Nonlinear mixed-effects modeling of the intravenous and oral kinetics of eleven perfluoroalkyl substances in female mice

Restrictions on long-chain per- and polyfluoroalkyl substances (PFAS) have prompted the recourse to structural analogs. However, limited toxicokinetic (TK) data hinders the understanding of how structure influences persistence. We investigated the TK of 11 PFAS in mice, including perfluoroalkyl carboxylic acids (PFCA), perfluoroalkane sulfonic acids (PFSA), and ether derivatives of PFCA, GenX, and PFO2OA. Following intravenous and oral administration of PFAS mixture, plasma, urine and feces were collected over 119 days. Data were analyzed using nonlinear mixed-effects modeling, and allometry was used to predict human clearance from mouse data. Plasma clearances of long-chain PFCA and PFSA were at least 150 times lower than those of shorter-chain analogs. Ether linkages reduced clearance by about 25-fold for GenX and 4-fold for PFO2OA compared to PFHxA. Renal excretion accounted for 70%-100% of total clearance for short-chain PFAS (PFBS, PFBA, PFHxA) and ether derivatives, as well as PFOA and PFDS. PFHxS and PFOS showed balanced renal and fecal clearance, while long-chain PFCA (PFNA, PFDA) were primarily eliminated via feces. Deep tissue compartments, identified for PFHxA and PFO2OA, contributed to prolonged terminal half-lives. Mean residence time ranged from less than 1 day for PFBA, PFHxA, GenX, and PFO2OA to up to 68 days for other PFAS, with PFBS showing nonlinear disposition, characterized by rapid initial clearance followed by a sharp decline. The allometric approach reliably predicted human plasma clearance from mouse clearance. These findings showed that mechanistic TK modeling is valuable for accurately estimating clearance, providing critical insights to support risk assessment of PFAS.