A Computational Workflow for Probabilistic Quantitative in Vitro to in Vivo Extrapolation

Area of Science:

• Toxicology and Pharmacology
• Computational Biology
• Pharmacokinetics

Background:

• Quantitative in vitro to in vivo extrapolation (QIVIVE) is crucial for risk assessment.
• Translating in vitro concentration-response to in vivo dose-response relationships requires robust methodologies.
• Deriving benchmark dose values (BMD) necessitates accurate extrapolation models.

Purpose of the Study:

• To develop a computational workflow for QIVIVE.
• To integrate physiologically based pharmacokinetic (PBPK) modeling with statistical methods for robust extrapolation.
• To quantify uncertainty in PBPK model parameters and structure during QIVIVE.

Main Methods:

• Integration of PBPK modeling, global sensitivity analysis (GSA), Approximate Bayesian Computation (ABC), and Markov Chain Monte Carlo (MCMC) simulation.
• Development of a statistical framework to accommodate parameter uncertainty, population variability, and model structural uncertainty.
• Application of the workflow to ethylene glycol monoethyl ether (EGME) and methoxyacetic acid (MAA) kinetics in rats.

Main Results:

• The workflow successfully translates in vitro concentration-response data to in vivo dose-response relationships.
• Posterior distributions of in vivo, population-based dose-response values are generated for a given route of exposure.
• Uncertainty from PBPK model parameters and structure is quantified within the posterior distribution of in vivo dose.

Conclusions:

• The developed computational workflow provides a rigorous statistical framework for QIVIVE.
• The method can be applied to extrapolate data across different organisms, including humans.
• Future development aims for a user-friendly, freely available modeling platform to simplify QIVIVE.