Diagnosis of Osteoarthritis Subtypes with Blood Biomarkers

Abstract Objective To identity osteoarthritis(OA) subtypes with gene expression of peripheral blood mononuclear cells. Methods Gene expression data (GSE48556) of Genetics osteoARthritis and Progression (GARP) study was downloaded from Gene Expression Omnibus. Principal component analysis and unsupervised clustering were analyzed to identify subtypes of OA and compare major KEGG pathways and cell type enrichment using GSEA and xCell. Classification of subtypes were explored by the utilization of support vector machine. Results Unsupervised clustering identified two distinct OA subtypes: Group A comprised of 60 patients (56.6%) and Group B had 46 patients (43.3%). A classifier including nine genes and CD4 + T cell and Regulatory T cell flow cytometry could accurately distinguish patients from each group (area under the curve of 0.99 with gene expression). Group A is typical degenerative OA with glycosaminoglycan biosynthesis and apoptosis. Group B is related to Graft versus host disease and antigen processing and presentation, which indicated OA has a new type of “Antigen processing and presentation” similarly as that of RA. Conclusion OA can be clearly classified into two distinguished subtypes with blood transcriptome, which have important significance on the development of precise OA therapeutics.