Understanding Peripheral Biological Variability in Posttraumatic Stress Disorder: Omics and Causal Approaches

Abstract

Posttraumatic stress disorder (PTSD) is the quintessential trauma-related psychiatric disorder that profoundly impacts long-term health. A growing body of literature links PTSD to elevated risks for cardiometabolic diseases and premature mortality, with stress-activated biological processes in both central and peripheral systems discussed as potential mechanisms. Recent studies assessing associations between PTSD and biological differences in markers obtained from peripheral tissues (e.g., inflammatory markers) have highlighted promising directions for improving our understanding of PTSD pathophysiology and health consequences. However, current knowledge remains limited regarding the relevance of different types of biological data and the utility of various methodological approaches. This dissertation examines links between PTSD symptoms and three types of biological data derived from peripheral blood samples: metabolomics, DNA methylation (DNAm), and triglycerides. Applying novel approaches in our investigations, we aimed to improve the characterization of PTSD-related biological differences and explore their implications for long-term health. In Chapter 1, we pursued the first large-scale, population-based metabolome-wide investigation of PTSD, using plasma metabolomics data from three subsamples (total N=2,835 women) nested within the Nurses’ Health Study II (NHSII). Building on prior metabolomic analyses of depression and anxiety, our results revealed shared and distinct metabolic alterations linked to PTSD versus depression or anxiety. Notably, we identified 29 metabolites (primarily elevated glycerophospholipids and glycerolipids) associated with persistent PTSD symptoms but not with remitted PTSD. Our findings highlight the considerable metabolic impact of chronic, non-remitted PTSD symptoms and provide insights into mechanisms that may underlie elevated cardiometabolic disease risk in individuals experiencing PTSD symptoms. In Chapter 2, we offered a proof-of-principle for using DNA methylation data from peripheral blood samples to identify subtypes of trauma-exposed civilians with meaningful differences in PTSD symptoms and health. Exploratory analyses were performed in a subsample of women in NHSII (N=707), with external validation in the Detroit Neighborhood Health Study (N=324), a longitudinal cohort of predominantly non-Hispanic African American men and women. While specific subtype classifications remain preliminary, our study demonstrated that peripheral DNAm profiles may capture both short-term and long-term information about PTSD symptom severity and physical health. This work supports the utility of a multivariate approach linking biological variation to symptom and health heterogeneity. In Chapter 3, we examined the bidirectional relationship between PTSD and circulating triglycerides, triangulating evidence from a summary data-based two-sample Mendelian randomization study and observational analyses in NHSII. Our findings supported an effect of PTSD on elevated triglyceride levels, whereas the impact of triglycerides on PTSD risk appeared limited. By carefully considering potential biases in each design and performing a series of sensitivity analyses to address these concerns, this study improved our understanding of the causal relationship between PTSD and triglycerides. Taken together, findings from these three studies highlight the importance of understanding peripheral biological differences linked to trauma and PTSD. Both broad, hypothesis-generating strategies and focused, hypothesis-testing investigations can offer valuable insights. A key direction for future work is to assess the extent to which these biological differences may mediate long-term health risks related to PTSD.