The role of metabolic perturbations in mediating the effects of ambient air pollution on lung cancer in the cancer prevention studies.

10541 Background: Exposure to ambient air pollution is an established risk factor for lung cancer. Investigation on the underlying mechanisms of air pollution carcinogenicity remains challenging due to the complex composition of air pollution and a lack of sensitive biomarkers to accurately measure exposure and corresponding response. To address the knowledge gaps, we applied high-resolution metabolomics to identify metabolic signatures of exogenous air pollution exposures and endogenous processes involved in lung carcinogenesis. Methods: A total of 683 matched lung cancer and control pairs within the established Cancer Prevention Study-II (CPS-II) Nutrition and CPS-3 cohorts were included in this study. The plasma metabolome was profiled with ultrahigh-performance liquid chromatography-tandem mass spectrometry. Assessment of exposure to six ambient air pollutants, including carbon monoxide (CO), nitrogen dioxide (NO 2 ), particulate matter (PM 10 ), fine particulate matter (PM 2.5 ), sulfur dioxide (SO 2 ), and ozone (O 3 ), was conducted using spatiotemporally-resolved models based on residential address at blood draw and calculated based on yearly average levels. We conducted metabolome-wide association studies with multiple linear regression models to assess associations of ambient air pollution and lung cancer with a meet-in-the-middle approach. Models controlled for potential confounders and covariates including age at blood draw, sex, race, body mass index, alcohol consumption, smoking status, passive smoke exposure, vegetable and fruit intake, and educational level. Metabolites significant at the FDR < 0.2 level in the air pollution model were analyzed in the lung cancer model. High-dimensional mediation analysis was used as a secondary analysis to compare results from the meet-in-the-middle analysis. Results: Among 1,204 metabolic features extracted from the blood samples, seven features were significantly associated with air pollution exposure and lung cancer incidence at the FDR < 0.2 level in the meet-in-the-middle analysis. Six features were significant at the p < 0.01 level through high-dimensional mediation analysis. All confirmed metabolites are enriched within peptide, lipid, and amino acid pathways. Gamma-glutamylglutamine and gamma-glutamylmethionine were statistically significantly associated with CO, NO 2 , and PM 10 exposure and lung cancer incidence in both analyses. Conclusions: This is the largest prospective metabolomics study examining biological perturbations associated with air pollution exposure and lung cancer incidence. Peptide metabolism may play an important role in mediating the association between air pollution and lung cancer. Findings from this study will lay the foundation for future studies to further understand the biological mechanisms underlying the carcinogenicity of air pollution.