Identification of potential pathogenic genes for lung adenocarcinoma induced by 6PPD-quinone through network toxicology and SHAP analysis

Abstract

Lung cancer continues to maintain a high mortality rate, with lung adenocarcinoma (LUAD) being its most common type. As an emerging environmental pollutant, 6PPD-quinone is attracting increasing attention. Through network toxicology, we constructed a prognostic model and validated it with molecular docking and dynamics simulations, revealing its underlying molecular mechanisms. RNA-seq and clinical data of LUAD were obtained from the TCGA database, and differential expression analysis was performed using DESeq2. Putative targets of 6PPD-quinone were retrieved from DIGEP-Pred 2.0 and SEA databases, while LUAD-associated genes were collected from GeneCards, CTD, and OMIM. Overlapping genes with DEGs were identified as potential targets. Protein–protein interaction networks were constructed via STRING and visualized with Cytoscape. The hub genes were screened using four algorithms (Degree, EPC, MCC, MNC). GO and KEGG enrichment analyses were conducted on both potential and hub genes. A prognostic model was built based on Cox regression, with risk scores calculated for survival and independent prognostic analyses. SHAP analysis was applied to interpret feature contributions. Molecular docking and molecular dynamics simulations assessed binding stability between 6PPD-quinone and key genes, and qPCR validated expression changes in BEAS-2B and A549 cells with or without 6PPD-quinone exposure. 6PPD-quinone-linked LUAD were used to identify 31 potential targets and screen 18 network hub genes. Through a seven-gene prognostic signature (GPI, FAM83D, KRT17, CTLA4, MAP3K8, SIX1, ABCC2) construction, clear stratification into high- and low-risk groups was achieved with significantly different survival rates (1-year AUC = 0.711). SHAP analysis highlighted GPI, FAM83D, KRT17, and ABCC2 as major pathogenic contributors. Molecular docking demonstrated stable binding between 6PPD-quinone and GPI, FAM83D, and ABCC2 (binding energies < −5 kcal/mol). Furthermore, MD simulations confirmed the stability of the ABCC2-ligand complex. qPCR validation confirmed that exposure to 6PPD-quinone markedly enhanced the expression of these genes in A549 cells. This study identifies GPI, FAM83D, and ABCC2 as potential pathogenic genes linking 6PPD-quinone exposure to LUAD, possibly through genetic dysregulation. The integrative network toxicology and SHAP-based approach provide a robust framework for uncovering environmental carcinogenesis mechanisms, offering new insights into the health risks of tire-derived pollutants.