Webber. R. and Mumford, R.A.. (1995) Quantification of a matrix metalloproteinase-generated aggrecan GI fragment using monospecific antipeptide serum.
EQUINE VETERINARY JOURNAL Equine vet. J. (1997) 29 ( 5 ) 335-342
335
Characterisation of equine matrix metalloproteinase 2 and 9; and identification of the cellular sources of these enzymes in joints P. D. CLEGG*, R. M. BURKEt, A. R. COUGHLANS, C. M. RlGGS and S. D. CARTER Department of Veterinary Clinical Science and Animal Husbandry, #Department of Veterinary Pathology, University Veterinary Teaching Hospital, University of Liverpool, Leahurst, Neston. S. Wirral. L64 7TE and ?Department of Biochemistry and Applied Molecular Biology, UMIST, PO Box 84 Manchester, M60 IQD, UK. Keywords: horse; joint disease; matrix rnetalloproteinase;gelatinase
Summary The cellular production by resident articular cells and infiltrating inflammatory cells of the gelatinase matrix metalloproteinases (MMP) was investigated by tissue culture methods and analysis of cell supernatants by gelatin zymography. Peripheral blood neutrophils in short term culture produced MMP-9, as did peripheral blood monoeytes in culture. Isolated articular chondrocytes in monolayer culture produced both MMP-2 and MMP-9, although articular cartilage maintained as explant culture produced MMP-2 alone. Synovial fibroblasts grown in monolayer culture produced MMP-2 alone, although synovial membrane in explant culture produced both MMP-2 and the active form of MMP-2. Lysis of blood polymorph neutrophils produced large quantities of MMP-9, but lysis of blood monocytes, synovial fibroblasts and articular chondrocytes produced little enzyme indicating that, unlike the other cell types, polymorph neutrophils store MMPs intracellularly. Equine MMP-2 was purified from synovial fibroblast cell culture supernatant, and equine MMP-9 from polymorph neutrophil cell culture supernatant, by gelatin-sepharose affinity chromatography. The 2 enzymes were identified from their molecular weights and by their respective N-terminal amino acid sequences which showed homology with the enzymes from other species. The demonstration that invasive cells and resident articular cells can produce enzymes which are capable of digestion of certain component molecules of the articular cartilage matrix, shows that therapeutic targeting of these enzymes could be a valid proposition in the prevention of cartilage destruction in osteoarthritis. Introduction Osteoarthritis is not a single disease entity, but a group of disorders characterised by biochemical deterioration of the articular cartilage, accompanied by changes in bone and soft tissues of the joint (McIlwraith and Vachon 1988). Clinically, this is manifested as joint pain and dysfunction. The essential
'Author to whom correspondence should be addressed.
pathological features include various degrees of local splitting and fragmentation (fibrillation) or complete erosion and loss of articular cartilage. The predominant and most widely accepted theory of osteoarthritis is that this breakdown of articular cartilage matrix occurs primarily by enzymatic degradation (Brandt and Mankin 1986) and cartilage fibrillation. The eventual erosive lesions most probably result from the inability of this biochemically compromised cartilage to withstand repeated loading over the course of the disease (Malemud et al. 1987). Proteolytic enzymes are classified into 4 distinct types depending on their catalytic mechanism: serine proteinases, cysteine proteinases, aspartic proteinases and metalloproteinases (Barrett and Saklatvala 1985). The matrix metalloproteinases (MMPs) are a group of zinc dependent endopeptidases which are centrally involved in the normal physiological turnover of the extracellular matrix. The MMPs can be divided into 3 groups: collagenases (MMP 1 and 8), Wpe N collagenases/gelatinases (MMP 2 and 9) and stromelysindproteoglycanases(MMP 3 and 10).The gelatinases represent a subgroup of the MMP family and consist of 2 distinct gene products with similar enzymatic activities: a 70 kDa gelatinase (MMP-2) expressed by most connective tissue cells (Collier et al. 1988), and a 92 kDa gelatinase (MMP-9) expressed by inflammatory phagocytes and tumour cells (Hibbs et al. 1985,1985; Wilhelm et al. 1989). MMP2 is also known as gelatinase A or 72 kDa ype IV collagenase, and MMP-9 as gelatinase B or 92 kDa type IV collagenase. The MMPs are thought to be important in joint disease as they are implicated in degradation of the matrix molecules of articular cartilage. MMP-2 and 9 have been demonstrated to degrade minor collagens found in articular cartilage (Smith et al. 1991; Hirose et al. 1992), as well as some of the articular cartilage proteoglycans such as aggrecan core protein (Fosang et al. 19921, cartilage link protein (Nguyen et al. 1993) and aggrecan (Lark et al. 1995). In the horse, relatively few data on MMPs and joint disease are available. Equine synoviocytes and chondrocytes produce MMP-3 on stimulation with interleukin-1 (IL-1) (May et al. 1992), as do equine articular cartilage explants (Moms and Treadwell 1994). Equine synoviocytes produce MMP collagenase activity, as do equine peripheral blood neutrophils (Spiers el al. 1994). No work has so far been reported on the gelatinase group of MMPs in equine joint disease. They have
Characterisation of MMP-2 and 9
336
been identified in synovial fluid from man (Hirose et al. 1992; Koolwijk et al. 1995) and dog (Coughlan et al. 1995). Matrix metalloproteinase gelatinase production has been shown in articular and inflammatory cells in man and other species (Hibbs et al. 1985, 1987; Okada et al. 1990; Campbell et al. 1991). The respective roles of the resident articular cells (chondrocytes and synovial lining cells) as well as the roles of infiltrating inflammatory cells (polymorph neutrophils and peripheral blood monocytes) in the proteolytic digestion of cartilage in joint disease is still debated. In this study the MMP gelatinase production, by different articular tissues and by inflammatory cells, has been characterised using both explant tissue culture and by cell culture of isolated cells from normal equine joints. The equine gelatinase MMPs have been purified from cell culture supernatants, identified by N-terminal amino acid sequencing and the sequences compared with those found in man and dog.
Materials and methods Isolation of and growth of polymorph neutrophils and peripheral blood monocytes
cells were seeded at 1 x 106/ml in Dulbecco's modified Eagles medium (DMEM) plus FCS (5%), amphotericin B (0.5 pg/ml), gentamicin (100 pg/ml) and HEPES buffer (20 mmolll) and incubated for 12 h at 37°C in 5% CO,. The cells were grown to confluence, and subcultured at a density of 5 x 104/ml into a 75cm3 flask. Once the cells had reached confluence, the media was changed to the same but without FCS. The cells had an obvious fibroblastic phenotype at this stage. The cells were grown for a further 3 days and the supernatant was then removed, centrifuged at 200g for 10 min and stored at -70°C until assayed. Synovial membrane explant culture Synovial membrane was obtained as described previously and diced into approximate 1 mm3 cubed explants. The tissue explants were washed repeatedly in DMEM plus amphotericin B (0.5 pg/ml), gentamicin (100 pg/ml) and HEPES buffer (20 mmoVI), and then 6 explants were placed in each well of a 24 well plate. One ml of the above media was added and the explants incubated for 3 or 6 days at 37°C in 5% CO,. The supernatant was removed and centrifuged at 200g for 10 min and stored at -70°C until assay. Isolation of chondrocytes
Equine polymorph neutrophils (PMN) and peripheral blood monocytes (PBM) were isolated from peripheral blood using a discontinuous Percoll density gradient (Pycock et al. 1987), with concentrations modified to 62.5 and 75%, as this was found to be more successful in the isolation of PMN. Cell type was confirmed by staining with eosin and methylene blue, while cell viability was assessed by trypan blue exclusion and was always in excess of 90%. Polymorph neutrophils were washed 3 times with Hanks buffered saline solution (HBSS). The washed cells were resuspended (1 X 107cells/ml) in HBSS, and incubated at 37°C in 5% CO, for 30 min in the presence or absence of Phorbol myristal acetate (PMA)(SO ng/ml), or lysed with 0.1% Triton X100. Cells were sedimented by centrifugation and the supernatants stored in aliquots at -70°C until further analysis (Hibbs et al. 1985; Hasty et al. 1986). Peripheral blood monocytes were washed 3 times in HBSS, resuspended in RPMI 1640 containing 25 mmol/l HEPES buffer, 50mgA gentamicin and 10% fetal calf serum (FCS) at a concentration of 5 x 106cells /mland incubated for 12 h at 37°C in 5% CO,. The nonadherent cells were removed by washing in RPMI media and the adherent cells (monocytes) allowed to spread out to confluence over about 36 h. Once the cells had reached confluence the media was changed to the same but without FCS. The cells were then incubated for 48 h either in the presence or absence of 20 pg/ml LPS. Adherent cells were incubated with trypsin to release them from adherence to the plastic surface and then lysed with 0.1% Triton X-100. In all cases the cells were sedimented by centrifugation and the supernatants stored in aliquots at -70°C until further analysis. Isolation and growth of synovial fibroblasts Normal synovial tissue was obtained at post mortem by aseptic dissection of the soft tissue lining of the cranial aspect of the metacarpophalangeal joint from horses (n = 3) subjected to euthanasia for a medical condition with no diagnosed orthopaedic disease. The synovial tissue was diced and placed in a tissue digestor and synovial fibroblasts obtained by sequential enzyme digestion with trypsin and collagenase (May et al. 1988). Isolated
Normal articular cartilage was shaved from the metacarpal condyles at post mortem from horses (n = 3) subjected to euthanasia for medical conditions with no diagnosed orthopaedic disease. The shavings were transferred to a tissue digestor and chondrocytes obtained by enzymatic digestion (May et al. 1989) with the modification of the above technique that a single digestion step with 0.1% clostridial collagenase was employed, with stimng over 20 h (Nixon et al. 1992). Isolated cells were seeded at 1 x lo6 cellslml. in a complete medium (Dulbecco's modified Eagles medium (DMEM), amphotericin B (0.5 pg/ml), gentamicin (100 pg/ml) and HEPES buffer (20 mmol/l) and incubated for 24 h at 37°C in 5% CO,. The media was then replaced with the above media supplemented with 10% FCS and the cells grown to confluence. The media was then replaced with the original media lacking FCS. The cells at this stage were spherical and did not show any evidence of a fibroblastic phenotype. The cells were grown for a further 3 days and the supernatant was then removed, centrifuged at 200 g for 10 min and stored at -70°C until assayed. Articular cartilage explant culture Articular cartilage was obtained as described previously, and explants prepared and incubated as for synovial membrane explant culture. Gelatin zymography Gelatinase enzyme activity was assayed in the cell culture supernatants using gelatin zymography. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed using a vertical gel apparatus' according to the method of Laemmli (1970), with the modification that gelatin was included in the resolving gel (Hibbs et al. 1985). Gelatin was co-polymerised into 0.75 mm thick, 7.5% polyacrylamide gels at a final concentration of 2.5 mg/ml. Cell culture supernatants samples were diluted 4: 1 in a Tris/HCl sample buffer containing 1.5% SDS, 5% glycerol and 0.005% Bromophenol blue, pH 6.8 and incubated at 37°C for 1 h prior to electrophoresis. Each
P. D.Clegg er al.
-
225kDa
331
225kDa
94kDa
-
85kDa
A
B
94kDa 85kDa
C
Fig I : Gelatin zymographic analysis of cell supernatants from short term (30 min) culture of blood polymorph neutrophils. Lane A) No stimulation. Lane B ) Stimulated with 50 ng/ml Phorbol myristal acetate. Lane C) Cells lysed with 0.2% Triton.
dilution (7.5 pl) was loaded in a well and samples electrophoresed at 200v for approximately 45 min. The gels were washed for 1 h at room temperature in 2.5% Triton X-100, rinsed in distilled water and incubated for 18 h at 37°C in a reactivation buffer containing 50 mMTris/HC1,50 mM CaCI,, 10 mM NaCI, 0.05% Brij35, pH 7.6. Following incubation, the gels were stained with 2% Coomassie brilliant blue R-200 followed by destaining in a mixture of 7% (v/v) glacial acetic acid, 30%(v/v) methanol and 63% (v/v) distilled water. The gels were dried in a cellophane sandwich. Molecular weight standard markers (Mark 12 wide range protein standards: Novex) were run on each gel which were stained prior to the gel being washed in Triton. The effects of including 10 mmoVl, 50 mmol/l and 100 mmoVl EDTA (an MMP inhibitor) or 50 mmoVl benzamidine or 50 mmoYl PMSF (serine proteinase inhibitors) in the standard reactivation buffer were also investigated.
Enzyme isolation and purification The gelatinase enzyme activity in the supernatants from PMN and from synovial fibroblasts was assayed by gelatin zymography. Enzyme purification was performed by gelatin Sepharose affinity chromatography by the method described by Hibbs et al. (1985). Relative protein amounts in the eluted fractions were measured by absorbancy of the eluted fractions at 280 nm. The purification of the isolates was checked by running the purified samples on SDS-PAGE (Laemmli 1970) followed by silver staining of the gel (Silver Staining Kit Protein2). The enzyme activity of the purified sample was assessed by gelatin zymography as described above.
A
B
Fig 2: Gelatin zymographic analysis ($cell supernatants from culture of peripheral blood polymorph monocytes. Lane A ) Stimulated with lipopolysaccharide for 48 h. Lane B ) Cells lysed with 0.2% Triton.
discarded and the residue allowed to evaporate. The pellet was then resuspended in 50 pl of 1:s sample buffer and incubated for 1 h prior to running on SDS-PAGE following the method of Laemmli (1970) at 200v for 45 min. The gels were then electroblotted onto a polyvinylidene difluoride (PVDF) membrane' (Matsuidaira 1987). The membrane was stained with Coomassie Blue R-250 in 50% methanol for 5 min and then destained in 50% methanol, 10% acetic acid to allow visualisation of the protein bands on the membrane. The membrane was rinsed in distilled water and air dried. The Nterminal amino acid sequence was obtained using a automatic peptide sequencer3. Sequences obtained were compared with known sequences using a computer database (SwissProt).
Results Polymorph neutrophil culture Polymorph neutrophil grown in culture with no stimulation produced strong gelatinolytic enzyme activity, as measured by gelatin zymography, at 94 kDa and 225 kDa and a weaker band at 85 kDa. On stimulation of PMN with PMA, increased amounts of enzyme activity were seen at 85,94 and 225 kDa. Lysis of the cells with PMN produced similar zymographic findings as with the PMA stimulated cells (Fig 1). Peripheral blood monocyte culture
N-terminal amino acid sequencing of MMP-2 and MMP-9
The purified enzymes from the affinity chromatography separation were concentrated 10-fold by trichloroacetic acid (TCA) precipitation. Briefly, to 0.75 ml of sample was added 0.75 ml of 20% TCA, and after mixing centrifuged at 12000g for 5 min. The supernatant was discarded and the pellet washed 3 times in diethyl ether (1.5 ml). Excess diethyl ether was
Peripheral blood monocyte grown in culture with no stimulation produced small amounts of gelatinolytic enzyme activity as measured by gelatin zymography at 85 kDa, 94 kDa and 225 kDa. On stimulation of the cells with LPS, slightly increased amounts of enzyme activity were seen at 94 and 225 kDa. Scant enzyme activity was noted at 94 and 225 kDa in supernatants from cells lysed with Triton X-100 (Fig 2).
Characterisation of MMP-2 and 9
338
94kDa
Fig 3: Gelatin zymographic analysis of cell supernatants from 2 separate monolayer cultures of isolated (passage 2) synovialjibroblasts (in serum-jree media). The cells were cultured for 48 h.
A
B
C
64kDa
Fig 5: Gelatin zymographic analysis of cell supernatants from 2 monolayer primary cultures of isolated articular chondrocytes (in serum-free media). The cells were cultured for 48 h.
Fig 4: Gelatin zymographic analysis of cell supernatants from tissue culture of equine synovial membrane explants (in serum-free media). Lane A ) lime = 1 days. Lane B) lime = 3 days. Lane C) lime = 6 days.
Synovial fibroblast culture Synovial fibroblasts grown in serum free culture media produced a single band of gelatinolytic activity at 64 kDa (Fig 3). Synovial membrane explant culture
Synovial membrane explants grown in serum free tissue culture produced a single band of gelatinolytic activity at 64 kDa at one and 3 days, and at 6 days a second band of gelatinolytic activity was detected at 59 kDa as well as the original 64 kDa band (Fig 4). Chondrocyte culture
Chondrocytes from articular cartilage grown in serum free culture media produced gelatinolytic activity at both 64 and 94 kDa (Fig 5). Articular cartilage explant culture
Articular cartilage explants grown in serum-free tissue culture medium produced a single band of gelatinolytic activity at 64 kDa, at both 3 and 6 days (Fig 6). Inhibitor studies
All enzyme bands (64,94 and 225 kDa) were partially inhibited by 10 mmoVl EDTA (a MMP inhibitor) and totally inhibited by 50 and 100 mmoM EDTA, but not by 50 mmoM Benzamidine nor by 50 mmol/l PMSF (specific serine proteinase inhibitors). This indicated that the enzymes were MMPs. Enzyme isolation and purification
Fractions obtained from chromatography on gelatin Sepharose examined by gelatin zymography revealed gelatinolytic activity in fractions eluted by DMSO. These fractions had a low protein content (c10 pg/ml;OD,,,); fractions of purified MMP-2 analysed by SDS-PAGE and silver stained revealed a large protein band at 64 kDa, and a faint band at 59 kDa, under non-reducing conditions, and a large protein band at 72 kDa, and a faint band at 65 kDa (Fig 7) under reducing conditions. Similar analysis of the purified MMP-9 revealed 2 protein bands at 94 kDa and 225 kDa
A
B
Fig 6: Gelatin zymographic analysis of cell supernatants from tissue culture of 2 equine articular cartilage explants (in serum-free media). lime = 3days.
(Fig 8) under nonreducing conditions, and a single protein band at 100 kDa under reducing conditions. N-terminal amino acid sequencing of MMP-2 and MMP-9
N-terminal amino acid sequence for the terminal 15 amino acids of MMP-2 and MMP-9 (Fig 9) were obtained. Equine MMP-2 had very close homology with the N-terminal amino acid sequences of MMP-2 from man, rat, mouse and dog with 13 out of 14 identified amino acids being identical. The sequence obtained from equine MMP-9, showed a lesser degree of homology with only 10 out of 15 amino acids being identical. Comparison with sequences obtained using a computer database of protein sequences (SwissProt) identified human MMP-9 as the fifth most highly conserved protein in comparison to equine MMP-9. None of the more highly conserved proteins were proteolytic enzymes, and the sequences which were more homologous were not from the N-terminal region of the proteins.
Discussion On the basis of electrophoretic mobility, substrate specificity, inhibitor profiles and N-terminal amino acid sequencing, we consider the gelatinolytic bands shown in this study to be the equine MMP-2 (64 kDa band), active MMP-2 (59 kDa band), MMP-9 monomer (94 kDa band), active MMP-9 (85 kDa band) and MMP-9 dimer (225 kDa band). The dimerisation occurs when MMP-9 is present in excess, relative to tissue inhibitor of metalloproteinase (TIMP) and may be important in the control of activation as, once the dimer has been formed, it can be activated by stromelysin, unlike the monomer MMP-9/TIMP complex. (Goldberg et al. 1992). Analysis of the N-terminal amino acid sequence of MMP-2 demonstrates a high degree of conservation
P. D. Clegg et a/.
339
116kDa 97kDa
-16kDa
72kDa 65kDa
97kDa
66kDa
66kDa
A
B
A
B
Fig 7: Purijied MMP-2 run on a 7.5% polyacrylamide gel under reducing conditions. Lane A = MMP-2, Lane 6 = Molecular weight markers.
Fig 8: Purified MMP-9 run on a 7.5% polyacrylamide gel under nonreducing conditions. Lane A = MMP-9, Lane B = Molecular weight markerLs.
between the sequence in the horse, and that in the dog, man and other species, though MMP-9 only demonstrates a 66% homology between the horse and human sequence. The sequencing assists in the confirmation of the identity of the individual bands as does the effect of inhibitors such as EDTA, benzamidine and PMSF on enzyme activity. Though there is only a moderate degree of N-terminal amino acid conservation between equine and human MMP-9, the identification of the protein as equine MMP-9 was considered reasonable as none of the proteins with a higher degree of sequence conservation identified on a computer database of protein sequences were proteolytic enzymes, their molecular weights were considerably different from equine MMP-9 and these conserved sequences were not N-terminal sequences but mid-protein. The other metalloproteinase which is secreted by PMNs, identified in other species, which has a similar molecular weight (85 kDa) to MMP9 is PMN collagenase (MMP-8). This MMP in man has an Nterminal amino acid sequence with no homology to equine MMP-9 (Hasty et al. 1990). MMPs are secreted as inactive zymogens, and activation requires loss of an N-terminal propeptide and, therefore, a loss of molecular weight. Activation of MMP-2 in vivo is poorly understood (Okada et al. 1990; Woessner 1991; Birkedal-Hansen et al. 1993), though recently a plasma membrane bound activator of MMP-2 has been identified and designated membrane typematrix metalloproteinase (MT-MMP) (Strongin et al. 1995). PMNs and PBMs both appear to produce MMP-9 as both the inactive zymogen and the lower molecular weight active form. MMP-9 activation is better understood and involves the action of MMPs including MMP-1, 2, 3 and 7, as well as the action of other proteases such as trypsin, tissue kallikrein and cathepsin G (Fridman et al. 1995; Sang et al. 1995). In the case of MMP-2, all the producing cells and tissues produced the higher molecular weight zymogen, when the molecular weights of equine MMP-2 were compared with those found in other species (BirkedalHansen et al. 1993), except in the case of the synovial membrane explant culture, where a lower molecular weight form of MMP2 was produced, as well as the zymogen, after being maintained
in culture over several days. The purification of MMP-2 using gelatin-sepharose chromatography appeared to cause an increase in amount of the active lower molecular weight form of MMP-2, by some unknown process. Inhibition of MMPs occur via tissue inhibitors of metalloproteinases (TIMPs) and a2-macroglobulin (BirkedalHansen et al. 1993). In this study the cellular production and effect of inhibitors on the enzyme could be disregarded as SDSPAGE splits the inhibitor from the enzyme as well as activating the enzyme (Birkedal-Hansen and Taylor 1982). Whereas MMP-2 and 9 are produced by both resident articular and invading inflammatory cells, their exact role and importance in the pathogenesis of joint disease still has to be fully investigated. It has been demonstrated in the murine air pocket model of cartilage breakdown that gelatinase activity was the most prominent of any MMP activity, and the emergence of MMP-2 and 9 activity coincided with collagen loss from the implanted cartilage (Trancart et al. 1992). In the past, there has been little evidence that these enzymes are able to break down type I1 collagen, which is the predominant type of collagen in articular cartilage, though recently MMP-2 has been shown to act as a collagenase and can hydrolyse triple-helical collagen when rendered free from TIMP (Aimes and Quigley 1995). MMP-2 and 9 have high affinity for denatured collagen and digestion of denatured collagen may be important in the pathogenesis of joint disease as MMP-2 and 9 could continue to digest degraded collagen once other collagenases have started the enzymatic breakdown. Type XI collagen is located specifically in cartilage tissue, either within type II collagen fibrils in the matrix (Mayne 1989) or on the surface of chondrocytes (Smith et al. 1985) and is digested by both MMP-2 and 9 (Smith et al. 1991; Hirose et al. 1992). Type X I collagen interacts strongly with the cartilage proteoglycan and could play a role in the stabilisation of the matrix and may fulfil the role of an exocytoskeleton (Hirose et al. 1992). Furthermore, type XI collagen is not broken down by any other collagenases (Eyre et al. 1989). MMP-9 also breaks down aggrecan core protein (Fosang et al. 1992) and cartilage link protein (Nguyen et al. 1993), while MMP-2 cleaves
Characterisation of MMP-2 and 9
340
Horse MMP 9 Human MMP-9
1-V-F-P-G-D-L-G-T-L-L-T-D-A-E-L-15 1-L-F-P-G-D-L-R-T-N-L-T-D-R-Q-L-15
Fig 9: N-ierminal amino acid sequence of MMP-2 of horse, dog (A.B. Coughlan, unpublished data), and man (Collier et al. 1988) (X) = undetermined residue. N-terminal amino acid sequence of MMP-9 of horse and man (Triebel et al. 1992).
aggrecan in v i m in a similar fashion to stromelysin (MMP-3) (Lark et al. 1995). MMP-2 and 9 may also have an important role in joint disease as they synergistically potentiate the action of interstitial collagenases to degrade collagen (Murphy and Docherty 1992). MMPs have also been shown to be important in the activation of other MMPs, although the role of MMP-2 and 9 and other MMPs in this biological activation is still incompletely understood (Birkedal-Hansen et al. 1993). These enzymes may also have a deleterious effect on the joint by a direct action on the synovium, which would lead to a potentiation of the inflammatory response. MMP-2 and 9 are inhibited by TIMPs, therefore the presence of these enzymes may be deleterious to the joint as they may act by ‘mopping up’ such local inhibitors, thereby facilitating damage by other activated MMPs. MMP-2 and 9 have been shown to have an important role in the binding and control of TIMPs, with both MMP-2 and 9 forming complexes with TIMP which do not involve the enzyme binding site (Goldberg et al. 1989). Isolated articular chondrocytes were shown to produce both MMP-2 and 9 though no MMP-9 dimer activity was seen when the cells were maintained in monolayer culture. As no MMP-9 dimer was detected, it can be presumed that there are TIMPs present to excess (Goldberg et al. 1992), indicating that equine chondrocytes may also be producing TIMPs. In this experiment the chondrocytes used were a primary culture and demonstrated a chondrocyte phenotype on examination via phase contrast microscopy. The chondrocytes began to demonstrate a fibroblastic phenotype when they were passaged and maintained in culture. In other species MMP-2 is expressed by most connective tissue cells (Collier et al. 1988), and MMP-9 expressed by inflammatory phagocytes and tumour cells (Hibbs et al. 1985, 1987; Wilhelm et al. 1989), though MMP-9 production has also been shown in human osteoarthritic cartilage and in normal human cartilage on exposure to IL-1 (Mohtai et al. 1993). It is of interest that normal equine cartilage explants only produced MMP-2, and no MMP-9 activity was observed. This is in contrast to the isolated articular chondrocytes which produce both MMP-2 and 9. This finding may be due to the inability of the larger MMP-9 molecule to escape from the cartilage matrix (Sledge 1993). Alternatively, the chondrocytes in explant cultures are surrounded by normal matrix and, as such, are in a much more normal physiological environment than cells cultured isolated in monolayers. Therefore, the enzyme production by such explant culture systems may be a more accurate reflection of in vivo events. Equine synovial fibroblasts grown in monolayer culture produce MMP-2 but not MMP-9. This characteristic is similar to fibroblasts obtained from other sites and in other species (Seltzer et al. 1981; Golds et al. 1983; Okada et al. 1990). Although the synovial fibroblasts used in this experiment were passage 2 cultures, with all the cells demonstrating a fibroblastic
of synovial membrane produced MMP-2 in culture, and with continual incubation a second gelatinolytic band of activity is seen at 59 kDa. It is probable that the 59 kDa band is the active form of MMP-2 which is produced in the intact tissue (Woessner er al. 1991). Until recently it was not known how MMP-2 was activated, when several groups demonstrated a membrane bound activator of MMP-2, membrane-type matrix metalloproteinase (MT-MMP) (Nomura et al. 1995; Strongin et al. 1995). MT-MMP is the main and possible only route of activation of MMP-2 and has been demonstrated in human articular cartilage (Buttner et al. 1996) and is expressed by human chondrocytes (Jawahar et al. 1996). It would be important to show whether MT-MMP is present in equine synovial membrane and to identify its role in equine joint pathology. Polymorph neutrophils produce large amounts of MMP-9 monomer and dimer, as is the case in man (Sopata and Dancewicz 1974; Hibbs et al. 1985). As the dimer occurs only when the enzyme is in excess in comparison to TIMPs it would appear unlikely that PMNs produce much inhibitory activity as well (Goldberg et al. 1992). MMP caseinase production by equine neutrophils in culture has been demonstrated (Spiers et al. 1994) and it is possible that this activity is due to MMP-9 (monomer and dimer) in the horse, as both have strong caseinolytic activity (P.D. Clegg, unpublished data). Equine synovial fluids from cases of articular sepsis have greatly increased levels of MMP-9 monomer and dimer, and these levels correlate with the synovial fluid white blood cell count, of which the cellular content is predominantly PMN (Clegg et al. 1997). This observation once again supports the therapeutic regime of joint debridement and flushing in cases of articular sepsis in order to remove the potentially deleterious enzymes and their cellular sources. Peripheral blood monocytes (PBM) produced gelatinolytic activity due to MMP-9 monomer and dimer. This activity was only marginally increased when the cells were stimulated with LPS. Monocyte/macrophage MMP-9 has been shown to be immunologically and biochemically identical to MMP-9 produced by PMN (Hibbs 1992). The stage of differentiation of PBM is important as MMP production increases on differentiation of these cells to macrophages (Shapiro et al. 1991). While infiltrating monocytes are rarely seen in joint disease, they may be of importance in the pathogenesis of joint diseases as it is considered that the synovial type A cells in the synovial membrane are derived from the circulating bone marrow-derived monocyte pool (Fox and Kang 1993). Lysates of equine chondrocytes, synovial fibroblasts and PBMs contained only small amounts of gelatinase enzyme in comparison to the amount of enzyme produced by cells in culture. In contrast, PMN lysates contained large amounts of MMP-9 monomer and dimer. This suggests that the enzymes are stored within the PMNs, whereas they are not stored in other cell types, where they are produced as required. It has been shown that human PMNs synthesise and store MMPs intracellularly in specific granules during maturation of the cell within the bone marrow (Bainton et al. 1971; Hibbs and Bainton 1989). This work allowed the cellular sources of the gelatinase MMPs to be identified, though the relative importance of different cell types in the pathogenesis of joint disease still is open to debate. Resident articular cells (chondrocytes and
P.D.Clegg et al. synovial fibroblasts) appear to be a significant source of the MMP gelatinases, though inflammatory cells can produce large amounts of MMP-9 as well. In articular sepsis the intense inflammatory cell infiltrate present within the joint is almost certainly the source of the MMP-9. Osteoarthritis in the horse has an inflammatory component, therefore the MMPs present in this disease may come from either the resident articular cells or the infiltrating inflammatory cells. It is still unknown what ability MMP molecules have to enter the matrix of articular cartilage from synovial fluid, if they are produced extrinsically from it in either the synovial membrane or inflammatory cells. Also, while chondrocytes are able to produce both MMP gelatinases, the ability of these enzymes to diffuse far from the chondrocyte in the articular cartilage is still not known, and therefore their effect on the cartilage as a whole is still open to conjecture. Therapy in osteoarthritis has often in the past been directed at symptomatic pain relief, although there have been a number of agents available which have claimed to have a chondroprotective action in preventing the progression of the catabolic process in articular cartilage. These agents have recently been reclassified as disease modifying osteoarthritis drugs (DMOADs) (Dieppe 1995). One potential route of action of these drugs is by inhibition of destructive enzymes, such as the MMPs. It has been demonstrated that polysulphated glycosoaminoglycans can inhibit casein degrading metalloproteinases of equine synovial cell origin (May er al. 1988). At present there are numerous synthetic MMP inhibitors under commercial development and these are entering clinical trials as therapy against various forms of arthritis in man (Buttle ef al. 1995). Characterisation of the catabolic enzymes involved in articular cartilage degradation in the horse may allow future assessment of novel therapeutic approaches in the treatment of equine joint disease. This work has identified and characterised the equine gelatinases MMP-2 and MMP-9. The results identify that various cell types, both resident articular cells and infiltrating inflammatory cells are potential sources of these enzymes. It is thought that the importance of each cell type in joint disease may vary dependent upon the inflammatory component of the disease within the joint.
Acknowledgements P.D. Clegg is in receipt of a Horserace Betting Levy Board Research Training Scholarship. A.R. Coughlan is in receipt of a Wellcome Trust Research Training Scholarship. We thank Prof R. Beynon and Dr D. Robertson of Department of Applied and Molecular Biology, UMIST for help in performing the amino acid sequencing of the enzymes.
Manufacturers' addresses 'Bio-Rad. Heme1 Hempstead, UK. *Pharmacia Biotech, St Albans, Herts, UK. 3Applied Biosystems Ltd, Warrington, UK.
References Aimes. R.T. and Quigley. J.P. (1995) Matrix metalloproteinase-2 is an interstitial
collagenase - Inhibitor free enzyme catalyses the cleavage of collagen fibrils and soluble native type-l collagen generating the specific 3/4- length and 1/4length fragments. J. Biol. Chem. 270,5872-5876.
Bainton, D.F., Ullyot, J.L., and Farquhar, M.G. (1971) The development of neutrophilic polymorphonuclear leukocytes in human bone marrow. J. Exp. Med. 134,907-934.
34 I Barrett, A.J. and SaMatvala, J. (1985) Proteinases in joint disease. In: Textbook of Rheumatology. Eds: W.N. Kelley, E.D. Harris, S. Ruddy and C.B. Sledge. W.B. Saunders. Philadelphia. pp 182-196. Birkedal-Hansen, H. and Taylor, R.E (1982) Detergent-activation of latent collagenase and resolution of its component molecules. Biochenf. Biophys. Res. Conim. 107, 1173-1178. Birkedal-Hansen. H., Moore, W.G.I., Bodden, M.K., Windsor, L.J., BirkedalHansen, B., Decarlo, A. and Engler, J.A. (1993) Matrix metalloproteinases; A Review. Critical Reviews Oral Biology Medicine 4, 197-250. Brandt, K.D. and Mankin, H.J. (1986) Workshop on the etiopathogenesis of osteoarthritis. J. Rheum. 13. 1126-1160. Buttle. D.J., Bramwell, H. and Hollander, A.P. (1995) Proteolytic mechanisms of cartilage breakdown: a target for arthritis therapy? J. Clin. Parhol: ntol. Puthol. 48, M167-Ml77. Buttner. E , Chubinskaya, S., Margerie. D., Huch, K.. Flechtenmacher. J.. Cole, A,. Kuettner, K. and Bartnik, E. (1996) Membrane type-matrix metalloproteinase (MT-MMP) is expressed in human articular cartilage. Transactions Orrh Res. Soc. 42, 173. Campbell, E.J.. Cury, J.D.. Shapiro, S.D., Goldberg, G.I. and Welgus, H.G. (1991) Neutral proteinases of human mononuclear phagocytes. J. Ininiunol. 146, 1286-1293. Clegg, P.D., Coughlan. A.R.. Riggs, C.M. and Carter, S.D. (1997) Matrix metalloproteinases 2 and 9 in equine synovial fluids. Equine vet. J. 29, 343348. Collier, I.E., Wilhelm, S.M., Eisen, A.Z., Manner, B.L., Grant, G.A., Seltzer, J.L.. Kronberger, A., He, C.,Bauer, E.A. and Goldberg, G.I. (1988) H-ras oncogenetransformed human bronchial epithelial cells (TBE- 1) secrete a single metalloprotease capable of degrading basement membrane collagen. J. Biol. Cheni. 263,6579-6587. Coughlan, A.R., Carter, S.D.. Robertson. D.H.L., Kelly, D.F.. Bennett, D., May, C. and Beynon, R.J. (1995) Zymographic analysis of synovial fluid metalloproteinases in canine osteoarthritis. Vet Conip. Orrh. Trarmt. 8. 62. Dieppe, P. (1995) Towards effective therapy for osteoarthritis. Rhermi. Europe. 24. 118-120. Eyre, D., Wu. J-J. and Wooley. D. (1994) All three chains of I alpha, 2 alpha, 3 alpha collagen from hyaline cartilage resist human collagenase. Biochenr. Biophys. Res. Conmiun. 118,724-729. Fosang.A.J., Neame, P.J., Last, K., Hardingham, T.E.. Murphy, G. and Hamilton, J. (1992) The interglobular domain of cartilage aggrecan is cleaved by PUMP, gelatinases, and cathepsin-B. J. Biol. Cheni. 267, 19470-19474. Fox, R.I. and Kang, H. (1993) Structure and function of synoviocytes. In: Arthritis and allied conditions. Eds: D.J. McCarty and W.J. Koopman. Lea and Fehiger, Philadelphia. pp 263-278. Fridman, R., Toth. M..Pena, D. and Mobashery, S . (1995)Activation of progelatinase B (MMP-9) by gelatinase A (MMP-2). Cancer Res. 55. 2548-2555. Goldberg, G.L. Manner, B.L., Grant, G.A., Eisen, A.Z., Wilhelm, S . and He, S. (1989) Human 72-Kilodalton type IV collagenase forms a complex with il tissue inhibitor of metalloproteases designated TIMP-2. Proc. nurl. Acad. Sci. O.S.A.86, 8207-8211. Goldberg, G.I., Strongin, A., Collier, I.E., Genrich, L.T. and Marmer, B.L. (1992) Interaction of 92kDa Type IV collagenase with tissue inhibitor of metalloproteinases prevents dimerization. complex formation with interstitial collagenase, and activation of the proenzyme with stromelysin. J. Biol. Cheni. 267,4583-4591. Golds, E.E., Santer, V., Killackey, I. and Roughley, P.J. (1983) Mononuclear cell factors stimulate the concomitant secretion of distinct latent proteoglycans. gelatin and collagen degrading enzymes from human skin fibroblasts and synovial cells. J. Rheum. 10, 861-867. Hasty, K.A., Hibbs. M.S., Kang,A.H. and Mainardi, C.L.(1986) Secreted formof human neutrophil collagenase. J. Biol. Cheni. 261, 5645-5650. Hasty. K.A.. Pourmotahbed, T.E. Goldberg, G.I., Thompson, J.P., Spinella, D.G.. Stevens, R.M. and Mainardi. C.L. (1990) Human neutrophil collagenase: A distinct gene product with homology to other matrix metalloproteinases. J. B i d . Chem. 265, 11421-11424. Hibbs. M. (1992) Expression of 92hDa phagocyte gelatinase by inflammatory and connective tissue cells. Matrix, Suppl. 1, 51-57. Hibbs. M. and Bainton, D.F. (1989) Human neutrophil gelatinase is a component of specific granules. J. clin. Invest. 84, 1395-1397. Hibbs, M., Hasty, K., Seyer. J., Kang, A. and Mainardi, C. (1985) Biochemical and immunological characterisation of the secreted forms of human neutrophil gelatinase. J. Biol. Chem. 260, 2493-2500. Hibbs, M., Hoidal, J. and Kang, A. (1987) Expression of a metalloproteinase that
Characterisation of MMP-2 and 9
342 degrades native type V collagen and denatured collagens by cultured human alveolar macrophages. J. din. Invest. 80, 1644-1650. Hirose,T..Reife, R.A..Smith,G.N., Stevens, R.M.,Mainardi, C.L. andHasty, K.A. (1992) Characterisation of rype V collagenase (gelatinase) in synovial fluid of patients with inflammatory arthritis. J. Rheum. 19,593-599. Jawahar, S . , Birkhead, J. and Vasios, G. (1996) Membrane type-matrix metalloproteinase is expressed in human chondrocytes. Transactions Onh. Res. SOC.42, 174. Koolwijk, P., Miltenburg, A.M.M.. van Erck, M.G.M., Oudshoorn. M.. Niedbala, M.J., Breedveld. F.C. and van Hinsbergh, V.W.M. (1995) Activated gelatinaseB (MMP-9) and urokinase-type plasminogen activator in synovial fluids of patients with arthritis. Correlation with clinical and experimental variables of inflammation. J. Rheunutol. 22,385-393. Laemmli. U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. Lond. 227,680-685. Lark, M.W.. Williams, H., Hoernner, L.A., Weidner, J., Avala, J.M., Harper, C.F., Christen, A., Olszewski, J., Konteatis, Z.. Webber. R. and Mumford, R.A. (1995) Quantification of a matrix metalloproteinase-generatedaggrecan GI fragment using monospecific antipeptide serum. Biochem. J . 307,245-252. McIlwraith, C.W. and Vachon, A. (1988) Review of pathogenesis and treatment of degenerative joint disease. Equine vet J., Suppl. 6,3-11. Malemud, C.J., Martel-Pelletier, J. and Pelletier, J.P. (1987) Degradation of extracellular matrix in osteoarthritis: 4 fundamental questions. J. Rheum.. Suppl. 14.20-22. Matsudaira. P. (1987) Sequence from picomole quantities of proteins elecaoblotted onto polyvinylidene difluoride membranes. J. B i d . Chem. 262, 10035-10038. May, S.A., Hooke. R.E. and Lees, P. (1988a) The effect of drugs used in the treatment of osteoarthrosis on stromelysin (proteoglycanase) of equine synovial cell origin. Equine vet. J., Suppl. 6, 28-32. May, S.A., Hooke, R.E. and Lees. P. (1988b) Bone fragments stimulate equine synovial lining cells to produce the inflammatory mediator prostaglandin E2. Equine vet. J., Suppl.6, 131-132. May, S.A., Hooke. R.E. and Lees, P. (1989) Identity of the E-series prostaglandin produced by equine chondrocytes and synovial cells in response to a variety of stimuli. Res. vet. Sci. 46,5457. May. S.A.. Hooke, R.E. and Lees, P. (1992) Interleukin-1 stimulation of equine articular cells. Res. vet. Sci. 52. 342-348. Mayne, R. (1989) Cartilage collagens: What is their function. and are they involved in articular disease? Arthritis Rheum. 32, 241-246. Mohtai, M., Smith, R.L.. Schurman. D.J., Tsuji, Y., Toni. EM.. Hutchinson, N.I., Stetler-Stevenson, W.G. and Goldberg. G.I. (1993) Expression of 92-M) type IV collagenase/gelatinase (gelatinase B9) in osteoarthritic cartilage and its induction in normal human articular Cartilage by interleukin-1. J. d i n . Invest. 92, 179-185. Moms, E.A. and Treadwell, B.V. (1994) Effect of interleukin-I on articular cartilage from young and aged horses and comparison with metabolism of osteoarthritic cartilage. Am. J. vet. Res. 55, 138-146. Murphy, G . and Docheny, A.J.P. (1992) The matrix metalloproteinases and their inhibitors. Am. J . Respi,: CellMol. Biol. 7. 120-125. Nguyen, Q.. Murphy, G.. Hughes, C.E.. Mort, J.S. and Roughley, P.J. (1993) Matrix metalloproteinases cleave at two distinct sites on human cartilage link protein.
Biochem. J. 295,595-598.
Nixon, A.J.. Lust, G. and Vernier-Singer, M. (1992) Isolation, propgation and cryoprreservationof equine articular chondmcytes. Am. J. vef. Res. 53,2364-2370. Nomura, H., Sato. H.. Seiki, M., Mai, M. and Okada, Y. (1995) Expression of membrane-type matrix metalloproteinase in human gastric carcinomas. Cancer Res. 55,3263-3266. Okada, Y.,Morodomi, T.. Enghild, J.J.. Suzuki, K., Yasui, A,, Nakanishi, I.. Salvesen. G. and Nagase, H. (1990) Matrix metalloproteinase 2 from human rheumatoid synovial fibroblasts. Purification and activation of the precursor and enzymatic properties. Eur J. Biochem. 194,721-730. Pycock. J.F., Allen, W.E. and Moms, T.H. (1987) Rapid, single-step isolation of equine neutrophils on a discontinuous Percoll density gradient. Res. vet. Sci. 42, 411-412. Sang, Q-X., Birkedal-Hansen. H. and Van Wart, H.E. (1995) Proteolytic and nonproteolytic activation of human neutrophil progelatinase B. Biochimica biophysica acfa. 1251, 99- 108. Seltzer. J.L., Adam, S.A., Grant. G.A. and Eisen, A.Z. (1981) Purification and properties of a gelatin-specific neutral protease from human skin. J . B i d . Chenz. 256, 4662-4668. Shapiro, S.D.. Campbell. E.J., Senior, R.M., and Welgus, H.G. (1991) Proteinases secreted by human mononuclear phagocytes. J. Rheunurtol. 18 Suppl. 27,9598. Sledge, C.B. (1993) Biology of the joint. In: Texbaok of Rheunzatolog!. Eds: W.N. Kelley, E.D. Harris, S . Ruddy and C.B. Sledge. W.B. Saunders. Philadelphia. pp 1-21. Smith. G.J., Williams, J. and Brandt, K (1985) Interactions of proteoglycans with the pericellular (1 alpha, 2 alpha. 3 alpha) collagens of cartilage. J . Biol.Cheni. 260, 10761- 10767. Smith, G.N., Hasty, K.A., Yu, L.P., Lamberson, K.S.. Mickler, E.A. and Brandt, K.D. (1991) Cleavage of l l p e XI collagen fibers by gelatinase and by extracts of osteoarthritic canine cartilage. Matrix 11,36-42. Sopata, I. and Dancewicz, A.M. (1974) Presence of a gelatin-specific proteinase and its latent form in human leucocytes. Biochim. Biophys. Acta. 370, 510-523. Spiers. S . , May, S.A.. Bennett. D. and Edwards, G.B. (1994) Cellular sources of proteolytic-enzymes in equine joints. Equine ver. J. 26, 43-47. Strongin, A.Y., Collier, I.. Bannikov, G . , Marmer, B.L., Grant, G.A. and Coldberg, G.I. (1995) Mechanism of cell surface activation of 72-kDa type IV collagenase. J. Biol. Chem. 270, 5331-5338. Trancart. M.M.. Chalmeigne, N., Girardot, C., Zarpanelian, C. and Prigent. D. (1992) Gelatinase is the main matrix metalloproteinase involved in granulomainduced cartilage degradation. h r . J. tissue reactions- expt and d i n . aspects. 14, 287-294. Triebel. S., Blaser, J., Reinke. H.. Knauper. V. and Tschesche. H. (1992) Mercurial activation of human PMN leucocyte type IV procollagenase (gelatinase). FEBS len. 298,280-284. Wilhelm, S.. Collier, I.. Manner. B., Eisen, A,, Grant, G . , and Goldberg. G. (1989) SV40-transformed human lung fibroblasts secrete a 92-kDa n p e IVcollagenase which is identical to that secreted by normal human macrophages. J. Biol. Chent. 264, 17213-17221. Woessner, J.F. (1991) Matrix metalloproteinases and their inhibitors in connective tissue remodelling. FASEB J. 5. 2145-2154.
Received for publication: 6.9.96 Acceaied: 10.3.97
