Objective. To determine if the T cell antigen receptor V, usage of unstimulated rheumatoid arthritis (RA) synovial fluid (SF) T cells is biased compared with those in peripheral blood (PB).Methods. Freshly isolated, matched synovial fluid and peripheral blood T cells were analyzed for V, gene expression using quantitative polymerase chain reaction (PCR) methods. Ten synovial fluid samples from the knees of 7 patients with RA were studied. The PCR assay used 26 V , primers with a constant region C, primer, and 2 C, primers that co-amplied a product that served as an internal standard. Cycle number and complementary DNA content were controlled to ensure the linear accumulation of PCR products. Labeled products were separated on 10% polyacrylamide gels and counted with a Betascope blot analyzer.Results. There were consistent differences between the V, gene usage of SF and PB T cells directly isolated from patients with RA, regardless of HLA-DR haplotype. In all synovial specimens, V, 2 was increased relative to the peripheral blood, while Vj13.1 and V,13.2 were decreased. V& and b 2 1 were increased in 9 of the 10 synovial samples. Analyses of bilateral SF specimens from 2 subjects and serial specimens from the same knee of 1 subject revealed virtually identical patterns in each Submitted for publication April 19, 1994; accepted in revised form June 28, 1994. patient. The SF V, bias was not solely due to differences in the proportion of CD4+ and CD8+ cells, because the CD4:CD8 ratios in SF and PB were similar. However, V , gene usage of separated CD4+ and CD8+ synovial T cells showed that V $ and V ' were more highly expressed on CD4 cells.Conclusion. Freshly isolated synovial T cells from inflamed (not end-stage) knees of patients with RA have a remarkably consistent biased V, gene usage compared with PB T cells. V, 2 and V, 6 are uniformly increased, and this increase is primarily in CD4+ T cells. The same V, bias in the SF T cells of several RA patients suggests that shared antigens may be stimulating the T cell response.