Abstract
A model for liver injury using compounds is widely employed in hepatotoxicity research. The method of continuous administration by mixing compounds into drinking water is convenient and useful for longitudional analysis of hepatotoxicity progression. However, commonly used models are limited. Therefore, this study aimed to validate the expansibility of a drinking water model by constructing and evaluating a model.
Through toxicological database searches and literature reviews, 4,4'-methylene dianiline (MDA) was identified as a candidate compound. A 28-day drinking water exposure test was conducted, resulting in elevated liver injury markers such as ALT and fibrosis markers. Pathological images, analyzed using anomaly detection with neural network, revealed a distinct pathology from existing model. To assess differences in toxicity manifestation mechanisms, multi-view data were obtained in the early stages of toxicity manifestation. Analysis of blood biochemistry values suggested a cholestasis-like liver injury. Immune cell behavior assessment revealed an increase in neutrophils similar to existing models, but distinctive behaviors were also observed. Analysis of RNA-seq data identified differential gene expression profiles, particularly highlighting variations in the fibrinolysis.
This study successfully established a novel drinking water liver injury model using MDA, demonstrating the expansibility of drinking water administration models. The model is anticipated to be valuable in various hepatotoxicological studies.
