Abstract EP33: Plasma Lipidomic Signatures And Potential Biomarkers For Risk Of Depression: A Longitudinal Study In A Large Sample Of Community-dwelling American Indians In The Strong Heart Study

Abstract

Background: Depression is an independent risk factor for CVD. Dyslipidemia, an established risk factor for CVD, has also been associated with depression, but a full spectrum of plasma lipidome for risk of depression is still lacking in any racial/ethnic groups. Objective: To identify lipids associated with risk of depression, independent of known clinical factors. Methods: We studied 2,498 fasting plasma samples from 1,249 American Indians attending two exams (2001-2003, 2006-2009, average 5-year apart) in the Strong Heart Study. Plasma lipids were repeatedly measured by LC-MS. Depressive symptoms were assessed using the Center for Epidemiologic Studies for Depression Scale (CES-D). Depression was defined as total CES-D score ≥ 16. All participants were free of overt depression at baseline. Generalized estimating equation was used to identify lipids associated with risk of depression in a discovery sample (N=875), followed by replication in an independent sample (N=374). Results from discovery and replication samples were combined by meta-analysis. The model adjusted for age, sex, study center, BMI, and presence/absence of chronic diseases (e.g., diabetes, hypertension, CVD, CKD). Longitudinal analysis was conducted by regressing changes in lipids on changes in the CES-D score and related psychometric measures (e.g., quality of life, perceived stress, and social support), adjusting for clinical factors and baseline lipids. Network analysis was performed to identify differential lipid networks associated with risk of depression. Multiple testing was controlled at FDR<0.05. Results: Of 1,249 non-depressive participants at baseline (mean age 41, 66% women), 201 individuals (141 in discovery, 60 in replication) developed incident depression after 5-year follow-up. Our lipidomics detected 1,542 lipids (518 known) in 14 lipid classes. Of the 518 known lipids, baseline levels of 9 lipids (e.g., PI(18:1/18:2), SM(d39:1), SM(d41:2)) were associated with a 28-40% decreased risk of depression in both discovery and replication samples. Meta-analysis identified 7 glycerophospholipids (e.g., PC(40:6), PE(16:0/22:5), PI(18:1/18:2)) and 8 sphingomyelins (e.g., SM(d36:3), SM(d40:2), SM(d41:2)) associated with incident depression. Longitudinal changes in long-chain unsaturated fatty acids (e.g., FA(16:1), FA(18:1)) were inversely, while changes in saturated acylcarnitines (e.g., AC(24:0), AC(26:0)) were positively, associated with changes in CES-D score. Network analysis identified two differential lipid modules associated with risk of depression. Conclusion: Novel plasma lipids and lipidomic signatures significantly predict risk of depression beyond clinical factors. Our findings shed light on the mechanisms through which dyslipidemia contributes to depression and provide instrumental data for risk prediction, stratification, and potential therapeutics.