Machine Learning-Enabled Network Biology Analysis for Biomarker Discovery: Uncovering Molecular Interactions between Endocrine-Disrupting Chemicals and Hormone-Sensitive Cancers

Endocrine-disrupting chemicals (EDCs) are important factors in causing hormonal and metabolic disorders. This study analyzed the potential of these EDCs-inducing hormone-sensitive cancers (HSCs), namely, breast cancer (BC), ovarian cancer (OC), endometrial cancer (EC) and prostate cancer (PC) through in silico approaches. The network analysis found seven hub proteins (IL6, TP53, EGFR, TNF, JUN, IL1B and PTGS2) and four clusters (C1, C2, C3 and C4). Enrichment analysis between the clusters and the individual EDCs found that most pathways were related to chemical carcinogenesis, metabolic regulation and signaling pathways. Further IL6, PTGS2 and TNF were found to be more connected within the network, and expression analysis found IL6 is differentially expressed (DE) in BC and OC; PTGS2 was DE in BC and PC; TNF was DE in BC, OC and EC. Additionally, these genes had a significant number of correlated genes. The toxicokinetic analysis found seven EDCs with systemic toxic availability and carcinogenicity. While these EDCs have been identified as potential carcinogens, their specific mechanisms in inducing HSCs remained unclear. This study attempted to elucidate these underlying mechanisms. The final validation derived from comparing gene ontologies of the EDCs with the clusters found pathways associated with carcinogenesis, cancer progression and metabolic regulation. The study reveals the potential molecular mechanisms involved between the EDCs and the proteins IL6, PTGS2 and TNF in causing cancer and suggests that avoiding products containing EDCs is important.