Use of in vitro metabolomics in NRK cells to help predicting nephrotoxicity and differentiating the MoA of nephrotoxicants.

We describe a strategy using an in vitro metabolomics assay with tubular rat NRK-52E cells to investigate the Modes of Action (MoAs) of nephrotoxic compounds. Chemicals were selected according to their MoAs based on literature information: acetaminophen, 4-aminophenol and S-(trichlorovinyl-)L-cysteine (TCVC), (covalent protein binding); gentamycin, vancomycin, polymycin B and CdCl2 (lysosomal overload) and tenofovir and cidofovir (mitochondrial DNA-interaction). After treatment and harvesting of the cells, intracellular endogenous metabolites were quantified relative to vehicle control. Metabolite patterns were evaluated in a purely data-driven pattern generation process excluding published information. This strategy confirmed the assignment of the chemicals to the respective MoA except for TCVC and CdCl2. Finally, TCVC was defined as unidentified and CdCl2 was reclassified to the MoA "covalent protein binding". Hierarchical cluster analysis of 58 distinct metabolites from the patterns enabled a clear visual separation of chemicals in each MoA. The assay reproducibility was very good and metabolic responses were consistent. These results support the use of metabolome analysis in NRK-52E cells as a suitable tool for understanding and investigating the MoA of nephrotoxicants. This assay could enable the early identification of nephrotoxic compounds and finally reduce animal testing.

In vitro-to-in vivo extrapolation (IVIVE) by PBTK modeling for animal-free risk assessment approaches of potential endocrine-disrupting compounds.

While in vitro testing is used to identify hazards of chemicals, nominal in vitro assay concentrations may misrepresent potential in vivo effects and do not provide dose-response data which can be used for a risk assessment. We used reverse dosimetry to compare in vitro effect concentrations-to-in vivo doses causing toxic effects related to endocrine disruption. Ten compounds (acetaminophen, bisphenol A, caffeine, 17α-ethinylestradiol, fenarimol, flutamide, genistein, ketoconazole, methyltestosterone, and trenbolone) have been tested in the yeast estrogen screening (YES) or yeast androgen-screening (YAS) assays for estrogen and androgen receptor binding, as well as the H295R assay (OECD test guideline no. 456) for potential interaction with steroidogenesis. With the assumption of comparable concentration-response ratios of these effects in the applied in vitro systems and the in vivo environment, the lowest observed effect concentrations from these assays were extrapolated to oral doses (LOELs) by reverse dosimetry. For extrapolation, an eight-compartment Physiologically Based Toxicokinetic (PBTK) rat model based on in vitro and in silico input data was used. The predicted LOEL was then compared to the LOEL actually observed in corresponding in vivo studies (YES/YAS assay versus uterotrophic or Hershberger assay and steroidogenesis assay versus pubertal assay or generation studies). This evaluation resulted in 6 out of 10 compounds for which the predicted LOELs were in the same order of magnitude as the actual in vivo LOELs. For four compounds, the predicted LOELs differed by more than tenfold from the actual in vivo LOELs. In conclusion, these data demonstrate the applicability of reverse dosimetry using a simple PBTK model to serve in vitro-in silico-based risk assessment, but also identified cases and test substance were the applied methods are insufficient.