Genetic Studies of Salivary Cortisol Profiles and their Influence on Chronic Disease Risk Factors.

Abstract

A growing body of work has examined the contribution of stress to various health outcomes. The hormone cortisol is likely a key mediator of the stress response that influences multiple physiologic systems that are involved in common chronic disease (e.g. cardiovascular system, immune system, metabolism). An individual’s daily cortisol response (e.g. waking, peak, end of day) has been shown to be patterned by race/ethnicity as well as socioeconomic factors. Despite evidence of associations of various risk factors with cortisol levels, considerable intra- and inter-individual variability in cortisol remains unexplained. Alone or through interaction with environmental features, genetic factors could contribute to unexplained variability in cortisol concentrations or cortisol responsivity. Furthermore, genetic factors may influence how cortisol affects a wide range of anthropometric, metabolic, and inflammatory processes underlying chronic disease risk. In this dissertation, both a gene-based approach and a genome-wide association study (GWAS) were used to investigate genetic contributions to cortisol variability and its physiological effects in a sample of European Americans, African Americans, and Hispanic Americans from the Multi-Ethnic Study of Atherosclerosis (MESA). The sequence kernel association test (SKAT) was used for gene-based analysis. In the gene-based analysis of six stress response genes, we found that three genes had significant influence on cortisol features in at least one ethnic group. Only one gene, SLC6A4, had a significant effect across ethnic groups in meta-analysis (p < 0.05). Extending this work to an analysis of gene-by-cortisol interaction effects on BMI, glucose, and inflammatory factors, we used SKAT and its interaction based extension to identify four genes (ADRB2, NR3C1, NR3C2, SLC6A4) that have significant evidence of interaction with cortisol features to influence anthropometric, metabolic, and inflammatory markers (p < 0.05). In GWAS, we found 18 regions with p < 1x10-6 across the seven cortisol features evaluated in the three ethnic groups. Meta-analyses across ethnic groups identified only five genomic regions with p < 5x10-6; none of the GWAS results replicated in meta-analysis. Overall, this dissertation illustrates that genetic analyses across ethnic groups can provide new insights into the role of genes in cortisol features and their relationship with chronic disease risk factors.