Unlocking the power of transcriptomic biomarkers in qualitative and quantitative genotoxicity assessment of chemicals

Traditionally, genotoxicity assessment starts with an in vitro test battery addressing both gene mutation and chromosome aberration. Despite the widespread use of this approach, a ‘standard battery’ is faced with several limitations including (i) the relatively low throughput, (ii) the high number of misleading positive results triggering unnecessary animal assays and (iii) the limited mechanistic information provided. Moreover, genotoxicity data are generally only used in a qualitative way, i.e. to classify chemicals as ‘genotoxic’ or ‘non-genotoxic’, and rarely for a quantitative analysis such as potency ranking or to derive a point of departure for risk assessment. To modernize genotoxicity assessment and reduce reliance on experimental animals, new approach methodologies (NAMs) providing dose-response data are needed. An interesting group of NAMs are transcriptomic signatures, or biomarkers, consisting of a subset of genes that robustly and consistently respond to chemicals from specific mechanistic classes. Previously, we developed a transcriptomic biomarker, GENOMARK, that consists of 84 genes to identify genotoxic substances in metabolically competent human HepaRG™ cells. The biomarker shows a very high prediction performance to classify genotoxicants based on gene expression data collected with microarray or RT-qPCR. Here, we additionally demonstrate how GENOMARK gene expression data can be used quantitatively, i.e. for potency ranking of genotoxic chemicals. Moreover, we compare its performance to that of TGx-DDI, another biomarker for genotoxicity. First, the cross-platform applicability of GENOMARK was investigated using TempO-Seq®, a high-throughput technique to generate gene expression data for either a large gene set (i.e. +S1500) or the entire genome. To this extent, HepaRGTM cells were exposed to 10 chemicals in an adequate concentration range and gene expression data for both the GENOMARK (84 genes) and the TGx-DDI (64 genes) biomarkers were extracted. Hazard classifications for genotoxicity were conducted using the GENOMARK and TGx-DDI prediction models. Potency ranking of the chemicals predicted as genotoxic was performed using benchmark dose (BMD) modeling in BMDExpress with a benchmark response value of 1 standard deviation. The results confirmed the applicability of GENOMARK on a high-throughput platform TempO-Seq® with a predictive accuracy of 100%. Interestingly, although consisting of a different subset of genes, both biomarkers demonstrated a high concordance (100%) in hazard classifications. In addition, the potency rankings of the genotoxicants were similar for both biomarkers. Overall, this work highlights the potential of transcriptomic biomarkers such as GENOMARK and TGx-DDI to facilitate a rapid and efficient human-relevant identification of genotoxicants while simultaneously providing information on potency to better inform risk assessment.