ARTHRITIS & RHEUMATISM Vol. 40, No. 8, August 1997, pp 1391-1399 0 1997, American College of Rheumatology 1391 MATRIX METALLOPROTEINASE 13 (COLLAGENASE 3) IN HUMAN RHEUMATOID SYNOVIUM OTSO LINDY, YRJO T. KONTTINEN, TIM0 SORSA, YANLI DING, SEPPO SANTAVIRTA, ARNOLDAS kEPONIS, and CARLOS LOPEZ-OTiN Objective. To show the eventual presence and extent of production of matrix metalloproteinase 13 (MMP-13, or collagenase 3) in rheumatoid synovial tissue samples and extracts, and to assess the inhibition characteristics of recombinant MMP-13. Methods. Immunohistochemical avidin-biotinperoxidase complex staininglmorphometry was used to analyze MMP-13-positive cells in situ. Neutral salt extraction of synovial tissue, electrophoresis of the extract in different buffer systems, and Western blotting were also used. The inhibitory properties of doxycycline, clodronate, pamidronate, and D-penicillamine for recombinant enzyme were determined with a soluble type I1 collagen assay. Results. MMP-13 was detected in fibroblast- and macrophage-like mononuclear cells in the synovial lining and stroma and in vascular endothelial cells. The overall expression of MMP-13 in these cells in the synovial stroma was high in rheumatoid arthritis (86 2 12%) compared with osteoarthritis (17 f 5%) patient samples (P = 0.0027). In a high-pH native electrophoresis gel, immunoreactivity to anti-MMP-1 and anti-MMP-13 were clearly separated, with anti-MMP-13-immunoreactive material migrating faster than anti-MMP-l-immunoreactive material. Finally, in contrast to MMP-1 and Supported by the Finnish Rheumatism Research Foundation, the Emil Aaltonen Foundation, the Center for International Mobility (CIMO), the Rector of the University of Helsinki, the Academy of Finland, and an Evo Clinical Research Grant from the Helsinki University Central Hospital, Helsinki, Finland; the European League Against Rheumatism, Switzerland; and the Comision Interministerial de Ciencia y Tecnologia (SAF94-0892) and Glaxo-Wellcome, Spain. Otso Lindy, MD, Tim0 Sorsa, DDS, Yanli Ding, DDS: University of Helsinki, Helsinki, Finland; Yrjo T. Konttinen, MD, Seppo Santavirta, M,D: Helsinki University Central Hospital, Helsinki, Finland; Arnoldas Ceponis, MD: Invalid Foundation, Helsinki, Finland; Carlos L6pez-Otin, PhD: Universidad de Oviedo, Oviedo, Spain. Address reprint requests to Otso Lindy, MD, Department of Medical Chemistry, PO Box 8 (Siltavuorenpenger 10 A), FIN-00014 University of Helsinki, Helsinki, Finland. Submitted for publication July 23, 1996; accepted in revised form March 17, 1997. MMP-8, MMP-13 was found to be relatively resistant to the inhibitory effects of doxycycline and clodronate in vitro. Conclusion. Due to its localization in synovial tissue, its substrate profile, increased expression, and relative resistance to known MMP inhibitors, MMP-13 is suggested to play a major role in the pathogenesis of tissue destruction in rheumatoid arthritis. The role of collagenases and other matrix metalloproteinases (MMP) in the turnover and catabolism of connective tissue proteins has raised a lot of interest since the initial discovery of tadpole collagenase (1). Today, the MMP superfamily includes 4 different groups of enzymes, namely, collagenases (fibroblast-type collagenase 1, or MMP-1, neutrophil-type collagenase 2, or MMP-8, and collagenase 3, or MMP-13), gelatinased type IV collagenases, stromelysins (matrilysin and macrophage metalloelastase), and membrane-type MMPs (currently 4 known members) (2-4). Several MMPs seem to be involved in the normal turnover of joint tissues, as well as the pathologic destruction of the rheumatoid joint (5). Interstitial collagenases are the only mammalian enzymes capable of cleaving the triple-helical domain of fibrillar or group 1 collagens, which include the major interstitial collagens type I, 11, and 111 (6). The collagenic structures of the inflamed joint may be attacked by either MMP-8 from the neutrophils in the synovial fluid (7) or by MMP-1 secreted by fibroblasts and macrophage-like cells in the synovium and/or pannus (8,9). The recent cloning of human collagenase 3/MMP-13 from mammary carcinoma tissue has brought up some interesting new issues (3). The collagenase cloned from mouse and rat osteoblasts (lO,ll), which was originally thought to be an evolutionary deviation from the MMP-1 of other mammals, was shown to have a counterpart in human tissue. Characterization of MMP-13 showed a substrate specificity favoring degradation of cartilage type I1 collagen and LINDY ET AL 1392 Table 1. Clinical characteristics of the study patients* Patient1 sexiage Diagnosis RA RA RA RA RA RA OA OA OA OA OA OA OA Osteochondritis dissecans Luxation of patella Rupture of meniscus Disease duration (years) >5 7 14 12 10 >5 5 13 5 >10 8 >5 10 5 113 113 Site of synovial biopsy Shoulder MCP joint Elbow PIP joint PIP joint Wrist Knee Knee Knee Knee Knee Knee Knee Knee Knee Knee * RA = rheumatoid arthritis; MCP = metacarpophalangeal; PIP proximal interphalangeal; OA = osteoarthritis. = joint tissue (12,13). MMP-13 was later cloned from human cartilage (14,15) and quite recently, Wernicke et a1 found that MMP-13 messenger RNA (mRNA) is also expressed in rheumatoid arthritis (RA) and osteoarthritis (OA) synovial tissue, but not in any other tissue examined (16). These observations suggest a synovial tissue/articular cartilage-specific expression of MMP-13. These observations prompted us to assess whether RA synovial MMP-13 mRNA is translated to the corresponding protein in situ, and in which cells and to what extent does this occur compared with OA and with some degenerative and/or traumatic joint changes. Because collagenases have been considered one of the prime targets for anticollagenolytic enzyme inhibitors, such as tetracyclines for MMP-8 (17-19) and bisphosphonates (20), we also tested the inhibitory potential of such compounds. PATIENTS AND METHODS Patients and samples. Synovial tissue specimens were obtained from 16 consecutive patients at the time of joint replacement surgery or diagnostic/therapeutic arthroscopy (Table 1). Six of these patients (all women, mean age 52.5 years, age range 37-73) had RA (21), with a mean disease duration of 8.8 years (range 5-14 years), and 7 of these patients (6 women and 1 man, mean age 65 years, age range 48-90) had OA (22), with a mean disease duration of 8 years (range 5-13 years). Samples were also available from 1 patient with osteochondritis dissecans, 1 with luxation of the patella, and 1 with a ruptured meniscus. Diagnosis of the last 3 conditions were proved by arthroscopy. Ten synovial samples were obtained from the knee, 2 from the proximal interphalangeal joint, 1 from the meta- carpophalangeal joint, 2 from the wrist, and l from the shoulder joint. All samples were snap-frozen, embedded in Tissue-Tek OCT compound (Lab-Tek Products, Elkhart, IN), and stored at -70°C until used. Cytospin preparation of neutrophils. To exclude crossreaction of the anti-MMP-13 antibody with MMP-8 collagenase, isolated leukocytes were used as a known MMP-8positive control. Polymorphonuclear leukocytes were isolated from the buffy coat layer of a 20-ml blood sample by a standard protocol in microfuge tube scale (23). A viability count was obtained in 0.2% trypan blue in phosphate buffered saline, and cells were adjusted to a concentration of 0.6 X 106/ml. Cytospin preparations were made at 6.0 X lo4 cells/slide. Slides were air dried for 10 minutes and fixed in -20°C methanol for 10 minutes. Immunohistochemical staining. Avidin-biotinperoxidase complex (ABC) staining (24) was utilized using Vectastain Elite ABC rabbit kit (Vector, Burlingame, CA). Six micrometer-thick cryostat sections were applied to gelatinformalin-coated slides, and fixed in acetone for 5 minutes at 4°C. Endogenous peroxidase activity in the tissue sections and cytospin preparations was inhibited with 0.3% H,O, in methanol for 30 minutes. Samples were incubated serially at 22°C in 1) normal goat serum (diluted 1:60) for 30 minutes, 2) primary rabbit anti-human MMP-13 antibodies (diluted 1500) for 60 minutes, 3) biotinylated goat anti-rabbit IgG (diluted 1:133) for 30 minutes, 4) AJ3C (diluted 1:lOO) for 30 minutes, or 5) 3,3'-diaminobenzidine tetrahydrochloride (DAB; Sigma, St. Louis, MO) in 0.006% H,O, solution in TBS (0.05M Tris, 0.15M NaC1, pH 7.4) followed by a wash in tap water. Consecutive sections or cytospin preparations were counterstained with Harris hematoxylin or were not counterstained. Slides with counterstaining were repeatedly washed in tap water. Finally, sections or cytospin preparations were dehydrated, cleared in xylene, and mounted in a synthetic mounting medium (Diatex, Becker, Marsta, Sweden). Morphometry was done using an ocular counting square (20 squares X 20 squares) and an oil immersion objective (1,000- magnification). At least 500 of each of synovial stromal cells, synovial lining cells, and endothelial cells were counted to estimate the ratio of positive cells. Lymphocytes were excluded from the cell counts (25). A 0.1% solution of bovine serum albumin (Sigma) in TBS was used for dilution of the primary antibodies and sera. Between steps (except after incubation in normal goat serum), the slides were washed 3 times for 5 minutes in TBS. Routine controls were omission or use of normal rabbit IgG (no. X903; Dako, Glostrup, Denmark) instead of the primary antibody and exposure of the tissue sections to DAB and H202 (for endogenous peroxidase). Double staining for MMP-13 and CD68. ABC-alkaline phosphatase-anti-alkaline phosphatase (APAAP) double staining was performed, and the color reactions were developed as described in detail elsewhere (24,26,27). Briefly, consecutive 5 Fm-thick cryostat sections from 2 of the synovial specimens used for MMP-13 staining were processed for staining of MMP-13 according to the protocol for the ABC technique (24) up to the step where the slides were rinsed in tap water. Then, the slides were washed with TBS (0.05M Tris, 0.15M NaCI, pH 7.4), and incubated at room temper- COLLAGENASE 3 IN RHEUMATOID SYNOVIUM ature in 1) normal rabbit serum (diluted 150) for 1 hour, 2) mouse anti-human CD68 IgG (dilution 1 5 0 ) for 1 hour, 3) rabbit anti-mouse IgG (dilution 1:25) for 1hour, or 4) MAAP solution (dilution 1:25) for 1 hour. Alkaline phosphatase was visualized after treatment with BCIP (Sigma no. B-0274) and nitroblue tetrazolium (Sigma no. N-6876) in alkaline phosphatase buffer for 10 minutes at room temperature, covered from light. Endogenous alkaline phosphatase activity was inhibited by including 0.003M levamisole (Sigma no. L-9756). Finally, the slides were dehydrated, cleared, and mounted. Reagents for the double-staining protocol were from Dako A& unless otherwise specified. Between each step, the sections were washed 3 times for 5 minutes with buffer. Preparation of the antiserum. The immunization procedure is described elsewhere (3). Briefly, recombinant human MMP-13 was produced in Escherichia coli, and 8M urea was used to solubilize the insoluble fraction of the cell lysate. Then, 1 ml of the 8M urea extract was electrophoresed through a 12% polyacrylamide gel, and the portion of the gel containing the recombinant protein was excised, ground, and incubated with 2 ml of deionized water at 37°C for 20 hours with sporadic vortexing. Next, 1 ml of sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE)-purified protein was used to immunize a New Zealand white rabbit according to the method described by Vaitukaitis (28). The rabbit was bled 6 weeks after the injection, and IgG were purified by chromatography through DEAE-cellulose (Whatman DE52; Whatman, Kent, UK), equilibrated, and eluted with 0.02M phosphate buffer, pH 7.2. The IgG concentration was 1 pglpl. Preliminary experiments examining the possible cross-reactivity with other MMPs, including MMP-1, MMP-8, and stromelysins, failed to reveal any significant crossreactivity. Absorption control tests. The specificity of immunoreaction was tested in 1 RA and 1 OA sample by preabsorbing the anti-MMP-13 with an excess of antigen, as described previously (29). Briefly, the maximum reproducible dilution (1:1,000) of anti-MMP-13 for the tissue was determined, and this working dilution was incubated with 0.1 nmoles of the antigen overnight at 4°C. This presaturated antibody dilution was then used in place of the primary antibody. Extraction of proteolytic enzymes. Synovial tissue specimens were stored at -70°C until used. No cartilage was included in the samples to be used for tissue extraction. Tissue pieces were homogenized in 3 volumes of TBS-1M MgCI,, extracted on ice for 1 hour, and clarified by centrifugation. The supernatant was dialyzed against 3 changes of TNC (0.005M CaCI, in TBS, pH 7.5) and stored at -70°C until analyzed (30). Calculation of the isoelectric point (PI). Calculations were performed using pK, values for amino acids. The amino acid sequences for MMP-1 and MMP-13 were retrieved from the Swiss-Prot protein databank, signal peptides were removed with a text editor program, and charges were calculated for procollagenases and collagenases using the Microsoft Excel 5.0 spreadsheet program. Possible sialic acid residues in glycosylated forms of MMP-l and MMP-13 were not taken into account in the calculations of the apparent PI (31-33). The charges calculated were used to choose the pH for the nondenaturing electrophoresis buffer, since molecules of approximately the same size but different charge would show different electrophoretic mobilities. 1393 Electrophoresis. Three buffer systems were used for separating MMP-1 and MMP-13 in the extracts. First, SDSPAGE was run as described by Laemmli (34), without reduction. Second, low pH Triton-acid-urea gels (TAU gels, 10%) were run essentially as described by Smith (35). Third, high-pH gels were run in am Ornstein-Davis buffer system using 7.5% gel (36,37). A minigel apparatus from Bio-Rad (Richmond, CA) was used for all gels. Protein blotting and immunostaining. Synovial tissue extracts were prepared for the different buffer systems, electrophoresed, and blotted onto nitrocellulose filter. Proteins from SDS-PAGE gels were blotted in routinely used Trisglycine buffer in a Bio-Rad Mini-TransBlot system (34,38). From the TAU gels, proteins were capillary blotted overnight at room temperature. After a brief soak in 2.3% (weighti volume) SDS-5% (v/v) 2-mercaptoethanol, blotting from Ornstein-Davis native gels was carried out with 0.1% SDS added to the transfer buffer (39). Nitrocellulose filters were blocked with 3% (wiv) gelatin, washed in TBS, and incubated overnight with rabbit anti-human MMP-13 (1500) or rabbit anti-human MMP-1 (1500). Goat anti-rabbit IgG horseradish peroxidase conjugate was diluted to 1500, and the blots were developed using 4-chloro-1-naphthol (Sigma) as a chromogen. Collagenase assay. Soluble rat tail tendon type I and bovine cartilage type I1 collagens were purified by pepsin extraction and selective salt precipitation at acidic and neutral pH, as described in detail elsewhere (40). Recombinant human MMP-13 was activated with 0.001M aminophenyl mercuric acetate (APMA) for 30 minutes and incubated with type I and type I1 collagen at a concentration of 1.5 pM for 3 hours at 22°C. The enzyme reaction was terminated by boiling for 5 minutes after addition of sample buffer without 2mercaptoethanol. Intact collagen a chains and their 3/4 (aA) and ?h(aB) degradation products were separated by 11% SDS-PAGE (41). The percentage of collagen degraded was estimated by image analysis (Molecular Analyst; Bio-Rad) of Coomassie blue-stained gels. Specifically, the gel was scanned on the hard disk of the computer, intact monomeric collagen bands were marked, and the marked areas were analyzed for pixel intensity. The scores of 3h fragments were multiplied by Y3 to compensate for the ?A fragment not scanned, and the percentage from the total (intact band score plus degraded band score) was calculated. Production and purification of human recombinant MMP-13. Human MMP-13 was produced using a vaccinia virus expression system as previously described (3). Vaccinia virus-expressing human MMP-13 was obtained using a plaque selection system (42). Plasmid pRB-col3 was obtained by inserting an Eco RIIHin dIII fragment containing the gene downstream of a vaccinia virus synthetic earlyilate promoter, in plasmid pBR21. Confluent monolayers of CV-1 cells in T25 flasks were infected with 1 plaque forming unit (PFU) per cell of vaccinia virus vRB12 clone a4 and transfected with 10 pg of plasmid pRB-col3. At 2 days postinfection, the progeny virus was harvested. The recombinant, termed VV-col3, was selected by 2 consecutive rounds of plaque purification on BSC-1 cell monolayers. For production of the recombinant protein, preconfluent BSC-1 cells in 900-cm2 roller bottles were infected with wild-type vaccinia virus (strain WR) or VV-col3, at a multiplicity of infection of about 5 PFUicell. The medium was harvested at 24 hours postinfection. 1394 LINDY ET AL Figure 1. Immunostaining for matrix metalloproteinase 13 (MMP-13) in rheumatoid arthritis (RA) and osteoarthritis (OA) synovial membranes. A, RA synovium, showing the lining (arrows). B, Consecutive section from A, representing a negative-staining control in which the primary antibodies were omitted from the staining sequence. C, Expression of MMP-13 in RA synovial stromal cells (arrows) and vascular endothelium (open circle). D, RA sublining tissue stained for MMP-13 (arrows). E, OA synovium,showing the lining (arrows). (Frozen sections; avidin-biotin-peroxidase complex stained, without counterstain; original magnification X 160 in A and E, X400 in B, C, and D). The recombinant protein was purified from the culture medium by chromatography onto an S-Sepharose fast flow column (2.5 X 10 cm; Pharmacia, Uppsala, Sweden) equilibrated in 0.02M Tris HC1, pH 7.2,0.005M CaCl,, 0.05% NaN,. The matrix was washed in the same buffer to background readings, followed by another wash with the same buffer with 0.2M NaCl to remove impurities. Finally, the column was eluted with the same buffer with 0.5M NaCl, and the fractions containing pro-MMP-13 were identified by analysis in SDS-PAGE. Assessment of the inhibition characteristics of the recombinant MMP-13. Briefly, autoactive MMP-13 was incubated with type I and type I1 collagen and analyzed as described above. Autoactivation refers to an apparently spontaneous conversion of pro-MMP-13 into its enzymatically active form in the absence of proteolytic activators or organomercurials (12). Collagenase activity was determined in the presence of doxycycline, clodronate, pamidronate, and D-penicillamine, at a wide range of concentrations. The concentrations were chosen to also cover those attained COLLAGENASE 3 IN RHEUMATOID SYNOVIUM 1395 Figure 2. Double-labeling for MMP-13 and CD68 (marker for monocyteimacrophages) in A, rheumatoid arthritis and B, osteoarthritis synovial membranes. Note that some stromal and lining cells (arrows) are positive for both anti-MMP-13 and anti-CD68, whereas other, apparently fibroblast-like cells (arrowheads) are positive only for MMP-13. (Frozen sections, avidin-biotin-peroxidase complex and alkaline phosphatase-anti-alkaline phosphatase double-stained, without counterstain; original magnification X 400). See Figure 1 for definitions. in vivo in patients receiving regular treatment with these drugs. Statistical analysis. BMDP 7.01 software was used for the statistical calculations. P values were estimated using the Mann-Whitney test for unpaired groups. RESULTS Immunohistochemistry results. All the synovial samples used in this study showed immunoreactivity to anti-MMP-13. Staining for MMP-13 was observed in synovial lining, endothelial, and stromal cells (Figure 1). Double labeling for MMP-13 and CD68 confirmed that MMP-13 was expressed in both synovial fibroblasts and monocyte macrophages, as well as in both A and B synoviocytes (Figures 2A and B). In addition, in RA, accumulations of mononuclear inflammatory cells, apparently lymphocytes, in perivascular and sublining areas were negative for both MMP-13 and CD68 (data not shown). The results of morphometric assessments are presented in Table 2. Although the majority of the lining, synovial stromal, and endothelial cells from both RA and OA patients were found to express MMP-13, the ratio of MMP-13-positive synovial lining and stroma1 cells was significantly higher in RA than in OA patients (for lining cells P = 0.0262; for stromal cells P = 0.0027). The ratio of MMP-13-positive endothelial cells in RA and OA patients did not differ statistically sig- nificantly (P = 0.3491). The control samples from patients with other joint derangements were excluded from statistical comparison due to the small number of cases. Cytospin preparations of neutrophils (known to be MMP-8-positive) did not stain with anti-MMP-13. Positive staining for MMP-13 was found in the ductal epithelium of the mammalian carcinoma specimens (not shown). Absorption control with purified recombinant MMP-13 abolished staining from strongly positive cells, leaving a light brown background. The antigen t o antibody ratio in the absorption test was calculated to give at least a 20-fold antigen excess over the binding sites. Detection of synovial tissue extract collagenases blotted onto nitrocellulose. Regular SDS-PAGE did not allow unequivocal separation of the 2 MMPs (data not shown). After calculation of the isoelectric points, 2 buffer systems were chosen for separation, both with a reduction step added in sample preparation. At low pH, where pro-MMP-1, MMP-1, pro-MMP-13, and MMP13, according to calculations, have different charges, the recovery of protein on nitrocellulose was low. In contrast, at p H 9.5, protein recovery was better and a clear separation of MMP-1 and MMP-13 could be achieved: the relative mobility for the fastest-migrating MMP-1 and MMP-13 bands was 0.4 and 0.56, respectively (Figure 3). A control sample of 300 ng of MMP-1 blotted from a high-pH native gel was not detected by antiMMP-13 antibody (data not shown). LINDY ET AL 1396 Table 2. Quantitative evaluation of matrix metalloproteinase 13 (MMP-13) staining* MMP-13-positive synovial cells (%) Patient Lining cells Endothelial cells Stromal cells 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 98 90 95 89 95 83 80 82 88 75 82 92 89 80 84 85 97 90 98 92 95 88 85 90 93 80 90 98 93 80 89 85 98 72 94 80 97 74 11 15 20 16 17 26 16 4 15 10 * Ratios were estimated for 500 cells per type. Inhibition characteristics of the recombinant MMP-13. Clodronate, pamidronate, and D-penicillamine were assessed over a range of concentrations that can be reached with regular treatment and/or that have been shown to inhibit other collagenases (17,18,20). Activity was measured using a soluble type I1 collagen assay (Figure 4). Clodronate, pamidronate, and D-penicillamine did not inhibit recombinant MMP-13 (Figure 5). Doxycycline had inhibitory potential starting at approximately 200 f l ,but the 50% inhibition concentration (IC5,J values were not yet reached at 500 p M (Figure 5). This differs from the IC,, values for MMP-8 (26 and for MMP-1 (280 rJ.M) (18). synovial lining and stroma in vivo. The results presented in this paper are in clear accordance with those reported by Wernicke et al, who cloned MMP-13 from an RA synovial sample and demonstrated expression of its mRNA in RA and OA synovial tissue, but did not find it in several normal human tissues (16). Our results demonstrate protein level expression of MMP-13 in RA and OA synovium. MMP-13 seems thus to be specifically expressed in synovial tissue, which makes it an enzyme of potential pathogenetic importance in the arthritides. The combination of high tensile strength provided by type I1 collagen and high swelling pressure of the proteoglycan matrix is important for the biomechanical function of cartilage (44,45). Different strategies have been used in various attempts to overcome the poor inherent healing capacity of articular cartilage: heterologous chondrocytes, periosteum, perichondrium, and osteochondral grafts, and, more recently, autologous chondrocytes (46) have been transplanted to treat focal articular cartilage defects. Because of the importance of type I1 collagen to the biomechanical properties of the cartilage and the poor inherent healing capacity of the tissue, the initial and specific cleavage of type I1 collagen must be a highly regulated event. Collagenases of the MMP family are the only mammalian enzymes which are able to cleave the al(I1) M), 40 M 0 D r 20 DISCUSSION In the original work by Freije and coworkers (3), MMP-13 mRNA was found in breast carcinomas, but not in normal human tissues. However, articular cartilage was not included in their normal tissue panel. Later, Flannery and Sandy reported MMP-13 mRNA expression in chondrocytes (43), a finding later confirmed by Mitchell et a1 (15) and by Reboul et a1 (14). Reboul and coworkers did not find MMP-13 mRNA in interleukinlp-stimulated synoviocyte cultures or in 1 OA synovial tissue sample. In light of the data reported by Reboul et al, the present data and those recently published by other investigators (16), it is possible that MMP-1 and MMP-13 are differently regulated under conditions used in synoviocyte cultures. Alternatively, in vitro cell cultures do not reproduce the conditions that prevail in the 40 w 80 I 0 2 4 6 8 10 12 14 PH Figure 3. Charge-pH plot of MMP-1 ( 0 ) and MMP-13 (W), demonstrating the charge difference. In native gels, MMP-1 and MMP-13 would migrate according to their intrinsic charge. For maximum separation, pH 3 and Ornstein-Davis buffer systems were chosen. At low pH, the recovery of protein on nitrocellulose was low (not shown). The blotted Ornstein-Davis gel shown demonstrates neutral salt extracts of 3 representative rheumatoid arthritis samples stained for MMP-1 (first 3 lanes) and MMP-13 (second 3 lanes). Note that the different net molecular charges allowed separation of the 2 MMP superfamily members at pH 9.5, which was used to demonstrate the specificity of the antisera. For clarity, pro-MMPs are not shown. See Figure 1 for definitions. COLLAGENASE 3 IN RHEUMATOID SYNOVIUM triple helix at the 775Gly-776Lxu peptide bond (47). Hitherto, only MMP-1 and MMP-8 have been considered to be responsible for thc degradation of type I1 collagen. MMP-8, which cleaves type I collagen faster than type I1 collagen, can be found in synovial fluid neutrophils (7,23). The rheumatoid changes are characterized by development of peripheral erosions associated with fibroblast- and macrophagc-like cells at the advancing edge of the pannus (48). These cells have been suggested to contain fibroblasttypc collagenase 1 (MMP-I) (8,9), which, however, is not particularly effective against monomeric type I1 collagen (30,49,50). Therefore, the occurrence of a third type of collagenasc, collagenase 3/MMP-13, in synovial tissue in both RA and OA is of particular interest. MMP-13 was observed in fibroblast- and macrophage-lke cclls both in thc synovial stroma and the lining. According to recent data, MMP-13 differs from MMP-I in that it clcavcs type I1 collagen more efficiently than other fibrillar collagens (12). In addition, MMP-13 is active against aggrecan, the major proteoglycan of the hyaline articular cartilage (13). Based on its substrate profile against the 2 major components of the hyaline articular cartilage and its now- Figure 4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of type I (first 5 lanes) and type I1 (second 5 lanes) collagen. Lane 1, Intact d(1) and u2(I) chains of type I collagen; lanes 2 and 4, time-dependent (4 hours and 12 hours) degradation of type I collagen by autoactivated recombinant MMP-13; lanes 3 and 5, the degradation was not increased by the presence of 0.001Morganomercurial aminophenyl mercuric acetate (APMA)-type pro-MMP activator; lane 6 , intact rrl(I1) chains; lanes 7 and 9, time-dependent (4 hours and 12 hours) degradation by autoactivated recombinant MMP-13; lanes 8 and 10, the degradation was the same in the presence of APMA. Arrowheads indicate intact monomeric collagen chains. The bar graph shows the percentage degradation of collagen substrate. See Figure 1 for other definitions. 1397 A & b B CC D b Figure 5. Inhibition of MMP-13-mediated degradstion of typc I1 collagen by doxycycline, clodronate, pamidronate, and D-penicillamine. The sodium dodecyl sulfate-polyacrylamide gels represent a range of roncuntrations that were selected to cover thc concentrations attained in vivo. as follows. A, For doxycycline, lane 1 contains 1.5 pM type I1 collagen alonc: lanes 2-10 contain type I I collagen with 160 ng 0 1 recornbinant MMP-I3 enzyme protein without and with 20.40, ti). 100, 200,300,100, and 5(W1 pA1 doxycycline, respectively. H, For clodronate, lane 1 contains type I1 collagen alone; lanes 2-8 contain type I1 collagen with recombinant MMP-13 without and with 50, IN).750.500, 750. and 1,WO JLM clodronate, respectivcly. C, For paniidronate. lane 1 contains type I1 collagen alone; lanes 2-8 contain type I1 collagcn with recombinant MMP-13 without and with 50, 100,250,500,750, and 1,000 phf pamidronate, respectively. D, For D-prnirillarninc, lane 1 contains type I1 collagen alone; lanes 2-6 contain type I1 collagen with recombinant MMP-13 without and with 50, 100, 200, and 500 JLM D-penicillamine, respectively. Arrowheads indicate intact monomeric collagen chains. demonstrated presence in synovial membrane, we suggest that MMP-13 may represent a particularly significant mediator of tissue destruction in the arthritides. Our immunohistochemical findings of the presence of MMP-13 in RA and OA synovial membranes were confirmed by demonstrating MMP-13 in synovial tissue extracts. Notably, no cartilage was included in our samples. Electrophoretic separation in native gels was efficiently used to separate collagenases that displayed similar apparent molecular weights in SDS-PAGE. Estimates of charge and isoelectric points are naturally only apparent, since the calculations cannot include the effects of the conformation or sialic acid content of MMP-1 and MMP-13. Because of the significant degree of glycosylation of MMP-1 and MMP-13 (10% and 25%, respectively), possible sialic acid end groups may contribute to the total charge and mobility of the collagenases. According to our results after electroblotting of 1398 the Ornstein-Davis gels, RA and OA synovial membranes contained both MMP-13 and MMP-1, which were clearly separated from each other under the conditions used. Separation in different gel systems also served in the characterization of the MMP-13 and MMP-1 antisera. Selection of the electrophoresis system, therefore, helped to show the presence of MMP-13 protein in RA and OA synovial tissue. MMPs have a great inherent tissue destruction potential. Therefore, many MMPs, including MMP-1, are not constitutively expressed (or only at a low level), but appear first upon appropriate stimulation of the producer cells with, for example, certain cytokines (5). Furthermore, they are usually not stored in the producer cells, but are rapidly secreted after their synthesis (5). This type of regulation may be important for the expression of these enzymes. Therefore, localization of MMP-13-positive cells in the vicinity of subsynovial lymphocyte infiltrates may reflect increased MMP-13 production due to as-yet-unidentified inflammatory cytokines produced by the mononuclear cells in rheumatoid synovitis in situ. Interestingly, there was no difference in MMP-1 and MMP-13 mRNA expression as analyzed by densitometry from Northern blots standardized for GAPDH expression (16). This suggests that, in addition to transcriptional events mediating the regulation of MMP-13 expression, posttranscriptional events can also be important. Because of the potential significance of MMPs in joint-destructive events, inhibitors of MMPs have received extensive interest (see ref. 51). In particular, drugs already widely used, including tetracyclines (51) and bisphosphonates (20), would seem to offer certain practical advantages over as-yet-unregistered drugs. Tetracyclines and bisphosphonates have been previously reported to inhibit several members of the MMP superfamily (51,20) and may share a common type of mechanism of action. They are now reported to be relatively ineffective against MMP-13 compared with their potency against MMP-1 and MMP-8. This type of inhibitory profile, in conjunction with its tissue-specific localization, substrate profile, and increased expression in RA, may further contribute to the pathogenetic significance of MMP-13. In conclusion, our findings suggest that MMP-13 is present in RA synovial tissue and that MMP-1 and MMP-8 are not the only collagenases involved. In fact, some of the earlier reports on synovial collagenase may have to be reassessed due to the possible presence of collagenase 3/MMP-13 activity. Because of its cellular localization in the fibroblast-like cells, apparent synovial LINDY ET AL tissue-specific localization, and substrate specificity, it may play a major role in local tissue-destructive events. Increased expression in RA may be a reflection of the catabolic state in the joint and could be explained by the more obvious inflammatory nature of RA compared with OA. Finally, the relative resistance of MMP-13 to several known collagenase inhibitors further adds to its pathogenetic significance. ACKNOWLEDGMENT Rabbit polyclonal anti-MMP-1 was kindly provided by Prof. H. Birkedal-Hansen,Department of Oral Biology, School of Dentistry, The University of Alabama at Birmingham. REFERENCES 1. Gross 3, Lapibre CM: Collagenolytic activity in amphibian tissues: a tissue culture assay. Proc Natl Acad Sci U S A 48:1014-1022, 1962 2. Nagase H, Barrett AJ,Woessner J F Jr: Nomenclature and glossary of the matrix metalloproteinases. Matrix Suppl 1:421-424, 1992 3. Freije JMP, Diez-Itza I, Balbin M, SBnchez LM, Blasco R, Tolivia J, M p e z - 0 t h C Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas. J Biol Chem 269:16766-16773, 1994 4. Puente XS, Pendas AM, Llano E, Velasco G, L6pez-0th C: Molecular cloning of a novel membrane-type matrix metalloproteinase from a human breast carcinoma. Cancer Res 56:944-949, 1996 5. Werb Z, Alexander CM: Proteinases and matrix degradation. In, Textbook of Rheumatology. Fourth edition. Edited by WN Kelley, ED Harris Jr, S Ruddy, CB Sledge. Philadelphia, WB Saunders, 1993 6. Prockop DJ, Kivirikko KI: Collagens: molecular biology, diseases, and potentials for therapy. Annu Rev Biochem 64:403-434, 1995 7. Harris ED Jr, DiBona DR, Krane SM: Collagenase in synovial fluid. J Clin Invest 48:2104-2113, 1969 8. Evanson JM, Jeffrey JJ, Krane SM: Human collagenase: identification and characterization of an enzyme from rheumatoid synovium in culture. Science 158:499-502, 1967 9. Dayer JM, Krane SM, Russell RGG, Robinson DR. Production of collagenase and prostaglandins by isolated adherent rheumatoid synovial cells. Proc Natl Acad Sci U S A 73:945-949, 1976 10. Quinn CO, Scott DK, Brinckerhoff CE, Matrisian LM, Jeffrey JJ, Partridge NC: Rat collagenase: cloning, amino acid sequence comparison, and parathyroid hormone regulation in osteoblastic cells. J Biol Chem 265:22342-22347, 1990 11. Henriet P, Rousseau GG, Eeckhout Y: Cloning and sequencing of mouse collagenase cDNA divergence of mouse and rat collagenases from the other mammalian collagenases. FEBS Lett 310: 175-178, 1992 12. Knauper V, Lbpez-0th C, Smith B, Knight G, Murphy G: Biochemical characterization of human collagenase-3. J Biol Chem 271:1544-1550, 1996 13. Fosang AJ, Last K, Knauper V, Murphy G, Neame PJ: Degradation of cartilage aggrecan by collagenase-3 (MMP-13). FEBS Lett 380:17-20, 1996 14. Reboul P, Pelletier J-P, Tardif G, Cloutier J-M, Martel-Pelletier J: The new collagenase, collagenase-3, is expressed and synthesized by human chondrocytes but not by synoviocytes: a role in osteoarthritis. J Clin Invest 97:2011-2019, 1996 15. Mitchell PG, Magna HA, Reeves LM, Loprestimorrow LL, COLLAGENASE 3 IN RHEUMATOID SYNOVIUM Yocum SA, Rosner PJ, Geoghegan KF, Hambor JE: Cloning, expression, and type I1 collagenolytic activity of matrix metalloproteinase-13 from human osteoarthritic cartilage. J Clin Invest 97:761-768, 1996 16. Wernicke D, Seyfert C, Hinzmann B, Gromnica-Ihle E: Cloning of collagenase 3 from the synovial membrane and its expression in rheumatoid arthritis and osteoarthritis. J Rheumatol 23:590-595, 1996 17. Golub LM, Ramamurthy NS, McNamara TF, Gomes B, Wolff M, Casino A, Kapoor A, Zambon J, Ciancio S, Schneir M Tetracyclines inhibit tissue collagenase activity: a new mechanism in the treatment of periodontal disease. J Periodontal Res 19651-655, 1984 18. Suomalainen K, Sorsa T, Golub LM, Ramamurthy N, Lee H-M, Uitto V-J, Saari H, Konttinen Y T Specificity of the anti-collagenase action of tetracyclines: relevance to their anti-inflammatory potential. Antimicrob Agents Chemother 36227-229, 1992 19. Smith GN, Brandt KD, Hasty KA: Activation of recombinant human neutrophil procollagenase in the presence of doxycycline results in fragmentation of the enzyme and loss of enzyme activity. Arthritis Rheum 39:235-244, 1996 20. Teronen 0,Konttinen YT, Salo T, Ingman T, Sorsa T Clodronate inhibits MMP-1 and MMP-8 in GCF and PISF (abstract 31). Bone 17:605, 1995 21. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 31:315-324, 1988 22. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, Christy W, Cooke TD, Greenwald R, Hochberg M, Howell D, Kaplan D, Koopman W, Longley S 111, Mankin H, McShane DJ, Medsger T Jr, Meenan R, Mikkelsen W, Moskowitz R, Murphy W, Rothschild B, Segal M, Sokoloff L, Wolfe F Development of criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 29:10391049, 1986 23. Konttinen YT, Lindy 0, Kemppinen P, Saari H, Suomalainen K, Vauhkonen M, Lindy S, Sorsa T: Collagenase reserves in polymorphonuclear neutrophil leukocytes from synovial fluid and peripheral blood of patients with rheumatoid arthritis. Matrix 11:296301, 1991 24. Hsu S-M, Raine L, Fanger H: Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 29:577-580, 1981 25. Konttinen YT, Reitamo S, Ranki A, Hayry P, Kankaanpaa U, Wegelius 0: Characterization of the immunocompetent cells of rheumatoid synovium from tissue sections and eluates. Arthritis Rheum 24:7i-79, 1981 26. Cordell JL. Falini B. Erber WN. Ghosh AK. Abdulaziz Z. MacDonald S, Pulford &,Stein H, Mason DY: Immunoenzymatic labeling of monoclonal antibodies using immune complexes of alkaline phosphatase and monoclonal anti-alkaline phosphatase (APAAP complexes). J Histochem Cytochem 32:219-229, 1984 27. Konttinen YT, Hukkanen M, Segerberg M, Rees R, Kemppinen P, Sorsa T, Saari H, Polak JM, Santavirta S: Relationship between neuropeptide immunoreactive nerves and inflammatory cells in adjuvant arthritic rats. Scand J Rheumatol 21:55-59, 1992 28. Vaitukaitis JL: Production of antisera with small doses of immunogen: multiple intradermal injections. Methods Enzymol 73:46-52, 1981 29. Van Noorden S: Tissue preparation and immunostaining techniques for light microscopy. In, Immunocytochemistry: Modern Methods and Applications. Second edition. Edited by JM Polak, S van Noorden. Bristol, Wright & Sons, 1986 30. Konttinen YT, Lindy 0, Suomalainen K, Ritchlin C, Saari H, 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 1399 Vauhkonen M, Lauhio A, Santavirta S, Sorsa T: Substrate specificity and activation mechanisms of collagenase from human rheumatoid synovium. Matrix 11:395-403, 1991 Segel IH: Biochemical Calculations. New York, Wiley and Sons, 1968 Ribeiro JM, Sillero A An algorithm for the computer calculation of the coefficients of a polynomial that allows determination of isoelectric points of proteins and other macromolecules. Comput Biol Med 20:235-242, 1990 Henriksson G, Englund AK, Johansson G, Lundahl P: Calculation of the isoelectric points of native proteins with spreading of pK, values. Electrophoresis 16:1377-1380, 1995 Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685, 1970 Smith BJ: Acetic acid-urea polyacrylamide gel electrophoresis of proteins. Methods Mol Biol 32:39-47, 1994 Davis BJ: Disc electrophoresis. 11. Method and application to human serum proteins. Ann N Y Acad Sci 121:404-427, 1964 Ornstein L: Disc electrophoresis. I. Background and theory. Ann N Y Acad Sci 121:321-349, 1964 Towbin H, Staehelin T, Gordon J: Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A 76:43504354, 1979 Waterborg JH, Harrington RE: Western blotting of histones from acid-urea-Triton- and sodium dodecyl sulfate-polyacrylamide gels. Anal Biochem 162:430-433, 1987 Miller EJ, Rhodes KR: Preparation and characterization of the different types of collagen. In, Structural and Pontractile Proteins: Part A. Methods in Enzymology. Edited by LW Cunningham, DW Fredriksen. New York, Academic Press, 1982 Turto H, Lindy S, Uitto V-J, Wegelius 0, Uitto J: Human leukocyte collagenase: characterization of enzyme kinetics by a new method. Anal Biochem 83557-569, 1977 Blasco R, Moss B: Selection of recombinant vaccinia viruses on the basis of plaque formation. Gene 158:157-162, 1995 Flannery CR, Sandy JD: Identification of MMP-13 (collagenase-3) as the major matrix metalloproteinase expressed by chondrocytes during retinoic acid-induced matrix catabolism (abstract), FortyFirst Annual Meeting of the Orthopedic Research Society, Orlando, FL, February 1995 Brandt KD, Mankin HJ: Osteoarthritis and polychondritis. In, Textbook of Rheumatology. Fourth edition. Edited by WN Kelley, ED Harris Jr, S Ruddy, CB Sledge. Philadelphia, WB Saunders, 1993 Poole AR: Cartilage in health and disease. In, Arthritis and Allied Conditions. Twelfth edition. Edited by DJ McCarty, WJ Koopman. Philadelphia, Lea and Febiger, 1993 Brittberg M, Lindahl A, Nilsson A, Ohlsson C, Isaksson 0, Peterson L Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation. N Engl J Med 3312389895, 1994 Krane SM: Mechanisms of tissue destruction in rheumatoid arthritis. In, Arthritis and Allied Conditions. Twelfth edition. Edited by DJ McCarty, WJ Koopman. Philadelphia, Lea and Febiger, 1993 Shiozawa S, Shiozawa K, Fujita T: Morphologic observations in the early phase of the cartilage-pannus junction: light and electron microscopic studies of active cellular pannus. Arthritis Rheum 26:472-478, 1983 Welgus HG, Jeffrey JJ, Eisen AZ: The collagen substrate specificity of human skin fibroblast collagenase. J Biol Chem 256:95119515, 1981 Docherty AJP, Murphy G: The tissue metalloproteinase family and the inhibitor TIMP: a study using cDNA and recombinant proteins. Ann Rheum Dis 49:469-479, 1990 Inhibition of matrix metalloproteinases: therapeutic potential. Edited by RA Greenwald, LM Golub. Ann N Y Acad Sci 732,1994
1/--страниц