Single-cell analysis revealed potential mechanisms of biomarkers related to polyamine metabolism in osteoarthritis.

Many studies have shown that polyamine metabolism is associated with the development of osteoarthritis (OA), but the exact mechanism is unclear. In this study, OA datasets were obtained from the public database to screen differentially expressed genes (DEGs). Key module genes for polyamine metabolism were identified through weighted gene co-expression network analysis (WGCNA). We identified 129 DEGs and 1210 key module genes associated with polyamine metabolism, revealing 29 candidate genes with significant involvement in key biological functions and pathways, including the PI3K-Akt and hippo signaling pathways. The potential biomarkers underwent feature selection through machine learning to determine their diagnostic effectiveness via receiver operating characteristic (ROC) curve. Specific, COL5A3, COL6A1, and PNPLA2, demonstrated diagnostic potential with area under the curve (AUC) value above 0.7. In addition, single-cell analyses showed significant differences in biomarkers between OA and control in both hypertrophic chondrocyte (HTC) and fibrocartilage chondrocytes (FC). Finally, human chondrosarcoma cell line (SW1353) was treated with interleukin-1β to establish an OA cell model for the verification of biomarker expression. RT-qPCR results showed that COL5A3, COL6A1, and PNPLA2 were significantly downregulated in the OA group, which was consistent with the results of public database. In conclusion, combined with database analysis and cellular experiments, this study confirmed that COL5A3, COL6A1 and PNPLA2 were potential biomarkers for OA, which provided support for exploring related molecular mechanisms and screening biomarkers of the disease.