1. The neutral collagenase released into the culture medium by explants of rheumatoid synovial tissue has been purified by ultrafiltration and column chromatography, utilising Sephadex G-200, Sephadex QAE A-50 and Sephadex G-100 superfine.2. The final collagenase preparation had a specific activity against thermally reconstituted collagen fibrils of 312 l g collagen degraded min-' mg enzyme protein-', representing more than a 1000-fold increase over that of the active culture medium.3. Electrophoresis in polyacrylamide disc-gels with and without sodium dodecyl sulphate showed the enzyme to migrate as a single protein band. Elution experiments from polyacrylamide gels and chromatography columns have provided no evidence for the existence of more than one collagenase.4. The molecular weight of the enzyme, as determined by dodecylsulphate-polyacrylamide gel electrophoresis, was 33 000.5. Data obtained from studies with the ion-exchange resin and from gel electrophoresis in acid and alkaline buffer systems suggested a basically charged enzyme.6. It did not hydrolyse the synthetic collagen peptide Pz-Pro-Leu-Gly-Pro-D-Arg and non-specific protease activity was absent.7. The collagenase attacked undenatured collagen in solution at 25 "C resulting in a 58 % loss of viscosity and producing the two characteristic products TCA (3/4) and TCB (1/4).8. At 37 "C and pH 8.0 both reconstituted collagen fibrils and gelatin were degraded to peptides of less than 10000 molecular weight.9. As judged by the release of soluble hydroxyproline peptides and electron microscopic appearances the enzyme degraded human insoluble collagens derived from tendon and soft juxta-articular tissues although rates of attack were less than with reconstituted fibrils.10. The data suggests that pure rheumatoid synovial collagenase at 37 "C and neutral pH can degrade gelatin, reconstituted fibrils and insoluble collagens without the intervention of non-specific proteases.11. The different susceptibilities of various collagenous substrates to collagenase attack are discussed.Knowledge of the action of purified human collagenase is central to our understanding of its role in vivo and to any considerations that its specific inhibition might offer a therapeutic approach to the management of diseases characterised by disordered collagen metabolism. Since the first purification of an animal collagenase from tadpole tailfin cultures by Nagai et al. in 1966 [l] a number of approaches to the purification of animal collagenases have been adopted which include gel filtration [2 -41, ion-exchange chromatography [5,6] and affinity chromatography [7-101. In the present study rheumatoid synovial collagenase from tissue culture medium has been purified by a procedure involving a series of chromatographic steps designed to facilitate a quick and effective separation of the enzyme. Since previous attempts to study the properties of the synovial enzyme have largely utilised impure preparations, or have been incomplete Eur.

Immunolocalization studies of rheumatoid tissues employing specific synovial collagenase antibody have demonstrated immunoreactive enzyme at the cartilage/pannus junction. Collagenase was not detected in chondrocytes or the cartilage matrix remote from the resorbing front, and relatively little enzyme was observed in the hypertrophied synovial membrane itself. These observations directly support the idea that synovial collagenase participates in cartilage erosion in rheumatoid arthritis.A fundamental feature of the joint damage that occurs in rheumatoid arthritis is progressive erosion of the articular cartilage by the hypertrophic synovium or pannus (1,2). This process involves both proteoglycan and collagen removal, but because the turnover of cartilage proteoglycan is significantly greater than the turnover of collagen, it is likely that irreversible cartilage damage is attributable to collagen breakdown. The observation that rheumatoid synovial tissue in culture produces large quantities of neutral collagenase (3) has led to speculation on the role of this enzyme in the joint erosion of rheumatoid arthritis (43). Highly purified synovial collagenase has been shown to degrade cartilage collagen (6,7), insoluble collagen fibers (8), and slices of human articular cartilage (9), and it is relatively unaffected by cartilage proteoglycans (10). However more definitive evidence of collagenase involvement has awaited the demonstration and localization of the enzyme directly in rheumatoid tissues. This report presents our findings with a specific anticollagenase antibody for immunolocalization studies of joint tissues in rheumatoid arthritis. MATERIALS AND METHODS MaterialsDulbecco's modified Eagle's medium containing HEPES buffer and 0.85 g/L sodium bicarbonate was supplied by Flow Laboratories, Irvine, Scotland. Rabbit anti-(sheep serum proteins) serum was purchased from Dakopatts, A/S, Denmark. FITC-labeled rabbit antibodies to sheep IgG were acquired from Wellcome Reagents Ltd, Beckenham, England, and unlabeled rabbit antibodies to sheep IgG were purchased from Miles-Yeda Ltd, Rehovot, Israel. Rheumatoid Synovial CollagenaseRheumatoid synovial tissue removed by synovectomy was cut into fragments and explanted into culture flasks containing Dulbecco's modified Eagle's medium with HEPES buf-

Chronic monoarticular allergic arthritis was induced in BALB/c mice using methylated BSA as antigen and Freund's complete adjuvant, together with Bordetella pertussis as a secondary adjuvant. The optimum conditions for induction of chronic persistent arthritis and the histological characteristics of the arthritic lesion are described. Both the synovitis and erosive progression of the arthritis could be suppressed by daily treatment with prednisolone (1-10 mg/kg) or dexamethasone (0.5-2.5 mg/kg) for 4 weeks commencing 2 weeks after the induction of arthritis. In contrast, daily treatment with the non-steroidal anti-inflammatory agents ibuprofen (50-100 mg/kg), flurbiprofen (1-9 mg/kg) or indomethacin (0.1-3 mg/kg) had no significant effect on either the synovitis or erosions as judged histologically. Synovial fluid differential leukocyte counts were altered by treatment with ibuprofen and indomethacin but not by flurbiprofen or the corticosteroids. The suppressive effect of the corticosteroids was not due to either suppression of antibody synthesis or alteration of the number of leukocytes in the peripheral circulation.

The effects of treatment with second-line antirheumatic drugs and cytotoxic agents on the severity of experimental monoarticular arthritis in BALB/c mice have been investigated. The arthritis was assessed histologically in terms of synovitis and erosions of cartilage and bone. Azathioprine (20 mg/kg) and sulphasalazine (10-30 mg/kg oral; 30-100 mg/kg i.p.) produced significant suppression of synovitis and erosions when administered daily for 4 weeks commencing 2 weeks after induction of the arthritis. Dapsone (1-10 mg/kg) and to a lesser extent methotrexate (2 mg/kg) produced some suppression of erosive disease when administered daily for 4 weeks commencing 2 weeks after induction of the arthritis but this failed to reach statistical significance. Chloroquine, D-penicillamine and sodium aurothiomalate all failed to have any effect on the disease with any of the treatment schedules used. Auranofin had no effect on the disease when treatment commenced 2 weeks after intra-articular injection but produced variable suppression at high doses when administered from the day of intra-articular injection.

1. An immunoglobulin G antibody has been isolated from the sera of sheep immunised with purified synovial collagenase obtained from tissue culture medium of rheumatoid synovium.2. In double diffusion gels and on immunoelectrophoresis the antibody gave a single precipitin line when run against crude and pure synovial enzyme preparations. Selective adsorption of collagenase from concentrated synovial tissue culture medium eliminated the precipitin reponse. N o immunoprecipitation was observed with whole serum or synovial tissue homogenates in double diffusion gels.3. Incubation of immune immunoglobulin G with synovial collagenase resulted in enzyme inhibition and the formation of immunoprecipitates. Fab' fragments of the antibody were similarly inhibitory but gave no precipitin response. Non-specific neutral proteinase activity released by synovial explants was not inhibited by antibody.4. Human collagenases derived from gastric mucosa, cornea, and skin were inhibited by the antibody and produced precipitin lines on immunodiffusion showing a reaction of complete identity with the synovial enzyme. Granulocyte collagenase preparations did not react with the antibody.5. The non-human collagenases derived from rabbit cornea and dog gingiva were inhibited by the antibody but those from guinea pig skin and the bacterium Clostvidium histolyticum were not. None of these collagenases gave precipitin responses on iminunodiffusion against the antibody.6. Synovial enzyme preparations inhibited or denatured by chemical reagents retained their capacity to react with the antibody, whereas inhibition by the serum proteins az-macroglobulin or PI -anticollagenase masked the antigenicity of the enzyme as judged by immunodiffusion experiments.These observations and their relevance to the use of the antibody in elucidating the physiological role of collagenase are discussed.Since the first description of a vertebrate neutral collagenase [ 1 ] many similar enzymes have been isolated from a variety of normal and diseased human tissues [2-41. Evidence for the involvement of collagenase in the catabolism of collagen is substantial, but precise definition of the enzyme's role in vivo is lacking, especially in conditions such as rheumatoid arthritis.The availability of a specific antibody to human collagenase would provide an important tool with which to obtain a more exact interpretation of its functional role in physiological and pathological conditions. Use of such an antibody both as a specific collagenase inhibitor in biological studies and as a specific label in immunofluorescent studies might permit a better understanding of the importance of neutral collagenase in those diseases characterised by disordered collagen metabolism. To this end we have immunised sheep with purified rheumatoid synovial collagenase and isolated a specific anti-collagenase immunoglobulin from the immune-sera. We describe here the specificity and characterization of the antibody so produced. MATERIALS AND METHODS Rheumatoid Synovial CollagenaseThe enzyme was ...