Rheumatoid arthritis is more than cytokines: Autoimmunity and rheumatoid arthritis

Josef S. Smolen and Gu¨nter SteinerAutoimmunity, characterized by autoantibodyexpression, is one of the hallmarks of rheumatoid arthri-tis (RA). This feature distinguishes RA from otherinflammatory and degenerative joint diseases, such aspsoriatic arthritis, reactive arthritis, or osteoarthritis(OA), where autoantibody responses are absent. Theprominence of serologic manifestations of RA suggeststhat, despite the occurrence of joint inflammation anddestruction in the various forms of arthritis, these dis-eases display significant differences in etiopathogenesis.The first autoantibody described in RA morethan 60 years ago was rheumatoid factor (RF) (1). Sincethen, an increasing number of autoimmune responsesfor RA have been detected, some of which appear to bemuch more specific for RA than RF. These autoanti-bodies include the antiperinuclear factor (APF) andantikeratin antibodies (AKA) (2), anticollagen antibod-ies (3), antibodies to nuclear antigens such as Epstein-Barr nuclear antigen and RA33 (4,5), anti-Sa (6), andanti-p68 (7). Many of the autoantigens targeted by theseantibodies have been characterized: RF is directedagainst IgG (8); APF and AKA are directed againstcitrullinated filaggrin and other proteins containing cit-rulline, a posttranslationally modified arginine residue(9); RA33 is identical to the heterogeneous nuclearribonucleoprotein A2 (hnRNP-A2) (10); and p68 hasrecently been identified as the heavy chain bindingprotein BiP (11).Neither the histologic features of the RA synovialmembrane nor the cytokine profile are pathognomonic forthe disease: synovial lining cell hyperplasia and cellularinfiltrates can be found in many disorders (12), and theproinflammatory cytokines tumor necrosis factor (TNF ) and interleukin-1, which are considered hallmarksof RA, can also be found in other disorders, such aspsoriatic arthritis (13), in which TNF blockade can be usedas successfully as in RA (14). Despite these similarities, thetypical bone and cartilage destruction distinguishes RAfrom most other joint disorders. The only other distinguish-ing characteristic of RA is autoimmunity.Is there a link between autoimmunity and RAjoint destruction? Although the role of various auto-immune responses in the pathogenesis of RA is cur-rently unknown, RF, anticollagen, anti-RA33, and anti-citrulline antibodies are all present in synovial fluid. Thisfinding suggests local production (15–18). Therefore, itis tempting to hypothesize that autoimmune responsesin the joint might fuel the inflammatory process andcontribute to destruction by virtue of immune complexformation and cell-mediated reactivities.* Indeed, RAsynovial fluid shows low complement levels (20), andimmune complexes have been found in the fluid as wellas trapped in cartilage of RA patients (21,22). Cellularreactivity against immunoglobulins, collagen, and RA33/hnRNP-A2 has been observed among T cells from RA(and less so from non-RA) patients (23–25). Such auto-reactive T cells appear to be almost exclusively of theTh1 type (25), like those found in the RA synovialmembrane (26–28).One piece of evidence for the pathogenicity inRA of autoimmune responses in general and autoanti-bodies in particular is the passive transfer of disease byautoantibodies or autoreactive cells and/or the inductionof autoimmunity and autoimmune disease by the puta-tive autoantigen (29). Although only collagen amongknown antigens targeted in RA (30) can elicit erosivearthritis in experimental animal models, RF can beobserved in the destructive arthritis of MRL/lpr mice(31) and, interestingly, anti-RA33 has been found in seraof both MRL/