Experimental exposure assessment of designed chemical mixtures in cell-based in vitro bioassays

Abstract

Cell-based bioassays are useful tools for the effect assessment of complex mixtures, but so far exposure assessment has not been performed for mixtures of chemicals. In the present study, cytotoxicity and activation of oxidative stress response were measured for three designed chemical mixtures with up to twelve components. The measurements of biological responses were complemented by concentration measurements using solid-phase microextraction to derive the freely dissolved concentrations of the mixtures (Cfree,mix). The tested mixtures showed slightly higher cytotoxic effects than predicted by the concentration addition model. Nominal and freely dissolved effect concentrations of the mixtures were very similar (within a factor of 1.5), but nominal concentrations (Cnom) and Cfree of the individual mixture components were only similar for the hydrophilic chemicals (e.g., caffeine, coumarin, lamotrigine). For hydrophobic (e.g., fluoranthene) and acidic chemicals (e.g., diclofenac, naproxen) Cfree was up to 648 times lower than Cnom. Chemicals were dosed in equipotent nominal concentration ratios and therefore contributed equally to the detected effects. Hydrophilic chemicals with low potency dominated Cnom,mix (up to 95%) and Cfree,mix (up to 99%). Several mixture components (e.g., diclofenac, ibuprofen, naproxen and warfarin) showed increasing free fractions with increasing Cnom,mix and therefore also a concentration-dependent contribution to Cfree,mix. Based on the findings of this study, we concluded that Cnom,mix will be sufficient for evaluating the toxicity of mixtures that contain chemicals with diverse physicochemical properties at low concentration levels. In contrast, for risk assessment purposes and quantitative in vitro to in vivo extrapolations, Cfree,mix is a better parameter because the in vitro responses can be related to freely dissolved concentrations in human plasma.