Combined transcriptome and metabolome approaches to characterize neurotoxicity

Abstract

In modern toxicology, the fusion of omics technologies with a high-throughput framework stands at the forefront of advancing our mechanistic understanding of toxicants. The TempOSeq technology (targeted transcriptomics) ensures such a high throughput and is already widely used for comparative toxicity studies. Despite its significant potential to unveil crucial information about the mode of action of compounds, metabolomics currently lacks throughput and has not been extensively utilized in toxicity experiments. This doctoral thesis exemplifies the transformative paradigm of dual omics studies in toxicology, demonstrated through two proof-of-concept toxicogenomics research papers and a broad review of how improved knowledge of chemical exposure can shape future concepts of safety research. In manuscript #1, we investigated the impact of proteasome inhibition (by MG-132) on human neurons (mature LUHMES cells), with focus on the associated metabolic and transcriptomic changes. Metabolic profiling was done over 12 h, in order to capture also short-lived responses, both early and late. This revealed an early transient accumulation of AMP, NADH and lactate, as well as down-regulation of glycolytic- and citric acid intermediates, altogether indicating an initial perturbation in energy metabolism. The transcriptome analysis indirectly confirmed the induction of proteostatic stress through the detection of cytoprotective counterregulations: up-regulation of proteasome subunits and molecular chaperones. The study suggests a co-occurrence of protective and maladaptive changes (e.g. GSH synthesis as a counter-regulation for increased oxidative stress) in response to proteasome dysfunction. In manuscript #2, we combined two omics approaches for the mechanistic understanding of 7 hits from an in vitro neurotoxicity screen. On the level of gene expression, the most common observed patterns were impaired cell differentiation and activation of stress-response pathways (ATF4, NRF2). Metabolomics profiling was performed for a subset of the test compounds, with a shared mode of action (c-I inhibitors MPP+, rotenone and berberine). This uncovered a complex metabolic signature, including increased levels of biomarkers associated with oxidative stress (5-oxo-proline, methionine sulfoxide) and neurodegeneration (saccharopine, aminoadipate, branched-chain ketoacids). The results of this study raise concerns about the potential neurotoxicity of berberine, a human drug and food supplement. In manuscript #3, we provide a framework for addressing uncertainties in toxicology, for a more reliable transfer of toxicological findings from model systems to humans. The aim is to move away from standardized safety factors towards quantitative factors, which can be modeled. This approach builds on the already existing concept of Gene x Environment interaction (G × E), which influences the toxicity outcome. We suggest that the originally rigid concept of "gene sequence" (G) should be expanded to include not only the primary DNA structure, but also the regulation of gene expression and function. Also, the concept of "environment" (E) needs some re-consideration in situations where exposure timing is pivotal. Finally, we suggest that prolonged or repeated exposure to the insult (chemical, physical, life style) affects gene expression, so that G and E are not always independent. The implementation of this modernized G × E concept should help improve the understanding of the selection and propagation of adverse outcome pathways (AOP) in different biological environments.