Cleavage of cartilage oligomeric matrix protein (thrombospondin-5) by matrix metalloproteinases and a disintegrin and metalloproteinase with thrombospondin motifs

Matrix Biology 22 (2003) 267–278

Cleavage of cartilage oligomeric matrix protein (thrombospondin-5) by matrix metalloproteinases and a disintegrin and metalloproteinase with thrombospondin motifs Sally C. Dickinsona, Mireille N. Vankemmelbekeb, David J. Buttlec, Krisztina Rosenbergd, Dick Heinegard ˚ d, Anthony P. Hollandera,* a Academic Rheumatology, University of Bristol, Avon Orthopaedic Centre, Southmead Hospital, Bristol BS10 5NB, UK Division of Microbiology and Infectious Diseases, University of Nottingham, Queen’s Medical Centre, Nottingham NG7 2UH, UK c Division of Genomic Medicine, Sheffield Childrens’ Hospital, Stephenson Wing, D-Floor, University of Sheffield, Sheffield S10 2TH, UK d Department of Cell and Molecular Biology, Section for Connective Tissue Biology, Lund University, SE-22184 Lund, Sweden b

Received 3 September 2002; received in revised form 5 March 2003; accepted 24 March 2003

Abstract Cartilage oligomeric matrix protein (COMP) is a pentameric glycoprotein present in cartilage, tendon and ligament. Fragments of the molecule are present in the diseased cartilage, synovial fluid and serum of patients with knee injuries, osteoarthritis and rheumatoid arthritis. Although COMP is a substrate for several matrix metalloproteinases (MMPs), the enzymes responsible for COMP degradation in vivo have yet to be identified. In this study we utilised well-established bovine cartilage culture models to examine IL-1a-stimulated COMP proteolysis in the presence and absence of MMP inhibitors. COMP was released from bovine nasal cartilage, in response to IL-1a, at an intermediate time between proteoglycans and type II collagen, when soluble MMP levels in the culture medium were undetectable. The major fragment of COMP released following IL-1a-stimulation migrated with an apparent molecular mass of approximately 110 kDa (Fragment-110) and co-migrated with both the major fragment present in human arthritic synovial fluid samples and the product of COMP cleavage by purified MMP-9. However, the broadspectrum MMP and ADAM inhibitor BB94 only partially inhibited the formation of Fragment-110 and failed to inhibit COMP release significantly. Therefore the results of these studies indicate a role for proteinases other than MMPs in the degradation of COMP in bovine cartilage. It was further demonstrated that purified COMP was cleaved by ADAMTS-4, but not ADAMTS-1 or -5, to yield a fragment which co-migrated with Fragment-110. Therefore this is the first demonstration of COMP as a substrate for ADAMTS-4, although it remains to be determined whether this enzyme plays a role in COMP degradation in vivo. 䊚 2003 Elsevier Science B.V. and International Society of Matrix Biology. All rights reserved. Keywords: Cartilage oligomeric matrix protein; ADAMTS; Aggrecanase; Matrix metalloproteinase; Metalloproteinase

1. Introduction The extracellular matrix of cartilage consists of several types of collagen, proteoglycan and other noncollagenous, non-proteoglycan macromolecules, all of which interact to form a highly specialised connective tissue (Poole, 1993). Degradation of aggrecan and type II collagen, the two major components of adult hyaline cartilage, are prominent features of arthritic disease which lead ultimately to joint failure (Hollander et al., *Corresponding author. Tel.: q44-117-959-5918; fax: q44-117959-6187. E-mail address: [email protected] (A.P. Hollander).

1995; Lohmander et al., 1993). However, non-collagenous proteins other than proteoglycans are also degraded during the disease process and may play important structural andyor regulatory roles during disease progression. Cartilage oligomeric matrix protein (COMP) is one of the more prominent non-collagenous components of cartilage, contributing approximately 1% of the wet weight of articular tissue (Hedbom et al., 1992). This pentameric disulfide-bonded glycoprotein is also present in tendon (DiCesare et al., 1994) and ligament (Muller et al., 1998) with the highest levels of the molecule detected in high weight-bearing tissues, including the

0945-053X/03/$30.00 䊚 2003 Elsevier Science B.V. and International Society of Matrix Biology. All rights reserved. doi:10.1016/S0945-053XŽ03.00034-9

268

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

Fig. 1. Matrix protein release in bovine cartilage cultures. Explants of bovine nasal (a–c) or articular (d–f) cartilage were cultured with 3 nM IL-1a, in DMEM, for 2 weeks, followed by a further 3 weeks with DMEM only (squares). Control explants were cultured in DMEM only for the entire culture period (circles). Medium was replenished weekly. Release of proteoglycan (a, d), COMP (b, e) and type II collagen (c, f) were measured as the cumulative release of glycosaminoglycans, COMP and epitope CB11B, respectively. Results are shown as a percentage of the total (residueqmedium), at different times of culture, and are the mean and S.E.M. of 5 experiments. *, P-0.02; NS, not significant, vs. control cultures (Mann–Whitney U-test).

fibrocartilagenous regions of tendon (Smith et al., 1997). As COMP can interact with both fibronectin (DiCesare et al., 2002) and collagen types I, II and IX (Rosenberg et al., 1998; Thur et al., 2001) in vitro, it may play a role in the assembly, organisation and maintenance of the cartilaginous matrix. Mutations in the gene for COMP can result in two musculoskeletal diseases, pseudoachondroplasia and multiple epiphyseal dysplasia (Briggs et al., 1995; Hecht et al., 1995), which are characterised by short stature and early onset osteoarthritis (OA). Fragments of COMP have been detected in the diseased cartilage, synovial fluid and serum of patients with knee injuries, post-traumatic and primary OA and

rheumatoid arthritis (RA) (DiCesare et al., 1996; Hummel et al., 1998; Neidhart et al., 1997; Saxne and Heinegard, 1992). However, the enzymes responsible for COMP degradation in vivo have yet to be identified. Purified COMP is a substrate for matrix metalloproteinases (MMPs) including interstitial collagenase (MMP1), collagenase-3 (MMP-13), stromelysin-1 (MMP-3) and gelatinase B (MMP-9). Several of the resulting fragments migrate with similar electrophoretic mobilities to those present in OA and RA synovial fluid samples (Ganu et al., 1998). In addition, a slight negative correlation (Rsy0.31) has been demonstrated between COMP concentration and gelatinase activity in equine synovial fluids (Misumi et al., 2001). However, the

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

broad-spectrum MMP inhibitors BB94 and CGS 27023A only partially inhibited COMP fragmentation in IL-1astimulated bovine articular cartilage even when used at concentrations sufficient to block MMP activity (Ganu et al., 1998, 1999). In addition, CGS 27023A only partially inhibited the release of COMP from IL-1atreated bovine nasal cartilage (Goldberg et al., 1995). As hydroxamate-based MMP inhibitors can also inhibit other metalloproteinases, including a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) -4 (aggrecanase-1), ADAMTS-5 (aggrecanase-2) and ADAM-10 (Amour et al., 2000; Tortorella et al., 2001; Vankemmelbeke et al., 2001), the inhibition of COMP degradation by BB94 and CGS 27023A, may be due to inhibition of metalloproteinases other than MMPs. Such enzymes are up-regulated in both arthritic disease and IL-1-stimulated culture models and are believed to play vital roles in cartilage matrix breakdown (Chubinskaya et al., 1998; Flannery et al., 1999; Malfait et al., 2002; Sugimoto et al., 1999). Due to the contradictory evidence for the role of MMPs in COMP degradation, the aim of this study was to perform a detailed analysis of COMP proteolysis in bovine nasal and articular cartilage model systems. 2. Results 2.1. IL-1a-induced matrix protein degradation We used well-established bovine cartilage culture models to study the degradation of COMP in comparison to proteoglycans and type II collagen. The time-course of matrix protein release from adult nasal cartilage explants is shown in Fig. 1a–c. In agreement with previous work (Goldberg et al., 1995; Kozaci et al., 1997), IL-1a induced a significant release of over 80% of the proteoglycan during the first week of culture, and over 90% by the end of week 2 (Fig. 1a). COMP was released more slowly than proteoglycan and stimulation with IL-1a had no significant effect during the first week of culture (Fig. 1b). However, in the second week, COMP release was significantly enhanced to over 50% of the total, and by the end of week 4 all of the COMP had been lost from the nasal cartilage explants. The degradation of type II collagen in IL-1a-stimulated explants was delayed until weeks 3–5 of culture, whereas there was negligible release from control cultures at any time point (Fig. 1c). As several differences between the behaviour of cultures of nasal and articular cartilage have previously been described (Kozaci et al., 1997; Price et al., 1999), we also examined matrix degradation in adult articular cartilage explants (Fig. 1d–f). Although IL-1a had a significant effect on proteoglycan loss at all time-points (Fig. 1d), release was more gradual than from equivalent nasal cartilage cultures and less than 70% of the proteo-

269

Fig. 2. MMP activity in conditioned medium from IL-1a-stimulated bovine nasal cartilage cultures. Explants were cultured with 3 nM IL1a for 2 weeks followed by a further 3 weeks with DMEM only. The medium collected each week was assayed for active MMPs using a quenched fluorescence substrate assay. Results are the mean and S.E.M. of 5 experiments. Control cultures (DMEM only) contained no detectable MMP activity (not shown). *, P-0.02 vs. control cultures (Mann–Whitney U-test).

glycan was lost by the end of the culture period. There was a marked release of COMP into the medium of control cultures at all time-points and the effect was not significantly increased by stimulation with IL-1a (Fig. 1e). Type II collagen was not released from either control or IL-1a-stimulated articular cartilage cultures at any time during the culture period (Fig. 1f). 2.2. MMP activity and matrix degradation To examine the relationship between MMP activity and matrix degradation, the conditioned medium samples assayed above for COMP, proteoglycan and type II collagen, were also assayed for total MMP activity using a quenched fluorescence substrate assay. In agreement with our previous findings (Brown et al., 1996; Kozaci et al., 1997), negligible MMP activity was detected in IL-1a-stimulated nasal cartilage conditioned medium during the first two weeks of culture (Fig. 2). However, during that time-period, over 90% of the proteoglycan and 50% of the COMP were released (Fig. 1a,b). MMP activity was only detected in weeks 3–5 of culture, which correlated with type II collagen breakdown. In medium from control nasal cartilage explants and control and IL-1a-stimulated articular cartilage explants, MMP activity was undetectable at all time-points (not shown). Although there was negligible type II collagen degradation under those conditions, there was extensive release of both proteoglycan and COMP. Therefore a clear correlation between MMP activity and type II collagen degradation was demonstrated, whereas both COMP and proteoglycans were released even in the absence of detectable MMPs.

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

270

Table 1 The 50% inhibition concentration (IC50) of matrix metalloproteinase inhibitors

Interstitial collagenase (MMP-1) Gelatinase A (MMP-2) Stromelysin-1 (MMP-3) Matrilysin (MMP-7) Neutrophil Collagenase (MMP-8) Gelatinase B (MMP-9) Collagenase-3 (MMP-13)

BB94

BB3437

BB3003

5 4 20 6 4 1 2

30 000 20 000 60 400 200 2000 100

)100 000 30 )100 000 )100 000 – 3000 –

IC50 values are expressed as nM concentrations and were supplied by Mr A. Galloway, British Biotech Pharmaceuticals Ltd, Oxford, UK.

2.3. Use of MMP inhibitors during explant culture To examine further the concept that COMP release from bovine cartilage explants may be mediated by proteolytic enzymes other than MMPs, selective and broad-spectrum MMP inhibitors were included in nasal and articular cartilage cultures. Each inhibitor was used at a concentration of 10 mM and their IC50 data against several MMPs are shown in Table 1. As the first significant effect of IL-1a on COMP release from nasal cartilage explants was observed in week 2 of culture (Fig. 1b), the effect of MMP inhibitors in this system was measured after 2 weeks. For articular cartilage cultures, any inhibition was measured after 1 week of IL-1a-stimulation, as the majority of COMP was released during the first week of culture (Fig. 1e). The effect of each inhibitor was determined by comparing the release of COMP in the presence of inhibitor to the release from explants cultured in IL-1a only. At a concentration of 10 mM, none of the inhibitors tested had a significant effect on COMP release from either nasal or articular cartilage explants (Table 2), although BB94, a broad-spectrum MMP inhibitor, partially inhibited COMP release from nasal cartilage. The cysteine proteinase inhibitors E64 and Ep453 (both used at a final concentration of 10 mM) and the isocoumarin derivative ACITIC (which inhibits serine proteinases with trypsin-like specificity; 0.1 mM), also had no significant effect on COMP release (not shown). 2.4. Analysis of COMP fragments In order to examine the fragments produced following IL-1a-stimulated release of COMP from bovine cartilage, conditioned medium samples were analysed by Western immunoblotting. For the reasons outlined above, week 1 conditioned medium was used from articular cartilage explants and week 2 medium from nasal cartilage explants. A polyclonal antibody raised to intact COMP detected two proteins in week 2 conditioned medium from control nasal cartilage explants (Fig. 3a, lane 2), which co-

migrated with the two major proteins present in purified bovine COMP (Fig. 3a, lane 1). COMP-120, which migrated with an apparent molecular mass of ;120 kDa, was intact COMP monomer, whilst COMP-100 appeared to be a fragment of ;100 kDa which was similar to previously reported degradation products of COMP (Morgelin et al., 1992; Vilim et al., 1997). In medium from IL-1a treated explants, COMP-120 was undetectable and COMP-100 only weakly so (Fig. 3a, lane 3). However, two major COMP fragments were detected, which migrated with apparent molecular masses of ;110 and ;90 kDa and were designated Fragment-110 and Fragment-90, respectively. The two fragments were not present in week 1 conditioned medium from IL-1a treated nasal cartilage explants, but were detected at all subsequent time points (not shown). An anti-peptide antibody raised to the 17 C-terminal residues of COMP, detected only COMP-120 in medium from control explants and Fragment-110 in medium from IL-1a-stimulated explants (Fig. 3b, lanes 2 and 3). When purified native COMP was cleaved with APMA-activated MMP-9, a major fragment was generated which co-migrated with Fragment-110 present in IL-1a-stimulated conditioned medium (Fig. 3a, lane 4). The fragment was also detected by the anti-peptide antibody, indicating it contained an intact C-terminus (Fig. 3b, lane 4). Only low levels of Fragment-90 were generated. Week 1 conditioned medium from control cultures of articular cartilage also contained COMP-120 and COMP-100 (Fig. 3c,d, lane 2). However, stimulation of articular cartilage explants with IL-1a did not deplete COMP-120 or COMP-100 and apparently generated only small amounts of Fragments-110 and -90 (Fig. 3c,d, lane 3). A similar pattern was observed at all subsequent time-points (not shown). Table 2 Inhibition of COMP release by proteinase inhibitors Inhibitor

10 mM BB94 10 mM BB3437 10 mM BB3003

Percentage inhibition of COMP release Nasal cartilage

Articular cartilage

33"15 y5"14 10"7

y4"9 0"5 4"3

Bovine articular or nasal cartilage explants were incubated with 3 nM IL-1a, in the presence or absence of the above inhibitors, for 1 or 2 weeks, respectively. COMP released into conditioned medium or remaining in cartilage residues was measured by inhibition ELISA. The effect of each inhibitor is expressed as the percentage inhibition of COMP release in the presence of inhibitor compared to culture in IL-1a only, after either 1 week (articular cartilage) or 2 weeks (nasal cartilage) of culture. Results show the mean"S.E.M. of 5 experiments. None of the inhibitors significantly blocked COMP release in comparison to explants cultured in IL-1a only (Mann–Whitney Utest).

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

271

Fig. 3. Analysis of COMP fragments in conditioned medium from bovine cartilage cultures. Medium samples from week 2 of culture of nasal cartilage explants (a, b) or from week 1 of culture of articular cartilage explants (c, d) were separated on 7.5% SDS-PAGE gels, under reducing conditions, and transferred to nitrocellulose membranes. Samples were probed with either a polyclonal antibody to COMP (a, c) or an anti-peptide antibody to the COMP C-terminus (b,d). In all panels, lane 1 contains standard bovine COMP, and lanes 2 and 3 medium from control and IL1a-stimulated explants, respectively. In some cases, MMP-9-cleaved COMP (lane 4) and medium from IL-1a-stimulated explants cultured in the presence of the synthetic MMP inhibitors BB94 (lane 5), BB3437 (lane 6) and BB3003 (lane 7), are also shown. Lane 8 contains extracts of nasal (a) or articular (c) cartilage explants stimulated with IL-1a for 2 or 1 weeks, respectively. The two molecules detected in standard COMP are shown as COMP-120 and -100, whilst the major products of IL-1a-induced COMP catabolism are shown as Fragments-110 and -90.

Both intact COMP monomer (COMP-120) and Fragment-110 were detected in extracts of articular and nasal cartilage explants stimulated with IL-1a for 1 or 2 weeks, respectively (Fig. 3a,c, lane 8), indicating that at least some of the COMP was degraded prior to release into the conditioned medium.

tors used at 10 mM) and ACITIC (a serine proteinase inhibitor used at 0.1 mM) also failed to inhibit COMP fragmentation in either the nasal or articular cartilage culture systems (not shown).

2.5. Inhibition of COMP fragmentation

As COMP was released from IL-1a-stimulated bovine nasal cartilage explants at a similar time to proteoglycan (Fig. 1), we investigated the possibility that COMP is a substrate for the aggrecanases ADAMTS-1, -4 andyor -5. When purified bovine COMP was incubated with 0.5U rhADAMTS-4 for 0–8 h at 37 8C, increasing amounts of Fragment-110 were detected by both the polyclonal antibody raised against intact COMP (Fig. 4a, lanes 1–5) and the anti-peptide antibody to the COMP C-terminus (not shown). The production of Fragment-110 was completely inhibited when ADAMTS-4 was pre-incubated with either BB-16 or epigallocatechin gallate (EGCG; Fig. 4a, lanes 6 and 7), potent inhibitors of ADAMTS-1, -4 and -5 at the concentrations used (Malfait et al., 2002; Vankemmelbeke et al., 2003). Under the conditions used, there was

IL-1a-stimulated bovine cartilage explants were cultured in the presence or absence of three different MMP inhibitors and the fragments of COMP released were analysed by Western immunoblotting. In 5 separate experiments, the broad-spectrum MMP inhibitor BB94 partially inhibited the IL-1a-induced COMP fragmentation observed in nasal cartilage conditioned medium (Fig. 3a, lane 5). BB94 also prevented the formation of the small amount of Fragment-110 released from IL-1astimulated articular cartilage explants (Fig. 3c, lane 5). However, neither BB3437 nor BB3003 inhibited the degradation observed in nasal (Fig. 3a, lanes 6 and 7) or articular cartilage cultures (Fig. 3c, lanes 6 and 7). Similarly E64, Ep453 (both cysteine proteinase inhibi-

2.6. Degradation of COMP by ADAMTS enzymes

272

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

Fig. 4. Cleavage of COMP by ADAMTS-1, -4 and -5. Purified bovine COMP was incubated at 37 8C with 0.5U rhADAMTS-4 (a) for 0 (lane 1), 2 (lane 2), 4 (lane 3), 6 (lane 4) and 8 h (lane 5). In lanes 6 and 7, rhADAMTS-4 was pre-incubated with 10 mM BB-16 and 5 mM EGCG, respectively. Purified bovine COMP was also incubated for 4 h at 37 8C with 1.5U rhADAMTS-5 (b) and 0.2U rhADAMTS-1 (c). All samples were separated on 7.5% SDS-PAGE gels, under reducing conditions, and transferred to nitrocellulose membranes. Proteins were detected using a polyclonal antibody raised against intact COMP and are labelled as in Fig. 3.

no formation of Fragment-110 when purified COMP was incubated for 4 h at 37 8C with either 1.5U rhADAMTS-5 (Fig. 4b) or 0.2U rhADAMTS-1 (Fig. 4c). 2.7. Analysis of COMP fragments in human synovial fluid samples Samples were obtained from patients with the active forms of nodal generalised OA (ns4), large joint OA (ns4), RA (ns3) and crystal pyruvate arthritis (ns 3). There were no differences in the COMP proteins detected in all patient groups, using the anti-peptide antibody to the COMP C-terminus, and representative samples are shown in Fig. 5. A fragment which comigrated with Fragment-110 in IL-1a-stimulated articular cartilage conditioned medium, was an abundant component of all the synovial fluid samples, whilst COMP-120 (intact COMP monomer) was also detected.

3. Discussion The results of this study indicate a role for proteinases other than MMPs in the degradation of COMP and demonstrate for the first time that COMP is a substrate for ADAMTS-4. Stimulation of bovine nasal cartilage explants with the catabolic cytokine IL-1a resulted in three sequential phases of matrix breakdown. An early phase of proteoglycan loss was followed by the more gradual release of COMP and then a final phase of type II collagen degradation. When total MMP activity released into the conditioned medium was measured using a quenched fluorescence substrate, the results clearly demonstrated that all detectable activity was seen only at the time of type II collagen degradation, after the onset of proteoglycan and COMP release. Therefore COMP, like aggrecan, appears not to be degraded by an MMP in this culture system. Further evidence for this comes from the fact that the broad-spectrum MMP inhibitor BB94,

Fig. 5. Western immunoblot analysis of human synovial fluid samples. Synovial fluids were separated on 7.5% SDS-PAGE gels, under reducing conditions, and transferred to nitrocellulose membranes. COMP was detected using an anti-peptide antibody to the COMP C-terminus. Typical results for synovial fluids from patients with nodal generalised OA (lanes 1 and 2), large joint OA (lanes 3 and 4), RA (lanes 5 and 6) and crystal pyruvate arthritis (lane 7) are shown. Week 1 conditioned medium from IL-1a-stimulated bovine articular cartilage cultures, is included for comparison (Med). Proteins are labelled as in Fig. 3.

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

which effectively inhibits MMPs-1, 2, 3, 7, 8, 9 and 13 at the concentration used, only partially inhibited COMP release from IL-1a-stimulated nasal cartilage. In addition, the more selective MMP inhibitors BB3437 (which inhibits MMPs-3, 7, 8 and 13 at the concentration used) and BB3003 (which inhibits MMP-2 with weaker activity against MMP-9 at the concentration used) also failed to inhibit COMP release. COMP was released from IL-1a-stimulated nasal cartilage during weeks 2–5 of culture as two major fragments. Fragment-110 resulted from N-terminal processing of the COMP molecule whereas Fragment-90 is believed to result from processing of COMP-100, a previously reported degradation product of COMP (Morgelin et al., 1992; Vilim et al., 1997). Although BB94 partially inhibited the COMP fragmentation in response to IL-1a, none of the three inhibitors had a significant effect on the process. Furthermore, in weeks 3–5 of culture, when high levels of active MMPs were detected in the nasal cartilage conditioned medium, there was no further breakdown of COMP Fragments-110 and 90 (not shown). Taken together, these results provide evidence that none of the known MMPs are involved in COMP release and degradation in the nasal cartilage culture system. The release of matrix proteins from IL-1a-stimulated articular cartilage explants, was quite different from the observed loss from nasal tissue. Previously we demonstrated that the early loss of proteoglycan from articular cartilage was not followed by subsequent type II collagen degradation (Kozaci et al., 1997). Here we demonstrate a second major difference between the nasal and articular cartilage model systems in that the majority of COMP was released from articular cartilage during the first week of culture from both control and IL-1astimulated explants. In addition, although Fragment-110 was first released during week 1 of culture, only low levels were observed and intact COMP monomer, COMP-120, was the major component detected. Fragment-110 is believed to be equivalent to the 67–80 kDa fragment detected by Ganu et al. under non-reducing conditions (the unfolding of disulfide loops in the COMP monomer ensures the fragment migrates at a higher apparent molecular mass following reduction) (Ganu et al., 1998). In order to detect such fragments, Ganu et al. used gram quantities of cartilage and concentrated resulting medium 3–4 fold, indicating they too detected only low levels of fragments equivalent to Fragment-110. The fact that large amounts of COMP were released even from control explants and IL-1a had very little effect on either the release or degradation of the molecule, suggests that COMP is only loosely bound within the articular cartilage matrix under these culture conditions, or that other, as yet unidentified, components are capable of competition for COMP binding sites in articular cartilage. It is unlikely that COMP released in

273

the presence of IL-1a represents newly synthesised molecules not incorporated into the ECM, as the cytokine has no effect on the basal level of COMP synthesis in cultured articular chondrocytes (Recklies et al., 1998). The small amounts of Fragment-110 that were released from IL-1a-stimulated articular cartilage explants, were inhibited by BB94 but not BB3437 or BB3003. In addition, negligible MMP activity was detected in conditioned medium throughout the culture period, indicating MMPs were not responsible for producing Fragment-110 in this system. Although the results presented here indicate that COMP was not degraded by any of the known MMPs in these bovine cartilage culture systems, it has been demonstrated both here and previously that COMP is a substrate for MMP-9 and other MMPs (Ganu et al., 1998; Stracke et al., 2000). Indeed such enzymes produce fragments which co-migrate with Fragment-110 in bovine cartilage cultures and human arthritic synovial fluid samples (Figs. 3 and 4; Ganu et al., 1998). However, 10 mM BB3437 and BB3003 failed to inhibit either the release or fragmentation of COMP from nasal cartilage and 10 mM BB94 only partially inhibited these processes. In addition, Ganu et al. demonstrated that although BB94 and CGS 27023A partially inhibited fragment formation after 3 days of culture, they had virtually no effect after 7 or 11 days (Ganu et al., 1999). It is therefore possible that the cleavage of COMP is mediated in part by an as yet unidentified proteinase which does not cleave the quenched fluorescence substrate used in the MMP assay and is not efficiently inhibited by the synthetic MMP inhibitors BB94, BB3437, BB3003 or CGS 27023A. Alternatively, the enzyme responsible for cleaving COMP may be retained within the cartilage matrix rather than released into the conditioned medium. ADAMTS enzymes are a family of proteins believed to be anchored to the ECM through interactions with the GAG side chains of aggrecan or other matrix components, via one or more thrombospondin type I motifs (Tortorella et al., 2000). It has recently been shown that the fragments of aggrecan present in arthritic synovial fluid samples and IL-1a-stimulated cartilage conditioned medium, mostly result from the activity of the enzymes ADAMTS-4 and -5 (Malfait et al., 2002; Tortorella et al., 2001). In addition, ADAMTS-1 can also cleave aggrecan and is up-regulated by IL-1 (Kuno et al., 1997, 2000), although a role for this enzyme in cartilage aggrecan turnover in vivo remains to be determined. As aggrecan and COMP were released at similar times in culture (Fig. 1), we investigated the possibility that COMP may also be a substrate for these enzymes. rhADAMTS-4 was shown to cleave purified COMP to yield a protein which co-migrated with Fragment-110, the fragment also detected following IL-1a-stimulation of bovine cartilage. The formation of the fragment was

274

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

prevented by BB-16 and EGCG, potent inhibitors of ADAMTS-4 (Malfait et al., 2002; Vankemmelbeke et al., 2003). Although it has previously been demonstrated that ADAMTS-4 cleaves aggrecan, versican and brevican (Matthews et al., 2000; Nakamura et al., 2000; Sandy et al., 2001; Tortorella et al., 1999), neither ADAMTS-1, -4 nor -5 cleave any of the major collagens or glycoproteins such as fibronectin, vitronectin or thrombospondin 1 (Rodriguez-Manzaneque et al., 2002; Tortorella et al., 2002). Therefore COMP represents the first substrate of ADAMTS-4 to be identified which is not a member of the large aggregating proteoglycan family. This opens new possibilities for enzyme recognition of substrate, in that COMP does not contain any of the GAG side chains suggested to play a role in aggrecan cleavage by ADAMTS-4 (Tortorella et al., 2000). Under the conditions used, neither rhADAMTS1 nor -5 cleaved COMP to produce Fragment-110, suggesting the cleavage was specific to ADAMTS-4. However, it is possible that the former two enzymes can cleave COMP but are much less efficient. Indeed ADAMTS-4 is more active than both ADAMTS-1 and -5 against purified aggrecan (Tortorella et al., 2002). If ADAMTS-4 is responsible for COMP degradation in these explant culture model systems, the fact that the degradation of COMP was delayed until the majority of the proteoglycan had been released from IL-1a-stimulated bovine nasal cartilage explants (Fig. 1), suggests that ADAMTS-4 cleaves aggrecan more efficiently than COMP. Fragments which contained an intact C-terminus and co-migrated with Fragment-110 in IL-1a-stimulated bovine cartilage conditioned medium, were an abundant component of synovial fluid samples from a range of arthritis patients. Such fragments may result from COMP degraded within the cartilage matrix and released, as is observed in IL-1a-treated bovine nasal cartilage cultures. Support for such a mechanism comes from the fact that fragments of COMP have been detected in diseased cartilage (DiCesare et al., 1996) and were also detected by Western immunoblotting in bovine nasal cartilage explants stimulated with IL-1a for 2 weeks (Fig. 3a, lane 8). Alternatively, COMP may be released from cartilage intact and cleaved in the synovial fluid by enzymes released from chondrocytes or synthesised by synovial cells. This mechanism is in agreement with the articular cartilage model system where COMP was released in a predominantly intact form even in response to IL-1a. A further possibility is that fragments present in synovial fluid samples result from degradation of COMP produced by the synovium itself. Although synovial fibroblasts can synthesise and secrete COMP (Dodge et al., 1998; Recklies et al., 1998), levels of the molecule in the synovium are several fold lower than in the synovial fluid (Skioldebrand et al., 2001). This indicates that in vivo, the transport is largely from the

joint compartment to the lining tissue. It would therefore seem likely that the COMP fragments present in synovial fluid originate primarily from cartilage, with a relatively small contribution from other connective tissues such as tendon, ligament and synovium (DiCesare et al., 1997, 1994; Muller et al., 1998). However, the enzymes responsible for degrading COMP may be produced by a variety of cell types. In this study we have demonstrated a role for proteinases other than MMPs in the release of COMP from bovine cartilage model systems and shown that purified COMP is a substrate for ADAMTS-4. However, it has also been demonstrated that COMP is a substrate for several MMPs in vitro. Identification of the enzymes responsible for COMP proteolysis in vivo, may provide new pharmacological targets for the prevention of cartilage degradation. If ADAMTS-4 andyor other ADAMTS enzymes prove to be responsible for both COMP degradation and aggrecan breakdown, specific inhibitors of this group of enzymes may provide vital protection for the ECM following cartilage injury and in joint disease. 4. Experimental procedures 4.1. Anti-COMP antibodies A polyclonal antibody to bovine COMP was raised in rabbits and traces of antibodies to fibronectin were removed by passing the antiserum over a fibronectin affinity column as described previously (Hedbom et al., 1992). An anti-peptide antibody to the C-terminal 17 amino acids of human COMP (CNDTIPEDYETHQLRQA) was raised in rabbits using a conventional immunisation scheme, where the natural N-terminal cysteine residue was used to couple the peptide to keyhole limpet haemocyanin. In Western immunoblot analysis of 4 M guanidine hydrochloride cartilage extracts, the antibody reacted only with COMP (data not shown). 4.2. Bovine cartilage explant cultures Bovine nasal septum cartilage, obtained from adult animals, was sliced thinly and washed with sterile PBS for 20 min. Disks of approximately 30–40 mg wet weight were prepared using a sterile belt punch. Fulldepth bovine articular cartilage pieces (f30 mg wet weight) were dissected from metacarpophalangeal joints, under sterile conditions, and washed with PBS. The nasal and articular cartilage explants were cultured in serum-free DMEM containing glutamine (2 mM), penicillin G (200 unitsyml), streptomycin (0.1 mgyml) and HEPES (10 mM) (all from Invitrogen, Paisley, UK) at 37 8C in a humidified atmosphere of 5% CO2 y95% air. The explants were cultured in 48-

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

well plates (Costar, High Wycombe, UK) with each well containing 2 pieces of cartilage and 400 ml of medium. We have shown previously that treatment of nasal cartilage with IL-1a for the first two weeks of culture only, is sufficient to stimulate subsequent type II collagen degradation (Kozaci et al., 1997). Explants were therefore cultured in DMEM containing 3 nM IL1a (a kind gift from Dr Craig Reynolds, National Cancer Institute, Frederick, MD) for 2 weeks, followed by DMEM only for a further 2 or 3 weeks. Control explants were cultured in DMEM only for the entire culture period. In some experiments, broad-spectrum or selective MMP inhibitors (kind gifts from Mr A. Galloway, British Biotech Pharmaceuticals, Oxford, UK) were included during culture. The inhibitors were all used at a concentration of 10 mM and their IC50 data against several MMPs are shown in Table 1. The cysteine proteinase inhibitors E64 and Ep453 (Sigma, Poole, Dorset, UK; both at 10 mM) and ACITIC (a kind gift from Prof. J.C. Powers, Georgia Institute of Technology, Atlanta, Georgia, USA; 0.1 mM), which inhibits serine proteinases with trypsin-like specificity, were also included in some cases. All inhibitors were dissolved in dimethyl sulfoxide (Sigma), the final concentration of which (0.1% (vyv)) had no effect on COMP, proteoglycan or type II collagen release (not shown). In all cases, DMEM, IL-1a and inhibitors were replenished at the end of each week and conditioned medium stored at y20 8C, together with undigested cartilage residues that remained at the end of the culture period. 4.3. Measurement of COMP release Nasal and articular cartilage residues collected at the end of culture were milled in liquid nitrogen, to obtain a fine particulate, and COMP was extracted with 4 M guanidine hydrochloride, containing proteinase inhibitors, as previously described (Hedbom et al., 1992). The resulting extracts were precipitated twice with 95% (vyv) ethanol to ensure all traces of guanidinium were removed. Extracts were dried at 37 8C and frozen until assay. The amount of COMP present in cartilage extracts and conditioned medium samples was measured by inhibition ELISA, using a modification of the assay for human COMP (Saxne and Heinegard, 1992). The modifications included the use of bovine COMP for coating microtitre plates and for standardization of the assay, and the use of the polyclonal anti-bovine COMP antiserum described above, for protein detection. The amount of COMP released into the medium during each week of culture was expressed as a percentage of total COMP in that culture well (mediumqresidue). 4.4. Measurement of type II collagen degradation Type II collagen remaining in the cartilage residues at the end of the experiments or released into the

275

medium during culture, was extracted by digestion with 1 mgyml proteinase K (EC 3.4.21.64; Sigma) at 56 8C for 15 h. The extracts were assayed by inhibition ELISA using a mouse IgG monoclonal antibody to denatured type II collagen (COL2-3y4m) as previously described (Hollander et al., 1994). The amount of type II collagen released into the medium during each week of culture was expressed as a percentage of total collagen in that culture well (mediumqresidue). Collagen release determined by this ELISA correlates very closely with release of hydroxyproline into the tissue culture medium (A.P. Hollander, unpublished observations). 4.5. Measurement of proteoglycan release Cartilage residues and medium samples were digested with proteinase K as described above. Proteoglycan in the digests was measured as sulfated glycosaminoglycan (sGAG) by colorimetric assay using 1,9-dimethylmethylene blue (Aldrich, Gillingham, UK) as previously described (Handley and Buttle, 1995). The amount of proteoglycan released into the medium during each week of culture was expressed as a percentage of total proteoglycan in that culture well (mediumqresidue). 4.6. Fluorimetric analysis of MMP activity The quenched fluorescence substrate 7-methoxycoumarin-4-ylAcetyl (Mca)–Pro–Leu–Gly–Leu–(3-w2,4(Dpa)–Ala– dinitrophenylx-L-2,3-diaminopropionyl) Arg–NH2 (from Dr C. Graham Knight, Strangeways Research Laboratory, Cambridge, UK), which is cleaved efficiently by all MMPs tested to date (Brown et al., 1996), was used to evaluate the total MMP activity in conditioned medium samples (Knight et al., 1992). The assays were performed at 37 8C, as previously described (Price et al., 1999), using a Perkin Elmer LS50B fluorimeter linked to a computer running the FLUSYS software (Rawlings and Barrett, 1990). The steady-state rate of cleavage was expressed as the amount of product produced per minute per cartilage explant. 4.7. Detection of COMP by western immunoblot analysis Samples of conditioned medium from control or IL1a-treated nasal and articular cartilage explants were separated by 7.5% SDS-PAGE, under reducing conditions, and transferred to nitrocellulose membranes (BioRad, Hemel Hempstead, UK), as previously described (Price et al., 1999). Proteins were detected with either the polyclonal rabbit anti-bovine COMP antibody or the goat anti-peptide COMP antibody described above, both diluted 1:150. Goat anti-rabbit (Southern Biotechnology, Birmingham, AL) or rabbit anti-goat (Vector Laboratories, Peterborough, UK) alkaline phosphatase-conjugated secondary antibodies were used, at dilutions of

276

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

1:1000 and 1:600, respectively. Alkaline phosphatase substrate solution, prepared from a commercial kit (BioRad), was incubated with the membranes until optimal colour development. The COMP present in human synovial fluid samples was also analysed by Western immunoblotting. The samples were collected by aspiration from the knee joints of patients with the active forms of nodal generalised OA (ns4), large joint OA (ns4), RA (ns3) or crystal pyruvate arthritis (ns3) (by Dr S.H. Till, University of Sheffield, UK). Synovial fluid (100 ml) was incubated for 2 h, at 37 8C, with 10 ml 300 unitsy ml bovine testicular hyaluronidase (EC 3.2.1.35), in 0.1 M NaCl, 0.1 M sodium acetate, 10 mM EDTA, 50 mM phenylmethylsulfonyl fluoride and 0.17 M N-ethylmaleimide (all from Sigma), pH 6.0. The samples were analysed by Western immunoblotting as described above. In order to determine the apparent molecular mass of COMP proteins, the relative migration positions of samples and molecular mass standards were analysed using an IS-1000 Digital Imaging System (Alpha Innotech Corporation, San Leandro, CA). 4.8. Cleavage of purified COMP with ADAMTS-1, -4 and -5 and MMP-9 Recombinant human (rh) ADAMTS-1, -4 and -5 were all purified from conditioned culture medium of HighFive cells (Invitrogen) transfected with rhADAMTS-1, -4 and -5 expression vectors, as previously described (Vankemmelbeke et al., 2003). Native COMP (6 mg), purified from adult bovine articular cartilage under non-reducing conditions (Rosenberg et al., 1998), was incubated at 37 8C with either 0.5U rhADAMTS4, 1.5U rhADAMTS-5 or 0.2U rhADAMTS-1, for between 0 and 8 h, in 0.1 M Tris–HCl, 0.1 M NaCl, 10 mM CaCl2 and 0.1% (wyv) CHAPS, pH 7.5. A unit of enzyme activity was defined as that which released 5 mg sGAG per hour at 37 8C, when the rhADAMTS enzymes were incubated with aggrecan entrapped in polyacrylamide as a substrate (Vankemmelbeke et al., 2001). In some cases the enzymes were pre-incubated with 10 mM BB-16 (a kind gift from Mr A. Galloway, British Biotech Pharmaceuticals, Oxford, UK) or 5 mM epigallocatechin gallate (EGCG; Sigma), at 4 8C for 30 min, prior to incubation with COMP. Purified human pro-MMP-9 (Biogenesis, Poole, UK), which had been activated with 2 mM p-aminophenylmercuric acetate (APMA; Sigma) in 0.1 M Tris–HCl, 0.2 M NaCl, 20 mM CaCl2 and 0.12% (vyv) Brij-35 (Sigma), pH 7.5, was incubated with native bovine COMP (enzyme: substrate weight ratio of 1:25) at 37 8C for 15 h. The MMP-9 was then inactivated by the

addition of EDTA to a final concentration of 20 mM. Degradation of COMP by the different enzymes was detected by Western immunoblotting as described above. 4.9. Statistical analysis of data Differences in amounts of proteoglycan, COMP and type II collagen in cartilage explants were determined by the 2-tailed Mann–Whitney U-test with P-0.05 taken as significant. References Amour, A., Knight, C.G., Webster, A., et al., 2000. The in vitro activity of ADAM-10 is inhibited by TIMP-1 and TIMP-3. FEBS Lett. 473, 275–279. Briggs, M.D., Hoffman, S.M.G., King, L.M., et al., 1995. Pseudoachondroplasia and multiple epiphyseal dysplasia due to mutations in the cartilage oligomeric matrix protein gene. Nat. Genet. 10, 330–336. Brown, C.J., Rahman, S., Morton, A.C., Beauchamp, C.L., Bramwell, H., Buttle, D.J., 1996. Inhibitors of collagenase but not gelatinase reduce cartilage explant proteoglycan breakdown despite only low levels of matrix metalloproteinase activity. J. Clin. Pathol. Mol. Pathol. 49, M331–M339. Chubinskaya, S., CS-Szabo, G., Kuettner, K.E., 1998. ADAM-10 message is expressed in human articular cartilage. J. Histochem. Cytochem. 46, 723–729. DiCesare, P.E., Carlson, C.S., Stolerman, E.S., Hauser, N., Tulli, H., Paulsson, M., 1996. Increased degradation and altered tissue distribution of cartilage oligomeric matrix protein in human rheumatoid and osteoarthritic cartilage. J. Orthop. Res. 14, 946–955. DiCesare, P.E., Carlson, C.S., Stollerman, E.S., Chen, F.S., Leslie, M., Perris, R., 1997. Expression of cartilage oligomeric matrix protein by human synovium. FEBS Lett. 412, 249–252. DiCesare, P.E., Chen, F.S., Moergelin, M., et al., 2002. Matrix–matrix interaction of cartilage oligomeric matrix protein and fibronectin. Matrix Biol. 21, 461–470. DiCesare, P.E., Hauser, N., Lehman, D., Pasumarti, S., Paulsson, M., 1994. Cartilage oligomeric matrix protein (COMP) is an abundant component of tendon. FEBS Lett. 354, 237–240. Dodge, G.R., Hawkins, D., Boesler, E., Sakai, L., Jimenez, S.A., 1998. Production of cartilage oligomeric matrix protein (COMP) by cultured human dermal and synovial fibroblasts. Osteoarthr. Cartilage 6, 435–440. Flannery, C.R., Little, C.B., Caterson, B., Hughes, C.E., 1999. Effects of culture conditions and exposure to catabolic stimulators (IL-1 and retinoic acid) on the expression of matrix metalloproteinases (MMPs) and disintegrin metalloproteinases (ADAMs) by articular cartilage chondrocytes. Matrix Biol. 18, 225–237. Ganu, V., Goldberg, R., Peppard, J., et al., 1998. Inhibition of interleukin-1a-induced cartilage oligomeric matrix protein degradation in bovine articular cartilage by matrix metalloproteinase inhibitors. Arthritis Rheum. 41, 2143–2151. Ganu, V., Melton, R., Wang, W., Roberts, D., 1999. Matrix metalloproteinase inhibitor CGS 27023A protects COMP and proteoglycan in the bovine articular cartilage but not the release of their fragments from cartilage after prolonged stimulation in vitro with IL-1a. Ann. N.Y. Acad. Sci. 878, 607–611.

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278 Goldberg, R.L., Spirito, S., Doughty, J.R., Ganu, V., Heinegard, D., 1995. Time dependent release of matrix components from bovine cartilage after IL-1 treatment and the relative inhibition by matrix metalloproteinase inhibitors. Trans. Orthop. Res. Soc. 20, 125. Handley, C.J., Buttle, D.J., 1995. Assay of proteoglycan degradation. Methods Enzymol. 248, 47–58. Hecht, J.T., Nelson, L.D., Crowder, E., et al., 1995. Mutations in exon 17B of cartilage oligomeric matrix protein (COMP) cause pseudoachondroplasia. Nat. Genet. 10, 325–329. Hedbom, E., Antonsson, P., Hjerpes, A., et al., 1992. Cartilage matrix proteins. an acidic oligomeric protein (COMP) detected only in cartilage. J. Biol. Chem. 267, 6132–6136. Hollander, A.P., Heathfield, T.F., Webber, C., et al., 1994. Increased damage to type II collagen in osteoarthritic articular cartilage detected by a new immunoassay. J. Clin. Invest. 93, 1722–1732. Hollander, A.P., Pidoux, I., Reiner, A., Rorabeck, C., Bourne, R., Poole, A.R., 1995. Damage to type II collagen in aging and osteoarthritis starts at the articular surface, originates around chondrocytes, and extends into the cartilage with progressive degeneration. J. Clin. Invest. 96, 2859–2869. Hummel, K.M., Neidhart, M., Vilim, V., et al., 1998. Analysis of cartilage oligomeric matrix protein (COMP) in synovial fibroblasts and synovial fluids. Br. J. Rheumatol. 37, 721–728. Knight, C.G., Willenbrock, F., Murphy, G., 1992. A novel coumarinlabelled peptide for sensitive continuous assays of the matrix metalloproteinases. FEBS Lett. 296, 263–266. Kozaci, L.D., Buttle, D.J., Hollander, A.P., 1997. Degradation of type II collagen, but not proteoglycan, correlates with matrix metalloproteinase activity in cartilage explant cultures. Arthritis Rheum. 40, 164–174. Kuno, K., Kanada, N., Nakashima, E., Fujiki, F., Ichimura, F., Matsushima, K., 1997. Molecular cloning of a gene encoding a new type of metalloproteinase-disintegrin family protein with thrombospondin motifs as an inflammation associated gene. J. Biol. Chem. 272, 556–562. Kuno, K., Okada, Y., Kawashima, H., et al., 2000. ADAMTS-1 cleaves a cartilage proteoglycan, aggrecan. FEBS Lett. 478, 241–245. Lohmander, L.S., Neame, P.J., Sandy, J.D., 1993. The structure of aggrecan fragments in human synovial fluid. Evidence that aggrecanase mediates cartilage degradation in inflammatory joint disease, joint injury and osteoarthritis. Arthritis Rheum. 36, 1214–1222. Malfait, A.-M., Liu, R.-Q., Ijiri, K., Komiya, S., Tortorella, M.D., 2002. Inhibition of ADAM-TS4 and ADAM-TS5 prevents aggrecan degradation in osteoarthritic cartilage. J. Biol. Chem. 277, 22201–22208. Matthews, R.T., Gary, S.C., Zerillo, C., et al., 2000. Brain-enriched hyaluronan binding (BEHAB)yBrevican cleavage in a glioma cell line is mediated by a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) family member. J. Biol. Chem. 275, 22695–22703. Misumi, K., Vilim, V., Clegg, P.D., Thompson, C.C.M., Carter, S.D., 2001. Measurement of cartilage oligomeric matrix protein (COMP) in normal and diseased equine synovial fluids. Osteoarthr. Cartilage 9, 119–127. Morgelin, M., Heinegard, D., Engel, J., Paulsson, M., 1992. Electron microscopy of native cartilage oligomeric matrix protein purified from the Swarm rat chondrosarcoma reveals a five-armed structure. J. Biol. Chem. 267, 6137–6141. Muller, G., Michel, A., Altenburg, E., 1998. COMP (cartilage oligomeric matrix protein) is synthesized in ligament, tendon, meniscus and articular cartilage. Conn. Tissue Res. 39, 233–244. Nakamura, H., Fujii, Y., Inoki, I., et al., 2000. Brevican is degraded

277

by matrix metalloproteinases and aggrecanase-1 (ADAMTS4) at different sites. J. Biol. Chem. 275, 38885–38890. Neidhart, M., Hauser, N., Paulsson, M., DiCesare, P.E., Michel, B.A., Hauselmann, H.J., 1997. Small fragments of cartilage oligomeric matrix protein in synovial fluid and serum as markers for cartilage degradation. Br. J. Rheumatol. 36, 1151–1160. Poole, A.R., 1993. Cartilage in health and disease. In: McCarty, D.J., Koopman, W.J. (Eds.), Arthritis and Allied Conditions. A Textbook of Rheumatology. Lea and Febiger, Philadelphia, pp. 279–333. Price, J.S., Wang-Weigand, S., Bohne, R., Kozaci, L.D., Hollander, A.P., 1999. Retinoic acid-induced type II collagen degradation does not correlate with matrix metalloproteinase activity in cartilage explant cultures. Arthritis Rheum. 42, 137–147. Rawlings, N.D., Barrett, A.J., 1990. FLUSYS: a software package for the collection and analysis of kinetic and scanning data from Perkin-Elmer fluorimeters. Comput. Appl. Biosci. 6, 118–119. Recklies, A.D., Baillargeon, L., White, C., 1998. Regulation of cartilage oligomeric matrix protein synthesis in human synovial cells and articular chondrocytes. Arthritis Rheum. 41, 997–1006. Rodriguez-Manzaneque, J.C., Westling, J., Thai, S.N.-M., et al., 2002. ADAMTS1 cleaves aggrecan at multiple sites and is differentially inhibited by metalloproteinase inhibitors. Biochem. Biophys. Res. Comm. 293, 501–508. Rosenberg, K., Olsson, H., Morgelin, M., Heinegard, D., 1998. Cartilage oligomeric matrix protein shows high affinity zincdependent interaction with triple helical collagen. J. Biol. Chem. 273, 20397–20403. Sandy, J.D., Westling, J., Kenagy, R.D., et al., 2001. Versican V1 proteolysis in human aorta in vivo occurs at the Glu441–Ala442 bond, a site that is cleaved by recombinant ADAMTS-1 and ADAMTS-4. J. Biol. Chem. 276, 13372–13378. Saxne, T., Heinegard, D., 1992. Cartilage oligomeric matrix protein: a novel marker of cartilage turnover detectable in synovial fluid and blood. Br. J. Rheumatol. 31, 583–591. Skioldebrand, E., Lorenzo, P., Zunino, L., Ekman, S., 2001. Concentration of collagen, aggrecan and cartilage oligomeric matrix protein in synovial fluid from equine middle carpal joints. Equine Vet. J. 33, 394–402. Smith, R.K.W., Zunino, L., Webbon, P.M., Heinegard, D., 1997. The distribution of cartilage oligomeric matrix protein (COMP) in tendon and its variation with tendon site, age and load. Matrix Biol. 16, 255–271. Stracke, J.O., Fosang, A.J., Last, K., et al., 2000. Matrix metalloproteinases 19 and 20 cleave aggrecan and cartilage oligomeric matrix protein (COMP). FEBS Lett. 478, 52–56. Sugimoto, K., Takahashi, M., Yamamoto, Y., Shimada, K., Tanzawa, K., 1999. Identification of aggrecanase activity in medium of cartilage culture. J. Biochem. 126, 449–455. Thur, J., Rosenberg, K., Nitsche, D.P., et al., 2001. Mutations in cartilage oligomeric matrix protein causing pseudoachondroplasia and multiple epiphyseal dysplasia affect binding of calcium and collagen I, II, and IX. J. Biol. Chem. 276, 6083–6092. Tortorella, M., Pratta, M., Liu, R.-Q., et al., 2000. The thrombospondin motif of aggrecanase-1 (ADAMTS-4) is critical for aggrecan substrate recognition and cleavage. J. Biol. Chem. 275, 25791–25797. Tortorella, M.D., Burn, T.C., Pratta, M.A., et al., 1999. Purification and cloning of aggrecanase-1: a member of the ADAMTS family of proteins. Science 284, 1664–1666. Tortorella, M.D., Liu, R.-Q., Burn, T., Newton, R.C., Arner, E., 2002. Characterization of human aggrecanase 2 (ADAM-TS5): substrate

278

S.C. Dickinson et al. / Matrix Biology 22 (2003) 267–278

specificity studies and comparison with aggrecanase 1 (ADAMTS4). Matrix Biol. 21, 499–511. Tortorella, M.D., Malfait, A.-M., Deccico, C., Arner, E., 2001. The role of ADAM-TS4 (aggrecanase-1) and ADAM-TS5 (aggrecanase-2) in a model of cartilage degradation. Osteoarthr. Cartilage 9, 539–552. Vankemmelbeke, M.N., Holen, I., Wilson, A.G., et al., 2001. Expression and activity of ADAMTS-5 in synovium. Eur. J. Biochem. 268, 1259–1268.

Vankemmelbeke, M.N., Jones, G.C., Fowles, C., Ilic, M.Z., Handley, C.J., Day, A.J., Knight, C.G., Mort, J.S., Buttle, D.J. Selective inhibition of ADAMTS-1, -4 and -5 by catechin gallate esters. Eur. J. Biochem., 270, 1–10. Vilim, V., Lenz, M.E., Vytasek, R., et al., 1997. Characterisation of monoclonal antibodies recognising different fragments of cartilage oligomeric matrix protein in human body fluids. Arch. Biochem. Biophys. 341, 8–16.