Machine learning-driven prediction models and mechanistic insights into cardiovascular diseases: deciphering the environmental endocrine disruptors nexus

Cardiovascular disease (CVD) persists as the foremost cause of global mortality, yet the mechanistic links between environmental pollutants and CVD pathogenesis remain poorly defined. This study addresses this gap by integrating machine learning-driven epidemiology with computational biology to systematically evaluate the role of endocrine-disrupting chemicals (EDCs) in CVD development. We analyzed data from the NHANES cohort to identify CVD-associated EDCs using advanced predictive modeling. Molecular docking and dynamics simulations were employed to characterize interactions between prioritized compounds and the NOX2-p22phox complex, a key regulator of oxidative stress. Structural and functional impacts on NADPH oxidase activity were assessed through residue-level binding analysis and reactive oxygen species (ROS) quantification. Machine learning identified 3-hydroxyfluorene (3-HF) as a novel environmental risk factor for CVD. Molecular simulations revealed that 3-HF selectively binds to the transmembrane domain of the NOX2-p22phox complex, forming stable interactions with residues critical for structural integrity (e.g. T135, H160). These interactions destabilized the protein complex, impairing NADPH oxidase assembly and suppressing ROS generation. Further analysis demonstrated that 3-HF-mediated oxidative stress disruption correlates with vascular dysfunction pathways implicated in CVD progression. This study establishes 3-HF as a redox-disrupting environmental contaminant contributing to CVD through NOX2-p22phox targeting. By bridging population-level exposure data with atomic-scale mechanistic insights, our work provides a transformative framework for environmental health risk assessment and preventive intervention design.