Metabolomics signatures of air pollution mixtures and lung cancer risk: A large-scale metabolome-wide association study in the Cancer Prevention Study cohorts.

Abstract

10571 Background: Despite the well-established link between air pollution exposure and lung carcinogenesis, the underlying mechanism remains underexplored. While metabolomics has provided valuable molecular insights, existing air pollution metabolomics studies only assess biological changes to single air pollutants, without examining the join effects of air pollutant mixtures. We sought to identify metabolic signatures mediating the association between ambient air pollution mixtures and lung cancer risk. Methods: A total of 603 matched lung cancer case-control pairs were selected from the Cancer Prevention Study cohorts. Controls were matched to cases on age at blood draw, sex, race/ethnicity, and blood draw date. Plasma metabolomics was conducted using ultrahigh-performance liquid chromatography-tandem mass spectrometry. Air pollution mixtures include carbon monoxide, nitrogen dioxide, particulate matter, sulfur dioxide, and ozone, matched to subjects’ residential address at blood draw. We conducted an untargeted metabolome-wide association study with multivariable-adjusted linear and conditional logistic regression models to assess associations between metabolites and individual air pollutants and lung cancer risk, respectively, adjusting for individual and lifestyle-related covariates. We used a meet-in-the-middle (MITM) approach to identify potential mediators. Quantile g-computation was used to examine the effects of air pollution mixtures on metabolome perturbations to analyze the collective effects of multiple air pollutants concurrently. Results: Among 1,138 extracted metabolic features, 162 were significantly associated with the combined mixture (false discovery rate < 0.2). Using MITM, no metabolites were associated with both air pollution mixture and lung cancer risk at FDR < 0.2. At raw p < 0.05, 16 metabolites were identified as potential mediators of the air pollution mixture and lung cancer risk. These metabolites, including 4-hydroxyphenylacetylglutamine, glutamate, and 3-aminoisobutyrate, were also significantly associated with the mixture at FDR < 0.2. Conclusions: Findings from this analysis demonstrate the feasibility of considering multiple air pollutants as mixtures when conducting air pollution-related metabolomics investigations. Future exploration of the underlying biological mechanisms will be critical for validating pathways that participate in air pollution carcinogenicity. As lung cancer cases in never smokers continue to increase, it is pertinent to study air pollution and other environmental risk factors on cancer incidence.