Identification of early liver toxicity gene biomarkers using comparative supervised machine learning

Abstract

Screening agrochemicals and pharmaceuticals for potential liver toxicity is required for regulatory approval and is an expensive and time-consuming process. The identification and utilization of early exposure gene signatures and robust predictive models in regulatory toxicity testing has the potential to reduce time and costs substantially. In this study, comparative supervised machine learning approaches were applied to the rat liver TG-GATEs dataset to develop feature selection and predictive testing. We identified ten gene biomarkers using three different feature selection methods that predicted liver necrosis with high specificity and selectivity in an independent validation dataset from the Microarray Quality Control (MAQC)-II study. Nine of the ten genes that were selected with the supervised methods are involved in metabolism and detoxification (Car3, Crat, Cyp39a1, Dcd, Lbp, Scly, Slc23a1, and Tkfc) and transcriptional regulation (Ablim3). Several of these genes are also implicated in liver carcinogenesis, including Crat, Car3 and Slc23a1. Our biomarker gene signature provides high statistical accuracy and a manageable number of genes to study as indicators to potentially accelerate toxicity testing based on their ability to induce liver necrosis and, eventually, liver cancer.

New approach methodologies: A quantitative in vitro to in vivo extrapolation case study with PFASs

“…Perhaps these gene expression biomarkers measured in the in vitro system are more sensitive as readouts compared to an in vivo change in blood cholesterol. In general, gene expression alterations in a biological system upon exposure to xenobiotics may serve as early indicators of eventual toxicity, and as such they may start at lower exposure levels prior to manifestation of the respective toxicity phenotype that may occur at higher doses ( Gatzidou et al, 2007 ; Joseph, 2017 ; Smith et al, 2020 ). As mentioned earlier, interpretation of changes in gene expression with regards to eventual adverse effects shall be done with caution.…”

Downregulation of CYP39A1 Serves as a Novel Biomarker in Hepatocellular Carcinoma with Worse Clinical Outcome

“…Another study revealed that higher hepatic mRNA levels of hepatic CYP39A1 were linked to higher serum cholesterol but protect against steatosis, steatohepatitis, and liver fibrosis in a subset of patients [45]. Furthermore, CYP39A1 could be served as liver toxicity gene markers [46]. Moreover, it was found that CYP39A1 mRNA was increased in models of hepatoprotection mice from cholestasis induced by lithocholic acid (LCA), which toward the formation of less toxic bile acids therefore leading less liver injury [47].…”

In vitro functional models for human liver diseases and drug screening: beyond animal testing

“…A toxicogenomic open database 231 was used to extract such features to predict liver toxicity, while the computational model was optimized to predict whether a drug compound can cause liver necrosis and identify target gene biomarkers as disease indicators. 228 Here, the model identified predictor biomarkers that are involved in liver metabolism and detoxification (i.e., Car3, Crat, Cyp39a1, Dcd, Lbp, Scly, Slc23a1, and Tkfc), carcinogenesis as well as transcriptional regulation (i.e., Ablim3). In the future, similar methods could be used to speed up the prediction of drug toxicity effects in humans.…”