Network toxicology and machine learning reveal key molecular targets and pathways of mono-2-ethylhexyl phthalate-induced atherosclerosis

Mono-2-ethylhexyl phthalate (MEHP), a major metabolite of diethylhexyl phthalate, is increasingly recognized as an environmental contaminant with potential cardiovascular toxicity. However, the molecular mechanisms underlying MEHP-induced atherosclerosis (AS) remain poorly understood. This study aimed to investigate the toxicological targets and pathways through which MEHP contributes to AS development using network toxicology approaches. DEGs associated with AS were identified from the GSE100927 dataset. MEHP targets were predicted using multiple databases including SEA, SwissTargetPrediction, and TargetNet. Common toxicological targets were identified through intersection analysis. Functional enrichment analysis, GSVA, and ssGSEA were performed. Machine learning algorithms including LASSO regression, RF, and SVM were employed to identify key targets. A nomogram model was constructed for AS risk prediction, and molecular docking analysis was conducted to validate protein-ligand interactions. Analysis identified 13,905 DEGs in AS, with 172 potential MEHP targets yielding 92 common toxicological targets. Enrichment analysis revealed involvement in calcium signaling, PPAR signaling, inflammatory response, and immune pathways. Machine learning identified three key targets: PDPK1, HDAC10, and HRH1. The nomogram model based on HDAC10 and HRH1 demonstrated excellent predictive performance. ssGSEA analysis revealed significant associations between key targets and immune pathways. Molecular docking confirmed strong binding affinities, with HRH1-MEHP showing the highest affinity. MEHP may promote AS via coordinated effects on calcium handling, receptor-mediated and transcriptional signaling, oxidative stress, apoptosis, and immune activation. HRH1, HDAC10, and PDPK1 emerge as mechanistic mediators and potential biomarkers, with an HDAC10/HRH1 nomogram offering translational utility for AS risk stratification; docking results provide testable hypotheses for mechanistic validation.