UC Berkeley

Background

Obesity is a metabolic condition characterized by high levels of abdominal adiposity and currently affects approximately 35% of all US adults. Obesity is commonly measured using the body mass index (BMI), measured in kg/m2, with a level ≥ 30 indicating obesity. There are significant sex and racial/ethnic disparities in obesity; the highest levels of age-adjusted obesity are seen in non-Hispanic Black women (~58%) compared to the lowest levels seen, in non-Hispanic White women (~32%). Obesity is associated with many negative health outcomes, such as atherosclerotic cardiovascular disease (CVD) and various cancers (e.g. breast, colon, and endometrial), and has also been shown to be an independent risk factor for all-cause mortality.

Given the multifactorial etiology of obesity, examining multiple pathways that incorporate biological, behavioral, and environmental effects on weight gain may provide insights that lead to the prevention of obesity. Previous studies have demonstrated modest effects of psychosocial factors (e.g. job-related demands, relationship demands, and personal health problems) on the risk of obesity. Additionally, genome-wide association studies (GWAS) on obesity have identified up to 97 genetic risk variants that are associated with a high BMI. The effects of BMI-associated genetic risk variants detected thus far have been weak, with the strongest predictor, the fat mass and obesity gene (FTO), contributing a 0.39 point increase in BMI for each copy of the risk allele, explaining only ~0.34% of total genetic variance.

Two types of research studies that might help expand on the limited findings of previous research and help incorporate the effect of psychosocial factors and genetic factors together are gene by environment interaction studies and epigenetic studies (i.e. DNA methylation (DNAm) studies). Animal models have shown that acute stressors (e.g. maltreatment or maternal neglect) alter DNAm in mice and rats. Less is known about the interaction of psychosocial factors, genetic risk factors for obesity, and epigenetic regulation of those genetic risk factors in humans. However, genetic susceptibility to obesity, together with high levels of external stressors, may increase the risk of obesity and account for a previously unexplained proportion of the variance in obesity.

Methods

This dissertation uses two analytical approaches to investigate the relationship between psychosocial risk factors and genetic risk factors and obesity: a systematic review and secondary data analyses. Chapter 2 presents a detailed systematic review of available studies of gene-environment interaction studies and epigenetic studies that focus on the effects of psychosocial and genetic risk factors for obesity. Chapters 3 and 4 use secondary data from a large population-based longitudinal cohort study, known as the Multi-Ethnic Study of Atherosclerosis (MESA).

Chapter 3 explores the interaction of three psychosocial factors (i.e. chronic burden of stress, everyday hassles, and depression) and an obesity genetic risk score on obesity. Obesity genetic risk is derived from the most recent meta-analysis, which established 97 independent genetic variants associated with body mass index (BMI). This analysis was conducted using interval-censored survival modeling using a Weibull distribution. Both multiplicative and additive effects were determined, so as to give a comprehensive assessment of genetic and psychosocial interactions.

Chapter 4 investigates the effect of three psychosocial factors (i.e. chronic burden of stress, everyday hassles, and Cohen’s perceived stress) on DNAm levels of obesity risk genes. I employed a two-level model for each unique gene (87 genes from 97 obesity GWAS SNPs), treating the CpG as the level-1 unit and the individual as the level-2 unit. For genes associated with psychosocial variables, we assess cross-sectional associations between DNAm and genetic expression levels. An association between DNAm and expression demonstrates the functional importance of DNAm as a gene regulator.

Significance

This dissertation examines the interactive relationship between genetic and non-genetic factors as they relate to obesity. Previous findings for BMI-associated genetic risk factors have been relatively weak. The effects of psychosocial factors on obesity have been primarily examined through mechanisms that involve behavior change, such as altered diet patterns or altered physical activity. Assessing the epigenetic effects and gene-environment interactions of psychosocial factors and obesity genetic risk may reveal pathways through which people develop a greater risk for obesity.

My findings provide evidence of an interaction between psychosocial factors and genetic risk for obesity in multiple subpopulations of the MESA cohort study. I identify several BMI-associated genes that are differentially methylated by levels of chronic stress. DNAm is significantly associated with genetic expression, revealing a functional mechanism by which exogenous factors affect genetic expression, not directly attributed to inherited genomic sequences. These findings suggest potential underlying biological mechanisms whereby psychosocial factors and genetic risk factors interact to cause obesity in a manner that is not mediated by altered behavioral patterns of energy intake and expenditure. As inherited genomic sequences are not easily modifiable to prevent negative health conditions, it is important to establish multiple systems where we can prevent additional cases of obesity by targeting a modifiable risk factor (e.g. psychosocial stress) that interacts with genes or affects their expression. Because obesity remains a major concern in the United States, investigators should continue to search for mechanisms through which it can occur, in order to help reduce the burden of obesity.