Integrating bioinformatics, molecular dynamics simulation and experimental verification to screen diagnostic biomarkers for polyamine metabolism-related osteoarthritis and predict potential drugs.

BACKGROUND: Osteoarthritis (OA) represents a prevalent chronic joint degeneration disorder regulated by multiple factors. Polyamine metabolism (PM) contributes substantially to OA development. However, the underlying mechanism remains unclear. The current study aimed to identify PM-related OA biomarkers and to discover potential therapeutic small-molecule compounds (SMCs). The goal was to lay the groundwork for future diagnostic and therapeutic approaches for OA. METHODS: The current study screened differentially expressed PM-related genes (DEPMGs) by integrating data from the Gene Expression Omnibus and GeneCards databases. Next, functional annotations were performed for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG). Weighted gene co-expression network analysis (WGCNA) was used to pinpoint crucial modular genes. Then, four machine learning algorithms were employed to determine the hub genes. Their diagnostic efficacy was evaluated via nomogram models and receiver operating characteristic curves. We screened SMCs with potential for treating OA through molecular docking and molecular dynamics simulations. Finally, to validate the expression of the aforementioned biomarkers, qRT-PCR and Western blot experiments were performed in a hydrogen peroxide (H2O2)-treated C-28/I2 human chondrocyte model. RESULTS: First, thirty DEPMGs were screened. GO and KEGG enrichment analyses revealed their involvement in biological processes, including the cellular response to lipopolysaccharide, chondrogenesis, and the IL-17 signaling pathway. Using WGCNA and machine learning, four hub genes were identified. Molecular docking and molecular dynamics simulations revealed that triptolide exhibited a strong binding affinity with the target protein and that the binding system demonstrated excellent stability. In vitro experiments revealed that the four hub genes were significantly upregulated in OA (p < 0.05), consistent with the bioinformatics predictions. CONCLUSIONS: This study initially identified four genes closely associated with polyamine metabolism-related genes: PLOD1, TSPO, SPP1, and COL6A1. These genes demonstrated potential value in the early diagnosis and precise intervention of OA. Triptolide was also found to have therapeutic potential for treating OA. These findings lay the groundwork for developing OA biomarkers and innovative therapeutic strategies.