Metformin dampens the progression of cholangiofibrosis induced by thioacetamide using deep learning

Abstract

Purpose: The consistent use of metformin has been linked to a reduced incidence of neoplastic diseases among diabetic populations. As a preventive intervention, metformin may offer a more favorable risk-benefit profile. Here, we explored the efficacy of metformin in the primary prevention of cholangiofibrosis, which can precede the carcinogen-induced development of cholangiocarcinoma (CCA). Our objective was to assess the potential of metformin to act as an intervention prior to the onset of these conditions. Methods: A rat model of thioacetamide (TAA)-induced cholangiofibrosis was utilized to assess the impact of metformin on the induction process of cholangiocarcinoma (CCA). Liver tissues were harvested and analyzed histologically using light microscopy, complemented by a deep-learning convolutional neural network for enhanced evaluation. Additionally, RNA sequencing (RNA-seq) was performed to investigate the genetic alterations associated with metformin treatment in this TAA-induced cholangiofibrosis model. Results: In the rat model, the TAA control group exhibited an increased incidence and average count of cholangiofibrosis cases in the liver, with rates of 100 % and an average of 12.0, compared to the metformin-treated group, which showed an incidence of 70 % and an average of 3.3. Notably, the progression from normal cholangioles to cholangiofibrosis was associated with the upregulation of several proteins critical for metabolic processes and the tumor microenvironment. These alterations were significantly mitigated by metformin treatment. Conclusions: Long-term metformin use may offer protective benefits against cholangiofibrosis, partially by regulating metabolic processes and improving the tumor microenvironment.

Keywords

• Cholangiofibrosis
• Metformin
• Thioacetamide
• Artificial intelligence