HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 42, 33027-33037, October 20, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/42/33027    most recent M910207199v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fosang, A. J. Articles by Hembry, R. M. Search for Related Content PubMed PubMed Citation Articles by Fosang, A. J. Articles by Hembry, R. M. Social Bookmarking                      What's this?

Generation and Novel Distribution of Matrix Metalloproteinase-derived Aggrecan Fragments in Porcine Cartilage Explants^*

Amanda J. Fosang^§, Karena Last, Heather Stanton, David B. Weeks, Ian K. Campbell^¶, Timothy E. Hardingham, and Rosalind M. Hembry^**

From the  University of Melbourne, Department of Paediatrics, Orthopaedic Molecular Biology Research Unit and Murdoch Childrens Research Institute, Royal Children's Hospital, Parkville, 3052, Australia, ^¶ The Walter & Eliza Hall Institute of Medical Research, Parkville, 3052, Australia,  Wellcome Trust Centre for Cell-Matrix Research, University of Manchester, Manchester, M13 9PT United Kingdom, and ^** University of East Anglia, School of Biological Sciences, Norwich, NR4 7TJ United Kingdom

Received for publication, December 23, 1999, and in revised form, July 3, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have studied aggrecan catabolism mediated by matrix metalloproteinases (MMPs) in a porcine cartilage culture system. Using antibodies specific for DIPEN³⁴¹ and ³⁴²FFGVG neoepitopes, we have detected MMP-derived fragments in conditioned medium and cultured cartilage, by radioimmunoassay, Western blotting, and immunolocalization. Radioimmunoassay revealed that the amount (pmol of epitope/mg of total glycosaminoglycan) of ³⁴²FFGVG epitope released from cartilage remained constant over a 5-day culture period and was not increased by IL-1 or retinoate. However, the proportion (pmol of epitope/mg of released glycosaminoglycan) of ³⁴²FFGVG epitope released was decreased upon stimulation, consistent with the involvement of a non-MMP proteinase, such as aggrecanase. The data suggest that in vitro MMPs may be involved in the base-line catabolism of aggrecan. Immunolocalization experiments showed that DIPEN³⁴¹ and ITEGE³⁷³ epitopes were increased by treatment with IL-1 and retinoate. Confocal microscopy revealed that ITEGE³⁷³ epitope was largely intracellular but with matrix staining in the superficial zone, whereas DIPEN³⁴¹ epitope was cell-associated and widely distributed in the matrix. Surprisingly, the majority of ³⁴²FFGVG epitope, determined by radioimmunoassay and Western blotting, was retained in the tissue despite the absence of a G1 domain anchor. Interleukin-1 stimulation caused a marked increase in tissue DIPEN³⁴¹ and ³⁴²FFGVG epitope, and the ³⁴²FFGVG fragments retained in the tissue were larger than those released into the medium. Active porcine aggrecanase was unable to cleave ³⁴²FFGVG fragments at the Glu³⁷³ Ala³⁷⁴ bond but cleaved intact aggrecan at this site, suggesting that ³⁴²FFGVG fragments are not substrates for aggrecanase. The apparent retention of large ³⁴²FFGVG fragments within cartilage, and their resistance to N-terminal cleavage by aggrecanase suggests that ³⁴²FF6V6 fragments may have a role in cartilage homeostasis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Aggrecan is the major proteoglycan present in articular cartilage, and it is the molecule that endows cartilage with its intrinsic capacity to bear load and resist compression. This weight-bearing capacity is dependent on the structure and organization of collagen and aggrecan within the extracellular matrix. Normal turnover of aggrecan is a conservative process in which the rate of breakdown and release of fragments from the tissue does not exceed the rate at which it is replaced by newly synthesized molecules. In pathology, an increased rate of degradation results in a net loss of aggrecan, and the tissue becomes thin and mechanically weak.

Aggrecan is lost from cartilage following proteolysis of the core protein. Aggrecanases are thought to be the enzymes primarily responsible for this proteolysis. Two ADAM¹ proteins with thrombospondin motifs, ADAMTS-4 (aggrecanase-1) (1) and ADAMTS-5 (aggrecanase-2, also known as ADAMTS-11) (2, 3) have been purified from bovine cartilage, and the corresponding human enzymes have been cloned and shown to exhibit the same specificity for aggrecan substrates as the purified bovine enzymes. The most widely documented activity of aggrecanase is hydrolysis of the Glu³⁷³ Ala³⁷⁴ bond (numbering based on the human sequence) (4) in the interglobular domain (IGD) (Fig. 1c); however aggrecanases also cleave within highly conserved regions of the chondroitin sulfate attachment region (5, 6). Fragments resulting from aggrecanase action and containing the ³⁷⁴ARGSV N terminus (Fig. 1c) have been detected in conditioned culture medium by sequence analysis of isolated fragments or by detection with specific neoepitope antibodies. Aggrecanase-derived ³⁷⁴ARGSV fragments are the major aggrecan products found in synovial fluids from arthritis patients (7, 8). Furthermore, the aggrecanase-derived ITEGE³⁷³ neoepitope at the C terminus of the G1 domain has been detected in human (9, 10), bovine (11, 12), pig (12), and rat (11) cartilage and in mice with experimental arthritis (13, 14).

In addition to cleavage by aggrecanase, there is convincing evidence for the direct involvement of the matrix metalloproteinase (MMP) family of enzymes in aggrecanolysis, albeit at much lower levels. The specific products of MMP activity have been detected in vivo (10, 15, 16) and in vitro (17-19) by sequencing and more recently by cleavage site-specific neoepitope antibodies (20, 21). MMPs cleave the aggrecan protein core at the Asn³⁴¹Phe³⁴² bond in the IGD and possibly at other more C-terminal sites as well. The ³⁴²FFGVG N-terminal neoepitope (Fig. 1c) has been found in human synovial fluids (16) and conditioned medium from porcine (22) and human osteoarthritic (12) cartilage cultures. The C-terminal neoepitope DIPEN³⁴¹ (Fig. 1c) has been immunolocalized in human articular and intervertebral disc cartilage (9, 10) and in experimentally induced arthritis models in mice (13, 14, 23-26), providing evidence that MMPs directly degrade aggrecan in vivo.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Schematic diagram showing cleavage sites and neoepitopes generated by aggrecanase and MMP cleavage in the aggrecan interglobular domain. a, aggrecan is immobilized in cartilage by binding to hyaluronan and link protein via its N-terminal G1 domain. b, aggrecan core protein with G1, G2, and G3 globular domains and extended regions comprising keratan sulfate (KS) and two chondroitin sulfate domains (CS-1 and CS-2). c, expansion of the G1-IGD-G2 region showing the immunoglobulin (Ig) fold motif (A loop) in G1 and the proteoglycan tandem repeat motif (PTR; B and B' loops) in G1 and G2. Cysteine residues and disulfide bonds in G1 and G2 are shown. Amino acids in the human sequence (4), amino acids flanking and bridging the aggrecanase and MMP cleavage sites, and the neoepitope sequences generated by cleavage are included. Antibodies specific for each neoepitope are boxed. The asterisks mark substituted residues reported in other species (54, 55).

While it is clear that both MMP and aggrecanase activities are involved in aggrecanolysis, there are currently no assays for quantitating levels of neoepitopes; thus, the relative abundance of these activities in human tissue, animal models, or culture systems under conditions of normal or stimulated turnover is not known. Furthermore, the exact relationship between MMP and aggrecanase activities remains poorly understood. A number of studies have examined aggrecanase-mediated loss of aggrecan fragments in cell and tissue culture systems, but there is no information regarding MMP-mediated loss of aggrecan fragments.

In this paper, we present a detailed analysis of MMP-mediated aggrecan catabolism in a cartilage explant system. We show that DIPEN³⁴¹ epitope is more widely distributed in the cartilage matrix, compared with the ITEGE³⁷³ epitope, which is mostly intracellular, presumably following endocytosis. We also show that the majority of MMP-derived ³⁴²FFGVG fragments remain in cartilage during a 5-day culture experiment, and only a proportion are released into the medium. Once created, ³⁴²FFGVG fragments are resistant to cleavage in the IGD by aggrecanase. The finding that ³⁴²FFGVG fragments are retained in cartilage matrix and resistant to aggrecanase digestion raises the possibility that they may have other roles in cartilage homeostasis.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Dulbecco's modified Eagle's medium (DMEM) was from Trace Biosciences (New South Wales, Australia). Fetal calf serum, penicillin, and streptomycin antibiotics were from CSL Ltd. (Victoria, Australia). Recombinant human interleukin-1 (IL-1) was from Genzyme Diagnostics. Retinoic acid and chondroitinase ABC lyase (Proteus vulgaris sp.) (EC 4.2.2.4) were from ICN Biochemicals. 4-(2-aminoethyl)benzene sulfonylfluoride was from Calbiochem-Novabiochem. An ECL-plus enhanced chemiluminescence kit and Na¹²⁵I (IMS 30) were from Amersham Pharmacia Biotech. Dried cells of Staphylococcus aureus, the substance P 7-11 fragment, poly-L-lysine, chondroitinase ABC lyase (C-2905) (P. vulgaris sp.) (EC 4.2.2.4) used for immunofluorescence studies, 4-aminophenylmercuric acetate, normal rabbit IgG (I5006), and papain (EC 3.4.22.2) were from Sigma. 1,9-Dimethylmethylene blue dye was from SERVA Feinbiochemica (Heidelberg, Germany). Vectashield mounting medium was from Vector Laboratories. Agarose type HSC was from Park Scientific (Northampton, UK). Keratanase (Pseudomonas sp.) (EC 3.2.1.103) was from Seikagaku (Tokyo, Japan). Cysteine proteinase inhibitor E-64, pepstatin, and a chemiluminescence blotting substrate kit were from Roche Molecular Biochemicals. Fluorescein isothiocyanate-conjugated sheep anti-rabbit IgG was from Silenus. Texas Red-conjugated donkey anti-rabbit IgG was from Jackson Immunoresearch Inc. Monoclonal antibody AF-28 specific for the N-terminal sequence ³⁴²FFGVG (27), anti-ITEGE³⁷³ and anti-DIPEN³⁴¹ rabbit sera (28), and polyclonal anti-G1 domain antiserum (29) were as described previously. Recombinant pro-MMP-13 (30) was a gift from Prof. G. Murphy and Dr. V. Knäuper (University of East Anglia), and recombinant TIMP-1 (31) was also a gift from Prof. G. Murphy.

Cartilage Explant Cultures-- Articular cartilage was dissected from the metacarpophalangeal joints of pigs aged 25-28 weeks, and approximately 100 mg was placed into wells of sterile 48-well culture plates. For all experiments, cartilage was cultured in a humidified incubator at 37 °C with 5% CO₂ for 5 days in serum-free DMEM containing 100 units/ml penicillin and 100 µg/ml streptomycin, 2 mM L-glutamine, and 20 mM Hepes. Each well contained 500 µl of medium that was collected daily, frozen for further analysis, and replaced with fresh medium. The cartilage was treated to stimulate aggrecan degradation, commencing on day 3 of culture. Treated tissue was cultured in medium containing either 10 ng/ml IL-1 or 1 µM retinoic acid for days 3-5 of the culture period. Control cartilage received medium with no additives for 5 days. We have labeled this culture regime A (see Fig. 5). In one experiment (Fig. 5), we compared a different culture regime in which 10-20 mg of cartilage was cultured for 3 days in three changes of 1 ml of DMEM containing 10% fetal calf serum and then washed and cultured serum-free in the presence or absence of 10 ng/ml IL-1 for a further 4 days. The medium was not changed during the 4-day IL-1 stimulus. This has been labeled culture regime B in Fig. 5.

At the end of the culture period, the tissue was blotted dry and weighed and then extracted with shaking for 48 h at 4 °C with 1 ml of 4 M guanidine hydrochloride (GdnHCl), 50 mM sodium acetate, pH 5.8, to isolate aggrecan and G1-enriched aggrecan fragments bound to hyaluronan. After GdnHCl extraction, the tissue was digested overnight at 60 °C with 1 ml of papain solution containing 125 µg/ml papain, 0.1 M sodium acetate, pH 5.5, 5 mM EDTA, 5 mM cysteine-HCl. The concentration of sulfated glycosaminoglycan in all samples (media, dialyzed GdnHCl extracts, and papain digests) was determined using the 1,9-dimethylmethylene blue assay (32). For some experiments, cultures were extracted with 4 M GdnHCl after each of days 1-5 for the analysis of G1 neoepitopes generated daily. Larger culture experiments containing approximately 1 g of tissue in 10 ml of medium or 400 mg of tissue in 3 ml were done under the same conditions (A) and used for the analysis of AF-28 epitope by radioimmunoassay.

AF-28 Radioimmunoassay-- An AF-28 radioimmunoassay, approximately 10-fold more sensitive than the competition enzyme-linked immunosorbent assay described previously (27), was developed using monoclonal AF-28 and MMP-derived G2 domain labeled with Na¹²⁵I. Purified G2 domain containing the ³⁴²FFGVG N terminus was used as both standard antigen and iodinated competitor in the assay. G2 domain was prepared by MMP digestion of pig G1-G2 (33) and isolated free of G1 by size exclusion chromatography on a Biosep-SEC S4000 column (Phenomenex) following overnight mixing with hyaluronan (17). The concentration of ³⁴²FFGVG neoepitope (pmol/ml) in the G2 peak eluting from the column was measured by competition enzyme-linked immunosorbent assay (27) and then used as standard antigen in the radioimmunoassay.

Approximately 5 µg of G2 protein was iodinated with 0.3 mCi of Na¹²⁵I by the chloramine T method (34). After a 45-s incubation at room temperature and quenching with sodium metabisulfite, unincorporated isotope was removed on a Biogel P6-DG column equilibrated in PBS. For the radioimmunoassay, 100 µl of sample or standard (in the range 0.14-18 pmol/ml of AF-28 epitope) was added to tubes containing 50 µl of diluted AF-28 and 50 µl of ¹²⁵I-G2 (approximately 30,000 cpm/50 µl) diluted in assay buffer containing 1% BSA in PBS. The tubes were vortexed and incubated overnight at 4 °C. Control tubes containing no antibody were included as blanks. Rabbit anti-mouse immunoglobulin (50 µl) diluted 1:100 in assay buffer was added, and the tubes were vortexed and allowed to stand at room temperature for 30 min. 50 µl of dried cells of S. aureus, washed and resuspended as a 10% solution in assay buffer, was then added to each tube to bind and precipitate immune complexes and allowed to stand for 15 min at room temperature. The supernatants were removed, the cell pellet was washed with 1 ml of assay buffer, and radioactivity in the pellets was counted on an LKB 1260 multigamma II -counter. The range of the assay was 1.7 ± 0.29 to 17.7 ± 4.38 pmol/ml, and the concentration of standard competitor that gave 50% inhibition was 5.42 ± 1.29 pmol/ml. The 7-11 fragment of substance P, with sequence FFGLM-NH₂ was not detected in the assay at concentrations up to 0.1 µM.

Immunofluorescence-- Full thickness cartilage excised from a 13-day-old pig and cultured cartilage explants from 25-28-week-old pigs were embedded in 7% gelatin and frozen in liquid nitrogen. Sections (8 µm) were cut on a cryostat, taken onto glass slides coated with poly-L-lysine, dried briefly, and fixed in formaldehyde freshly prepared from paraformaldehyde (4% in PBS, pH 7.4; 30 min, room temperature). After washing in PBS (three times for 5 min each) to remove fixative, the sections were treated with chondroitinase ABC (0.01 units/100 µl/section) in 0.1 M Tris-HCl, 30 mM sodium acetate, pH 8.0, with proteinase inhibitors (20 µg/ml E64, 0.5 mM 4-(2-aminoethyl)benzene sulfonylfluoride, 5 µM pepstatin, 10 mM EDTA) for 1 h at 37 °C to remove chondroitin sulfate chains. The sections were washed again, permeabilized for 5 min at room temperature with 0.1% Triton X-100 in PBS, rewashed, and then stained by indirect immunofluorescence using either rabbit anti-DIPEN³⁴¹ (IgG, 0.05 µg/ml), rabbit anti-ITEGE³⁷³ (IgG, 5 µg/ml), or normal rabbit IgG (5 µg/ml) as control. IgGs from rabbit antisera were purified using protein A-Sepharose followed by fractionation on ovalbumin-coupled Sepharose to remove reactivity against the ovalbumin carrier conjugated to the peptide immunogen (28). The secondary antibody was affinity-isolated fluorescein isothiocyanate-conjugated sheep anti-rabbit IgG (1:500). All antibodies were diluted in PBS with 1% BSA. Sections were mounted in Vectashield, viewed by epifluorescence on an Olympus IX70 inverted microscope with narrow band fluorescein isothiocyanate filter set and photographed on Eastman Kodak Co. EPH P1600 film.

For confocal analysis, sections were stained as above using a Texas Red-conjugated donkey anti-rabbit IgG (1:200) as secondary antibody. Sections were viewed on a MRC 600 confocal microscope (Bio-Rad) with a krypton/argon laser using the 568-nm laser line. Serial 1-µm optical sections through the cells were collected with a confocal aperture of 0.5 and kalman averaging over 10 scans at slow scan speed. Images were processed with Bio-Rad Comos software and printed on a Codonics NP1600 printer (Laserlines Ltd., Banbury, UK).

The following control studies were done to establish antibody specificity in immunolocalization. Frozen sections of cartilage from a 13-day-old pig with and without chondroitinase treatment were incubated with either anti-DIPEN³⁴¹ or anti-ITEGE³⁷³ followed by secondary antibody. No fluorescence was observed, indicating that neither neoepitope was present in normal young intact cartilage (see also Fig. 6, a and f). To create MMP cleavage neoepitopes, sections were incubated with 50 µl of 4-aminophenylmercuric acetate-activated MMP-13 for 30 min at 37 °C. Following washing and indirect immunolocalization with the anti-DIPEN³⁴¹ and anti-³⁴²FFGVG antibody, the cartilage sections showed intense immunofluorescence throughout the depth of cartilage. Sections stained with normal rabbit IgG were negative. Sections stained with anti-DIPEN³⁴¹ antibody preabsorbed with an 800-fold molar excess of DIPEN³⁴¹-conjugated BSA were also negative. The DIPEN³⁴¹-conjugated BSA was equivalent to an 800-24,000-fold molar excess of peptide epitope over anti-DIPEN³⁴¹ IgG, since 1 mol of BSA carried 1-30 mol of DIPEN³⁴¹ peptide, as determined by the size and breadth of the DIPEN³⁴¹-BSA conjugate on SDS gels.

Digestion with Porcine Aggrecanase-- To determine whether MMP-derived aggrecan fragments could be cleaved by aggrecanase, we designed an experiment using ³⁴²FFGVG fragments as substrate and conditioned medium from stimulated cultures as a source of aggrecanase. The ³⁴²FFGVG substrate for these experiments was present in 15× concentrated day 1 conditioned medium (substrate medium). Purified A1D1 from pig articular cartilage was included as a control substrate. The concentration of sulfated glycosaminoglycan on ³⁴²FFGVG substrate or A1D1 substrate in each digest was 50 µg/20 µl. Porcine aggrecanase was present in concentrated conditioned medium from day 4 and 5 cultures co-stimulated with 10 ng/ml IL-1 and 1 µM retinoate (stimulated enzyme source). Concentrated medium from day 4 and 5 untreated cultures was included as a control (control enzyme source). 7.5 µl of control or stimulated enzyme source was present in each digest, and 3.5 µM TIMP-1 was added to some digests to control for any exogenous MMP activity that may have been present in the concentrated media.

Analysis of DIPEN and ITEGE G1 Fragments-- For Western blot analysis of G1 species, 4 M GdnHCl extracts were dialyzed exhaustively against distilled water and freeze-dried. Samples were deglycosylated overnight at 37 °C in buffer containing 50 mM Tris acetate buffer, pH 7.2, with 0.005 units of keratanase and 0.005 units of chondroitinase ABC in the presence of 10 mM EDTA, 20 µg/ml E64, 5 µM pepstatin, and 1.25 mM 4-(2-aminoethyl)benzene sulfonylfluoride. All samples for DIPEN³⁴¹ and ITEGE³⁷³ analysis were loaded on SDS gels on the basis of equal tissue weights. For experiments to test the susceptibility of aggrecanase-G1 fragments to digestion by MMPs, aliquots of dialyzed extracts were digested overnight with recombinant MMP-13 (30) in buffer containing 10 mM calcium chloride, 100 mM sodium chloride, 50 mM Tris-HCl, pH 7.5, at 37 °C for 21 h. EDTA and 1,10-phenanthroline were added to final concentrations of 10 mM and 2 mM, respectively, and the samples were then deglycosylated as above.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Pattern of MMP-derived Aggrecan Fragments Released from Cartilage Cultures-- To investigate the pattern of MMP-derived aggrecan fragments released from cultured cartilage, aliquots of media from each day of culture, containing equal amounts of sulfated glycosaminoglycan, were electrophoresed on composite gels (35) and analyzed by toluidine blue staining and Western blotting with monoclonal AF-28, which recognizes the N-terminal sequence ³⁴²FFGVG. Immunoreactivity to ³⁴²FFGVG was detected in media from all days of culture, commencing on day 1. The presence of ³⁴²FFGVG epitope in unstimulated cultures (Fig. 2b; Fig. 2, d and f, days 1 and 2) suggests that MMPs may be involved in the unstimulated, or "base-line" release of aggrecan from cartilage in culture. Treatment of cartilage with IL-1 or retinoate appeared to reduce the amount of ³⁴²FFGVG epitope released on days 4 and 5. IL-1-treated cartilage released one predominant band of ³⁴²FFGVG immunoreactivity on composite gels that was of fast relative mobility (Fig. 2d, open arrow). In contrast, retinoate treatment generated at least two ³⁴²FFGVG bands on composite gels that were sharp and of slower relative mobility (Fig. 2f, closed arrows). These different fragmentation patterns suggest that IL-1 and retinoate stimulate different catabolic pathways for the degradation of aggrecan and that IL-1 induces more extensive C-terminal processing giving rise to smaller (high mobility) fragments. The majority of fragments released by IL-1 treatment and detected by toluidine blue (Fig. 2c, arrowhead) did not co-migrate with ³⁴²FFGVG epitope (Fig. 2d, open arrow), indicating that ³⁴²FFGVG epitope is not present on the larger glycosaminoglycan-containing fragments released by IL-1 treatment.

View larger version (45K): [in this window] [in a new window]   Fig. 2.   Pattern of 342FFGVG aggrecan fragments released from cartilage cultures. Porcine cartilage slices were cultured in serum-free DMEM for days 1 and 2 (lanes 1 and 2) and then stimulated with either 10 ng/ml IL-1 (c and d) or 1 µM retinoate (e and f) for days 3-5 (lanes 3-5) or unstimulated (a and b). Aliquots of media containing 5 µg of sulfated glycosaminoglycan were electrophoresed on composite gels and analyzed for MMP-derived 342FFGVG neoepitope (b, d and f). The pattern of total proteoglycan fragments was shown by toluidine blue staining (a, c and e). Bands marked with arrows are described under "Results."

Quantitation of MMP-derived Aggrecan Fragments Released from Cartilage Cultures-- The amount of ³⁴²FFGVG epitope released each day from cartilage cultures was measured by AF-28 radioimmunoassay. Fig. 3 shows the results of two separate experiments in which ³⁴²FFGVG epitope has been expressed per total glycosaminoglycan (Fig. 3, b and e) or per glycosaminoglycan released (Fig. 3, c and f). The amount of total glycosaminoglycan (Fig. 3, a and d) and ³⁴²FFGVG neoepitope (Fig. 3, b and e) released into the culture medium was similar in both experiments. The results show that the amount of ³⁴²FFGVG epitope released per day remains relatively constant between treatments (Fig. 3, b and e) but that the proportion of released fragments with ³⁴²FFGVG N termini is decreased following IL-1 or retinoate stimulation (Fig. 3, c and f). These results are consistent with the Western blot data in Fig. 2 and indicate the predominance of a non-MMP proteinase, such as aggrecanase, in the stimulated release of aggrecan fragments.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Quantitation of 342FFGVG aggrecan fragments released from cartilage cultures. Concentrations of sulfated glycosaminoglycan (a) and 342FFGVG neoepitope (b and c) released from cartilage explants into medium were determined in two separate experiments. For experiment 1, quadruplicate cultures containing 0.4 g of cartilage in 3 ml of medium were maintained for 5 days, and 12 ml of medium/day was concentrated 6-fold for radioimmunoassay of 342FFGVG neoepitope. For experiment 2, triplicate cultures containing 1 g of cartilage in 10 ml of medium were maintained for 5 days, and 30 ml of medium per day was concentrated 15-fold for radioimmunoassay.

MMP-derived ³⁴²FFGVG Fragments Are Present in GdnHCl Extracts-- The low abundance (1% or less of total glycosaminoglycan) of ³⁴²FFGVG fragments detected in the conditioned medium prompted us to look for ³⁴²FFGVG fragments in the GdnHCl extracts. Surprisingly, a substantial proportion of ³⁴²FFGVG epitope remained in the tissue after 5 days in culture and was detected in the GdnHCl extracts by radioimmunoassay (Table I). In some cases, the amount of ³⁴²FFGVG epitope remaining in the tissue was as much as 10 times greater than the amount of ³⁴²FFGVG released into the conditioned medium. The results showed that IL-1 caused an increase in tissue ³⁴²FFGVG epitope, but there was no increase with retinoate.

To further examine the generation and distribution of fragments remaining in the tissue, control and stimulated cartilage was extracted with 4 M GdnHCl on each of days 1-5, and the extracts were analyzed for ITEGE³⁷³, DIPEN³⁴¹, and ³⁴²FFGVG neoepitopes (Fig. 4). The gels were loaded on an equal tissue weight basis. Tissue treated with IL-1 or retinoate for days 3-5 contained increased amounts of ITEGE³⁷³ epitope (Fig. 4, b and c, lanes 3-5), compared with control tissue (Fig. 4a, lanes 3-5), although in IL-1-treated cultures, ITEGE³⁷³ epitope appeared to be reduced on day 5 compared with day 4 (Fig. 4b, lanes 4 and 5). These results confirm studies from other laboratories showing that IL-1 and retinoate stimulate aggrecanase activity in explant cultures. The DIPEN³⁴¹ blots revealed an increase in DIPEN³⁴¹ G1 domain on days 3 and 4 following IL-1 treatment (Fig. 4e, lanes 3 and 4) and a further marked increase on day 5 (Fig. 4e, lane 5). The increased DIPEN³⁴¹ epitope corresponded with increased ³⁴²FFGVG epitope (Fig. 4h, lanes 3-5) in extracts of IL-1-stimulated tissue. There was no increase in tissue DIPEN³⁴¹ and ³⁴²FFGVG epitope following retinoate treatment (Fig. 4, f and i, compared with d and g).

View larger version (48K): [in this window] [in a new window]   Fig. 4.   DIPEN341, 342FFGVG, and ITEGE373 neoepitopes in GdnHCl cartilage extracts. Cartilage slices were extracted with 4 M GdnHCl after days 1 (lane 1), 2 (lane 2), 3 (lane 3), 4 (lane 4), and 5 (lane 5) of culture. The extracts were dialyzed, deglycosylated, and analyzed on 5% SDS gels with Western blotting for ITEGE373 (a-c) or DIPEN341 (d-f) neoepitopes. 342FFGVG (g-i) epitopes were detected on 4% SDS gels (an arrow marks the top of 4% gels). Samples were loaded by equal tissue weight equivalent to 1-3 mg.

Our finding that DIPEN³⁴¹ and ³⁴²FFGVG epitope increased following IL-1 treatment (Fig. 4) differed from previous reports that showed no expression or regulation by IL-1 of MMP-derived neoepitopes in cultured pig cartilage (12). To address this discrepancy, we compared epitope generation under conditions where tissue was cultured with 10% serum for 3 days, followed by 4 days of serum-free IL-1 stimulation, and without a change of medium during days 4-7 (12) (Fig. 5, culture condition B), with our 5-day serum-free cultures, which received daily changes of IL-1 on days 3-5 (Fig. 5, culture condition A). Under both conditions of culture, IL-1 induced an increase in DIPEN³⁴¹ (Fig. 5a) and ³⁴²FFGVG (Fig. 5b) epitope. A small ³⁴²FFGVG fragment was identified in the B cultures (Fig. 5b, lane 4). The size of this fragment was similar to ³⁴²FFGVG fragments with ³⁷³ITEGE C termini generated by MMP-8 digestion of pig G1-G2 (27) and fragments with DITVQ³⁵⁴ C termini generated by MMP-14 digestion of pig G1-G2 (21). The small ³⁴²FFGVG species lacked an ITEGE³⁷³ C terminus (results not shown) and in this respect is similar to fragments found in human synovial fluids (16). Our data show that culture conditions do not account for the differences between the present results and those of Little et al. (12). We are continuing to address these issues in collaborative studies with Little and co-workers.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   The effect of culture conditions on the generation of MMP neoepitopes by IL-1. Cartilage slices were cultured alone (lanes 1 and 3) or in the presence of IL-1 (lanes 2 and 4). The cartilage was extracted with 4 M GdnHCl after 5 days of culture (condition A; see "Experimental Procedures") or 7 days of culture (condition B). The extracts were dialyzed, deglycosylated, and analyzed on 5% SDS gels with Western blotting for DIPEN341 (a) or 342FFGVG (b) neoepitopes. Samples were loaded on the basis of equal total proteoglycan. c, native pig G1-G2 (3 µg) was digested for 18 h at 37 °C with 176 µg/ml MMP-1, in a total volume of 10 µl. Dilutions of digested G1-G2 were electrophoresed on 5% SDS gels and analyzed by Western blotting with anti-DIPEN341 antiserum () or monoclonal anti-342FFGVG (). Bands were quantitated by densitometric scanning.

The sensitivity and detection range of the ³⁴²FFGVG and DIPEN³⁴¹ antibodies was compared by digesting pig G1-G2 with MMP-1 and analysis of sample dilutions by Western blotting. A single band with an approximate mass of 100 kDa was detected by anti-³⁴²FFGVG antibody, and a single band with approximate mass of 60 kDa was detected by anti-DIPEN³⁴¹ antibody (data not shown). The ³⁴²FFGVG and DIPEN³⁴¹ bands were analyzed by densitometric scanning. Fig. 5c shows nanograms of digested G1-G2 substrate plotted against densitometric volume and reveals that the slope of the line for ³⁴²FFGVG epitope is 3,762.88 ± 469.93 and that the slope for DIPEN³⁴¹ is 341.27 ± 48.39. This result may provide an explanation for why ³⁴²FFGVG epitope in stimulated cultures appears significantly greater than DIPEN³⁴¹ epitope in Fig. 4, e and h, and Fig. 5, a and b, when we predict that they would be approximately equal.

Detection of MMP- and Aggrecanase-derived G1 Domains by Immunofluorescence-- The tissue distribution of DIPEN³⁴¹ G1 domain in control and stimulated cartilage was investigated by immunofluorescence and compared with the distribution of ITEGE³⁷³ G1 domain. Cartilage cultured for 5 days in serum-free DMEM and stained with anti-DIPEN³⁴¹ had no immunofluorescence (Fig. 6a), showing that normal cartilage does not contain detectable levels of DIPEN³⁴¹ epitope and that the culture procedure does not result in neoepitope production. Explants frozen after 1 day of stimulation with IL-1 (day 3) had bright staining of superficial cartilage matrix and cells (Fig. 6b). Hypertrophic chondrocytes and interterritorial matrix also stained weakly (Fig. 6c), but midzone matrix was negative (Fig. 6c, arrow). Adjacent sections stained with normal rabbit IgG were negative throughout (data not shown). On day 4, the superficial matrix staining extended further into the midzone, and the matrix staining surrounding hypertrophic chondrocytes was more extensive and intense (Fig. 6d). Explants stimulated with IL-1 and frozen on day 5 had matrix staining throughout the sections (data not shown). In contrast, sections from explants cultured with retinoate and frozen on day 5 showed only weak immunofluorescence of hypertrophic chondrocytes and adjacent matrix (Fig. 6e).

View larger version (74K): [in this window] [in a new window]   Fig. 6.   Immunolocalization of DIPEN341 and ITEGE373 G1 domains in cultured cartilage. Porcine cartilage explants were cultured in serum-free DMEM for 2 days and then cultured for a further 3 days in either serum-free DMEM (control, a and f) serum-free DMEM containing 10 ng/ml IL-1 (b-d and g-i), or serum-free DMEM containing 1 µM retinoate (e and j). Explants were removed on day 3 (b, c, g, and h), day 4 (d and i), and day 5 (a, e, f, and j), frozen, sectioned, and stained by indirect immunofluorescence with either anti-DIPEN341 (a-e) or anti-ITEGE373 (f-j). Bar, 50 µm. a, section from an explant cultured in serum-free DMEM for 5 days and stained with anti-DIPEN341. There is no immunofluorescence present. b, section from an explant removed 1 day after the addition of IL-1 (day 3) and stained with anti-DIPEN341. Immunofluorescence is present in superficial cartilage matrix and associated with chondrocytes. c, as in b, hypertrophic chondrocytes and lower matrix stain, but midzone matrix and cells are negative (arrow). d, section from an explant stimulated with IL-1, frozen on day 4, and stained with anti-DIPEN341. The interterritorial matrix staining is more extensive than on day 3. e, section of an explant removed on day 5 after retinoate treatment, stained with anti-DIPEN341. The hypertrophic chondrocytes have cell-associated immunofluorescence, and the interterritorial matrix stains weakly, but midzone matrix and cells are negative (arrow). f, as in a but stained with anti-ITEGE373. No immunofluorescence is visible. g, as in b but stained with anti-ITEGE373. Superficial cartilage matrix has immunofluorescence, and chondrocytes below have weak intracellular staining. h, hypertrophic region below g showing strong cellular staining of chondrocytes. i, a similar region to h in an IL-1-treated explant on day 4 stained with anti-ITEGE373. Chondrocytes have intracellular staining and are surrounded by haloes of matrix immunofluorescence. j, midzone region of a retinoate-treated explant frozen on day 5. The chondrocytes have intracellular fluorescence, but there is no staining of surrounding matrix.

Sections from explants cultured in serum-free medium for 5 days and stained with anti-ITEGE³⁷³ had no immunofluorescence (Fig. 6f). IL-1-treated explants frozen on day 3 had bright immunofluorescence of superficial cartilage matrix (Fig. 6g), and chondrocytes in all zones were stained; hypertrophic cells were particularly bright (Fig. 6h). On day 4, hypertrophic chondrocytes in IL-1-treated explants were surrounded by haloes of matrix staining (Fig. 6i), but although by day 5 all chondrocytes stained brightly, no midzone matrix staining was visible (data not shown). All chondrocytes in explants cultured in retinoate for 3 days showed weak immunostaining for ITEGE³⁷³ (Fig. 6j) but limited matrix staining, similar to the day 3 IL-1-treated explants.

To further define the cellular distribution of both epitopes, sections from IL-1-treated explants frozen on day 3 were examined by confocal microscopy. Sections stained with anti-DIPEN³⁴¹ showed matrix staining adjacent to the chondrocyte lacuna (Fig. 7A, arrow) and cell-associated staining (Fig. 7A, arrowhead). Since the chondrocyte cell surface was not visible, it was not possible to determine whether this staining was on the matrix, on the cell surface, or intracellular. In contrast, chondrocytes stained with anti-ITEGE³⁷³ had clearly defined immunofluorescent intracellular vesicles, probably secondary lysosomes (Fig. 7B), strongly suggesting that the ITEGE³⁷³ epitope is endocytosed. Experiments to confirm this and to determine the precise location and fate of the DIPEN³⁴¹ epitope are in progress.

View larger version (45K): [in this window] [in a new window]   Fig. 7.   Localization of DIPEN341 and ITEGE373 G1 domains by confocal microscopy. Porcine cartilage explants were cultured in serum-free DMEM for 2 days and then cultured for a further day in serum-free DMEM containing 10 ng/ml IL-1, removed on day 3, frozen, sectioned, and stained by indirect immunofluorescence with either anti-DIPEN341 (A) or anti-ITEGE373 (B). 1-µm optical sections through chondrocytes and surrounding matrix were taken by confocal microscopy. A, sections stained with anti-DIPEN341 show matrix staining adjacent to the chondrocyte lacuna (arrow) and cell-associated staining (arrowhead). Bar, 10 µm. B, chondrocytes stained with anti-ITEGE373 have clearly defined immunofluorescent intracellular vesicles, probably secondary lysosomes. Bar, 5 µm.

Size Distribution of MMP-derived Aggrecan Fragments Released from Explant Cultures-- The results in Figs. 2-5 show that ³⁴²FFGVG epitope is present on a mixed population of size fragments, indicating variable sites of C-terminal processing. Gel permeation chromatography was used to assess the size distribution of ³⁴²FFGVG-containing fragments in conditioned medium and GdnHCl extracts. Concentrated extracts or pooled day 4 and 5 media for each treatment were eluted on a Sepharose CL-2B column under associative conditions. The column fractions were divided into three pools designated A (K_av = 0-0.22), B (K_av = 0.25-0.58), and C (K_av = 0.61-1.00) based on the profile of sulfated glycosaminoglycans (Fig. 8, a and c). The amount of ³⁴²FFGVG epitope in each pool was assayed by radioimmunoassay. As expected, stimulation of cultured cartilage with IL-1 or retinoate caused an increased release into the medium of glycosaminoglycan-containing fragments, compared with control tissue (Fig. 8a). Radioimmunoassay of AF-28 epitope present in the column pools showed that the average size of ³⁴²FFGVG-containing fragments released from stimulated cultures was smaller than ³⁴²FFGVG fragments released from control cultures (Fig. 8b). The ³⁴²FFGVG epitope released from stimulated cultures was evenly distributed between pools B and C, and no ³⁴²FFGVG fragments were present in pool A. In contrast, the majority (68%) of fragments released from control cultures were in the K_av 0.25-0.58 range, with 14% eluting in pool A. The distribution of ³⁴²FFGVG fragments in Fig. 8b is consistent with the Western blot analyses (Fig. 2), which showed that larger ³⁴²FFGVG fragments are released under control conditions and that smaller ³⁴²FFGVG fragments are released from stimulated tissue.

View larger version (38K): [in this window] [in a new window]   Fig. 8.   Size distribution of 342FFGVG-containing fragments released or extracted from cartilage cultures. GdnHCl extracts (c and d) or pooled day 4 and 5 conditioned media (a and b) from control (), IL-1 (), or retinoate-treated () cultures were concentrated and fractionated on a Sepharose CL-2B column under associative conditions, and the fractions were analyzed for sulfated glycosaminoglycan (a and c). The column fractions were then divided into pools A, B, and C as shown, dialyzed, and concentrated, and the amount of 342FFGVG epitope in each pool was determined by radioimmunoassay (b and d).

Aggrecan fragments extracted with 4 M GdnHCl from cartilage after 5 days in culture eluted as a broad peak on Sepharose CL-2B with K_av of approximately 0.25 (Fig. 8c). The average size of ³⁴²FFGVG fragments in the GdnHCl extracts was larger than that of fragments released into the medium. 20-30% of fragments extracted from IL-1- or retinoate-stimulated cultures were present in pool A (Fig. 8d), compared with the medium samples, which contained no ³⁴²FFGVG fragments in pool A (Fig. 8b).

MMP-derived Aggrecan Fragments Are Resistant to Aggrecanase Cleavage-- The survival in cartilage of large ³⁴²FFGVG fragments suggested that these fragments were not degraded by aggrecanase, although there was continuing loss from the tissue of aggrecan, presumably by aggrecanase activity. We therefore designed an experiment to determine whether an ³⁴²FFGVG fragment could be cleaved by aggrecanase. For these experiments, conditioned medium from control cultures containing large ³⁴²FFGVG fragments (Fig. 2, days 1 and 2) was concentrated and used as substrate. Conditioned medium from day 4 and 5 cultures stimulated with IL-1 and retinoate together (data not shown) was concentrated and used as a source of aggrecanase; concentrated medium from day 4 and 5 unstimulated cultures served as an enzyme control. Purified whole aggrecan (A1D1) was also digested and served as a positive and negative control for the presence of aggrecanase activity in the stimulated and control enzyme source, respectively. A small amount of the aggrecanase-derived G1 neoepitope ITEGE³⁷³ was present in the stimulated enzyme source (Fig. 9b, lane 12) but not in the control enzyme source (Fig. 9b, lane 11). There were no active enzymes present in the control enzyme source, since no new ³⁴²FFGVG or ITEGE³⁷³ epitope was generated from intact A1D1 (Fig. 9, lanes 2 and 3) or the ³⁴²FFGVG substrate (Fig. 9, lanes 7 and 8). When porcine aggrecanase (stimulated enzyme source) was incubated with articular A1D1, a strong ITEGE³⁷³ signal was detected on Western blots (Fig. 9b, lanes 4 and 5). This aggrecanase activity was not inhibited by exogenous TIMP-1 added at 3.5 µM (Fig. 9b, lane 5). As expected, only a small amount of ITEGE³⁷³ epitope was detected following incubation of aggrecanase with the ³⁴²FFGVG substrate, since the majority of the substrate already lacked G1 domains (Fig. 9b, lanes 9 and 10). Based on their size, the ITEGE³⁷³ fragments present in Fig. 9b, lanes 9 and 10, represent the G1 domain derived from intact aggrecan (present in low abundance) lost by passive release (36) into the culture medium. Most importantly, when aggrecanase was incubated with the ³⁴²FFGVG substrate, there was no detectable loss of AF-28 epitope (Fig. 9c, lanes 9 and 10), indicating that the active aggrecanase was unable to cleave the ³⁴²FFGVG substrate. In a separate experiment, the ³⁴²FFGVG substrate was digested with either aggrecanase, atrolysin C, or cathepsin B, and the survival of the ³⁴²FFGVG epitope was monitored by Western blotting after composite gels. The intensity of the Western blotting signal was unchanged by aggrecanase digestion but was completely removed by atrolysin C or cathepsin B digestion (data not shown). These results show that MMP-derived aggrecan fragments with ³⁴²FFGVG N termini are not substrates for aggrecanase cleavage in the IGD and indicate that the structural changes conferred on aggrecan by MMP removal of its G1 domain renders it resistant to subsequent aggrecanase cleavage in the IGD.

View larger version (99K): [in this window] [in a new window]   Fig. 9.   MMP-derived 342FFGVG fragments are resistant to aggrecanase digestion. Purified articular A1D1 (50 µg) (lanes 1-5) or 342FFGVG fragments present in conditioned media from day 1 cultures (50 µg) (lanes 6-10) were incubated overnight with control enzyme source (lanes 2 and 3 and lanes 7 and 8) or stimulated enzyme source (lanes 4 and 5 and lanes 9 and 10). Control and stimulated enzyme source was also incubated without substrate (lanes 11 and 12). TIMP-1 at a final concentration of 3.5 µM was included in tubes 3, 5, 8, and 10. The samples were electrophoresed on composite gels for toluidine blue staining (a) and 342FFGVG epitope (c) or SDS-PAGE following deglycosylation for ITEGE373 epitope (b).

The converse experiment was also done, in which aggrecanase-derived G1 fragments were tested as substrates for MMPs. Aliquots of dialyzed extracts were digested with MMP-13 and analyzed by Western blotting for a shift in the size of the G1 product and a change in G1 C-terminal neoepitopes (Fig. 10). Anti-G1 antiserum detected fragments of M_r 68,000 that were reduced to approximately M_r 50,000 following MMP digestion (Fig. 10). In each extract, ITEGE³⁷³ immunoreactivity was abolished by digestion with MMP (Fig. 10b, lanes 2, 4, and 6), which generated DIPEN³⁴¹ neoepitope on smaller G1 fragments (Fig. 10c, lanes 2, 4, and 6). In these experiments with exogenous MMP, all of the ITEGE³⁷³ G1 was converted to DIPEN³⁴¹ G1, suggesting that the aggrecanase-derived G1 domain is a viable substrate for MMPs.

View larger version (83K): [in this window] [in a new window]   Fig. 10.   MMP digestion of aggrecanase-derived ITEGE373 G1 fragments. Dialyzed guanidine extracts of 5-day cultured cartilage were incubated with (lanes 2, 4, and 6) or without (lanes 1, 3, and 5) 100 µg/ml MMP-13 overnight at 37 °C, deglycosylated, electrophoresed on 5% SDS gels, and analyzed by Western blotting with anti-G1 domain (a), anti-ITEGE373 antisera (b) or anti-DIPEN341 antisera (c).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We describe several new findings relating to the generation and distribution of MMP-derived aggrecan fragments in a cartilage culture system. First, we present direct quantitation of aggrecan catabolites from one turnover pathway and show that in pig cartilage explants, MMPs account for 1% or less of the total aggrecan released. This is in agreement with estimates made by Little et al. (12) and is compatible with other reports that MMP processing does not appear to occur in cultures of rat chondrosarcoma cells (11), bovine chondrocytes (11, 37), or bovine cartilage (12) systems. Our results support the conclusion that MMPs are not significantly responsible for aggrecanolysis. Second, we provide the first evidence that ³⁴²FFGVG fragments appear to be selectively retained in the cartilage matrix. These fragments may be retained for a role in cartilage homeostasis and suggest that although 1% MMP-mediated aggrecan loss may not impact significantly on load-bearing capacity, MMP-meditated catabolism may initiate other important physiological events. Third, we demonstrate that native ³⁴²FFGVG fragments are resistant to aggrecanase, and in our companion paper (56) we provide evidence to suggest that at least part of the  ... FVDIPEN³⁴¹ sequence flanking the MMP cleavage site may be required for binding of aggrecanase to its substrate. Our previous studies suggested that MMP and aggrecanase activities in pig cartilage explants were mutually exclusive (38, 39). This can now be explained by our finding that FFGVG³⁴² fragments are not substrates for aggrecanase and, consequently, that ³⁷⁴ARGSV fragments cannot be derived from ³⁴²FFGVG fragments. Fourth, we show that the DIPEN³⁴¹ epitope is widely distributed in the cartilage matrix, whereas the ITEGE³⁷³ epitope is often intracellularly localized in chondrocytes. The different distribution of the G1 neoepitopes suggests that MMP and aggrecanase activities may be spatially distinct. Fifth, we present novel data from Western blotting, radioimmunoassay, and immunolocalization showing that DIPEN³⁴¹ and ³⁴²FFGVG epitopes are significantly increased in pig cartilage explants by IL-1 treatment.

We were surprised to find such high relative proportions of ³⁴²FFGVG epitope in cartilage extracts, since aggrecan fragments are removed from cartilage. The diffusion of fragments from regions containing high concentrations of aggrecan can be rapid due to large thermodynamic chemical potential gradients or effective concentration gradients (40). Ilic et al. have characterized aggrecan in human (41) and bovine (6) cartilage matrix by sequencing the predominant components and found that only large molecules with intact N termini are retained in the tissue. In the same studies, ³⁷⁴ARGSV fragments of varying sizes and fragments derived from the chondroitin sulfate-rich region were rapidly lost from the tissue and recovered in the conditioned medium. While most degraded components are quickly released in culture, we now show that the majority of ³⁴²FFGVG fragments are retained in the tissue. IL-1 stimulation of pig cartilage significantly increased ³⁴²FFGVG fragments in tissue but without a corresponding increase in ³⁴²FFGVG epitope in the medium. The ratio of medium ³⁴²FFGVG to tissue ³⁴²FFGVG was in the range 1:4.5 for control, 1:10.8 for IL-1-stimulated, and 1:1.8 for retinoate-stimulated cartilage (Table I). Given that the ratio of medium/tissue fragments would be dramatically lower for ³⁷⁴ARGSV fragments released from human and bovine cartilage cultures (6, 41), and assuming that this would be similar for the pig, the retention of ³⁴²FFGVG fragments appears to be selective. The ³⁷⁴ARGSV fragments present in conditioned media of bovine and human cartilage cultures range in size (6, 41), so size selection does not appear to be the criterion for retention. Instead, it is possible that ³⁴²FFGVG fragments are specifically recognized and retained in the tissue by binding of the neoepitope sequence to matrix molecules or cells. Preliminary immunolocalization experiments did not detect the ³⁴²FFGVG epitope in tissue cultured with or without IL-1, consistent with the hypothesis that the epitope sequence was involved in binding elsewhere and was therefore unavailable for detection by AF-28 antibody. We have conducted some preliminary experiments to determine whether ³⁴²FFGVG fragments bind cartilage cells or matrix. The preliminary results indicate that cultured pig chondrocytes do not bind ¹²⁵I-labeled ³⁴²FFGVG G2 domain at 4 or 15 °C but that there may be some binding of ¹²⁵I-labeled FFGVGGEEDITVQTY 15-mer to cartilage slices, in a highly regionalized manner.²

Our results show that not all ³⁴²FFGVG fragments are retained in cartilage (Figs. 2, 3, and 8), and indeed other studies have shown that ³⁴²FFGVG fragments generated by MMP-3 digestion of tissue or 4-aminophenylmercuric acetate activation of endogenous MMPs leads to release of ³⁴²FFGVG-aggrecan (15, 42). Similarly, ³⁴²FFGVG fragments are present in human synovial fluid (16) and conditioned medium of IL-1-stimulated human osteoarthritic cartilage (12). It is interesting to consider the possibility that the ³⁴²FFGVG-aggrecan described in this study may represent the "non-aggregating" large proteoglycan. A number of studies have shown that cartilage extracts from mature animals contain a proportion of extractable aggrecan that does not aggregate with hyaluronan and link protein (43-46). In addition, we have extracted human and porcine cartilage with 4 M guanidine and found ³⁴²FFGVG fragments that distribute in the A2 and A3 fractions of cesium chloride density gradients,² indicating that these fragments are large but nonaggregating. Thus, while the proportion of ³⁴²FFGVG fragments generated in cultured tissue may only be 1% or less of the total, an accumulation of this product in vivo might represent a significant pool in mature tissue.

The 32-amino acid keratan sulfate containing fragment Phe³⁴²-Glu³⁷³, which migrates on SDS gels with an approximate M_r of 35,000 (27), was not detected in our experiments. This fragment can only come from processing of ITEGE³⁷³ G1 to DIPEN³⁴¹ G1, since the ³⁴²FFGVG fragment is resistant to aggrecanase cleavage at the Glu-Ala bond. Based on Western blotting of sequential daily extracts (Fig. 4) it is possible that ITEGE³⁷³ to DIPEN³⁴¹ processing may occur following IL-1 stimulation, since the ITEGE³⁷³ epitope was reduced concomitant with increased DIPEN³⁴¹ epitope. Alternatively, ITEGE³⁷³ epitope could be decreased by release of ITEGE³⁷³-G1 into the culture medium or internalization by chondrocytes (Fig. 7B). However, the marked increase in large molecular weight ³⁴²FFGVG epitope in the tissue suggests that the majority of the increased DIPEN³⁴¹ epitope derives from MMP cleavage of aggrecan and that MMP processing of G1 contributes minimally to the pool of DIPEN³⁴¹ fragments. In addition to the concomitant increase in the ³⁴²FFGVG epitope, we have also shown³ that increased DIPEN³⁴¹ epitope is not derived from cathepsin activity (47) in this system. We cannot exclude the possibility that our inability to detect the 32-mer fragment in guanidine extracts on Western blots is due to its internalization by chondrocytes. Indeed, we do not know whether the internalized ITEGE³⁷³ epitope seen in Fig. 7B represents intact G1, Phe³⁴²-Glu³⁷³ 32-mer fragment, an intermediate, or a combination of all of these. Similarly, we cannot exclude the possibility that ³⁴²FFGVG fragments may also be internalized, causing an underestimate of MMP activity in the AF-28 radioimmunoassay.

Several kinetic studies have shown that in normal mature articular cartilage there are at least two metabolic pools of aggrecan that turn over at different rates (48-52). These studies have determined long and short half-lives for different populations of aggrecan and shown the presence of a metabolically active pool with a short half-life and a metabolically inactive pool with a correspondingly longer half-life. Furthermore, it has been suggested that the metabolically active pool is located in the pericellular and territorial matrix surrounding chondrocytes, while the inactive pool is located in the interterritorial matrix, more remote from the cells (49). If the half-life of aggrecan were a function of its microanatomical location in cartilage, this would support the hypothesis that turnover pools reflect the different microdistribution of degradative enzymes. The immunolocalization data (Figs. 6 and 7) show that when the DIPEN³⁴¹ epitope was present, both cells and matrix were stained; however, ITEGE³⁷³ staining was mostly cell-associated and intracellular throughout the depth of the tissue, often in the absence of any matrix staining. One interpretation of this finding is that expression of active aggrecanase may be restricted to the pericellular region and therefore contribute to catabolism of the active pool. Active MMPs, as detected by DIPEN³⁴¹ staining, may have a wider distribution and therefore make a greater contribution toward turnover of aggrecan in the inter-territorial matrix during normal conservative turnover. Based on this hypothesis, we speculate that the pericellular location of aggrecanase would increase the opportunity for internalization of aggrecanase-generated ITEGE³⁷³ G1 domain, while the inter-territorial location of MMPs would limit the opportunity for internalization of DIPEN³⁴¹ G1 domain.

Because it is a qualitative technique, immunolocalization of DIPEN³⁴¹ and ITEGE³⁷³ epitopes may not reflect their true tissue content. For example, in control tissue, neither epitope is detected in 8-µm sections (Fig. 6, a and f), but tissue extracts of approximately 1-3 mg (Fig. 4) contain these epitopes in low abundance, relative to IL-1-treated tissue. Certainly, ITEGE³⁷³ epitope is not restricted to an intracellular location, since it was present in the superficial layer of IL-1-treated cartilage (Fig. 6g) and is released in large amounts from cartilage and also found in human synovial fluids (53). The lack of ITEGE³⁷³ matrix staining in deeper zones does not exclude the possibility of its presence there altogether but suggests either that the epitope is much less abundant in matrix than cells or that it may be more reactive in cells than in matrix. This could be due to variable accessibility of the antibodies or altered presentation of epitope.

In summary, our data suggest that in pig cartilage explants MMPs are involved in the conservative base-line turnover of aggrecan, which may amount to catabolism of around 1% of total aggrecan in the tissue. We suggest that because the ³⁴²FFGVG products of MMP catabolism are selectively retained in cartilage matrix and resistant to destruction by aggrecanase, they may have a role in conveying signals or organizing a microenvironment. Finally, since detection of ³⁴²FFGVG fragments in human synovial fluids (16) and IL-1-stimulated cultures of human osteoarthritic cartilage (12) suggests that their abundance may be greater in human tissue compared with cultured pig cartilage, further studies are needed to resolve the relative involvement of MMPs and aggrecanase in human growth and development and in disease.

	ACKNOWLEDGEMENTS

We thank Prof. Gillian Murphy (University of East Anglia, Norwich, UK) for recombinant TIMP-1, Dr. Vera Knäuper and Prof. Gillian Murphy (University of East Anglia) for pro-MMP-13, and Dr. Rose Maciewicz (AstraZeneca Pharmaceuticals, UK) for helpful discussions.

	FOOTNOTES

^* This work was supported by the National Health and Medical Research Council (Australia), the Royal Children's Hospital Research Institute (Melbourne), and the University of Melbourne Collaborative Research Program.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom correspondence should be addressed. Tel.: 61-3-9345-6628; Fax: 61-3-9345-7997; E-mail: fosang@cryptic.rch.unimelb.edu.au.

Supported by the Medical Research Council UK; recipient of a Royal Society (UK) study travel award to undertake a University of Melbourne Collaborative Research Program in the Department of Paediatrics.

Published, JBC Papers in Press, July 5, 2000, DOI 10.1074/jbc.M910207199

² A. J. Fosang, K. Last, and H. Stanton, unpublished results.

³ H. Stanton and A. Fosang, unpublished results.

	ABBREVIATIONS

The abbreviations used are: ADAM, a Disintegrin and Metalloproteinase-containing family of proteins; ADAMTS, A Disintegrin and Metalloproteinase with Thrombospondin motif-containing family of proteins; DMEM, Dulbecco's modified Eagle's medium; GdnHCl, guanidine hydrochloride; IGD, interglobular domain of aggrecan; MMP, matrix metalloproteinase; IL-1, interleukin-1; TIMP, tissue inhibitor of metalloproteinases; PBS, phosphate-buffered saline; BSA, bovine serum albumin.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES