Mode of Action Assignment of Chemicals Using Toxicogenomics: A Case Study with Oxidative Uncouplers

Abstract

A criterion frequently used to group chemicals in risk assessment is "mode of toxic action" (MoA). Routinely, structure-based approaches are used for the MoA categorization of chemicals, but they can produce conflicting results or fail to classify compounds. Biological activity-based approaches such as toxicogenomics which provide an unbiased overview of the transcriptomic changes after exposure to a compound may complement structure-based approaches in MoA assignment. Here, we investigate whether toxicogenomic profiles as generated after in vitro exposure of an established cell line (C3A hepatoma cells) are able to group together chemicals with an uncoupling MoA, and to distinguish the uncouplers from chemicals with other MoAs. In a first step, we examined whether chemicals sharing the same uncoupling of oxidative phosphorylation (OXPHOS) MoA produce similar toxicogenomic profiles and can be grouped together. In a next step, we tested whether the toxicogenomic profiles discriminate between OXPHOS and chemicals displaying a (polar) narcotic MoA. Experimentally, cells were exposed in vitro to equipotent concentrations of the test compounds and gene expression profiles were measured. The resulting toxicogenomic profiles assigned OXPHOS to one cluster and discriminated between the OXPHOS and the (polar) narcotics. In addition, the toxicogenomics data revealed that one and the same chemical can display multiple MoAs, which may help to explain conflicting results of MoA classification from structure-based approaches. The results strongly suggest the feasibility of MoA grouping of chemicals by using in vitro cell assay-based toxicogenomic profiles.

AOP Report: Uncoupling of Oxidative Phosphorylation Leading to Growth Inhibition via Decreased Cell Proliferation

“…Classical uncouplers, such as 2,4-DNP, carbonyl cyanide-p-trifluoromethoxyphenyl hydrazone, pentachlorophenol, and 3,5-di-t-butyl-4-hydroxybenzylidinemalononitrile have been widely used as positive controls in (eco)toxicological tests (see Schultz et al, 2002 ), whereas the hazards of “new” uncouplers, such as triclosan, emodin, and metabolites of polybrominated diphenyl ethers (PBDEs) are under extensive assessments (see Sugiyama et al, 2019 ). Predictive approaches such as quantitative structure–activity relationships (QSARs; see Dreier et al, 2019a ), biophysical models (see Ebert & Goss, 0014), and classification approaches such as toxicogenomics (see Hawliczek-Ignarski et al, 2017 ) have been developed to identify new organic uncouplers. There are also endogenous uncouplers for regulating energy homeostasis and body temperature, such as uncoupling proteins (reviewed in Rousset et al, 2004 ) and fatty acids (reviewed in Kadenbach, 2003 ).…”

In vitro or not in vitro: a short journey through a long history

“…The structure-based approaches for MOA classification, although being highly useful, also have inherent limitations. Therefore, it has been suggested to enhance the structure-based approaches by biological response profiles, with in vitro assays being the appropriate technological tool to provide the response data [ 46 , 75 , 115 , 123 ].…”

A Qualitative Modeling Approach for Whole Genome Prediction Using High-Throughput Toxicogenomics Data and Pathway-Based Validation

“…Different suites of genes that serve as transcriptional biomarkers of genotoxicity have been identified ( Amundson et al, 2001 ; Iida et al, 2003 ; Dickinson et al, 2004 ; Ellinger-Ziegelbauer et al, 2008 ; Boehme et al, 2011 ; Li et al, 2015 ). Toxicology is now moving toward use of higher-throughput in vitro methods as a basis for screening compounds for subsequent testing and these screening analyses for transcriptomic changes are playing an increasingly prominent role in early stages of testing ( Li et al, 2012 ; Hawliczek-Ignarski et al, 2017 ).…”