Abstract
BACKGROUND: Osteoarthritis (OA) is a common age-related joint disease characterized by joint destruction and impaired quality of life. Angiogenesis plays a vital role in OA progression. This study aimed to identify key angiogenesis-related genes (ARGs) in OA using transcriptomic and machine learning methods. METHODS: The GSE55235 dataset (10 OA and 10 healthy synovial tissue samples) was analyzed for differentially expressed genes (DEGs), integrated with weighted gene co-expression network analysis (WGCNA), and ARGs to identify differentially expressed ARGs (DE-ARGs). Candidate genes were identified through three machine learning algorithms and evaluated using ROC curve analysis. Gene set enrichment analysis (GSEA), immune cell infiltration analysis, and therapeutic agent prediction were performed. Synovial samples from 5 OA patients and 5 matched controls were collected for RT-qPCR validation of biomarkers. RESULTS: From 1552 DEGs, 11 DE-ARGs were identified, and six candidate genes were selected using machine learning. Four genes-COL3A1, OLR1, STC1, and KCNJ8-showed AUC >0.8 in both GSE55235 and GSE1919, indicating strong diagnostic value. GSEA linked biomarkers to the "lysosome" pathway, and eosinophils and Th2 cells were significantly associated with biomarkers. Potential therapeutic agents included bisphenol A, tetrachlorodibenzo-p-dioxin, and valproic acid. Clinical validation confirmed that COL3A1, OLR1, and STC1 expression levels were consistent with database findings. CONCLUSION: The study identified COL3A1, OLR1, STC1, and KCNJ8 as key angiogenesis-related biomarkers in osteoarthritis, which could serve as potential diagnostic tools and therapeutic targets. The research underscores the importance of angiogenesis in osteoarthritis progression and suggests that targeting angiogenesis-related pathways may offer new treatment strategies.