Network toxicology and interpretable modeling identify nicotine-associated neuroendocrine targets and a three-gene prognostic signature in breast cancer.

BACKGROUND: Nicotine, a major tobacco metabolite and persistent environmental pollutant, is epidemiologically associated with breast cancer risk, yet its non-classical carcinogenic molecular mechanisms and prognostic value remain unclear. This study was designed to systematically elucidate the potential targets and molecular mechanisms underlying the prognostic impact of nicotine on breast cancer, alongside developing an interpretable prognostic risk model. METHODS: This study integrated network toxicology, interpretable modeling, and molecular docking. Potential nicotine targets were intersected with breast cancer differentially expressed genes, and hub genes were identified via protein-protein interaction (PPI) network. A prognostic model was built using Cox regression analysis and interpreted by the SHapley Additive exPlanations (SHAP). The binding potential of nicotine to targets was predicted through in silico molecular docking. RESULTS: Thirty-three nicotine-breast cancer common targets and 21 hub genes were identified. A three-gene prognostic model (TACR1, NPY1R, ADRB1) was constructed via Cox regression based on nicotine-related candidate targets, with the risk score (RS) formula: RS =-0.07 × NPY1R - 0.17 × ADRB1 - 0.19 × TACR1. This model could potentially stratify patients into high- and low-risk subgroups (P<0.001), and the RS may serve as an independent prognostic indicator. SHAP analysis identified TACR1 as the main contributor. Molecular docking computationally predicted favorable binding interactions between nicotine and the three candidate targets, and enrichment analysis pointed to G protein-coupled receptor (GPCR) and neuroactive ligand-receptor interaction pathways. CONCLUSIONS: This study computationally prioritized potential nicotine-associated breast cancer targets and constructed a three-gene prognostic model; in silico docking analysis suggested the potential binding of nicotine to these targets. We hypothesize that nicotine might function as a potential neuroendocrine disruptor to modulate breast cancer progression via GPCR signaling pathways. We propose a unique nicotine-associated three-GPCR-gene signature that provides new clues for environmental health risk assessment and prognostic biomarker development in breast cancer, yet further experimental validation is needed.