Evaluating neonatal cord serum metabolome in association with adolescent cardiometabolic risk factors

Background: Risk factors for cardiometabolic disease may have fetal origins, but the biological pathways linking gestational conditions to these risk factors are only partially understood. Methods: Among 145 Cincinnati-based HOME Study mother-child dyads, we detected 14,384 cord serum metabolic features using liquid chromatography high-resolution mass spectrometry. We measured cardiometabolic risk factors, including visceral fat, serum triglyceride, high-density lipoprotein cholesterol (HDL), leptin, adiponectin, insulin concentration, glucose, and systolic blood pressure (SBP) at age 12 years. Using sparse Partial Least Squares Regression (sPLS-R), we simultaneously modeled the association of metabolic features with all 8 risk factors. We prioritized features with the highest sPLS-R-derived CM risk factor correlations for metabolic pathway enrichment analysis. Results: We identified two groups of cardiometabolic risk factors in adolescents maximally associated with neonatal metabolic features. The first was visceral fat, triglycerides, HDL, insulin, and leptin; the second was glucose and SBP. The 178 metabolic features with the highest sPLS-R-derived feature-outcome correlations were enriched in 31 pathways related to short-chain fatty acid, vitamins C and B3, and amino acid metabolism, as well as glycolysis and gluconeogenesis. Conclusions: We identified 31 pathways that may help elucidate underlying mechanisms between fetal environmental stressors and the development of cardiometabolic risk factors. Impact: Using non-targeted metabolomics, we identified neonatal metabolic features linked to two groups of cardiometabolic risk factors in adolescents, suggesting distinct early-life CM risk trajectories and adolescent subphenotypes. One cardiometabolic group was characterized by higher visceral fat, triglycerides, insulin, leptin, as well as lower HDL; the other group was related to elevated glucose and systolic blood pressure. Using a variable selection and data-dimension reduction technique, these two groups were associated with 178 metabolic features and 31 biological pathways related to short-chain fatty acid, vitamins C and B3, and amino acid metabolism, as well as glycolysis and gluconeogenesis.