Journal Article

Mechanistic Exploration of Aristolochic Acid I-Induced Hepatocellular Carcinoma: Insights from Network Toxicology, Machine Learning, Molecular Docking, and Molecular Dynamics Simulation

Lin, Hangqi, Liu, Bolin, Cheng, Peifeng, Ying, Xinwang, Tu, Tiantaixi, Yang, Ye, Xie, Qingfeng, Zheng, Tongtong

2025

Overview

This study explores how aristolochic acid I (AAI) drives hepatocellular carcinoma (HCC). We first employ network toxicology and machine learning to map the key molecular target genes. Next, our research utilizes molecular docking to evaluate how AAI binds to these targets, and finally confirms the stability and dynamics of the resulting complexes through molecular dynamics simulations. We identified 193 overlapping target genes between AAI and HCC through databases such as PubChem, OMIM, and ChEMBL. Machine learning algorithms (SVM-RFE, random forest, and LASSO regression) were employed to screen 11 core genes. LASSO serves as a rapid dimension-reduction tool, SVM-RFE recursively eliminates the features with the smallest weights, and Random Forest achieves ensemble learning through decision trees. Protein–protein interaction networks were constructed using Cytoscape 3.9.1, and key genes were validated through GO and KEGG enrichment analyses, an immune infiltration analysis, a drug sensitivity analysis, and a survival analysis. Molecular-docking experiments showed that AAI binds to each of the core targets with a binding affinity stronger than −5 kcal mol−1, and subsequent molecular dynamics simulations verified that these complexes remain stable over time. This study determined the potential molecular mechanisms underlying AAI-induced HCC and identified key genes (CYP1A2, ESR1, and AURKA) as potential therapeutic targets, providing valuable insights for developing targeted strategies to mitigate the health risks associated with AAI exposure.

• Publisher: MDPI AG,MDPI
• Subject: Aristolochic acid / aristolochic acid I / Aristolochic Acids - toxicity / Artificial intelligence / Carcinoma, Hepatocellular - chemically induced / Carcinoma, Hepatocellular - genetics / Carcinoma, Hepatocellular - metabolism / Comparative analysis / CYP1A2 protein / Cytochrome P450 / Decision trees / Development and progression / Ensemble learning / ESR1 protein / Genes / Health aspects / Health risks / Hepatocellular carcinoma / Hepatoma / Humans / Learning algorithms / Liver cancer / Liver Neoplasms - chemically induced / Liver Neoplasms - genetics / Liver Neoplasms - metabolism / Machine Learning / Molecular docking / Molecular Docking Simulation / Molecular dynamics / Molecular Dynamics Simulation / Molecular modelling / Mutation / network toxicology / Physiological aspects / Plant toxins / Protein Interaction Maps / Proteins / Sensitivity analysis / Support vector machines / Therapeutic targets / Toxicity / Toxicology / Traditional Chinese medicine