Abstract / Summary
This study employed an integrated in silico strategy combining network toxicology, molecular docking, molecular dynamics (MD) simulations, machine learning, and single‑cell RNA sequencing (scRNA‑seq) analysis to systematically investigate the molecular mechanisms through which per‑ and polyfluoroalkyl substances (PFAS) contribute to the pathogenesis of preeclampsia (PE). Network toxicology analysis identified 135 potential common targets of PFAS and PE, with ALB , IL6 , TNF , IL1B , and PPARG emerging as central hub genes. Functional enrichment analyses revealed that PFAS exposure is significantly associated with pathways involving nuclear receptor activity, monocarboxylic acid transport, inflammatory response, and lipid metabolism—particularly the PPAR signaling pathway, fatty acid degradation, and the IL‑17 signaling pathway. Molecular docking demonstrated strong binding affinities between four representative PFAS compounds (PFOS, PFOA, PFNA, PFHxS) and key targets such as TNF and ALB. MD simulations further indicated that PFHxS exhibited the highest conformational stability when bound to TNF, suggesting its potential for prolonged biological impact. Machine learning modeling identified a predictive gene signature (TRPM1, PALM, HIST1H4D, ANK1, RPP14) capable of distinguishing PE samples with an AUC of 0.789. Single‑cell transcriptomic analysis validated the heterogeneous expression of these biomarkers across placental cell types and revealed altered intercellular communication networks in PE. Collectively, this multi‑scale computational study elucidates how PFAS may disrupt placental function through synergistic modulation of inflammatory, lipid metabolic, and vascular regulatory pathways, providing novel insights into the environmental etiology of preeclampsia and highlighting potential biomarkers for risk stratification.