A targeted transcriptomics approach for the determination of mixture effects of pesticides

While real-life exposure occurs to complex chemical mixtures, toxicological risk assessment mostly focuses on individual compounds. There is an increasing demand for in vitro tools and strategies for mixture toxicity analysis. Based on a previously established set of hepatotoxicity marker genes, we analyzed mixture effects of non-cytotoxic concentrations of different pesticides in exposure-relevant binary mixtures in human HepaRG hepatocarcinoma cells using targeted transcriptomics. An approach for mixture analysis at the level of a complex endpoint such as a transcript pattern is presented, including mixture design based on relative transcriptomic potencies and similarities. From a mechanistic point of view, goal of the study was to evaluate combinations of chemicals with varying degrees of similarity in order to determine whether differences in mechanisms of action lead to different mixtures effects. Using a model deviation ratio-based approach for assessing mixture effects, it was revealed that most data points are consistent with the assumption of dose addition. A tendency for synergistic effects was only observed at high concentrations of some combinations of the test compounds azoxystrobin, cyproconazole, difenoconazole, propiconazole and thiacloprid, which may not be representative of human real-life exposure. In summary, the findings of our study suggest that, for the pesticide mixtures investigated, risk assessment based on the general assumption of dose addition can be considered sufficiently protective for consumers. The way of data analysis presented in this paper can pave the way for a more comprehensive use of multi-gene expression data in experimental studies related to mixture toxicity.