OpenTox Virtual Conference 2022
The evaluation of genotoxicity by using novel prediction models based on biomarker genes in human HepaRG™ cells
List of Authors:
Anouck Thienpont1, Stefaan Verhulst2, Leo A. van Grunsven2, Vera Rogiers1*, Tamara Vanhaecke1* and Birgit Mertens3*
- Department of In Vitro Toxicology and Dermato-Cosmetology, Vrije Universiteit Brussel (VUB), Laarbeeklaan 103, 1090 Brussels, Belgium
- Liver Cell Biology research group, Vrije Universiteit Brussel (VUB), Laarbeeklaan 103, 1090 Brussels, Belgium
- Department of Chemical and Physical Health Risks, Sciensano, Juliette Wytsmanstraat 14, 1050 Brussels, Belgium
*Equally contributing last authors
Abstract:
Transcriptomics-based biomarkers are promising new approach methodologies (NAMs) to detect molecular events underlying the genotoxic mode of action of chemicals. Previously, we developed the GENOMARK biomarker, consisting of 84 genes selected from whole genomics DNA microarray profiles of 24 (non-)genotoxic reference chemicals covering different modes of action in metabolically competent human HepaRG™ cells. The biomarker was combined with a prediction model to classify the compounds as genotoxic or non-genotoxic using the expression data for the 84 genes. Recently, we created two new prediction models based on supervised machine learning algorithms (i.e. support vector machine (SVM) and random forest (RF)) taking into account an extended reference dataset of 38 chemicals. The novel models demonstrated a high performance for predicting misleading positive compounds based on qPCR data and for known (non-)genotoxic compounds based on an existing gene expression dataset generated with RNA-Seq. In addition, we also investigated the quantitative application of the GENOMARK gene expression data. More specifically, the potencies of two known in vivo genotoxic chemicals, aflatoxin B1 (AFB1) and ethyl methanesulfonate (EMS), were compared by using the benchmark dose (BMD) modelling approach. Briefly, three different batches of HepaRGTM cells were exposed to AFB1 or EMS in a concentration range around the IC10 and gene expression data were collected with qPCR. Next, the covariate approach with a benchmark response of 10% was applied on the GENOMARK genes with a fold change higher than either 1.5 or 2. Quantitative BMD modelling confirmed that AFB1 is more potent than EMS in HepaRGTM cells. Currently, data for additional compounds are being generated with TempO-Seq to explore the applicability of GENOMARK for quantitative genotoxicity assessment of data generated with high-throughput sequencing approaches. The preliminary results of this study together with our previous data suggest that GENOMARK does not only allow hazard identification but also provides quantitative genotoxicity information highlighting its potential use for risk assessment.