Rheumatoid arthritis is a common cause of chronic disability for which current therapies are of limited value in controlling the disease process and outcome. Our initial approach to understanding the pathogenesis of RA and defining a novel therapeutic target was to investigate the role of cytokines by blocking their action with antibodies on cultured synovial-derived mononuclear cells in vitro. These investigations suggested that neutralization of TNF alpha with antibodies significantly inhibited the generation of other pro-inflammatory cytokines also over-produced, such as, IL-1, GM-CSF, IL-6 and IL-8. The implication that blockade of a single cytokine, TNF alpha might have far-reaching effects on multiple cytokines and thereby exert significant anti-inflammatory and protective effects on cartilage and bone of joints, was tested in arthritic DBA/1 mice immunized with collagen II. Impressive amelioration of joint swelling and joint erosions in this model encouraged clinical trials with a monoclonal anti-TNF alpha antibody. The cA2 chimeric anti-TNF alpha high-affinity antibody was initially tested in an open-label study at a dose of 20 mg/kg on 20 patients, with substantial and universal benefit. Subsequently, a randomized placebo-controlled double-blind trial was performed on 73 patients comparing a single intravenous injection of placebo (0.1% human serum albumin) with two doses of cA2. Using a composite disease activity index, at 4 weeks post infusion, 8% of patients receiving placebo improved compared with 44% receiving 1 mg/kg cA/2 and 79% receiving 10 mg/kg. Between 2 to 4 repeated cycles of cA2 were administered to 7 patients and all patients showed improvement of a similar magnitude with each cycle. These data support our proposition that TNF alpha is implicated in the pathogenesis of RA, and is thus a key therapeutic target. Monoclonal anti-TNF alpha antibodies control disease flares and are candidate agents for longer-term control of RA, although repeated therapy with cA2 is associated with anti-idiotypic responses in 50% of patients and a trend toward shortening of the duration of response. In the DBA/1 arthritic mice, synergy of action of anti-TNF and anti-CD4 is observed together with suppression of an anti-globulin response, indicating one way in which benefit might be augmented in the future.