OBJECTIVE: Clinical trials using interferon-beta (IFNbeta) in the treatment of rheumatoid arthritis have shown conflicting results. We undertook this study to understand the mechanisms of IFNbeta in arthritis at a physiologic level. METHODS: Collagen-induced arthritis (CIA) was induced in IFNbeta-deficient and control mice. The role of IFNbeta was investigated in both the priming and effector phases of the disease. The effect of IFNbeta deficiency on synovial cells, macrophages, and fibroblasts from preimmunized mice was analyzed by flow cytometry, immunohistochemistry, and enzyme-linked immunosorbent assay. Differences in osteoclast maturation were determined in situ by histology of arthritic and naive paws and by in vitro maturation studies of naive bone marrow cells. The importance of IFNbeta-producing fibroblasts was determined by transferring fibroblasts into mice at the time of CIA immunization. RESULTS: Mice lacking IFNbeta had a prolonged disease with a higher incidence compared with control mice. IFNbeta deficiency was found to influence the effector phase, but not the priming phase, of arthritis. Compared with control mice, IFNbeta-deficient mice had greater infiltration of CD11b+ cells and greater production of tumor necrosis factor alpha in vivo, and their macrophages and fibroblasts were both more activated in vitro. Moreover, IFNbeta-deficient mice generated a greater number of osteoclasts in vitro, and mice immunized to induce arthritis, but not naive mice, had a greater number of osteoclasts in vivo compared with control mice. Importantly, IFNbeta-competent fibroblasts were able to ameliorate arthritis in IFNbeta-deficient recipients. CONCLUSION: Our data indicate that IFNbeta is involved in regulating the activation state of osteoclasts and stromal cells, including macrophages and fibroblasts, but that it has little effect on T cells.