1356 NATIVE TYPE I1 COLLAGEN-INDUCED ARTHRITIS I N THE RAT The Effect of Complement Depletion by Cobra Venom Factor KEITH MORGAN, ROY B. CLAGUE, MARY J . SHAW, STEPHEN A . FIRTH, TREVOR M. TWOSE, and P. J . LENNOX HOLT Antibodies to native bovine type I t collagen may be important in initiating arthritis in rats immunized with this antigen. The cross-reactivity of these antibodies with native rat type I1 collagen was higher in rats that developed arthritis than in those that did not. Depletion of serum C3 levels by cobra venom factor delayed the onset of arthritis until C3 levels were returning to normal; therefore, complement may he involved in initiation of the arthritis, and this arthritis may be an example of an immune complex-mediated disease. An inflammatory polyarthritis can he induced in rats (1-3) and mice (4) after the intradermal injection of heterologous or homologous native type 11 collagen emulsified in complete or incomplete Freund's adjuvant. In view of the onset of arthritis at least 10 days after primary immunization of rats, the presence of From the Department of Kheumatology, University of Manchester Medical School, Oxford Road. Manchester. England and Imperial Chemical Industries Ltd. Mereside. Alderley kirk. Macclesfield, Cheshire. Supported by grants from the Arthritis and Rheumatism Council, Medical Research Council. and the North Western Kegional Hospital Board. Keith Morgan. BSc. PhD: Senior Scientific Officer; R o y 13. Clague. MD, MRCP (UK): Consultant Kheumatologisl: Mar; J . Shaw, BSc: Research Technician; Stephen A . Firth, HSc: Research Technician, Department of Rheumatology. University of Miinchester Medical School; Trevor M. Twose. HSc. Phl): Research I3ioScientist, Imperial Chemical Industries. Lld. Cheshire; P. J. I.ennox Holt, MD, FRCP: Consultant Rheumatologist, Department of Rheumatology, University of Manchester Medical School. Address reprint requests to Dr. K. Morgan, Department of Rheumatology, University of Manchester Medicid School. Oxford Road, Manchester MI3 9PT, England. Submitted for publication Octoher 24. 1980: accepted in revised form March 19, 1981. Arthritis and Rheumatism, Vol. 24, Nn. 1 1 (November 1981) high serum antibody levels to native type I1 collagen before the onset of arthritis in rats (9,and the finding of higher serum antibody levels to native type I1 collagen in arthritic than in nonarthritic rats (3,6,7), we suggest that these antibodies may have a role in the induction of the arthritis. Previous investigators (8) have noted a lack of species specificity of the antibodies produced by immunization with collagen. There have been no previous studies in this model to determine the crossreactivity of the serum antibodies produced by heterologous collagen with native rat type I1 collagen. We now present further evidence of the importance of humoral immunity to native type I1 collagen in the induction of this arthritis. First, we show that serum antibodies to native bovine type 11 collagen cross-react with native rat type I1 collagen to a greater extent in arthritic than in nonarthritic rats. Second, we show that rats treated with cobra venom factor (CVF) have a delayed onset of arthritis associated with depleted levels of complement ( C 3 ) . A possible pathogenetic mechanism for native type I1 collagen-induced arthritis is discussed. MATERIALS AND METHODS Rats. Outbred, female Sprague Dawley rats (150-250 gm) from the Manchester Medical School Closed Colony were used. They were given food and water ad libitum. Collagens. Native bovine type I1 collagen was extracted from articular or nasal cartilage by pepsin solubilization after previous treatment wth 2M magnesium chloride (9). Native pepsin-soluble rat type I1 collagen was extracted from a transplanted rat chondrosarcoma (10) by the same method. The collagens appeared pure for type by polyacrylamide gel electrophoresis and were negative for uronic acid, so there seemed to be no proteoglycan contamination (1 1). NATIVE TYPE I1 COLLAGEN-INDUCED ARTHRITIS The collagens were stored in a lyophilized state and dissolved in 0.45M NaCII0.02M Tris buffer (pH 7.5) at a concentration of 1 mg/ml before use. Native bovine type I1 collagen was emulsified at a ratio of 1:1 with Freund’s incomplete adjuvant (ICFA, Miles Laboratories Ltd). Cobra venom factor. Cobra venom factor (CVF) was prepared by DEAE-cellulose chromatography followed by Sephacryl S200 chromatography from the lyophilized venom of Nuju naju (Thailand Cobra, Sigma Cat. No. V9125) by the method described by Ballow and Cochrane (12). On polyacrylamide gradient gels (Pharmacia), the protein migrated as a single band (MW 190,000 daltons). This is the low molecular weight anticomplementary factor, not the high molecular weight factor described by Ballow and Cochrane (12), which has a molecular weight of almost lo6 daltons. Injection of the purified CVF into a rat caused rapid depletion of complement to levels that were not detectable in a standard hemolytic assay. To inactivate any phospholipase A2 activity, the cobra venom factor was incubated with 0.2 mM p-bromophenacylbromide (PBPB) at pH 8.0 for 30 minutes at 37°C (13) and the PBPB was subsequently removed by chromatography on Sephadex G25 (PD 10 mini-column, Pharmacia). This CVF was used for the rats in group 4. Treatment of rats. Experiment I: cross-reuctivity. Rats were injected intradermally on the back at several sites with a total of 1 ml of emulsion that contained 500 pg of native bovine type I1 collagen (day 0) followed by a booster dose of 500 pg of the antigen alone intraperitoneally on day 21. The rats were bled (= 1 ml) by cardiac puncture under light ether anesthesia on days 14, 21, 28,42, 56, and 70 after primary immunization. The blood was allowed to clot for 1 hour at room temperature, was centrifuged at 1500g for 7 minutes, and the serum was removed for storage at -70°C. Class-specific antibodies to native type I1 collagen (rat or bovine) were detected in sera by radioimmunoassay as described below. Rats were observed daily for the onset of arthritis and nonarticular lesions. Experiment 2: treatment with CVF. Rats were immunized as in experiment 1 and bled on day 0 (day of immunization) and various days thereafter until day 28. Immunized rats were injected intraperitoneally with either 5 units (250 111) of CVF or 250 p1phosphate-buffered saline (PBS; control group) in the morning and 6 hours later on days 9 and 12. Rats were examined daily for the onset of arthritis. Radioimmunoassay. Class-specific IgM and IgG serum antibodies to native bovine or rat type I1 collagen were measured by a solid-phase double-antibody radioimmunoassay as previously described (3). For the crossreactivity experiment, sera were measured for antibodies to both antigens in one assay. The results were expressed as milligrams of antigen-specific IgM and IgG antibody per liter of serum. Measurement of C3 levels. Serum C3 levels were measured by single radial immunodiffusion (14). The test sera were diluted 1 to 5 with PBS and allowed to diffuse into 1.5% agarose in PBS containing a 1 to 20 dilution of anti-rat BIC/BIA (C3) (Nordic Immunochemicals Ltd). The C3 levels were expressed as a percentage of the day 0 levels for each rat. Statistics. Experiment I: cross-reactivity. The anti- 1357 body levels to native rat type 11 collagen were plotted against the antibody levels to native bovine type I1 collagen in the same serum samples. Because the standard deviation of the anti-rat antibody levels increased in the samples that had higher anti-bovine antibody levels, a weighted linear regression was performed on the data to give an estimate of the slope (15). The slopes for arthritic rats were compared to the slopes for nonarthritic rats at each sample time by the Student’s t-test (16). Experiment 2: CVF. Antibody levels to native bovine type I1 collagen of CVF-treated and PBS-treated groups were compared after logarithmic transformation by a twoway analysis of variance in repeated measures (across times). When there were unequal numbers of rats in the two groups, an “unweighted means” analysis was used (17). The Student’s t-test was used to compare the groups at each time separately if the analysis of variance indicated a significant interaction with time and at later times if the numbers were reduced because rats had died under anesthesia. RESULTS Experiment 1: cross-reactivity. In a group of 34 rats immunized by native bovine type I1 collagen emulsified in ICFA, 22 (65%) developed an inflammatory polyarthritis in one or more hind limbs (onset: median day 12, range 10-18). Five rats developed swellings related to the cartilaginous endplates in the distal region of the tail, and the ears of 5 rats became erythematous and swollen, but none of these lesions were observed before day 30. When the levels of antibodies to native rat type I1 collagen were plotted against the levels of antibodies to native bovine type I1 collagen in the same sera, a straight line through the origin was produced (e.g., day 14 samples, Figure 1). Thus, the levels of antibodies to native rat type I1 collagen were a fixed proportion of the antibody levels to native bovine type I1 collagen, and the slope of this line estimates this proportion. Therefore, a proportion of the antibodies raised to native bovine type I1 collagen may cross-react with native rat type 11 collagen. For each sample date, the levels of IgM and IgG antibodies to rat collagen were plotted against the antibody levels to bovine collagen in the same samples for arthritic and nonarthritic rats separately, and the slopes were compared. For IgM antibodies at days 14 and 21, the slope was greater in arthritic rats than in nonarthritic rats, but this difference was not statistically significant (Table 1). In regard to the IgG antibodies, the slopes plotted for arthritic rats were greater than those for nonarthritic rats (except for days 56 and 70). These differences were statistically significant on day 14, P < 0.05; day 21, P < 0.001; and day 28, P < 0.02 MORGAN ET AL 1358 Table 2. Serum C3 levels in rats after treatment by cobra venom factor* C3 levels (96 of initial values) 0 Normal rats I 2 3 Rats immunized 9 days previously by type 11 collagen ICFA I 2 3 4 5 f **: ** 100 100 Hours after first CVF 3 6 24 loo 25.4 21.4 57.1 3.1 3.8 9.2 3.8 4.6 100 100 100 loo 100 17.9 15.7 60.8 22.5 36.9 12.5 5.7 40.5 5.4 3.1 5.4 4.3 3.8 I .6 2.3 4.1 * Five units given after initial bleed followed by 5 units after 6 hours. 30 60 120 90 150 IgG to native bovine type I I collagen (rng.1-l 1 Figure 1. Levels of serum IgG antibodies to native rat type I1 collagen plotted against IgG antibodies to native bovine type II collagen in the same sera (day 14 after immiiniration). (Table I ) . Thus, arthritic rats had serum antibodies, raised by immunization to native bovine type 11 collagen, that had greater cross-reactivity to rat type I1 collagen than did nonarthritic rats. Experiment 2: treatment with CVF. Cobra venom factor given intraperitoneally to 3 rats caused a reduction in serum C3 levels 3 hours after injection (Table 2). At 6 hours, the levels had fallen to less than 10% of their original values, and at 24 hours, they were less than 5% of their original values. Similar results were found in a group of 5 rats injected with native bovine type 11 collagen 9 days before. Arthritis has never been observed before day i0 in these SpragueDawley rats (personal observations); therefore, we decided to inject CVF on day 9 in order to deplete serum C3 levels before the expected onset of arthritis. Effect of CVF on the onset of native type I1 collagen-induced arthritis. Table 3 shows the number of rats that developed arthritis and the time of onset of the arthritis, after primary immunization with native type 11 collagen in rats treated with CVF or PBS. In all four groups, treatment with CVF caused a Table 3. Effect of cobra venom factor on incidence and onset of native type I1 collagen-induced arthritis Table 1. Comparison of slopes for arthritic and nonarthritic rats* Day IgM 14 21 IgG 14 21 28 42 56 70 Nonarthritic (n = 12)t Arthritic (n = 22) 0.89 2 0.02 0.77 ? 0.02 0.91 2 0.02 0.83 ? 0.03 0.81 2 0.04 0.80 5 0.04 0.68 f 0.05 0.63 2 0.06 0.61 5 0.07 0.66 5 0.08 Treatment 0.91 2 0.02 0.96 f 0.02 0.85 t 0.04 0.68 ? 0.04 0.61 2 0.04 0.55 2 0.04 Slope comparison$ i =: -0.6 f z: -1.38 f =: -2.35 i == -3.66 f =: - 71.65 f =: f = f =: -0.80 0.1 I - I .47 NS NS P < 0.05 P < 0.001 P < 0.02 NS NS NS * Slopes generated by plotting levels of antibodies to rat type I1 collagen against those to bovine type 11 collagen in the same samples. t Slope f standard error. $ N S = not significant. Group 1 PBS CV F Group 2 PBS CVF Group 3 PBS CV F Group 4 PHS CVF* Total PHS CVF No. of rats No. with arthritis Time of onset after immunization (days) 12 9 7 2 12-14 22 5 S 3 5 10-13 17-22 5 6 3 2 11-18 20 16 21 9 12 11-14 19-22 38 22 41 21 10-18 17-22 * Group 4: this CVF had phospholipase-A2 activity removed as described. NATIVE TYPE I1 COLLAGEN-INDUCED ARTHRITIS 1359 treated and PBS-treated rats, and there was no statistically significant difference between these groups. The IgM and IgG antibody levels to bovine type I1 collagen were measured less frequently in the other groups of rats, but again there was no difference between the levels in CVF-treated and PBS-treated rats. Relationship between C3 levels and the onset of arthritis. The onset of arthritis was delayed in CVFtreated rats (median, day 21) compared with PBStreated rats (median, day 12). Serum C3 levels were depleted in CVF-treated rats from day 9 and returned to normal between days 17 and 21. Thus, the onset of arthritis in the CVF-treated rats occurred only when the serum C3 levels had started to return to normal. Ua", alter lmmvnlralan Figure 2. Levels of serum 1gM and IgG antibodies to native bovine = type 11 collagen and serum C3 levels in PHS-treated rats. ~e IgM; a= IgG: C - -O = C3. delay in the onset of arthritis in comparison with PBStreated rats. The delay varied between groups. but there was no overlap of onset between PBS-treated and CVF-treated rats in any one group. The overall incidence of arthritis was similar in PBS-treated (22 of 38) and in CVF-treated rats (21 of 41). Effect of CVF on serum C3 levels and the levels of serum antibody to native type I1 collagen. The rats from group I (Table 3) were bled at frequent intervals after primary immunization with native type I1 collagen. Serum IgM and IgG antibody levels to this antigen and serum C3 levels were measured in these animals (Figures 2 and 3). There was no significant reduction in serum C3 levels of PBS-treated rats (Figure 2). However, serum C3 levels in CVF-treated rats were greatly reduced after CVF injection on day 9 and remained low for several days (Figure 3). By day 17. the C3 levels had returned to 75-95% of their initial values in 4 rats. although the levels remained low for longer in the others (mean 36.9% ? 12.4% on day 17). By day 19, the levels were high in all rats (mean 89.6% ? 5.3%), and by day 21, normal levels had been restored. The serum C3 levels in all the other groups of rats were similar; all CVF-treated rats showed a marked reduction in C3 levels with subsequent recovery, but no PBS-treated rats showed significant reductions in C3 levels. The 1gM and 1gG antibody levels to native bovine type I1 collagen were very similar in both CVF- DISCUSSION An inflammatory polyarthritis can be induced in some rats after immunization by heterologous (bovine, human, or chick) or homologous native type I1 collagen (1-3). Previous studies (3,6,7) have demonstrated that serum antibody levels to the native type I1 collagen used in the immunization procedure are higher in arthritic than in nonarthritic rats. N o previous studies on the cross-reactivity of these antibodies with rat native type I1 collagen have been reported. We have extracted native collagen from a transplanted rat chondrosarcoma, and this collagen has previously been CVF - CVF nay, alle. Imm""tZdll0" Figure 3. Levels of serum IgM and IgG antibodies to native bovine type 11 collagen and serum C3 levels in CVF-treated rats. = IgM; " 4 = IgG; c - a= C3. 1360 biochemically identified as cartilage type 11 collagen (10). We have shown that antibodies raised in rats to native bovine type I1 collagen cross-react to a large extent with rat native type I1 collagen. The extent of this cross-reactivity varied; it was greatest soon after immunization. More importantly, we have shown that the cross-reactivity of these antibodies with native rat type I1 collagen was greater in arthritic than nonarthritic rats. Further evidence that this experimental arthritis is induced by an autoimmune response to a component of rat articular cartilage is supplied by: I ) the demonstration that serum antibodies from rats with arthritis induced with native type I1 collagen react with homologous native type I1 collagen, and 2) the previous demonstration ( I ) that arthritis can be induced by immunization with homologous native type I1 collagen. Injection of CVF, given at least 24 hours before the expected onset of arthritis and followed by further injections 3 days later, delayed the onset of native type I1 collagen-induced arthritis by about 9 days in comparison to PBS-treated rats, but had no apparent effect on the eventual incidence of arthritis (Table 3). This effect was demonstrated in several groups of rats and was independent of any possible phospholipase A? activity (Table 3). Although an immunosuppressive effect has been shown to occur if CVF is administered at the time of immunization (18,19), no significant reduction in the humordl immune response to native type 11 collagen occurred when CVF was injected on days 9 and 12 after immunization. Since there was no apparent suppression of antibody formation in the rats treated with CVF, the CVF may not have produced a nonspecific suppression of arthritis by toxic action on cells or by alteration of cellular responses to external stimuli (20) via phospholipase A2 activity. CVF is known to activate the complement cascade and subsequently deplete complement components (21). Serum C3 levels were depleted in the CVF-treated rats between days 9 and 17 and showed a temporal relationship to the delayed onset of the arthritis. Therefore, complement may be directly involved in the induction of this arthritis. Cell-mediated immunity to native type I1 collagen was not measured in CVF-treated rats, but this work is in progress. Treatment with CVF, however, has been shown to have no effect on T cell function in mice according to several in vivo and in vitro tests (22). The role of T lymphocytes in this arthritis is MORGAN ET AL unclear at present, but a T cell response may be required to sustain the synovitis (23). In preliminary histologic studies investigators have shown the predominant infiltration of neutrophils and mast cells within 24 hours of the onset of the arthritis (personal observations). The infiltration of neutrophils in the synovium early in the arthritis has been described in both the rat (7) and mouse (4) models. Thus, neutrophils appear to be an important cellular component early in the genesis of the arthritic lesion. These findings could be explained as follows. The binding of IgM and IgG antibodies to the rat native type 11 collagen in the articular cartilage would form intraarticular immune complexes and subsequently activate the complement cascade (24). The action of the activated complement components, especially C5a, would lead to infiltration of neutrophils (25). mast cells, and other components of the inflammatory response from the blood vessels in the periarticular and articular tissues (e.g., synovium), because the articular cartilage is avascular. Phagocytosis of immune complexes by neutrophils would be accompanied by degranulation of lysosomes and release of hydrolytic enzymes and collagenase accompanied by break-down of the cartilage matrix and collagen (26,27). Thus, the arthritic lesion would be initiated by immune complex formation and would persist until total destruction of the cartilage occurred. Thus, the finding of high levels of serum antibody to native type I1 collagen before the onset of arthritis ( 9 ,the occurrence of higher levels in arthritic than nonarthritic rats (3,6,7), the greater cross-reactivity of these antibodies with homologous native type 11 collagen in arthritic and nonarthritic rats, the predominant early infiltration of neutrophils and mast cells, and the apparent dependence of the induction of the arthritis on the presence of complement components may be evidence of an immune complex-mediated disease. Some of the previous experimental models of inflammatory arthritis (28,29) may be examples of immune complex-mediated diseases (23,30), although in these instances, the antigen must be injected intraarticularly about 14 days after systematic immunization by the antigen in complete Freund's adjuvant. The foreign antigen is trapped in the articular cartilage in these models, whereas in native type I1 collageninduced arthritis, the antigen presumably forms an integral part of the articular cartilage. In these studies, further evidence is provided for the possible importance of the formation of intraar- NATIVE TYPE I1 COLLAGEN-INDUCED ARTHRITIS ticular immune complexes in the induction of arthritis. In addition, we may suggest that humoral immunity to native type I1 collagen in humans (31,32) may be important as a component in the destruction of articular cartilage in some patients with rheumatoid arthritis. ACKNOWLEDGMENTS We are grateful to Ms Jacqueline Weiss and Dr. R. Brown for analysis of the collagen types and for advice on their preparation. We would like to thank Drs. M. Billingham and M. Smith for the gift of the rat chondrosarcoma. We would also like to thank Miss Linda Hunt (Faculty of Medicine Computational Group) for statistical advice. Mrs. Doreen Ward for typing the manuscript, and the Department of Medical Illustration for preparing the figures. REFERENCES 1 . Trentham DE, Townes AS. Kang AH: Autoimmunity to 2. 3. 4. 5. 6. 7. 8. 9. 10. type I1 collagen: and experimental model of arthritis. J Exp Med 146357-867. 1977 Stuart JM, Cremer MA, Dixit SN, Kang AH, Townes AS: Collagen-induced arthritis in rats: comparison of vitreous and cartilage-derived collagens. Arthritis Rheum 22:347-352. 1979 Morgan K , Clague RB. Shaw MJ. Holt PJI,: Native type I1 collagen-induced arthritis in the rat. 1 . Incidence and humoral response to collagen. A n n Rheum Dis 39:285290. 1980 Courtenay JS, Dallman MJ, Dayan AD, Martin A. Mosedale B: Immunisation against heterologous type I 1 collagen induces arthritis in mice. Nature 283:666-668. I980 Clague RB. Morgan K. Shaw M J , Holt PJL: Native type I1 collagen-induced arthritis in the rat. 2. Relationship between the humoral immune response to native type I1 collagen and arthritis. J Kheumatol 7:775-782. 1980 Trentham DE, Townes AS, Kang AH, David JK: Humoral and cellular sensitivity to collagen in type I1 collagen-induced arthritis in rats. J Clin Invest 61 :89-96, I978 Stuart JM, Cremer MA, Kang AH, Townes AS: Collagen-induced arthritis in rats: evaluation of early immunological events. Arthritis Kheum 22:1344-1351, 1979 Steffen C , Timpl K . Wolff I: Immunogenicity and specificity of collagen. V. Demonstration of three different antigenic determinants on calf collagen. Immunology IS: 135-144, 1968 Trelstad KL, Rubin 1). Gross J: Osteogenesis imperfecta congenita: evidence for a generalised molecular disorder of collagen. Lab Invest 36:501-508. 1977 Smith BD, Martin GK. Miller EJ. Swarm R: Nature of the collagen synthesised by a transplanted chondrosarcoma. Arch Biochem Biophys 166:181-186. 1975 1361 1 1 . Bitter T, Muir HM: A modified uronic acid carbazole reaction. Anal Biochem 4:330-334, 1962 12. Ballow M, Cochrane CG: Two anticomplementary factors in cobra venom: hemolysis of guinea pig erythrocytes by one of them. J Immunol 103:944-952. 1969 13. Shaw JO, Roberts MF, Ulevitch KJ, Henson P. Dennis EA: Phospholipase A? contamination of cobra venom factor preparations. Am J Pathol 91:517-528, 1978 14. Mancini G , Vaerman JP, Carbonara AO. Heremans JF: A single-radial-diffusion method for the immunological quantitation of proteins. Proteins of Biological Fluids, 1 Ith Colloquium, Bruges. Edited by H Peeters, Oxford, Pergamon Press, 1964, pp 370-373 15. Edwards DW: Statistical Adjustment of Data. New York, Dover Publishing lnc., 1964, pp 32-33 16. Siegal S: N o n Parametric Statistics for the Behavioural Sciences. International Student Edition. New York, McGraw-Hill Kogakusha Ltd. 1956 17. Viner BJ: Statistical Principles in Experimental Design. Second edition. New York. McGraw Hill Book Co. 1962, pp 599-603 18. Pepys MB: Role of complement in induction of the allergic response. Nature (New Biology) 237: 157-159, I972 19. Pepys MB: Role of complement in induction of antibody production in vivo. J Exp Med 140:126-145, 1974 20. Tu AT: Venoms: Chemistry and Molecular Biology. New York, John Wiley and Sons Inc. 1977, p 52 21. Pepys MB: Studies in vitro of cobra factor and niurine C3. Immunology 28:369-377. 1975 22. Pepys MB: Role of complement in the induction of immunological responses. Transplant Rev 3293-120, 1976 23. Brackertz D. Mitchell GF, Vadas MA. Mackay IR: Studies on antigen-induced arthritis in mice. 111. Cell and serum transfer experiments. J Immunol 118:16451648. 1977 24. Jasin HE, Cooke TD: The inflammatory role of immune complexes trapped i n joint collagenous tissues. Clin Exp Immunol 33:416-424. 1978 25. Cochrane GG: Immune Complex-Mediated Tissue Injury, Mechanisms of Immunopathology. Edited by S. Cohen. PA Ward. K T McCluskey. New York, John Wiley & Sons Inc, 1979, p 42 26. Henson PM: Interaction of cells with immune complexes: adherence, release of constituents and tissue injury. J Exp Med (suppl) 134:114s-135~. 1971 27. Weissmann G: Lysosomal mechanisms of tissue injury in arthritis. N Engl J Med 286:141-146. 1972 28. Dumonde DC. Glynn LE: The production of arthritis in rabbits by an immunological reaction to fibrin. Br J Exp Pathol 43:373-383, 1962 29. Brackertz D, Mitchell GF. Mackay IR: Antigen-induced arthritis in mice. 1. Induction of arthritis in various strains of mice. Arthritis Kheum 20:841-850, 1977 MORGAN ET AL 1362 30. Cooke TD, Hurd ER, Ziff M, Jasin HE: The pathogenesis of chronic inflammation in experimental antigeninduced arthritis. 11. Preferential localisation of antigenantibody complexes to collagenous tissues. J Exp Med 135:323-338, 1972 31. Andriopoulos NA, Mestecky J , Miller EJ, Bradley EL: Antibodies to native and denatured collagens in sera of patients with rheumatoid arthritis. Arthritis Rheum 191613-617. 1976 32. Clague RB, Shaw MJ, Holt PJL: Incidence of serum IgG antibodies to native type I and type 11 collagens in patients with inflammatory arthritis. Ann Rheum Dis 39:201-206, 1980 A 8, R Symposia Issues Beginning in 1983 Arthritis and Rheumatism will no longer publish conference proceedings. However, a conference series to be issued jointly by the American Rheumatism Association and a private publisher is tentatively planned. Planners of future symposia who wish to have their proceedings published should contact Alan Myers, MD, Chair of the ARA General Publications Committee.