Host: The Japanese Society of Toxicology
Name : The 51st Annual Meeting of the Japanese Society of Toxicology
Date : July 03, 2024 - July 05, 2024
<Purpose> Centrilobular hepatocyte hypertrophy is frequently observed in toxicity tests using experimental animals, but its toxicological significance is unclear. Although cytochrome P450 induction is suggested to be involved in its development, it has not been demonstrated experimentally. In this study, we investigated the role of enzyme induction in centrilobular hepatocyte hypertrophy using gene expression data and machine learning techniques.
<Methods> Liver histopathological findings of rat repeated-dose toxicity tests and their hepatic gene expression data were obtained from Open TG-GATEs and Toxygates, respectively. A prediction model was developed using LightGBM with gene expression data as the explanatory variables and the presence or absence of centrilobular hepatocyte hypertrophy as the objective variable. The contribution of each gene to the prediction was evaluated using the mean absolute SHAP value. The top 100 genes were extracted for gene enrichment analysis.
<Results and Discussion> A developed model was evaluated using a 5-fold cross-validation and its ROC-AUC value was 0.886. The gene that contributed most to the prediction was Cyp2b1, and “Metabolism of xenobiotics by cytochrome P450” was most enriched. Among the genes included in this term, Cyp2b1, Akr7a3, Ephx1, Gsta3, and Ugt2b17 were observed in the top 10 genes with mean absolute SHAP values. These results suggest that the induction of drug-metabolizing enzymes, including CYP2B1, largely contributes to the development of centrilobular hepatocyte hypertrophy in rats.
