T cells regulate the disease process in rheumatoid arthritis (RA) on multiple levels and represent a logical choice for anti-inflammatory therapy. In the inflamed joint they promote neoangiogenesis and lymphoid organogenesis, and stimulate synoviocyte proliferation and development of bone-eroding osteoclasts. The design of T-cell-targeted therapies for RA needs to take into account the uniqueness of T-cell generation, turnover and differentiation in affected patients. Patients accumulate 'old' T cells that respond to alternate regulatory signals because of an accelerated immune aging process; any therapeutic interventions that increase the replicative stress of T cells should, therefore, be avoided. Instead, therapeutic approaches that raise the threshold for T-cell activation are more promising. As a rule, antigen-derived signals synergize with co-stimulatory signals to stimulate T cells; such co-stimulatory signals are now targeted in novel immunosuppressive therapies. An example is abatacept (soluble cytotoxic-T-lymphocyte-associated protein 4-immunoglobulin), which binds with high affinity to CD80/CD86 and effectively suppresses inflammatory activity in RA. The therapeutic benefits gained by disrupting T-cell co-stimulation indicate that the pathogenesis of RA is driven by a more generalized abnormality in T-cell activation thresholds rather than a highly selective action of arthritogenic antigens.