(Auto)immunity to cartilage matrix proteins - a time bomb?

Geng and colleagues consolidate and detail the role of cartilage oligomeric matrix protein (COMP) as a (potential) autoantigen in experimental and human arthritis, a finding also supported by the detection of COMP fragments and anti-COMP antibodies in rheumatoid arthritis serum and/or synovial fluid and by synovial B-cell responses against COMP. The reactivity to COMP is yet another example of how, in addition to collagen II and the large aggregating proteoglycan, cartilage-specific proteins can induce arthritis and contribute to autoimmunity. Progression of cartilage damage and degradation in disease is believed to promote the autoimmune reaction to cartilage components. However, Geng and colleagues show that anti-COMP mAbs bind in vivo to undamaged cartilage, as previously also observed for anti-collagen II antibodies. Whether this autoimmunity also involves modifications of cartilage matrix proteins, such as citrullination, remains to be further investigated. Latent, subpathogenic (auto)immune reactions directed against cartilage matrix proteins may thus eventually contribute to the outbreak of human arthritis.

In a previous issue of Arthritis, Research & Th erapy, Geng and collaborators from the laboratory of Rikard Holm dahl expand on the topic of cartilage oligomeric matrix protein (COMP) as an autoantigen in arthritis [1]. Th ey convincingly show that mice immunized with mouse recombinant full-length COMP or COMP fragments produce a rapid and strong IgG response to these proteins/fragments beginning on day 14. Th e response continues over day 35, given that onset of COMP-induced arthritis occurs on days 36 to 38 [2] and peaks on day 50.
Th e authors then generated mAbs by immunizing mice with the native form of recombinant rat COMP and by subsequent application of the classic hybridoma technique [3], of which 18 mAbs were cross-reactive with mouse COMP and were further analyzed. Th ey next showed that some of the mAbs against COMP bound to cartilage in vivo following injection into neonatal mice, and could thus be found in the right place for the induction of the pathogenetic cascade. After thorough screening of the epitope specifi cities of the diff erent anti-COMP anti bodies (with four antigenic domains in COMP, but a preferential response to the epidermal growth factor-like domain), the authors fi nally showed that combinations of the mAbs were capable of inducing arthritis upon in vivo injection, either in combination with sub-arthritogenic doses of a mAb directed against collagen II or, strikingly, just by themselves. In the latter case, however, the arthritis was less severe.
In conjunction with previous reports from the same group [2,4], these results consolidate and detail the role of COMP as a (potential) autoantigen in experimental and human arthritis -a fi nding supported not only by detection of COMP fragments and anti-COMP antibodies in rheumatoid arthritis serum and/or synovial fl uid, but also by synovial B-cell responses against COMP.
Th e reactivity to COMP is a further example, next to collagen II [5] and the large aggregating proteoglycan in cartilage [6], of how cartilage-specifi c proteins can induce arthritis and contribute to autoimmunity. Progression of damage to and degradation of the cartilage in disease is generally believed to promote the autoimmune reaction to cartilage components. However, Geng and colleagues' present paper shows that anti-COMP mAbs bind in vivo to undamaged cartilage, as previously also observed for anti-collagen II antibodies [7]. Whether this autoimmunity also involves modifi cations of cartilage matrix proteins, such as citrullination, remains to be further investigated.
Th e potential importance of autoimmunity to cartilage matrix proteins is further supported by the stunning and somewhat unexpected success of pure anti-B-cell therapy with, for example, anti-CD20 antibodies, in view of

Abstract
Geng and colleagues consolidate and detail the role of cartilage oligomeric matrix protein (COMP) as a (potential) autoantigen in experimental and human arthritis, a fi nding also supported by the detection of COMP fragments and anti-COMP antibodies in rheumatoid arthritis serum and/or synovial fl uid and by synovial B-cell responses against COMP. The reactivity to COMP is yet another example of how, in addition to collagen II and the large aggregating proteoglycan, cartilage-specifi c proteins can induce arthritis and contribute to autoimmunity. Progression of cartilage damage and degradation in disease is believed to promote the autoimmune reaction to cartilage components. However, Geng and colleagues show that anti-COMP mAbs bind in vivo to undamaged cartilage, as previously also observed for anti-collagen II antibodies. Whether this autoimmunity also involves modifi cations of cartilage matrix proteins, such as citrullination, remains to be further investigated. Latent, subpathogenic (auto)immune reactions directed against cartilage matrix proteins may thus eventually contribute to the outbreak of human arthritis. decade-long pathogenetic hypotheses favoring T-cell dominance [8,9]. Strikingly, such immune activation and/ or (auto)immunity is detectable both systemically and in the joint already before the onset of disease or early in experimental arthritis [10,11] and human arthritis [12,13], suggesting that these responses may be mounted before or in parallel to the fi nal pathogenetic cascade. Latent, subpathogenic (auto)immune reactions directed against cartilage matrix proteins may thus be a time bomb eventually contributing to the outbreak of human arthritis.
In summary, the data from Geng and colleagues provide further evidence and detailed antibody specifi city information about the contribution of COMP to arthritis. Th ey prepare the ground for future studies not only relevant to rheumatoid arthritis but also to other autoimmune diseases, given that some of the mAbs are not only cross-reactive between mouse and rat but also with human. We are looking forward to seeing the future fruits of this favorable research.