Although the cause of rheumatoid arthritis is uncertain, the mechanisms by which destruction of joint tissues may occur have been studied extensively. The inflammatory responses in rheumatoid arthritis are probably mediated by a variety of different agents which include prostaglandins, leukotrienes, kinins and other peptide mediators, complement components, and immune complexes. The ultimate destruction of proteoglycans and collagen within cartilage similarly depend upon the release of the appropriate degradative enzymes. At one time these were thought to be predominantly lysosomal acid proteinases but emphasis has recently shifted to neutral metallo-proteinases which include specific enzymes capable of degrading collagen or proteoglycans at neutral pH. Under normal conditions these proteinases are in latent form due in part to the presence of a tissue inhibitor of metallo-proteinases (TIMP). During studies of human joint tissues in culture, it has become apparent that products of one cell type may influence the behaviour of other cells. Thus, monocytes and macrophages may produce mediators, such as interleukins, one of which has been called mononuclear cell factor (MCF), which when added to cultures of human articular chondrocytes or synovial cells, markedly enhances production of prostaglandins and metallo-proteinases while depressing the amount of TIMP. Cultured human synovial tissue produces factors with similar properties, which may in turn be related to mediators such as catabolin, which can be produced by synovium and other connective tissues and which stimulate chondrocytes to degrade their own matrix. The production of these mediators may not only be relevant to rheumatoid arthritis but also to other diseases. Thus, MCF is capable of stimulating prostaglandin production by gingival cells and cells derived from human bone. Moreover MCF is itself capable of inducing bone resorption. Since both normal and diseased tissues are capable of producing and responding to these mediators, these potential degradative interactions must be kept in check in vivo. Glucocorticosteroids may play a role in the natural suppression of these mechanisms, since in vitro they are capable of inhibiting the production of factors as well as their effects on target tissues. Since these factors probably have anabolic activity as well, they may be involved in connective tissue repair after injury. Such intercellular mediators may play important roles in the control of connective tissue turnover, not only in disease states but also in the normal processes of growth and differentiation.