Glucose-6-phosphate and joint antigen reactivity in two mouse models of autoimmune arthritis

Pristane-induced arthritis (PIA) is a murine model that strongly resembles rheumatoid arthritis (RA). In this study, we have examined a library of candidate joint antigens for potential immunopathogenic reactivity in PIA using a variety of immunological techniques. T cell responses to sixteen different joint components were evaluated using proliferation assays, and sera were assayed by ELISA for antibodies to the same joint components. Cultured cells were assayed for antigen-specific cytokine gene activation using ELISA. Mice developed pristane arthritis between four and six months post injection. ELISA revealed positive antibody responses to all antigens tested between 4 and 12 months. A positive antibody response to glucose-6-phosphate isomerase (GPI) was detected at 2 months. Cytokine gene assays revealed that IFNγ and TNFα responses to antigen stimulation were increased throughout the study, whereas elevated IL-4 and IL-6 were detected only infrequently. These results indicate that PIA animals develop a broad reactivity to numerous joint antigens, rather than a dominant T cell or antibody response to a single joint antigen. Reactivity to joint antigens usually appeared prior to the onset of clinical disease, and the cytokine pattern in responding cells generally indicated a Th1 response to joint antigens. However, antibodies against GPI appear prior to onset of arthritis or the appearance of other autoantibodies, and also occurred at higher titers than the other antibodies. This finding suggests that GPI may be a particularly relevant antigen in the pathogenesis of PIA. We have shown that antibodies against glucose-6-phosphate isomerase (GPI) develop prior to antibodies against collagen type II (CII) in mice with collagen-induced arthritis (CIA) and that GPI, given orally or i.d. can not prevent or induce arthritis. Sera from CIA mice were harvested and anti-GPI and anti-CII antibodies extracted. The antibodies were then injected iv into naive mice, and three weeks after antibody treatment, mice were immunized with type II collagen. Neither anti-GPI nor anti-CII antibodies given iv caused clinically detectable arthritis. However, both anti CII and anti-GPI antibody treatment caused a significant decrease in arthritis after CII injection. The findings suggest that a non-arthritogenic dose of anti-GPI antibody can tolerize mice against CIA.