Mechanistic insights into endocrine-disrupting chemicals in endometriosis and polycystic ovary syndrome: integrating network toxicology and molecular dynamics

Background: Endocrine-disrupting compounds (EDCs) are widely believed to play an important role in the occurrence and development of a variety of endocrine-related diseases, especially in common gynecological endocrine diseases such as endometriosis and polycystic ovary syndrome (PCOS). Although studies have shown that EDCs can affect reproductive health by interfering with hormone synthesis, metabolism, and signal transduction pathways, the specific molecular mechanisms and toxicological effects mediated by EDCs in the above diseases are still unclear. Methods: This study systematically assessed the potential effects of 16 common EDCs on endometriosis and PCOS by integrating multi-omics analysis, network toxicology, machine learning, molecular docking, and molecular dynamics simulations. Using the ChEMBL, STITCH, and SwissTargetPrediction databases, a total of 508 EDC-associated target genes were identified. Differential expression analysis and weighted gene co-expression network analysis based on GEO datasets yielded 122 disease-associated genes. Cross-sectional analysis further narrowed this list to 14 candidate genes, and five core targets were selected using support vector machines, least absolute shrinkage and selection operator, and random forest algorithms. Gene set enrichment analysis and Kyoto Encyclopedia of Genes and Genomes pathway analysis revealed relevant biological pathways, while the CIBERSORT algorithm was applied to evaluate associations between core genes and immune cell infiltration. The two genes with the highest diagnostic efficiency, PARP1 and CKS1B, were subjected to molecular docking with all 16 EDCs. Subsequently, CKS1B and PARP1 were further analyzed via 100-ns molecular dynamics simulations with the four compounds exhibiting the lowest binding energies to validate binding stability and elucidate their interaction mechanisms. Results: We identified 14 key genes potentially involved in EDC-induced endometriosis and PCOS, and five core targets were further prioritized using multiple machine learning algorithms. Functional enrichment analysis indicated that these genes were primarily associated with the mitotic spindle, transforming growth factor β (TGF-β) signaling pathway, and the Hedgehog signaling pathway in endometriosis, while in PCOS, they were mainly enriched in the Hedgehog signaling pathway. Immune infiltration analysis revealed significant correlations between these core genes and multiple immune cell types. Molecular docking demonstrated that benzo[a]pyrene, PFOA, PFOS, and DDT exhibited high binding affinity for CKS1B and PARP1. Subsequent molecular dynamics simulations further confirmed the binding stability and specificity of these complexes under simulated physiological conditions. Conclusion: This study systematically constructed a network linking EDCs with endometriosis and PCOS, elucidating their potential molecular mechanisms and toxicological pathways. Among the identified targets, CKS1B and PARP1 emerged as key genes potentially involved in disease pathogenesis, offering new insights into the molecular basis of EDC-induced reproductive disorders.