Identification of Toxicological Targets of Benzo[a]pyrene and Its Metabolite BPDE in Lung Adenocarcinoma Revealed by Integrated Network Toxicology, Machine Learning and Multi-Omics

Background/Objectives: Benzo[a]pyrene (BaP) is a ubiquitous polycyclic aromatic hydrocarbon generated by incomplete combustion and a well-established environmental carcinogen, with the lung as a primary target organ via inhalation. Its metabolite benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE) is highly genotoxic; however, the molecular mechanisms linking BaP/BPDE exposure to lung adenocarcinoma (LUAD) remain incompletely understood. This study aimed to identify BaP/BPDE-associated pathogenic targets in LUAD. Methods: We developed an integrative toxicology framework combining chemical–protein interaction prediction, transcriptomic comparisons between human LUAD and normal lung tissue, network toxicology, explainable machine learning, single-cell and spatial transcriptomics, and molecular docking. Results: Integration of chemical–protein interaction predictions with transcriptomic data identified 135 putative BaP/BPDE-related targets associated with LUAD. An interpretable machine learning approach prioritized 11 core genes, including 4 potential novel contributors: GABRA2, GSTA3, EDAR, and MAOB. Single-cell and spatial transcriptomic analyses revealed cell-type-specific and spatially heterogeneous expression of the identified genes in LUAD, including their preferential enrichment in epithelial cells and heterogeneous distribution across tumor, interface, and stromal regions. Molecular docking further suggested potential physical interactions between BaP/BPDE and candidate target proteins. Conclusions: This study provides an integrated view of BaP/BPDE-associated molecular alterations in LUAD and highlights candidate molecular targets for future experimental validation and potential chemoprevention in high-exposure populations.