Autoantibody production in rheumatoid arthritis and the formation of IgG immune complexes (ICs) in the synovium are thought to be involved in the activation and infiltration of hematopoietic cells (1, 2). The binding and crosslinking of ICs to Fc receptors specific for IgG (FcγR) on leukocytes triggers the activation and regulation of a variety of cellular responses, including the release of inflammatory cytokines and chemokines (3–7). Two activating FcγR have been described in the mouse: FcγRI and FcγRIII (8). The high-affinity FcγRI is capable of binding monomeric IgG2a and IgG2a ICs. The low-affinity FcγRIII binds polymeric IgG of different IgG isotypes with different affinities (9). Based on these differences, it has been proposed that FcγRI prefers the IgG2a IC, whereas FcγRIII favors IgG1 and IgG2b ICs.
In proteoglycan-induced arthritis (PGIA), immunization with human PG leads to the development of autoreactive T cells and autoantibodies to murine PG (10). These autoantibodies gain access to the joint and bind cartilage PGs, forming ICs. We have recently shown that FcγR are critically involved in the development of arthritis. In FcR γ-chain−/− mice, joint inflammation is completely suppressed despite the activation of autoreactive T cells and B cells (11). Similar to PGIA, in several other murine models of arthritis, FcγR γ-chain expression is critical to the development of disease (12–16). Although these models have demonstrated the importance of FcγR expression, there is no clear consensus on which one, FcγR, FcγRI, or FcγRIII, is critical. In antigen-induced arthritis (AIA), neither FcγRI nor FcγRIII is involved in joint swelling and leukocyte infiltration, but cartilage damage is reduced in FcγRI−/− mice (17). In IC-mediated arthritis (ICA), FcγRIII is the principal activating FcγR that mediates joint inflammation; however, both FcγRI and FcγRIII are involved in cartilage destruction (18). In contrast, in the K/BxN model and collagen-induced arthritis (CIA), a significant reduction in joint swelling is observed in FcγRIII−/− mice (14, 19). This disparity in the requirement for FcγR suggests that different mechanisms may be involved in the development of disease.
One of the major differences between PGIA and other models of arthritis is the dependence on interferon-γ (IFNγ). IFNγ is an important proinflammatory cytokine released by Th1 cells, and it is central to the development of PGIA (20). We found that arthritis is suppressed and disease onset is substantially delayed in mice treated with anti-IFNγ antibodies or in IFNγ−/− mice (21). One of the important functions of IFNγ is the induction of FcγRI expression and the enhancement of IgG2a secretion via class switch (17, 22). Increased expression of FcγRI could facilitate enhanced binding of IgG2a ICs. Based on this information, we would speculate that FcγRI might be involved in PGIA. However, despite the high levels of IFNγ in PGIA, the PG-specific autoantibody response is dominated by the IgG1 isotype (23). If the IgG1 ICs containing PG-specific autoantibodies are important for arthritis, then FcγRIII may be involved in PGIA. To distinguish between the requirements for FcγRI and FcγRIII expression in PGIA, we assessed the development of PGIA in FcγRI−/− and FcγRIII−/− mice.
- Top of page
- MATERIALS AND METHODS
In previous studies we demonstrated the necessity for FcγR in PGIA (11). We have also shown that IFNγ is critically important for the induction of disease (21). Since IFNγ regulates FcγRI and IgG2a antibody isotype expression and FcγRI preferentially binds to IgG2a IC, we speculated that FcγRI might be the FcγR involved in PGIA. However, despite the high levels of IFNγ in PGIA, the dominant PG-specific antibody isotype is IgG1 (20). Since IgG1 IC preferentially binds to FcγRIII, these data also suggested the possibility that FcγRIII could be involved in PGIA (9). In this study, we demonstrated that FcγRIII expression is critical for the development of PGIA, whereas FcγRI expression is not necessary. Upon immunization with PG, FcγRI−/− mice developed disease with similar kinetics and severity as in littermate controls, whereas FcγRIII−/− mice failed to develop arthritis (Figures 1 and 2). Thus, the requirement for FcγRIII correlates with the dominant IgG1 autoantibody isotype produced in PGIA. In addition to antibody isotype, the cell population responsible for inflammation may determine the relevant FcγR. In PGIA, neutrophils are the dominant cell population in the inflamed joint, and FcγRIII is the only activating FcγR expressed on murine neutrophils (17). Thus, it is plausible that the loss of FcγRIII on neutrophils may inhibit neutrophil binding to IgG1 IC in the joint, thus preventing their retention and/or activation in the joint.
These findings corroborate the results of experiments in K/BxN CIA and ICA models, which also demonstrate a significant reduction in ankle swelling in FcγRIII−/− mice (14, 19). However, these data differ from findings in the AIA model, where neither FcγRI nor FcγRIII is involved in joint swelling and leukocyte infiltration, although reduced cartilage damage is observed in FcγRI−/− mice (17). The disparity in the requirement for FcγR in AIA suggests that other mechanisms are involved in the development of arthritis.
There are several explanations as to why inhibition of disease is dependent on FcγR expression. PGIA begins as a systemic immune response to human PG that proceeds to an autoreactive response to murine PG prior to any signs of joint involvement (29, 30). FcγR may contribute to systemic immunity by regulating B cell and/or T cell priming. With regard to FcγR control of B cell priming, evidence suggests that FcγR expression may facilitate antibody production (27, 31). However, we were unable to detect a deficiency in the secretion of PG-specific antibody in either FcγRI−/− or FcγRIII−/− mice (Figure 3), indicating that antibody production in PGIA is independent of FcγR expression. FcγR are also involved in T cell priming through the efficient uptake of antigen by antigen-presenting cells (27, 32–35). However, in FcγRI−/− and FcγRIII−/− mice, the PG-specific recall response was equivalent to that in WT mice, indicating that FcγR expression is not critical for T cell priming in this model. It is possible that FcγR expression may be important for T and B cell priming only under conditions in which the concentration of antigen is limiting. Since the development of PGIA requires multiple immunizations with PG, the increase in antigen concentration could overcome any requirement for FcγR in PG-specific T and B cell priming.
The absence of a requirement for FcγRIII in the initiation of an immune response to PG suggests that FcγRIII is important in the effector phase. To confirm the importance of FcγRIII in the effector phase, we successfully blocked transfer of arthritis with an antibody to FcγRIII (Figure 4). Since spleen cells from PG-immune mice contain T and B cells that can induce arthritis, the inhibition with anti-FcγRIII/FcγRII antibody suggests that FcγRIII expression is necessary for inflammation to be maintained in the joint.
In the absence of FcγRIII expression, there was a significant decrease in cytokine and chemokine RNA transcripts in the joint, whereas in FcγRI−/− mice levels similar to those in WT mice were detected (Figure 5). However, these transcripts are only partially reduced in FcγRIII−/− mice despite the fact that there was no evidence of inflammation in the paws. These data indicate that FcγRIII is only partially involved in activation of these proinflammatory mediators and that other mechanisms also contribute to cytokine and chemokine expression in the joint. The reduction in cytokine and chemokine mRNA corresponds to the absence of joint inflammation in FcγRIII−/− mice and suggests that FcγRIII may control activation of cells that are recruited to the joint.
We propose that in PGIA, joint inflammation is initiated by the activation of autoreactive B cells and the release of autoantibodies, followed by IC deposition on joint cartilage surfaces. These ICs instigate some degree of inflammatory cytokine and chemokine expression that occurs independent of FcγR expression (11). The release of chemotactic agents within the joint stimulates the influx of FcγRIII-bearing neutrophils capable of binding ICs that are either within the synovial fluid or bound to joint tissues. The crosslinking of FcγRIII leads to further expression of cytokines and chemokines within the joint, thereby maintaining or amplifying the response.