Exploring the association between per- and polyfluoroalkyl substances exposure and the risk of stroke : a systematic investigation using NHANES data analysis, network toxicology and molecular docking approaches

Abstract

Background: Epidemiologic evidence regarding the association between per- and polyfluoroalkyl substances (PFAS) exposure and stroke risk remains limited and inconclusive. Consequently, the current study sought to further examine this association and clarify the underlying molecular mechanisms. Materials and Methods: This cohort study analyzed data from 8081 participants of the 2003–2012 National Health and Nutrition Examination Survey (NHANES), employing multistage weighted logistic regression, weighted quantile sum (WQS) modeling, and partial least squares discriminant analysis (PLS-DA) to systematically evaluate the association between per- and polyfluoroalkyl substances (PFAS) exposure and stroke. Restricted cubic spline analysis was subsequently used to examine the nonlinear dose-response relationships. To investigate the underlying mechanisms, we integrated data from six databases (e.g., ChEMBL and GeneCards) to identify common molecular targets of PFAS and stroke. A protein-protein interaction (PPI) network was then constructed to identify core genes, while the binding interactions between PFAS and key targets were evaluated through molecular docking and dynamics simulations. Finally, functional enrichment analysis was performed on these core genes using the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. Results: After adjusting for potential confounders, six individual PFAS compounds, including perfluorooctane sulfonic acid (PFOS) (Odds Ratio [OR] = 1.59, 95% CI: 1.09–2.31), exhibited a significant positive association with the risk of stroke. The WQS model revealed a significant positive association for the PFAS mixture (OR = 1.027, 95% CI: 1.017–1.036), with PFOS contributing the highest weight (0.379). These findings were corroborated by the PLS-DA model, and the association remained significant in all subgroup analyses. The network toxicology analysis identified 183 common targets between PFOS and stroke, while the subsequent PPI network analysis identified six core genes, including AKT1 and HSP90AA1. GO and KEGG enrichment analyses demonstrated that these targets were markedly enriched in pathways associated with lipid and atherosclerosis metabolism, in addition to the PI3K-Akt and MAPK signaling pathways. Furthermore, molecular docking and molecular dynamics simulations supported potential interactions between PFOS and core targets such as AKT1. This suggests that PFOS may contribute to stroke pathogenesis by disrupting pathways involved in inflammatory regulation and apoptosis. Conclusions: This study identified a positive association between PFOS exposure and stroke risk, suggesting that the PI3K/AKT signaling pathway, along with its key effector molecule AKT1, may play a crucial role in mediating PFOS-induced stroke, thereby offering a theoretical foundation for the prevention and management of PFOS-associated stroke.