To reassess the role of autoantibodies to type II collagen in the pathogenesis of diseases, we studied antibodies from patients with rheumatoid arthritis (RA) and from patients with relapsing polychondritis for species specificity and collagen type specificity, using an improved enzyme‐linked immunosorbent assay. Antibodies were found in the sera of 15% of the RA patients and 50% of the relapsing polychondritis patients, as well as in the cartilage of 69% of the RA patients examined. Reaction with both homologous and heterologous type II collagens was common. Analysis of 19 selected RA sera revealed that autoantibodies were generally associated with specific antibodies to some species of heterologous type II collagen. In contrast, antibodies found in 4% of the non‐RA controls were specific for either bovine or chick type II collagen. These findings indicate that autoantibody formation in RA and relapsing polychondritis may occur as a result of an immune response to heterologous type II collagen. However, since RA and relapsing polychondritis patient sera differed in their reactivity with the cyanogen bromide‐digested peptides, it is possible that the clinical manifestation of collagen autoimmunity might be influenced by the epitope specificity of the antibodies.

Well over six decades since its first description, the Rheumatoid Factor (RF)—autoantibodies recognizing Fc (constant) portion of IgG through their own Fab (antigen binding variable segments)—is believed to have come of age. Autoimmune pancreatitis is a unique form of pancreatitis, biologically characterized by an elevated serum IgG4 concentration. Given the fact that IgG4 myeloma proteins can act as RF, we initially hypothesized that IgG4 in autoimmune pancreatitis might do likewise, hence potentially contributing to disease pathogenesis. Indeed Western blotting clearly showed that IgG4 binds to IgG1 κ, IgG2 κ, IgG3 κ myeloma proteins, as well as to IgG Fc, in line with a typical RF activity. Further experiments however unraveled the unexpected fact that unlike hitherto known RF, IgG4 does not engage IgG Fc through its Fab, but its very own Fc. These data therefore collectively describe a Novel RF (NRF) in autoimmune pancreatitis. In the future, the relevance of NRF, beyond autoimmune pancreatitis, in both diagnosis/prognosis as well as pathophysiology of autoimmune and other systemic diseases where IgG4's role seems paramount, needs to be systematically assessed.

To determine whether sex-linked factors are important in the susceptibility and resistance of cynomolgus monkeys to type I1 collagen-induced arthritis, we immunized 5 males and 10 females with chick type I1 collagen (CII) and studied their clinical and immune responses. All 10 females developed overt arthritis and produced antibodies to CII and cross-reactive antibodies to monkey type I1 collagen (MkII). In contrast, only 1 male monkey developed arthritis, which was transient and mild in severity. Examination of antisera obtained from the male monkeys disclosed high titers of antibody to CII, but little antibody to MkII. The cyanogen bromide peptide recognition patterns varied markedly from animal to animal, implying that monkeys are capable of recognizing multiple arthritogenic determinants on CII.The observations that certain strains of rats (I-4), mice (5-9), and nonhuman primates (10,ll) develop a destructive polyarthritis after immunization with heterologous or homologous type I1 collagen have stimulated interest in the potential role that autoimmunity to collagen might play in the pathogenesis of human arthritic diseases, such as rheumatoid arthritis and relapsing polychondritis. In rodents, susceptibility to collagen-induced arthritis is closely associated with the formation of antibodies that cross-react with homologous type 11 collagen (2,4,7,12). The critical role that anticollagen antibodies play in the initiation of collagen-induced arthritis has been underscored by studies showing that this disease can be passively transferred to naive animals with purified antibodies to heterologous or homologous type I1 collagen (13-15). The recent observation that collagen-induced arthritis can be passively transferred to mice with anti-type 11 collagen antibodies purified from rheumatoid serum supports the hypothesis that anticollagen antibodies might be pathogenic in humans (16).Investigations into the location of arthritogenic epitopes on the cyanogen bromide (CB) peptides of chick type I1 collagen (CII) have shown that the CBI 1 peptide is of particular importance (17). This conclusion stems from the discovery that antibodies produced by arthritic DBAil mice immunized with native or denatured CII react most intensely with CB 11 and only variably with the other CB peptides. When tested for cross-reactivity with mouse type I1 collagen (MsII), antibodies to CB I 1 consistently showed the

CII from several species and disparate regions of the CII molecule were able to induce autoantibody-mediated arthritis in outbred cynomolgus monkeys. The strong anti-MkCII response suggests that epitope spreading or induction of broad-based CII cross-reactivity occurred in these animals. Autoantibody levels to MkCII were higher in CIA-susceptible monkeys than in resistant monkeys, despite comparable antibody levels in response to the various immunizations of CII. These results closely parallel the type of anticollagen responses found in sera from rheumatoid arthritis patients. Perhaps this can be accounted for by similar major histocompatibility complex heterogenicity associated with an outbred population, or maybe this is a primate-specific pattern of reactivity to CII.