Abstract

Human health effects from chronic exposure to mixtures of pesticide residues are little investigated. We compared standard histopathology and serum biochemistry measures and multi-omics analyses in an in vivo subchronic toxicity test of a mixture of six pesticide active ingredients frequently detected in foodstuffs (azoxystrobin, boscalid, chlorpyrifos, glyphosate, imidacloprid and thiabendazole). Sprague-Dawley rats were administered with the pesticide mixture with each ingredient at its regulatory permitted acceptable daily intake. Analysis of water and feed consumption, body weight, histopathology and serum biochemistry showed little or no physiological effects from exposure to the pesticide mixture. In marked contrast, analysis of the host-gut microbiome axis using serum and caecum metabolomics revealed that nicotinamide and tryptophan metabolism were affected, which suggested the initiation of a cell danger response, including adaptation to oxidative stress. Only limited effects were detected on the caecum microbiota by shotgun metagenomics. Further analyses of in vitro bacterial cultures showed that growth of Lactobacillus rhamnosus and Escherichia coli strains was negatively impacted by the pesticide mixture at concentrations that were not inhibitory when exposure was to a single agent. Transcriptomics of the liver showed that 257 genes had their expression changed. Gene functions affected included those involved in the regulation of response to hormones and correlated with previously reported transcriptome changes following administration of nicotinamide. Genome-wide DNA methylation analysis of the same liver samples showed that 4255 CpG sites were differentially methylated (> 10% difference). Overall, we demonstrated that unlike standard blood biochemical and organ histological analysis, in-depth molecular profiling using a combination of high-throughput -omics methods in laboratory animals exposed to low concentrations of pesticides reveals metabolic effects on the gut-liver axis, which can potentially be used as biomarkers for the prediction of future negative health outcomes. Our data suggest that adoption of multi-omics as part of regulatory risk assessment procedures will result in more accurate outcome measures, with positive public health implications.