bioRxiv2024
Toxicokinetics describes the absorption, distribution, metabolism, and elimination of chemicals by the body. Predictions from toxicokinetic models provide key information for chemical risk assessment. Traditionally, these predictions extrapolate from experimental animal species data (for example, in rats) to humans. More recently, toxicokinetics has been used for extrapolation from in vitro "new approach methods (NAMs)" for toxicology to in vivo. Chemical-specific in vivo toxicokinetic data are often unavailable for the thousands of chemicals in commerce. Therefore, large amounts of in vitro data measuring chemical-specific toxicokinetics have been collected. These data enable "high-throughput toxicokinetic" or HTTK modeling. The httk R package provides a library of chemical-specific data from peer-reviewed HTTK studies. httk further provides a suite of tools for parameterizing and evaluating toxicokinetic models. httk uses the open-source language MCSim to describe models for compartmental and physiologically based toxicokinetics (PBTK), MCSim can convert the model descriptions into a high-speed C code script. New models are integrated into httk using the open-source package development functionality in R, a model documentation file (R script), and the HTTK model description code file (C script). In addition to HTTK models, httk provides a series of functionalities such as unit conversion, model parameterization, Monte Carlo simulations for uncertainty propagation and biological variability, in vivo-derived data for evaluating model predictions, and other model utility functions. Here, we describe in detail how to add new HTTK models to httk and take advantage of the pre-existing data and functionality in the package. As a demonstration, we describe the integration of the gas inhalation PBTK model into httk. Modern modeling approaches, as exemplified by httk, allow for clear communication, reproducibility, and public scrutiny. The intention of httk is to provide a transparent, open-source tool for toxicokinetics, bioinformatics, and public health risk assessment. Author SummaryWe describe the integration and evaluation of new physiologically based toxicokinetic (PBTK) models into an open-source R package. Adding a new model to the R package allows a modeler to use the existing tools and data for in vitro to in vivo extrapolation (IVIVE). Integration with the R statistical analysis environment further allows model assessment. This workflow is designed to create a more transparent and reproducible approach to toxicokinetic models developed for various exposure scenarios. Here, we demonstrate the model integration and evaluation workflow with an inhalation model. Additionally, we provide an evaluation of the overall package performance as new models, data, and functionality are added over time. Our results show that transparent development of models, and use of existing data within the open-source R package format, allows for improvement of in vitro to in vivo extrapolation estimations. IVIVE is vital for advancement of 21st century human health risk assessment.
pharmacology and toxicology↗