Abstract: BACKGROUND: At present, the pathogenesis of osteoarthritis is still unclear, and there is a lack of effective means to control the disease. Research on osteoarthritis is mostly concentrated in the field of immunity, and there are few studies in the field of oxidative stress. 
OBJECTIVE: To explore the roles of oxidative stress and immune infiltration in osteoarthritis and to predict related miRNAs and therapeutic agents. 
METHODS: The GSE55235 dataset (10 samples of osteoarthritis and 10 healthy control samples) and the GSE55457 dataset (10 samples of osteoarthritis and 10 healthy control samples) were obtained from the GEO database for merging to obtain their differentially expressed genes that were combined with oxidative stress genes to get the differentially expressed genes of oxidative stress. The differentially expressed genes of oxidative stress were analyzed for KEGG and GO enrichment, and the osteoarthritis pathways and biological processes were evaluated using GSEA enrichment analysis. The protein-protein interaction network was constructed using the STRING online website and Cytoscape software, and the Degree algorithm was run to get the key genes. The GSE1919 dataset was obtained from the GEO database as a validation dataset, and the key genes were analyzed by variance analysis and receiver operating characteristic curve analysis to get the core genes. In addition, immune infiltration was evaluated by CIBERSORT and the correlation between core genes and immune cells was explored. miRNA prediction of core genes was performed using TargetScan and target drugs were predicted using the DSigDB database. 
RESULTS AND CONCLUSION: Sixty-five differentially expressed genes and five core genes (IL1B, CXCL8, MYC, NFKBIA, JUN) associated with oxidative stress were identified. Enrichment analysis showed that differentially expressed genes associated with oxidative stress were concentrated in the pathways of oxidative stress, interleukin-17, osteoclast differentiation, fluid shear stress and atherosclerosis. The area under the receiver operating characteristic curve for the five core genes exceeded 0.85, indicating their excellent specificity and sensitivity in diagnosing bone and joint conditions, as well as their close association with immune cells. The predicted miRNA was has-miR-3937, and the therapeutic small-molecule drugs were metformin, ionomycin and celecoxib. To conclude, oxidative stress and immune infiltration exist in osteoarthritis, and immune infiltration is involved in activating oxidative stress. The core genes and predicted miRNAs can be used as novel markers for the diagnosis of osteoarthritis, and small molecule drugs are predicted to treat osteoarthritis.

Key words: osteoarthritis, oxidative stress, immune infiltration, differentially expressed gene, biomarker, miRNA, drug prediction