Unraveling the association between triclosan and breast cancer: A multidimensional research strategy integrating machine learning, mendelian randomization, and molecular dynamics simulations.

OBJECTIVE: The association between triclosan and breast cancer remains controversial, with early studies suggesting antitumor effects and recent evidence indicating pro-carcinogenic risks, but the key targets and molecular mechanisms remain unclear. This study integrated computational toxicology, network toxicology, machine learning, Mendelian randomization, and in vitro experiments to determine the direction of triclosan's effect on breast cancer and systematically elucidate its core molecular mechanisms. METHODS: ProTox3 predicted carcinogenicity. Triclosan targets (SwissTargetPrediction/PharmMapper) were intersected with breast cancer DEGs (GSE42568/GSE70947) to build a PPI network. Three ML algorithms (RF, LASSO, SVM-RFE) with SHAP screened hub genes. Two-sample MR assessed causality; docking and 100 ns MD verified binding. MCF-7 cells were treated with triclosan to assess proliferation (CCK-8, EdU, Ki-67), colony formation, migration, cell cycle progression, and NR3C1 expression, with siRNA knockdown further confirming the mediation. RESULTS: ProTox3 predicted moderate endocrine-related carcinogenicity. Intersection of 2,732 DEGs with triclosan targets yielded 98 candidates. Five hub genes (ADH5, PDK2, NR3C1, PTGS2, FABP4) were identified; RF AUC = 0.939. MR confirmed NR3C1 inversely causal on breast cancer risk (OR = 0.730, 95% CI:0.619-0.861, P < 0.001), and high NR3C1 predicted longer survival (HR = 0.86, 95% CI:0.77-0.95, P = 0.0042). MD showed stable binding (free energy: -116.676 ± 1.895 kJ/mol). In vitro, triclosan promoted proliferation (CCK-8/EdU, P < 0.01), colony formation/migration (P < 0.01), increased S-phase (P < 0.05), and reduced NR3C1 (P < 0.05); knockdown enhanced this, confirming downregulation. CONCLUSION: Triclosan stably binds and suppresses NR3C1, removing NR3C1-dependent restraint on proliferation, colony formation, migration, and cell cycle progression. NR3C1 negatively correlates with risk and is a favorable prognostic marker. This work provides a multi-level evidence chain from binding to prognosis, clarifying triclosan's pro-cancer mechanisms and offering clues for risk evaluation and intervention of endocrine-disrupting chemical-associated breast cancer.