Assessing Toxicokinetic Uncertainty and Variability in Risk Prioritization.

Area of Science:

• Toxicology and Pharmacology
• Computational Biology
• Biochemistry

Background:

• High(er) throughput toxicokinetics (HTTK) uses in vitro data for in vitro-in vivo extrapolation (IVIVE).
• Previous HTTK studies focused on biological variability, neglecting measurement uncertainty's impact.
• Accurate IVIVE is crucial for comparing in vitro bioactivity to human exposure estimates.

Purpose of the Study:

• To develop Bayesian methods for estimating uncertainty in key HTTK parameters.
• To quantify the impact of measurement uncertainty on IVIVE.
• To expand chemical coverage for HTTK analyses.

Main Methods:

• Bayesian statistical methods were employed to estimate chemical-specific uncertainty for unbound fraction in plasma (fup) and intrinsic hepatic clearance (Clint).
• New experimental measurements for fup (418 chemicals) and Clint (467 chemicals) were generated.
• A revised fup protocol measured unbound chemical across varying protein concentrations, and Monte Carlo simulations propagated uncertainty.

Main Results:

• Bayesian methods provided chemical-specific uncertainty estimates for fup and Clint.
• New data increased HTTK chemical coverage for ToxCast libraries to 57%.
• The revised fup protocol significantly reduced unmeasurable results (44% to 9.1%) and uncertainty (median CV from 0.4 to 0.1).

Conclusions:

• The developed methods successfully quantify measurement uncertainty in HTTK parameters.
• Incorporating uncertainty into IVIVE provides more robust risk assessment.
• These advancements can enhance risk-based chemical prioritization by integrating in vitro data with exposure assessments.