Citation:

Cho, E., A. Rowan-Carroll, A. Williams, C. Corton, H. Li, A. Fornace, C. Hobbs, AND C. Yauk. Development and Validation of the TGx-HDACi Transcriptomic Biomarker to Detect Histone Deacetylase Inhibitors in Human TK6 Cells. Archives of Toxicology. Springer, New York, NY, 95(5):1631-1645, (2021). https://doi.org/10.1007/s00204-021-03014-2

Impact/Purpose:

Early gene expression changes in response to a toxicant provide insight into the molecular initiating events (MIE) and key events (KE) involved in its mode of action (MOA). Transcriptomic signatures, or biomarkers, typically consist of a panel of genes that robustly and consistently respond to stressors belonging to specific mechanistic classes. Such biomarkers provide a pragmatic method for efficiently extracting mechanistic information from high-content transcriptomic data and can be used to detect early molecular perturbations that are predictive of chemical hazards. Therefore, increasing the diversity of available transcriptomic biomarkers will facilitate rapid screening of chemicals to identify potential MOAs and prioritize follow-up tests in toxicological assessment. The 18 known human HDACs are organized into four classes based on the sequence homology shared with yeast deacetylases. Classes I, II, and IV HDACs are Zn2+-dependent and class III HDACs are NAD+-dependent in their catalytic activities (Seto and Yoshida 2013). Class II is further divided into classes IIa and IIb. Protein expression and activity levels differ across the four main classes and the two sub-classes within class II based on tissue and cellular localization (Morris and Monteggia 2014; Wright and Menick 2016; Millard et al. 2017; Rajan et al. 2018). Class I HDACs are ubiquitously expressed and are located in the nucleus, while classes II and IV contain both nuclear and cytoplasmic HDACs with tissue-specific expression patterns (Park and Kim 2020). Along with other agents that modulate the PTM of histones, HDACi are considered epigenotoxicants. Several chemicals present in the environment, such as methoxyacetic acid, resveratrol, and butyrate, have been identified as HDACi (Wade et al. 2008; Ventrelli et al. 2013; Chang et al. 2014). The downstream effects of HDACi exposure are broad due to the diverse roles that HDACs play in the cell; the effects of HDACi include activation of pro-apoptotic genes, disruption in cell cycle, and inhibition of DNA repair (Bose et al. 2014). Dysregulation of PTM of histones has implications in developmental toxicity, neurotoxicity, and carcinogenesis (Audia and Campbell 2016). In the present study, we identified and tested a transcriptomic biomarker of HDACi, named TGx-HDACi, to address the limited tools available for identifying toxicants that operate through the inhibition of HDAC as the MIE. To do this, we expanded on the original TGx-DDI sample sets in TK6 cells, with a long-term vision to enable rapid screening for genotoxicity and epigenotoxicity using a single high-throughput transcriptomic analysis. We first constructed a reference compound set containing HDACi and non-HDACi chemicals. The HDACi chemical set was limited to the inhibitors of the classical, Zn2+-dependent human HDACs that constitute HDAC classes I (HDACs 1 to 3 and 8), II (HDACs 4 to 7, 9, and 10) and IV (HDAC 11); these HDACi contain Zn-binding groups and inhibit HDACs by chelating the Zn from the active site (L. Zhang et al. 2018). Sirtuins, the NAD+-dependent class III HDACs, are not affected by these HDACi and, thus, require a different class of chemicals for inhibition (Seto and Yoshida 2013). Gene expression profiling was applied to cells exposed to HDACi and non-HDACi chemicals using TempO-Seq (BioSpyder). The nearest shrunken centroid (NSC) method was applied to the whole transcriptome profiles of the reference compounds to derive the TGx-HDACi biomarker (Tibshirani et al. 2002). The performance of TGx-HDACi was evaluated by classifying an external validation compound set containing HDACi, non-HDACi, and non-HDACi epigenetic modulators (inhibitors of histone acetylase, histone methyltransferase, and histone demethylase), using different statistical analyses.

Description:

Transcriptomic biomarkers can be used to inform molecular initiating and key events involved in a toxicant’s mode of action. To address the limited approaches available for identifying epigenotoxicants, we developed and assessed a transcriptomic biomarker of histone deacetylase inhibition (HDACi). First, we assembled a set of ten prototypical HDACi and ten non-HDACi reference compounds. Concentration–response experiments were performed for each chemical to collect TK6 human lymphoblastoid cell samples after 4 h of exposure and to assess cell viability following a 20-h recovery period in fresh media. One concentration was selected for each chemical for whole transcriptome profiling and transcriptomic signature derivation, based on cell viability at the 24-h time point and on maximal induction of HDACi-response genes (RGL1, NEU1, GPR183) or cellular stress-response genes (ATF3, CDKN1A, GADD45A) analyzed by TaqMan qPCR assays after 4 h of exposure. Whole transcriptomes were profiled after 4 h exposures by Templated Oligo-Sequencing (TempO-Seq). By applying the nearest shrunken centroid (NSC) method to the whole transcriptome profiles of the reference compounds, we derived an 81-gene toxicogenomic (TGx) signature, referred to as TGx-HDACi, that classified all 20 reference compounds correctly using NSC classification and the Running Fisher test. An additional 4 HDACi and 7 non-HDACi were profiled and analyzed using TGx-HDACi to further assess classification performance; the biomarker accurately classified all 11 compounds, including 3 non-HDACi epigenotoxicants, suggesting a promising specificity toward HDACi. The availability of TGx-HDACi increases the diversity of tools that can facilitate mode of action analysis of toxicants using gene expression profiling.

URLs/Downloads:

Record Details: