Prenatal fipronil sulfone exposure is associated with altered fetal liver-function biomarkers: Evidence integrating epidemiology, network toxicology, and molecular docking.

Fipronil and its metabolites (FPM) are widely detected in environmental and biological matrices, but human evidence of developmental hepatotoxicity remains scarce. We quantified six fipronil-related compounds, five liver-function biomarkers, and four inflammatory cytokines in umbilical cord serum from 187 neonates. Fipronil sulfone was detected in all samples (median: 101.6 pg/mL). In covariate-adjusted linear models, each 10-fold increase in fipronil sulfone was associated with higher alanine aminotransferase (ALT) [β = 3.73, 95% CI: (1.58-5.88)], aspartate aminotransferase (AST) [β = 13.13, 95% CI: (2.15-24.12)], gamma-glutamyl transferase (GGT) [β = 51.50, 95% CI: (3.74-99.25)], and alkaline phosphatase (ALP) [β = 93.52, 95% CI: (44.61-142.44)]. The associations with ALT and ALP remained statistically significant after FDR correction. Bayesian kernel machine regression and machine-learning models consistently prioritized fipronil sulfone as an important mixture component associated with these biomarkers. Exploratory BBMD analysis estimated benchmark concentrations associated with a 10% change in liver-function biomarkers. Mediation analyses indicated that interleukin-1β partially mediated the associations with ALT and AST (16.25% and 19.75%, respectively), whereas tumor necrosis factor-α mediated the associations with ALT, AST, and ALP (11.21%-20.91%). Network toxicology highlighted inflammatory, oxidative-stress, PI3K-Akt, MAPK, and TNF signaling pathways, and molecular docking predicted favorable binding of fipronil sulfone to TNF, ALB, and INS. These findings associate cord-serum fipronil sulfone concentrations with altered cord-serum liver biomarkers and identify inflammatory signaling as a potential intermediate pathway. Longitudinal and experimental studies are needed to confirm temporality and biological causation.