We investigated the cytosolic proteome of inflamed synovial tissue by hierarchical clustering analysis and validated the feasibility of this proteome analysis by identifying proteins that were differentially expressed between rheumatoid arthritis (RA), spondyloarthropathy (SpA), and osteoarthritis (OA). Synovial biopsy samples were obtained from 18 patients undergoing needle arthroscopy for knee synovitis associated with RA (n = 6) and SpA (n = 6), and for joint effusion of the knee associated with OA (n = 6). Cytosolic proteins were extracted from the tissue and subjected to two-dimensional gel electrophoresis. Protein expression patterns were statistically analyzed and used for hierarchical cluster analysis. Proteins of interest were independently identified by matrix-assisted laser desorption/ionization- and electrospray ionization-mass spectrometry. Hierarchical cluster analysis of the complete match set, containing 640 spots, remarkably segregated SpA from RA and OA. Next, we used a subset of spots that was statistically, differentially expressed (P < 0.01), between RA and SpA, SpA and OA, or RA and OA, in both Student's t-test and Mann-Whitney U-test. The dendrograms revealed distinct clustering of RA versus SpA and RA versus OA. Spots that were differentially expressed between the groups were identified by tandem mass spectrometry. Fructose bisphosphate aldolase A and alpha-enolase showed higher expression levels in SpA than in OA (P < 0.01). Calgranulin A myeloid related protein-8 (MRP-8) was markedly up-regulated in RA and SpA patients in comparison to OA patients where this spot was below detection limit. The analysis of the cytosolic proteome of synovial tissue is a useful approach to identify disease-associated proteins in chronic inflammatory arthritis.