potential for epigenetic regulation via multiple pathways. The potential for epigenetic regulation at the OXTR gene adds further support to our proposal that OXTR x stress interactions could be important in the etiology and/or pathogenesis of stress-related disorders, many of which are co-morbid with alcohol abuse or dependence.
Showing the most recent 10 out of 34 publications
