Integrated network toxicology, machine learning and molecular docking reveal the mechanism of benzopyrene-induced periodontitis
Wenjie Wen, Li Rui, Pengpeng Zhuo, Chao Deng, Donglin Zhang
Abstract
BACKGROUND: Environmental pollutants, particularly from air pollution and tobacco smoke, have emerged as significant risk factors. Benzopyrene (BaP), a Group 1 carcinogen, is ubiquitously present in these pollutants, yet its molecular mechanisms in periodontitis remain largely unexplored. METHODS: We investigated these mechanisms through an integrated approach combining network toxicology, machine learning, and molecular docking analyses. Data from SwissTargetPrediction, CTD databases, and GEO datasets were analyzed to identify potential targets. Three machine learning algorithms (Support Vector Machine, Random Forest, and LASSO regression) were applied for core target identification, followed by Molecular docking analyses. RESULTS: We identified 11 potential targets associated with BaP-induced periodontitis, primarily involved in cellular response to lipopolysaccharide, endoplasmic reticulum function, and cytokine activity, particularly in IL-17 and TNF signaling pathways. Machine learning analysis identified three core targets: CXCL12, CYP24A1, and HMGCR. Molecular docking demonstrated strong binding affinities between BaP and these targets (binding energies <-5.0 kcal/mol). A diagnostic nomogram based on these core targets achieved high prediction accuracy (AUC = 0.922). CONCLUSIONS: This first comprehensive analysis of BaP-induced periodontitis using an integrated computational approach elucidates potential molecular mechanisms and identifies specific therapeutic targets. The diagnostic nomogram developed offers a promising tool for clinical periodontitis risk assessment, providing new perspectives on understanding the impact of environmental pollutants on periodontal health.