Abstract: BACKGROUND: There are still no sensitive markers for monitoring the occurrence or progression of osteoarthritis. The detection of changes in peripheral blood gene expression profiles during the active period of osteoarthritis is helpful to explore the precise targets for diagnosis and treatment in the blood and explain the pathogenesis. 
OBJECTIVE: To simultaneously analyze the differences in gene expression profiles of peripheral blood mononuclear cells and synovium between osteoarthritis and normal human using bioinformatics methods, to explore the diagnosis and treatment targets and therapeutic drugs of osteoarthritis in the blood from the molecular level, and to provide a new perspective on the study of osteoarthritis.
METHODS: We found GSE48556 and GSE55235 data sets from GEO and ArrayExpress databases. Blood samples from 105 patients with osteoarthritis and 24 normal controls and synovial samples from 10 patients with osteoarthritis and 10 normal controls were included. We used the R language limma package to screen differentially expressed genes in blood and synovium between osteoarthritis and normal controls and to identify the intersected genes. The gene ontology and KEGG pathways of differentially expressed genes were analyzed by DAVID. We used STRING and Cytoscape software to construct a protein-protein interaction network, used Mcode plug-in for module analysis, and used Cytohubba to screen out key genes. We imported differentially expressed genes into the online platform of Connectivity Map to screen out the top 10 small molecule drugs that are negatively related.
RESULTS AND CONCLUSION: A total of 142 differentially expressed genes were obtained, of which 75 were up-regulated genes and 67 were down-regulated genes. The gene ontology enrichment analysis of all differentially expressed genes mainly focused on T cell activation, leukocyte migration, leukocyte-cell adhesion, positive regulation of cell adhesion, and regulation of lymphocyte activation. The kyotoencyclopedia of genes and genomes was mainly enriched in the MAPK signaling pathway, Rap1 signaling pathway, Ras signaling pathway, B cell receptor signaling pathway, and chemokine signaling pathway. The protein-protein interaction network and related plug-ins were used to screen out 10 genes including mitogen-activated protein kinase 1, vascular endothelial growth factor A, growth factor receptor binding protein 2, early growth response protein 1 and ras-related c3 botulinum toxin substrate 2, which were highly related to cell apoptosis and immune inflammation in osteoarthritis. Connectivity Map screened out some small molecules with potential therapeutic effects, such as indirubin and beclomethasone. To conclude, the bioinformatics analysis can indicate that the differences of gene expression in peripheral blood mononuclear cells and synovium between osteoarthritis patients and normal controls are concentrated in the biological events of apoptosis and inflammation, so that the blood expression profile becomes an effective breakthrough for monitoring target markers of osteoarthritis and studying their potential molecules for drugs.

Key words: osteoarthritis, peripheral blood mononuclear cell, synovium, gene expression profile, apoptosis, inflammation, Connectivity Map, bioinformatics