In vitro and computational approaches to predict developmental toxicity: Integrating PBPK models with cell-based assays

The field of chemical safety assessment is experiencing a transformative shift away from animal testing toward new approach methodologies (NAMs). While in vitro screening platforms for teratogenic potential have advanced significantly, their practical application in public health risk assessment has been limited by the lack of quantitative frameworks connecting in vitro results to human exposure levels. This study investigated whether prenatal developmental toxicity assays could quantitatively predict developmental toxicity thresholds for nine known reproductive toxicants (bisphenol A, caffeine, carbamazepine, carbaryl, deltamethrin, ethanol, nicotine, thalidomide, and toluene) and one non-teratogen (acetaminophen). We employed a three-tiered physiologically based pharmacokinetic (PBPK) model to translate in vitro concentrations from ReproTracker, Stemina DevTOX quickPredict™, and developmental neurotoxicity (DNT) battery assays into human equivalent doses (HEDs). These in vitro-derived HEDs were then compared against HEDs based on rodent developmental studies' lowest observed effect levels (LOAELs), as well as human clinical doses and exposure predictions. Our findings demonstrate that the integration of PBPK modeling with Reprotracker, Stemina DevTOX quickPredict™ and DNT assays advances the reduction of animal testing while enhancing human-relevant developmental toxicity assessment. Moving forward requires parallel improvements in our understanding of developmental toxicity mechanisms and computational methods that effectively translate in vitro results to human risk evaluation.