Integrative analysis of thiamethoxam induced hepatocellular carcinoma toxicity mechanisms

Abstract

Abstract Neonicotinoid (NEO) pesticides play a crucial role in agricultural production. However, their potential risks to human health and the environment cannot be overlooked. To gain a comprehensive understanding of the toxicity and mode of action of NEOs, thiamethoxam (THX), which exhibits the highest potential for carcinogenicity and hepatotoxicity, was selected as the subject of this study. We identified 61 intersection genes between THX targets and hepatocellular carcinoma (HCC)-related genes. These genes were then uploaded to the Metascape database for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. The GO analysis indicated that the significant biological processes mainly involved the response to xenobiotic stimuli, cellular response to chemical stress, cellular response to biotic stimuli, and response to toxic substances. The KEGG enrichment analysis pinpointed several key pathways, primarily including the cell cycle and Glycolysis/Gluconeogenesis. Subsequently, the intersection genes were imported into the Gene Expression Profiling Interactive Analysis (GEPIA) and Gene Expression Omnibus (GEO) databases to analyze expression differences, leading to the identification of 15 significantly differentially expressed core genes (SDECGs). By applying the Support Vector Machine (SVM) machine—learning model, we screened out five feature genes (CYP2C19, CYP3A4, FBP1, THBS4, CYP7A1) and constructed a nomogram. Molecular docking of THX with these five feature genes showed binding energies of less than −5 kcal/mol. This study offers a theoretical foundation for understanding the underlying mechanisms of THX-induced HCC. The findings provide a scientific basis for the safety assessment of THX in agricultural applications and contribute to the establishment of pesticide safety standards.

Keywords

• Hepatocellular carcinoma
• Neonicotinoid pesticide
• Thiamethoxam
• Network toxicology
• Machine learning
• Molecular docking