Application of Open-Source PBPK Models in Rat-to-Human Pharmacokinetic Extrapolation of Oral Nicotine Exposure

Abstract Physiologically Based Pharmacokinetic (PBPK) models are often developed using animal data and applied to predict chemical movement and concentration in humans. However, differences in physiology and exposure routes between animal experiments and human exposures may impact the predictions and interpretation of PBPK model results. Data are needed to parameterize PBPK models, potentially requiring chemical-specific adjustment to model inputs. Since data may be limited for a chemical or exposure of interest, in silico approaches such as chemical structure-based modeling can help fill the gaps. This case study assesses generalized, open-source PBPK models for interspecies kinetic extrapolation of nicotine using both in vivo data from a rat oral gavage study and in silico predictions. Nicotine is used as a data-rich example chemical because PK data are available from different exposure routes in both humans and animals. Rat nicotine plasma data were obtained after oral gavage dosing of nicotine (up to 8 mg/kg/day over 7 days) and used to develop human nicotine models for both oral ingestion and buccal (mouth tissue) absorption. As an open-source buccal tissue absorption model was not available, we mimicked a buccal exposure by modifying the open-source model for the intravenous exposure route. An 8 mg/kg/day human dose using oral ingestion and buccal absorption resulted in an approximately 2- and 4-fold higher predicted maximum plasma concentration (Cmax) than an 8 mg/kg/day gavage exposure in rats, respectively, highlighting the impact of species and exposure route on model predictions. This study demonstrates that a generalized and open-source PBPK model can extrapolate the plasma kinetic profiles of nicotine between species using in vivo and in silico data after accounting for differences in exposure routes. In silico-informed model parameterizations provided similar results to rat in vivo-based parameterizations, highlighting the potential use of in silico approaches when data are limited.