Variation in predicted internal concentrations in relation to PBPK model complexity for rainbow trout.

Area of Science:

• Environmental toxicology
• Pharmacokinetics
• Computational chemistry

Background:

• Environmental toxicity assessment traditionally uses exposure concentrations.
• There is a growing need to incorporate internal tissue partitioning for more accurate assessments.
• Physiologically based pharmacokinetic (PBPK) models offer a framework for this internal partitioning analysis.

Purpose of the Study:

• To evaluate the accuracy of PBPK models based on different tissue constituents and chemical descriptors.
• To compare the performance of a simple "Kow-lipid content" PBPK model with a more elaborated model.
• To identify factors influencing chemical accumulation in tissues.

Main Methods:

• Utilized data from 150 experiments for 28 chemicals from US EPA databases.
• Developed and compared two PBPK models: a simple "Kow-lipid content" model and an elaborated model with five sorptive phases and LSER parameters.
• Analyzed chemical distribution tendencies using molecular descriptors.

Main Results:

• The elaborated PBPK model reduced prediction errors (differences between predicted and measured values > 1 order of magnitude) to 7% of experiments, compared to 37% for the simple model.
• Polar tissue constituents (water, polar lipids, proteins) are crucial for the accumulation of polar compounds with H-bond donating groups.
• Storage lipids and water govern the accumulation of non-polar compounds based on their polarity.

Conclusions:

• Elaborated PBPK models provide significantly higher accuracy in environmental toxicity assessments.
• Further improvements can be achieved by incorporating metabolic biotransformation and membrane transport.
• Understanding tissue partitioning behavior is key for accurate chemical risk assessment.