Studies in animal models of osteoarthritis (OA) have been used extensively to gain insight into the pathogenesis of OA, but early studies largely ignored inflammation except as a secondary phenomenon. Synovitis has often been noted as a feature in experimental OA, and more recent work has established a central role for inflammatory cytokines as biochemical signals which stimulate chondrocytes to release cartilage-degrading proteinases. Thus, proteinase inhibitors, cytokine antagonists and receptor blocking antibodies, and growth/differentiation factors have been considered as potential therapeutic agents and targets for gene therapy. Although there is some disagreement, it is generally accepted that IL-1 is the pivotal cytokine at early and late stages, while TNF-alpha is involved primarily in the onset of arthritis. Other cytokines released during the inflammatory process in the OA joint may be regulatory (IL-6, IL-8) or inhibitory (IL-4, IL-10, IL-13, IFN-gamma). Furthermore, studies in animal models have illustrated the potentially beneficial effects of anticytokine therapy with monoclonal antibodies or receptor antagonists, although local rather than systemic delivery would be necessary for the largely localized OA in humans. Transgenic or knockout mice have also provided insights into general mechanisms of cytokine-induced cartilage degradation but have not directly addressed OA pathogenesis. Similarly, animals with spontaneous or transgenic modifications in cartilage matrix components, growth/differentiation factors, or developmentally regulated transcription factors have provided information about potential gene defects that predispose to OA without addressing the role of inflammatory mediators in cartilage destruction. Although the multiple etiologies of human OA indicate that it is more complex than any animal model, the use of appropriate, well-defined animal models will establish the feasibility of novel forms of therapy.