Exploring the toxicological impact of perfluorooctanoic acid and perfluorooctane sulfonate on glioblastoma through network toxicology, machine learning, and multi-dimensional bioinformatics analysis

Exposure to perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) has been associated with the development of various malignant tumors. However, their roles and molecular mechanisms in glioblastoma (GBM) are still unclear. This study combined network toxicology, machine learning, immune infiltration analysis, single-cell RNA sequencing (scRNA-seq), molecular docking, Mendelian randomization (MR), and molecular dynamics (MD) simulation to explore the potential toxicological targets and mechanisms of PFOA/PFOS in GBM. Five core target genes (ANXA5, AURKA, CDK2, EIF4EBP1, and ODC1) were identified. Their predictive potential was validated using three external independent datasets, with AUC values mostly above 0.90. Gene Set Enrichment Analysis (GSEA) revealed significant enrichment of the phosphatidylinositol, ErbB, and MAPK signaling pathways. Furthermore, the expression levels of core targets exhibited strong correlations with immune cell infiltration, particularly with macrophages and NK cells. ScRNA-seq analysis revealed that the core targets were predominantly expressed in MES‑like and AC‑like malignant cells, suggesting their potential roles in regulating the functional phenotypes of GBM cell subpopulations. Molecular docking confirmed the strong binding affinity of PFOA/PFOS with five core targets. MR analysis revealed a significant association between ODC1 expression and GBM risk (OR = 2.16, 95%CI: 1.129-4.115; P = 0.0198), while MD simulation further verified the sustained binding interactions between ODC1 and PFOA/PFOS. We also proposed a novel adverse outcome pathway (AOP) framework linking PFOA/PFOS exposure to GBM, offering critical toxicological insights. Overall, these findings provide valuable evidence for the potential toxicological impact of PFOA/PFOS on GBM, highlighting the necessity for further mechanistic investigations. • The first exploration of network toxicology, machine learning, and multidimensional bioinformatics analysis elucidated the toxicological effects of PFOA/PFOS on GBM. • Five core target genes (ANXA5, AURKA, CDK2, EIF4EBP1, ODC1) and three signalings pathways (phosphatidylinositol, ErbB, and MAPK) may play crucial roles in the development and malignant progression of GBM associated with PFOA/PFOS exposure • A novel adverse outcome pathway (AOP) framework was constructed to provide mechanistic insights linking PFOA/PFOS exposure to GBM pathogenesis.