(Received for publication, June 2, 1997)
From the Connective Tissue Biology Laboratories,
School of Molecular and Medical Biosciences, University of Wales,
Cardiff CF1 3US, Wales, United Kingdom and ¶ Molekularbiologie, TA
Rheumatologie Immunology, H 528, Hoechst AG, Werk Kalle-Albert,
Rheingaustrasse 190, 65174 Wiesbaden, Germany
This paper describes the first report of the production and use of an artificial recombinant protein substrate to study "aggrecanase" activity. The substrate (rAgg1) is composed of the complete interglobular domain (IGD) of human aggrecan flanked by the "marker" sequences FLAGTM at the amino terminus and the human immunoglobulin G1 constant region at the carboxyl terminus. The expressed protein occurs as large multimolecular aggregates (>120 kDa) that, upon reduction, consist of a major isoform of 72 kDa (containing the IGD) and a minor 39-kDa species that through alternative splicing has had the IGD deleted. Using this recombinant substrate we developed a novel agarose cell culture system containing either rat chondrosarcoma or bovine chondrocytes that could be used in studies of the biochemical characterization of aggrecanase activities. These studies showed the following. (i) rAgg1 is a suitable substrate for aggrecanase proteolysis. (ii) Aggrecanase activity was specifically induced by exposing chondrocytes to retinoic acid. (iii) A considerable time period was required to synthesize and/or activate aggrecanase, with considerable differences in that found in rat chondrosarcoma versus bovine chondrocyte culture systems. (iv) Aggrecanase cleavage of the aggrecan IGD does not require the presence of the G1 or G2 globular domains or keratan sulfate post-translational modification in the IGD. (v) Aggrecanase is a diffusible activity that does not require association with the chondrocyte plasma membrane or immediate pericellular matrix for its action.
Mechanisms of proteoglycan breakdown in connective tissue are complex and involve multiple agents and pathways. In studies investigating the catabolism of aggrecan, the experimental systems used have included monolayer cultures of primary chondrocytes (1), established chondrocyte cell lines (2, 3), and explant cultures using cartilage from a variety of anatomical sites and animal species (4-6). Cytokines such as interleukin-1 and tumor necrosis factor have been extensively used as agents that promote the degradation of the extracellular matrix (2, 7, 8). In particular, these two cytokines have been shown to target the catabolism of aggrecan, the major proteoglycan present in hyaline articular cartilages. Several studies have now led to a number of important discoveries, which have defined at least seven specific cleavage sites along the protein core of aggrecan (9-12). Amino acid sequence analysis of cartilage proteoglycan breakdown products has defined the two major sites of proteolytic cleavage in aggrecan, both found within the interglobular domain (IGD)1 of the aggrecan core protein. These occur between amino acid residues Asn341-Phe342 and Glu373-Ala374 (human sequence enumeration (13)).
Several matrix metalloproteinases (MMP1, -2, -3, -7, -8, -9, and -13) have now been shown to cleave aggrecan at the former cleavage site. The second cleavage site was first described in the early 1990s (10-12). Now, however, almost 6 years later, the agent(s) responsible for cleavage at the Glu373-Ala374 peptide bond still remain to be identified. This uncharacterized activity, termed "aggrecanase" (12, 14), appears to have specificity for Glu-Xaa peptide bonds where Xaa is Ala, Gly, or Leu (9, 11, 15). The majority of aggrecan catabolites found in the synovial fluids of patients with arthritis appear to be generated by aggrecanase activity (15, 16). In a recent study, Fosang et al. (17) reported the presence of aggrecan catabolites in synovial fluid that contained the amino-terminal sequence FFG ... , which indicates that matrix metalloproteinases are also involved in aggrecan catabolism.
Culture systems have been manipulated with a variety of agents that enhance proteoglycan catabolism in efforts to discover the agent(s) responsible for aggrecan breakdown. However, these studies have been unsuccessful in defining any specific agent responsible for the degradation of aggrecan (18). Nonetheless, experimental data using purified aggrecan and modified aggrecan as a substrate for testing purified enzyme preparations in vitro have yielded some information on the specific cleavage sites of the enzymes themselves, but this work has not directly helped in determining the identity or biochemical properties of aggrecanase (14, 19, 20).
In previous studies (2, 21) we have developed a number of monoclonal antibodies (neoepitope antibodies) that recognize unique proteoglycan degradation products that have been generated by specific proteinases including aggrecanase. These antibodies have proved useful as tools in studying the mechanisms of the breakdown of proteoglycan in cartilage explant culture systems. The purpose of this current work was to further refine culture systems that would facilitate the identification of the agents responsible for the cleavage of aggrecan at the 374ARGSVI ... site in the IGD of aggrecan. In this study, we describe the development of an in vitro culture system that has enabled us to study the activity of aggrecanase against an artificial recombinant substrate in a cell culture system that is free of inherent endogenous proteoglycans and other extracellular components. The use of previously characterized neoepitope antibodies to key cleavage sites in the IGD of aggrecan facilitated the monitoring of products generated by the proteolytic action of aggrecanase in this system.
Alkaline phosphatase-conjugated second antibody and substrate used in Western blot analysis were obtained from Promega as the Protoblot Western blot AP system (catalog no. W3920). Nitrocellulose (0.2-µm pore size) was obtained from Schleicher and Schuell. Monoclonal antibody M1, which recognizes the FLAGTM epitope, and anti-FLAG M1 affinity gel were both obtained from Eastman Kodak Co. Anti-human IgG monoclonal antibody was obtained from Cappel (Durham, NC). Monoclonal antibody BC-3 was prepared as ascitic fluid (2). The Rx cell line was kindly provided by Dr. James Kimura, Bone Research Center, Henry Ford Hospital, Detroit, MI.
Preparation of the rAgg1 Genetic ConstructThe genetic
construct for rAgg1 codes for the signal sequence of the lymphocyte
glycoprotein T1/Leu-1 (CD5), the 8-amino acid-long FLAG epitope (22),
the 127-amino acid-long interglobular domain of the human cartilage
large aggregating proteoglycan aggrecan (Thr331-Gly457, assuming Val20 is
the native amino terminus (13)), a 4-amino acid-long spacer (GDPE), and
a 232-amino acid polypeptide sequence from the human IgG1 hinge, CH2
and CH3 constant regions (see Fig. 1),
thus giving rise to a secreted fusion protein with an expected
molecular mass of 41,059. Aggrecan cDNA sequences encoding the
interglobular domain were amplified by the reverse transcriptase
polymerase chain reaction (23) with synthetic oligonucleotides
complementary to sequences flanking this region. Total RNA was
extracted from human cartilage tissue obtained from joint replacement
surgery according to Adams et al. (24). Oligonucleotides
were designed to contain additional sequence information regarding the
8-amino acid FLAG epitope and the GDPE to allow the creation of
restriction enzyme cleavage sites at the 5 and 3
extremities of the
amplified cDNA segments to facilitate subsequent insertion into the
CD5-IgG1 expression vector, modified to contain a 3
NheI
site (25). The primer encoding the FLAG epitope, the amino-terminal
start of the interglobular domain, and including an NheI
site was synthesized with the following sequence: 5
-CGC GGG GCT AGC
CGA CTA CAA GGA CGA CGA TGA CAA GAC AGG TGA AGA CTT TGT GGA C. A
reverse primer encoding the carboxyl-terminal end of the interglobular
domain, the GDPE spacer, and a splice donor site containing a
BamHI site had the sequence: 5
-CGC GGG GGA TCC CCT CCC CCT
GGC AAA TGC GGC TGC CC. The polymerase chain reaction product was
digested with NheI and BamHI and ligated to a
NheI and BamHI cut vector CD5-IgG (25).
Production and Purification of rAgg1 in COS Cells
The genetic construct was transfected into COS cells via DEAE-dextran (26). Cells grown in DMEM containing 10% FBS were trypsinized, seeded onto fresh dishes 12 h after transfection, and allowed to grow for 5 days. On the 3rd day, fresh medium containing 10% FBS was added. Supernatants were harvested, centrifuged to remove nonadherent cells and debris, pooled, and stored at 4 °C. Fusion proteins were affinity-purified via anti-FLAG M1 affinity gel (Kodak) according to the manufacturer's protocol. The yield was 1-3 µg of fusion protein/ml of culture supernatant.
Culture of Rat Chondrosarcoma Cells and the Recombinant IGD Substrate (rAgg1) in Agarose in the Presence or Absence of Retinoic AcidRat chondrosarcoma cells were plated in 75-cm2
flasks and maintained in DMEM medium containing 5% FBS and 50 µg/ml
gentamicin. Confluent monolayers in 75-cm2 flasks were
harvested by trypsinization (0.25% trypsin (w/v) in DMEM containing
0.44 mM EDTA) for 15-20 min at 37 °C with agitation followed by digestion in 0.05% collagenase (w/v) in DMEM for 1-2 h.
The cells were then resuspended in DMEM at 4 × 106
cells/ml. 24-well culture plates were coated (200 µl/well) with a 1%
(w/v) solution of FMC Seaplaque-agarose in DMEM. The agarose was
solidified by incubation at 4 °C for 30 min (27) and then re-equilibrated to 37 °C. The chondrocyte cell suspension (described above) was diluted with a 2% (w/v) solution of Seaplaque-agarose in
DMEM such that the final cell concentration was 2 × 106 cells/ml. Aliquots (200 µl) of the agarose cell
suspension containing 0.4 × 106 cells were layered
over the previously prepared agarose plugs. Immediately following this,
50 µg of rAgg1 was added to the appropriate wells and mixed by
agitation. The plates were then incubated at 4 °C for 15 min to
solidify the agarose and re-equilibrated at 37 °C. Experimental
medium (with or without retinoic acid) was then added to triplicate
wells containing the recombinant IGD construct, as follows: control
cultures, DMEM plus gentamicin (50 µg/ml) and treated cultures, DMEM
plus gentamicin (50 µg/ml) plus 106 M
retinoic acid (2). Agarose cell cultures were then maintained at
37 °C in 5% CO2 for 96 h after which media and
agarose cell matrix extracts were further analyzed.
Agarose cultures containing only the rat chondrosarcoma cells (i.e. no rAgg1) were set up as described above. Experimental medium (with or without retinoic acid) containing 20 µg of rAgg1 was added to duplicate cultures. Thus, in this experiment the chondrocytes (in agarose) were partitioned from the multimolecular aggregate rAgg1 substrate in the medium. Agarose cultures were maintained as described above, and media were harvested at 16, 24, 48, 72, and 96 h for further analysis as described below.
Culture of Bovine Chondrocytes in Agarose in the Presence of Medium Containing rAgg1 Substrate in the Presence or Absence of Retinoic AcidBovine articular chondrocytes were isolated from the metacarpophangeal joints of 2-3-year-old cows using established procedures (28, 29). Briefly cartilage was removed from the joint surfaces under sterile conditions. The tissue was digested in 0.1% Pronase in DMEM containing 5% FBS for 1 h at 37 °C with agitation, washed, and further digested with 0.04% collagenase in DMEM containing 5% FBS overnight at 37 °C with agitation. Cells were then filtered through a 0.2-µm Nitex filter and washed before cell numbers were established. Cells were then plated in agarose (without substrate) as described above for the rat chondrosarcoma cells. Experimental medium (with or without retinoic acid) containing 20 µg of rAgg1 was added to duplicate cultures. Agarose cultures were maintained as described above and media were harvested at 24, 48, 72, 96, and 120 h for further analysis as described below. In a specific experiment conditioned media samples were obtained from control and retinoic acid-treated cultures at 120 h; these were used to demonstrate the soluble nature of aggrecanase activity in retinoic acid-treated cultures.
Analysis of Experimental Media from Control and Retinoic Acid-stimulated Agarose CulturesMedia (96 h) from cultures where both the cells and the rAgg1 were embedded in agarose were dialyzed exhaustively against distilled H2O, lyophilized, and reconstituted in an equal volume of SDS-PAGE sample buffer containing 10% (v/v) mercaptoethanol. Samples were subjected to SDS-PAGE, transferred to nitrocellulose, and immunolocated with either polyclonal or monoclonal antibodies using procedures described below. Media from cultures containing only rAgg1 (with or without retinoic acid) from the cells embedded in agarose were dialyzed exhaustively against distilled H2O, lyophilized, and reconstituted in SDS-PAGE sample buffer (at a concentration of 1 µg/µl rAgg1) containing 10% (v/v) mercaptoethanol. Samples were subjected to SDS-PAGE and immunolocation as described below.
SDS-Polyacrylamide Gel Electrophoresis and Western Blot AnalysesSamples containing native recombinant IGD substrate and its catabolites were electrophoresed on 10% polyacrylamide slab gels in SDS using procedures described by Laemmli (30). After electrophoresis, the fractionated proteins were electrophoretically transferred to nitrocellulose and immunolocated with a 1:1000 dilution of anti-FLAG monoclonal antibody M1 or monoclonal antibody BC-3 (anti-ARG ... ) and a subsequent alkaline phosphatase-conjugated second antibody using procedures previously described (2). In general, the immunoblots were incubated with substrate for 5-15 min at room temperature to achieve optimum color development.
BC-3 Antibody SpecificityFor experiments establishing antibody BC-3 specificity duplicate lanes containing 1 µg of rAgg1 substrate and product generated from cultures exposed to retinoic acid for 96 h were subjected to SDS-PAGE and transferred to nitrocellulose. The lanes were separated and one replicate was immunolocated with antibody BC-3 (1:10,000) alone, and the second replicate was immunolocated with BC-3 antibody (1:10,000) that had been preincubated for 1 h at 37 °C with 2.8 mg/ml synthetic peptide containing the amino-terminal sequence ARGSVI ... that was prepared with free amino and carboxyl termini on an Applied Biosystems 431A automated synthesizer using Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry with the peptide purified on a C8 reverse-phase column, and the structure was confirmed by electrospray mass spectrometry and amino acid analysis (Advanced Biotechnology Center, Charing Cross and Westminster Medical School). Second antibody and substrate development were as described above.
Transfection
of the recombinant IGD construct into COS-7 cells resulted in the
cellular expression and subsequent secretion of the recombinant IGD
product into the medium. The use of an anti-FLAG M1 affinity gel
facilitated its purification from the transfected cell medium. One
liter of transfected COS cell medium yielded 1-3 mg of rAgg1. SDS-PAGE
analysis (Fig. 2) indicated that under
non-reducing conditions (lane 1), rAgg1 formed large molecular mass aggregates (124 and >200 kDa) due to the presence of an
uneven number of cysteine residues in the human immunoglobulin domain
of the molecule. This aggregate formation facilitated immobilization of
the recombinant IGD substrate in the agarose cultures (see below).
Under reducing conditions the recombinant IGD construct occurred as two
molecular weight components having estimated molecular masses of 72 and
39 kDa (see Fig. 2, lane 2). Densitometric analysis of a
Coomassie Blue-stained gel of the separated recombinant IGD construct
showed that the ratio of the 72- to 39-kDa band was approximately 10:1
(results not shown). Amino-terminal amino acid sequence analysis
indicated that the 72-kDa form consisted of the correct FLAG and IGD
peptide sequence. However, amino acid sequence analysis of the 39-kDa
form gave the sequence DYKDDDDKTEPKSXD ... , indicating
that the IGD sequence had been deleted (see Fig. 1). This deletion most
likely occurs through the use of a previously reported alternate splice
donor site contained in the nucleotide sequence corresponding to the
amino terminus of the IGD (31) and the splice acceptor site of the
hinge region of the genomic DNA segment encoding the human IgG1 gene
(see Fig. 1). A disparity occurs in the calculated molecular mass and
the observed SDS-PAGE migration of the 72-kDa isoform of reduced rAgg1
that contains the IGD. We have performed enzymatic deglycosylation
analyses that indicate the presence of both O- and
N-glycosylation amounting to approximately 10 kDa (data not
shown). In addition, we have also used MALDI (matrix assisted laser
desorption ionization) spectroscopy to obtain a more accurate molecular
weight of the reduced 72-kDa rAgg1 substrate. These results indicate a
molecular mass of approximately 55 kDa (data not shown). Collectively,
these data more closely reflect the calculated molecular weight and indicate that the 72-kDa isoform has an anomalous migration in SDS-PAGE
analyses.
Immunochemical Analysis of Recombinant IGD Metabolites from Agarose Cultures Containing rAgg1 Treated with and without Retinoic Acid
rAgg1 was immobilized in agarose cell cultures containing
rat chondrosarcoma chondrocytes and maintained in medium with or without retinoic acid for 96 h. SDS-PAGE and Western blot analysis of rAgg1 and its metabolites released into the culture media are shown
in Fig. 3. Immunolocation with anti-FLAG
monoclonal antibody M1 (recognizing an epitope in the FLAG domain at
the amino terminus of the recombinant substrate) showed no differences
in the pattern of rAgg1 metabolites present in medium from either
control cultures or retinoic acid-treated cultures (Fig. 3, lanes
1 and 2). However, immunolocation with monoclonal
antibody BC-3 (anti-ARG ... ) showed positive reactivity only in
the retinoic acid-treated culture medium where it reacted with a band
migrating at approximately 66 kDa that was not recognized by the
anti-FLAG monoclonal antibody M1 (Fig. 3, lane 4,
lower arrow). Immunolocation with BC-3 showed no reactivity
in control cultures without retinoic acid treatment (Fig. 3, lane
3). This result indicates that aggrecanase catabolism of the
72-kDa form of rAgg1 has occurred through the removal of a small
peptide (~6 kDa), containing the FLAG epitope from the amino terminus
of the rAgg1 construct. As expected, there was no evidence that the
39-kDa isoform had been catabolized since it did not contain the IGD
sequence.
Immunoblot analyses with antibody BC-3 were also performed on media obtained from cultures with and without retinoic acid exposure where no rAgg1 was added. No glycosaminoglycan was detected in the medium from these cultures. In addition, Western blot analysis of the medium showed no immunostaining with antibody BC-3. This indicates that insufficient amounts of endogenous rat aggrecan were synthesized by the cells maintained in serum-free culture for 4 days for it to act as an endogenous substrate for aggrecanase activity (results not shown). This conclusion is also supported by the observation that retinoic acid essentially completely inhibits aggrecan synthesis in monolayer cultures (3).
Time Course of Release of Aggrecanase-generated Catabolites from rAgg1 Using Rat Chondrosarcoma CulturesIn a subsequent
experiment we used this culture system to investigate whether retinoic
acid-induced aggrecanase activity could diffuse from the agarose
containing the cells and catabolize rAgg1 sequestered in the medium
(probably because of its large aggregate properties). For these
experiments, the chondrocytes were suspended in agarose alone. The
rAgg1 substrate was added to control or retinoic acid-containing
medium, and culture media were harvested at 16, 24, 48, 72, and 96 h. Thus, to catabolize the rAgg1 aggregates the aggrecanase activity
would have to diffuse through the agarose and into the medium. The
results of immunolocation analyses of the rAgg1 medium metabolites with
antibody BC-3 are shown in Fig. 4A. These results indicate
that aggrecanase-generated products occur after 48 h of retinoic
acid exposure (Fig. 4A, lanes 4-6). At 72 and
96 h a BC-3 positive slower migrating band of approximately 121 kDa was evident. This band probably represents a minor aggrecanase product of aggregated rAgg1 that results from incomplete reduction. Data supporting this conclusion are indicated in Fig. 4C
where only 1 µg of rAgg1 (96 h) was loaded, and this minor component was not detected. No evidence of aggrecanase activity was evident in
control cultures after 96 h (Fig. 4, A and
C, lane 1). These results suggest that
aggrecanase can occur as a non-cell-associated proteolytic activity
that is present in the conditioned medium.
Additional experiments were also performed to discriminate better the undigested substrate from the 66-kDa aggrecanase-generated product. Fig. 4B shows immunolocation analysis of rAgg1 medium metabolites (1 µg) from cultures exposed to retinoic acid for 96 h (Fig. 4B, lane 1) that was subsequently reprobed with anti-FLAG M1 monoclonal antibody (Fig. 4B, lane 2) to identify undigested substrate containing the FLAG sequence at the amino terminus. These results clearly show the molecular mass shift from the 72-kDa undigested rAgg1 isoform to the 66-kDa aggrecanase-generated BC-3 positive product.
In addition, we also performed experiments to verify the specificity of the BC-3 antibody (Fig. 4C). Lane 1 of Fig. 4C shows BC-3 positive immunoreactivity on rAgg1 metabolites (1 µg) present in medium from cultures treated with retinoic acid for 96 h. However, this BC-3 reactivity is completely removed by preincubation of the antibody with 2.8 mg/ml synthetic peptide containing the amino-terminal sequence ARGSVI ... . This result confirms the specificity of monoclonal antibody BC-3 as previously reported (2).
Investigation of Aggrecanase Catabolism of rAgg1 Using Freshly Isolated Bovine Chondrocytes in Agarose CultureSimilar
experiments with freshly isolated bovine articular chondrocytes
substituting for Rx cells in agarose culture with and without retinoic
acid exposure were also performed. The results of immunolocation
analyses of media with antibody BC-3 at 24, 48, 72, 96, and 120 h
of culture are shown in Fig. 5. The
results indicate that for bovine chondrocytes rAgg1 catabolites
produced by aggrecanase are evident between 96 and 120 h of
culture in the presence of retinoic acid. The result also suggests that
aggrecanase activity is diffusible and takes approximately 48-72 h
longer to be expressed in bovine chondrocytes compared with the Rx rat chondrosarcoma cell cultures (Fig. 4, lanes 5 and
6).
An experiment was performed to demonstrate the diffusible properties of
aggrecanase. Conditioned medium (120 h) from both control and retinoic
acid-treated bovine cultures were collected and centrifuged to remove
any insoluble components, and 1 ml of the supernatant was incubated
with 5 µg of rAgg1 for 16 h at 37 °C. Replicate samples
without rAgg1 substrate were treated similarly (as controls). After
incubation, the samples were processed as described under
"Experimental Procedures," and the Western blots were immunolocated
with antibody BC-3 (Fig. 6). No
immunopositive BC-3 reactivity was seen in the control medium (Fig. 6,
lane 1) or in media that were incubated without rAgg1
(results not shown). However, medium from retinoic acid-treated
cultures showed strong BC-3 reactivity (Fig. 6, lane 2)
indicating that aggrecanase was a soluble proteolytic activity present
in the conditioned medium.
These results are the first report of the production and use of a recombinant protein substrate for the study of aggrecanase activity. In addition, we have developed a novel cell culture system to monitor aggrecanase activity without the complication of endogenous aggrecan acting as a substrate. The unique aggregating properties of this artificial substrate (rAgg1) facilitated its immobilization in the agarose cell culture system and also its sequestration in the medium compartment for other studies demonstrating the diffusible properties of this activity. Furthermore, we were able to use freshly isolated chondrocytes without the need to establish an endogenous extracellular matrix to act as a substrate for monitoring aggrecanase activity. The results obtained with the recombinant IGD construct demonstrate that there is not a need for a G1 and/or a G2 domain on the IGD of aggrecan for cleavage at the aggrecanase site. The rAgg1 construct showed no positive staining on Western blot analysis with an anti-keratan sulfate antibody 5-D-4 (results not shown). Therefore, there appears to be no requirement for keratan sulfate chains within the IGD that may give rigidity to this region in the native aggrecan molecule.
The induction of aggrecanase activity in rat chondrosarcoma cells (2, 3) and bovine monolayer cultures (3) has been described previously. Our studies demonstrate that aggrecanase activity is only expressed several hours after the chondrocytes are exposed to retinoic acid. Interestingly, the same time period (48 h) is required for aggrecanase induction from both rat chondrosarcoma monolayer cell cultures (3) and the same cells maintained in agarose cultures (this study). This finding indicates that the presence of the agarose matrix is not impeding the diffusion of aggrecanase into the conditioned media. This suggests that time may be required for synthesis and processing of the proteolytic activity. Alternatively, expression of aggrecanase activity may require time for the involvement of additional enzyme cascades to produce the active form(s) of the enzyme. This time-dependent expression of aggrecanase was demonstrated using two different animal sources of chondrocytes exposed to retinoic acid. In addition, there were different times required for the appearance of aggrecanase activity (synthesis and/or activation) using chondrocytes from the two different animal species.
Finally, our studies provide evidence that aggrecanase activity can occur in the medium free from association with the cell surface of the chondrocyte. Additional evidence supporting this conclusion came from the specific experiment where we isolated the medium prior to the addition of the rAgg1 substrate and then incubated the retinoic acid-treated conditioned media with rAgg1 in the absence of cells. After 16 h of cell-free incubation with rAgg1 BC-3, positive products were observed (Fig. 6, lane 2) that were identical to those in Fig. 5, lane 6 (120 h). Thus, the aggrecanase activity produced by the cells in culture can be manifested in the complete absence of cells. The identity of aggrecanase is still not known. Our findings of the diffusible properties of this enzyme still leave open the possibility that it could belong to several families of proteases, including matrix metalloproteinases that express their activity at physiological pH. This newly developed culture system will prove useful for further studies of the molecular mechanisms involved in aggrecan degradation by aggrecanase. It is also a novel experimental approach that could be considered for studies of the catabolism of other matrix macromolecules given the availability of recombinant substrates and neoepitope antibodies.