T and B cells remain important in the development of rheumatoid arthritis. Recent data indicate that the involvement of T cells is more complex than recognizing an arthritogenic antigen in synovia. Attempts to characterize and isolate proliferating synovial T cells have not led to a better understanding of the postulated antigen-recognition events in synovia. Experimental systems in which implants of rheumatoid synovia in severely immunocompromised mice were analyzed have shown that the T cell infiltrate does not persist, emphasizing that tissue inflammation is not a functional unit but is highly dependent on the peripheral T cell repertoire. In this compartment of circulating T cells, rheumatoid arthritis (RA) patients appear to have major abnormalities. CD4+ T cells in the blood are clonally expanded, and these clonotypes penetrate into the tissue but are not enriched compared with those in the circulation. Studies of CD8+ circulating cells also demonstrate clonal expansion. CD8 clonotypes appear to be present in healthy individuals, however, less frequently than in RA patients. Taken together, the role of T cells in RA seems to not be restricted to the joint, and their pathogenetic role may extend beyond antigen recognition in the joint. Convincing evidence has been provided that rheumatoid factor (RF) immunoglobulins can be somatically mutated, suggesting antigen selection as a driving force of RF production. However, most of the immunoglobulins with RF specificity in RA patients are in germline configuration. Emerging data support the concept that RF+ B cells normally do not secrete antibodies and antigen nonspecific stimuli may contribute to the induction of RF release in RA.(ABSTRACT TRUNCATED AT 250 WORDS)