Epigenetic Biomarkers of Long-term Psychosocial Stress and their Relationships with Cardiometabolic Risk Factors

Long-term exposure to psychosocial stress has been linked to an array of negative health outcomes, such as cardiovascular disease and its cardiometabolic risk factors, which include hypertension, dyslipidemia, obesity, and diabetes. Recently, epidemiological studies have begun to assess the complex relationship between psychosocial stress and cardiometabolic risk, as well as the impact of epigenomic, sociodemographic, and biological factors on this association. Nevertheless, the underlying mechanisms through which epigenomic biomarkers, including individual DNA methylation sites and epigenetic clocks, influence the effect of stress on cardiometabolic risk are not fully understood. Thus, this dissertation aims to characterize the biological mechanisms through which long-term psychosocial stress impacts cardiometabolic risk factors in a multi-ancestry population of older adults from the Health and Retirement Study (HRS). In Aim 1, we conducted an epigenome-wide association study (EWAS) to investigate the relationship between long-term psychosocial stress and DNA methylation. For identified CpG sites, we performed formal mediation testing to examine whether smoking, alcohol use, physical activity, and body mass index (BMI) mediated the relationship between stress and DNA methylation. Nine CpG sites were associated with psychosocial stress (all p<9E-07; FDR q<0.10). Additionally, health behaviors and/or BMI mediated 9.4% to 21.8% of the relationship between stress and DNA methylation at eight of the nine CpGs. Several of the identified CpGs xvii were in or near genes associated with cardiometabolic traits, psychosocial disorders, inflammation, and smoking. In Aim 2, we conducted epigenome-wide mediation analyses using the high-dimensional multiple testing (HDMT) method to examine whether DNA methylation mediates the relationship between psychosocial stress and cardiometabolic risk factors. We found that DNA methylation partially mediated the associations between psychosocial stress and BMI, waist circumference (WC), high-density lipoprotein cholesterol (HDL-C) and C-reactive protein (CRP), with the overall mediation effects across probes explaining between 35.8% and 46.3% of these associations. A subset of the mediating CpG sites were associated with the expression of genes enriched in pathways related to cytokine binding and receptor activity, chemotaxis, and neuron development. Finally, in Aim 3, we examined whether psychosocial factors, including loneliness and psychosocial stress, and depressive symptoms were associated with epigenetic age acceleration. For identified associations, we performed formal mediation testing to characterize the role of depressive symptoms on the relationship between psychosocial factors and epigenetic aging. We discovered that psychosocial stress, loneliness, and depressive symptoms were each associated with at least one measure of epigenetic age acceleration (FDR q<0.05). Further, sex and educational attainment were found to modify the effects of both stress and loneliness on epigenetic aging, with females and individuals without a college degree appearing more sensitive to these effects. Depressive symptoms partially mediated the relationships between psychosocial factors and epigenetic age acceleration, with the mediation effects explaining between 24.0% and 39.9% of these associations.
Together, these studies help to enhance our understanding of the biological pathways through which long-term psychosocial stress contributes to cardiometabolic disease risk in a diverse population of older adults. Further, results from this dissertation may have a significant public health impact by highlighting the importance of individual- and community-level interventions to manage the root cause of stress before it results in complex physical and mental health conditions.