Effective exposure of chemicals in in vitro cell systems: A review of chemical distribution models.

Nominal effect concentrations from in vitro toxicity assays may lead to inaccurate estimations of in vivo toxic doses because the nominal concentration poorly reflects the concentration at the molecular target in cells in vitro, which is responsible for initiating effects and can be referred to as the biologically effective dose. Chemicals can differentially distribute between in vitro assay compartments, including serum constituents in exposure medium, microtitre plate plastic, headspace and extracellular matrices. The partitioning of test chemicals to these extracellular compartments reduces the concentration at the molecular target. Free concentrations in medium and cell-associated concentrations are considered better proxies of the biologically effective dose. This paper reviews the mechanisms by which test chemicals distribute between in vitro assay compartments, and also lists the physicochemical properties driving the extent of this distribution. The mechanisms and physicochemical properties driving the distribution of test chemical in vitro help explain the makeup of mass balance models that estimate free concentrations and cell-associated concentrations in in vitro toxicity assays. A thorough understanding of the distribution processes and assumptions underlying these mass balance models helps define chemical and biological applicability domains of individual models, as well as provide a perspective on how to improve model predictivity and quantitative in vitro-in vivo extrapolations.

Dermal absorption study OECD TG 428 mass balance recommendations based on the EFSA database.

The European Food Safety Authority (EFSA) guidance (EFSA, 2017) for dermal absorption (DA) studies recommends stringent mass balance (MB) limits of 95-105%. EFSA suggested that test material can be lost after penetration and requires that for chemicals with <5% absorption the non-recovered material must be added to the absorbed dose if MB is <95%. This has huge consequences for low absorption pesticides. Indeed, one third of the MBs in the EFSA DA database are outside the refined criteria. This is also true for DA data generated by Cosmetics Europe (Gregoire et al., 2019), indicating that this criterion is often not achieved even when using highly standardized protocols. While EFSA hypothesizes that modern analytical and pipetting techniques would enable to achieve this criterion, no scientific basis was provided. We describe how protocol procedures impact MB and evaluate the EFSA DA database to demonstrate that MB is subject to random variation. Generic application of "the addition rule" skews the measured data and increases the DA estimate, which results in unnecessary risk assessment failure. In conclusion, "missing material" is just a random negative deviation to the nominal dose. We propose a data-driven MB criterion of 90-110%, fully in line with OECD recommendations.