Using High-Resolution Metabolomics for Identifying Metabolic Profiles Associated with Ambient Air Pollution in the Cancer Prevention Study-II Nutrition Cohort

Background. Mechanisms underlying ambient air pollution toxicity remains largely unknown despite the established links between air pollution and numerous adverse health effects. There is growing interest in applying high-resolution metabolomics (HRM) to study air pollution and health. Broad application of HRM is limited, however, due to concerns over statistical power and study generalizability. Methods. Using previously collected data, we conducted a cross-sectional metabolomics wide association study among 1,263 postmenopausal women within the American Cancer Society Cancer Prevention Study-II Nutrition Cohort. Modeled estimates of PM2.5 (1999-2004), NO2 (2006), and O3 (2002-2004) concentrations were linked to the participant residence at enrollment. Metabolic profiling on serum samples collected between 1998-2001 was measured by mass spectrometry-based platforms. We followed an untargeted metabolome-wide association study framework to evaluate metabolite and metabolic pathway perturbations associated with chronic exposures to air pollution. Results. We detected and analyzed 1,186 metabolites in serum samples, with an average technical intraclass correlation coefficient of 87% and coefficient of variance of 23%. Sixty-five unique metabolites were significantly associated with at least one air pollutant (false positive discovery rate <0.05), after controlling for covariates, including sociodemographic, smoking status, and year of blood draw. We observed perturbations in several inflammatory and oxidative stress related metabolic pathways associated with lipid, xenobiotic, amino acid, and vitamin metabolism. We subsequently confirmed the chemical identity of 43 unique metabolites with level 1 evidence, including diacylglycerol, taurine, docosapentaenoate, alpha- carboxyethylhydroxychroman, and palmitoyl ethanolamide. Conclusions. In this large cross-sectional analysis of air pollution and serum metabolites in postmenopausal women, we identified several novel metabolic markers and pathways associated with long-term exposures to air pollution and replicated others. These results support future development of sensitive metabolic markers of air exposures, as well as corresponding investigation into the specific molecular mechanisms and disease etiologies indicated by these markers. Keywords: Air Pollution, Metabolomics