Involvement of CD44 in induction of matrix metalloproteinases by a COOH-terminal heparin-binding fragment of fibronectin in human articular cartilage in culture.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 48, No. 5, May 2003, pp 1271–1280 DOI 10.1002/art.10951 © 2003, American College of Rheumatology Involvement of CD44 in Induction of Matrix Metalloproteinases by a COOH-Terminal Heparin-Binding Fragment of Fibronectin in Human Articular Cartilage in Culture Tadashi Yasuda,1 A. Robin Poole,2 Makoto Shimizu,1 Takefumi Nakagawa,1 Sohel M. Julovi,1 Hirokazu Tamamura,3 Nobutaka Fujii,3 and Takashi Nakamura1 effect. Treatment of cartilage with anti-CD44 antibody or HSPG resulted in significant inhibition of HBFN-f– stimulated MMP production. Preincubation with peptide V blocked binding of the anti-CD44 antibody to chondrocytes in cartilage. Conclusion. Interaction of the peptide V sequence in HBFN-f with glycosaminoglycans, such as those in CD44, plays an important role in HBFN-f–stimulated MMP production in articular cartilage. Because CD44 is up-regulated in osteoarthritic and rheumatoid arthritic cartilage, the role of the interaction between CD44 and HBFN-f in these pathologies should be of relevance and should be studied further. Objective. To investigate the mechanism of induction of matrix metalloproteinases (MMPs) by a 40-kd COOH-terminal heparin-binding fibronectin fragment (HBFN-f) containing III12–14 and IIICS domains in human articular cartilage in culture. Methods. Human articular cartilage was removed from macroscopically normal femoral heads and cultured with HBFN-f. MMP secretion into conditioned media was analyzed by immunoblotting (MMPs 1 and 13) and by gelatin zymography (MMPs 2 and 9). Type II collagen cleavage by collagenase was monitored in culture by immunoassay. Involvement of specific peptidebinding domains in HBFN-f and the involvement of CD44 were assessed with synthetic peptides and an anti-CD44 antibody. Immunofluorescence histochemistry was performed using fluorescein isothiocyanate– conjugated anti-CD44 antibody. Results. HBFN-f stimulated production of MMPs 1, 2, 9, and 13 in association with type II collagen cleavage by collagenase in human articular cartilage. Peptide V (WQPPRARI) of HBFN-f, which can bind cell surface heparan sulfate proteoglycan (HSPG), blocked MMP induction by HBFN-f, while the scrambled peptide V (RPQIPWAR) had no effect. Peptide CS-1 of 25 amino acids in IIICS of HBFN-f caused no significant The extracellular matrix of articular cartilage is composed mainly of proteoglycans and collagens, and its integrity provides the mechanical properties of cartilage (1). Progressive destruction of cartilage, which results from an imbalance between the anabolic and catabolic processes, is a common feature of rheumatoid arthritis (RA) and osteoarthritis (OA). Interleukin-1␤ (IL-1␤) and tumor necrosis factor ␣ have been shown to promote cartilage degradation by stimulating the production of matrix metalloproteinases (MMPs) (2). Fibronectin is a component of normal cartilage matrix (3). It consists primarily of 3 types of homologous repeating segments (designated I, II, and III). Fibronectin contains NH2-, gelatin-, cell-, and COOH-terminal heparin-binding domains. The central cell-binding region has an RGD sequence in domain III10, which is recognized by several cell surface integrin family members (4). Several sites in the heparin-binding domain that lie COOH-terminal to the central cell-binding domain also interact with the cell surface. Regions of domain III12–14 support cell attachment with varying affinities (5–7). The IIICS, or variable (V) region, contains the 1 Tadashi Yasuda, MD, PhD, Makoto Shimizu, MD, Takefumi Nakagawa, MD, Sohel M. Julovi, MD, Takashi Nakamura, MD, PhD: Kyoto University Graduate School of Medicine, Kyoto, Japan; 2 A. Robin Poole, PhD, DSc: Joint Diseases Laboratory, Shriners Hospital for Children, and McGill University, Montreal, Quebec, Canada; 3Hirokazu Tamamura, PhD, Nobutaka Fujii, PhD: Kyoto University Graduate School of Pharmaceutical Sciences, Kyoto, Japan. Address correspondence and reprint requests to Tadashi Yasuda, MD, PhD, Department of Orthopaedic Surgery, Kyoto University Graduate School of Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan. E-mail: email@example.com. Submitted for publication September 5, 2002; accepted in revised form January 24, 2003. 1271 1272 YASUDA ET AL ␣4␤1 integrin-binding sites, CS-1 and CS-5 (8–10). The COOH-terminal heparin-binding domain of fibronectin is known to bind CD44 (11), a principal hyaluronan receptor (12). Elevated levels of fibronectin fragments are found in OA cartilage (13–15) and in OA synovial fluid (15,16). The central cell-binding, NH2-terminal heparinbinding, and NH2-terminal gelatin-binding fragments of fibronectin have been shown to stimulate proteoglycan breakdown (17) and release of catabolic cytokines (18) in cultured articular cartilage explants. In addition to those fibronectin fragments, we recently found that the 40-kd COOH-terminal heparin-binding fibronectin fragment (HBFN-f) containing both the heparin-binding III12–14 and IIICS domains can stimulate type II collagen cleavage by collagenase following proteoglycan degradation, in association with the production of MMPs 3 and 13 in bovine articular cartilage explant cultures (19). Thus, increased levels of fibronectin fragments are thought to be involved in cartilage destruction in OA and RA through the induction of cytokines and MMPs. Currently, there is little information regarding the mechanisms whereby MMPs are induced by fibronectin fragments. MMP production by the central cell-binding fragment of fibronectin is probably mediated by ␣5␤1 integrin because the anti-␣5␤1 integrin antibody and the RGD-containing peptide induce MMP-1 and gelatinase in synovial fibroblasts (20). Recent studies using antisense oligonucleotides to an ␣5 integrin subunit have also shown the involvement of ␣5 integrin in cartilage chondrolysis induced by 29-kd NH2terminal heparin-binding and 50-kd NH2-terminal gelatin-binding fragments in addition to the cell-binding fragment of fibronectin (21). However, it remains unclear how other fibronectin fragments, including HBFN-f, can induce MMP production in cartilage. In this study, we investigated the mechanism of action of HBFN-f on MMP production in human articular cartilage. We found that a specific amino acid sequence of HBFN-f mediated the induction of collagenase and gelatinase in human articular cartilage, and that this involves binding of HBFN-f to CD44. MATERIALS AND METHODS Antibodies and reagents. Anti-human MMP-1 that reacts with 53-kd and 51-kd bands of proenzyme (M4177), anti-human MMP-2 that recognizes the 72-kd band (M4677), anti-human MMP-9 that reacts with the 92-kd band (M5427), and anti-human MMP-13 that recognizes the latent proenzyme at 60 kd (M4052) were obtained from Sigma (St. Louis, MO). Alkaline phosphatase–conjugated goat anti-rabbit IgG was purchased from Southern Biotechnology (Birmingham, AL). OS/37, a monoclonal anti-human CD44 antibody, was obtained from Seikagaku Kogyo (Tokyo, Japan). Mouse IgG1 was obtained from ICN Biomedicals (Aurora, OH). A 40-kd proteolytic COOH-terminal heparin-binding fragment of human plasma fibronectin containing type III12–14 segments and IIICS (HBFN-f) generated with ␣-chymotrypsin digestion was obtained from Gibco BRL (Gaithersburg, MD). The purity of the protein preparation was confirmed using the same method as used in our previous studies (19). Recombinant human IL-1␤ was purchased from R&D Systems (Minneapolis, MN). Purified human MMPs 2 and 9 were obtained from Chemicon (Temecula, CA). Heparan sulfate purified from porcine intestinal mucosa was purchased from Sigma. Synthetic peptides CS-1 (DELPQLVTLPHPNLHGPEILDVPST), peptide V (WQPPRARI), and the scrambled peptide V (RPQIPWAR) were obtained from Takara Shuzo (Kusatsu, Japan). The sequence for the scrambled peptide V is identical to that used in previous studies (22,23). Articular cartilage explant culture. Adult human articular cartilage was obtained from the femoral head after replacement surgery for femoral neck fracture. No significant arthritic changes such as fibrillation, were identified macroscopically in the articular cartilage. The cartilage was placed in a 24-well plate (⬃80 mg/well) and kept in 2 ml of serum-free Dulbecco’s modified Eagle’s medium (DMEM) containing 100 units/ml penicillin, 100 g/ml streptomycin, and 10 mM HEPES (all from Gibco BRL) in a humidified 5% CO2 atmosphere at 37°C. In this study, a serum-free cartilage explant culture system was used because chondrocytes in cartilage explants under serum-free conditions are still viable 40 days after treatment with fibronectin fragments (24), and serum supplementation decreases the chondrolytic activity of fibronectin fragments (17). The cartilage was precultured for 2 days, and medium was changed on day 0. Thereafter, medium was replaced every 4 days. Fresh HBFN-f or IL-1␤ was added (starting on day 0) at each medium change, as indicated below. Assays of endotoxin levels in 100 nM HBFN-f solution with the endotoxin assay kit (Sigma) showed minimal levels of detection of 6–8 ng/ml (19), significantly lower than the concentrations (1.0 g/ml) that are required to alter cartilage metabolism (25). In some experiments, heparan sulfate was added with or without HBFN-f beginning on day 0. In another set of experiments, following preincubation with peptide V, the scrambled peptide V, or CS-1, or following preincubation with antibody OS/37 or subclass-matched mouse IgG1 for 2 hours, articular cartilage was coincubated with HBFN-f from day 0. The cartilage explants and conditioned media were harvested on days 4, 8, and 12, and stored at ⫺20°C. Immunoblot analysis. Conditioned media were heated with sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer at 80°C for 20 minutes. Proteins were separated by SDS-PAGE under reducing conditions, and then transferred onto nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany). Gel loading was standardized based on the DNA contents of the cartilage explants. Membranes were blocked in phosphate buffered saline (PBS) containing 5% nonfat dry milk and 0.1% Tween 20 and incubated with the first antibody (concentration 1:1,000) overnight at 4°C. After incubation with alkaline phosphatase–conjugated second antibody (concentration CD44 IN MMP INDUCTION BY FIBRONECTIN FRAGMENT 1273 1:1,000) for 3 hours at room temperature, immunoreactive bands were visualized using BCIP and nitroblue tetrazolium. The presence of heparan sulfate itself had no significant effect on immunoblotting (data not shown). Protein band intensity was evaluated by densitometry using NIH Image (National Institutes of Health, Bethesda, MD). Gelatin zymography. Serum-free conditioned media were collected from human articular cartilage explant culture. Samples were prepared in nondenaturing loading buffer and separated on a 10% SDS–polyacrylamide gel impregnated with 1 mg/ml of gelatin. The amount of sample applied was determined, as above, by the DNA content in the cartilage explants. After electrophoresis, gels were washed with 50 mM Tris HCl (pH 7.5) containing 0.1M NaCl and 2.5% Triton X-100, and then rinsed with 50 mM Tris HCl (pH 7.5). To visualize proteinases, gels were incubated at 37°C in an incubation buffer containing 50 mM Tris HCl (pH 7.5), 200 mM NaCl, 10 mM CaCl2, and 0.02% NaN3. Gels were subsequently fixed and stained in Coomassie blue fixation solution. The areas of gelatin degraded by gelatinase were measured on the scanned digital image of stained gels. The presence of heparan sulfate itself had no significant effect on gelatin zymography (data not shown). Extraction and assay for cleavage of type II collagen by collagenase. Cartilage explants were digested to extract cleaved type II collagen as previously described (19,26). Briefly, the harvested cartilage explant was incubated overnight with 1.0 mg/ml ␣-chymotrypsin at 37°C to cleave and solubilize denatured collagen. After inhibition of ␣-chymotrypsin activity with N-tosyl-L-phenylalanine-chloromethyl ketone (Sigma), the samples were centrifuged, and the supernatants were recovered. The Col2-3/4Cshort epitope (hereafter referred to as Col2-3/4C) generated by cleavage of type II collagen by collagenase (26) was measured in ␣-chymotrypsin extracts by immunoassays. The release of the Col2-3/4C epitope into media was also measured by immunoassay (19,27,28). From the measurement of Col2-3/4C in ␣-chymotrypsin extracts and medium, the total amount of epitope present in both the tissue and medium for the period between the last medium change and the harvest was calculated. The remaining explant residues were digested overnight with 1.0 mg/ml proteinase K at 56°C for DNA assay (see below). Immunofluorescence histochemistry. After blocking with 1% bovine serum albumin for 24 hours, articular cartilage slices from the femoral head were incubated with fluorescein isothiocyanate (FITC)–conjugated antibody OS/37 (Seikagaku Kogyo) at 5 g/ml or subclass-matched FITC-conjugated mouse IgG1 (KPL, Gaithersburg, MD) at 5 g/ml for 24 hours at 37°C. In some experiments, after preincubation with peptide V at 100 M for 1 hour, articular cartilage was then incubated with FITC-conjugated OS/37 at 5 g/ml. Thereafter, following an extensive wash with DMEM, cartilage slices were subjected to cryostat sectioning at 6 m and fixed with 4% paraformaldehyde in PBS for 20 minutes. After counterstaining with propidium iodide (KPL), the sections were examined by confocal microscopy (FluoView; Olympus, Tokyo, Japan). Assay for DNA. DNA content was measured in proteinase K digests of articular cartilage explants, as previously described (29). Statistical analysis. Comparisons between 2 groups were performed by Student’s t-test. P values less than 0.05 were considered significant. RESULTS Induction of collagenase and gelatinase by HBFN-f in human articular cartilage culture. Consistent with our previous study, which showed that HBFN-f can induce significant type II collagen degradation in bovine articular cartilage by day 12 under serum-free conditions (19), HBFN-f stimulated the production of collagenases (MMPs 1 and 13) and gelatinases (MMPs 2 and 9) in human articular cartilage under serum-free conditions. Control cultures without treatment secreted basal levels of MMP-1, with barely detectable levels of MMP-13 production, as shown by Western blotting of culture media. When cartilage was incubated with HBFN-f at 1, 10, and 100 nM for 4 days, the fragment induced MMPs 1 and 13 only at 100 nM (Figure 1A). Immunoblot analysis demonstrated that HBFN-f at 100 nM also induced MMPs 2 and 9, while gelatin zymography revealed that the fragment caused increased secretion of latent and active forms of MMPs 2 and 9 (Figure 1B). Treatment with 100 nM HBFN-f from days 0 to 12 resulted in enhanced secretion of MMPs 1 and 13 into media during days 0–4, 4–8, and 8–12 (Figure 2A). IL-1␤ at 2 ng/ml also induced these collagenases; data from days 8 to 12 are shown in Figure 2A. MMP-1 levels (mean ⫾ SD) induced by HBFN-f reached a plateau by day 4, whereas MMP-13 levels increased with time, 20 ⫾ 13% (n ⫽ 3) during days 0–4 and 85 ⫾ 5% (n ⫽ 3) during days 4–8 compared with the protein band intensity during days 8–12 (calculated as 100%). While basal levels of MMP-2 (4 ⫾ 1%, n ⫽ 3) were observed in control cultures, HBFN-f enhanced levels of proMMP-2 and active MMP-2 (50 ⫾ 8% and 80 ⫾ 7%, n ⫽ 3) during days 0–4 and 4–8, respectively, compared with the levels during days 8–12 (calculated as 100%) (Figure 2B). Similarly, HBFN-f induced a significant secretion of both proMMP-9 and active MMP-9 with time, while no detectable MMP-9 was found in control cultures (Figure 2B). HBFN-f–increased levels of MMP-9 were 20 ⫾ 9% and 50 ⫾ 8% during days 0–4 and 4–8, respectively (n ⫽ 3), when the levels during days 8–12 were calculated as 100%. Induction of the release of collagenase-generated cleavage epitope of type II collagen by HBFN-f in human articular cartilage culture. HBFN-f at 100 nM induced an enhanced release into medium of the Col2-3/4C epitope generated by collagenase cleavage of type II collagen during days 0–8 (Figure 3). This release de- 1274 YASUDA ET AL to bind other receptors. Therefore, the CS-1 synthetic peptide was tested to determine the effect of its sequence in HBFN-f–stimulated MMP induction. Treatment with 10 M CS-1 alone resulted in a slightly enhanced secretion of MMP-1, compared with the control (Figure 4). When CS-1 was used in conjunction with HBFN-f, secretion of MMP-1 was enhanced. HBFN-f–induced levels of MMP-1 in the presence of 10 M CS-1 were 110 ⫾ 18%, 120 ⫾ 23%, and 140 ⫾ 27% during days 0–4, 4–8, and 8–12, respectively (n ⫽ 3), compared with HBFN-f–induced levels in the absence of CS-1 during days 8–12 (calculated as 100%). However, CS-1 at 10 M enhanced secretion of MMP-1 only, with no effect on MMPs 13, 2, or 9 until day 12. Treatment with lower (0.1 and 1 M) or higher (100 M) concentrations of CS-1 peptide yielded similar results (data not shown). Effects of peptide V derived from HBFN-f on HBFN-f–stimulated MMP induction in human articular cartilage explant culture. Peptide V derived from the III14 repeat of HBFN-f has a known heparin-binding activity and promotes focal adhesion formation (31). Figure 1. Induction of matrix metalloproteinases (MMPs) by COOH-terminal heparin-binding fibronectin fragment (HBFN-f) in human articular cartilage explant culture. A, HBFN-f at 1, 10, or 100 nM was added from day 0. Secreted levels of MMPs 1 and 13 during days 0–4 in conditioned media were analyzed by immunoblotting. B, HBFN-f at 100 nM was added from day 0. Secreted levels of MMPs 2 and 9 in conditioned media during days 4–8 were evaluated by immunoblotting and gelatin zymography. Purified MMPs 2 and 9 were used as positive controls. Two separate experiments were performed, with similar results. pends upon secondary cleavage of the denatured ␣ chain bearing the epitope that involves MMP activity (28). Levels of Col2-3/4C were similar in cultures with or without treatment with 100 nM HBFN-f on days 4 and 8. Consequently, HBFN-f caused increased generation of the Col23/4C epitope with time during days 0–8 in human articular cartilage culture. Our previous studies, using the specific inhibitor of MMP-13, suggest that collagenase 3 is a major player in type II collagen cleavage caused by HBFN-f (19). Enhancement of MMP-1 induction by the CS-1 domain of HBFN-f in human articular cartilage explant culture. Binding of HBFN-f to the ␣4␤1 integrin involves the CS-1 sequence in the IIICS domain of HBFN-f (8). Although normal chondrocytes scarcely express ␣4␤1 integrin (30), CS-1 may have the potential Figure 2. Time course of HBFN-f–induced MMP production in explant culture. Articular cartilage was incubated with HBFN-f at 100 nM from day 0. Control cultures contained no additives. Interleukin-1␤ (IL-1␤) at 2 ng/ml was used as a positive control for day 12. Conditioned media were collected on days 4, 8, and 12. A, Secreted levels of MMP-1 and MMP-13 in conditioned media were analyzed by immunoblotting. B, Secreted levels of MMPs 2 and 9 in conditioned media were evaluated by gelatin zymography. Three separate experiments were performed, with similar results. Each lane represents a 4-day accumulation of material and not the total accumulation from day 0 because conditioned media were collected every 4 days. See Figure 1 for other definitions. CD44 IN MMP INDUCTION BY FIBRONECTIN FRAGMENT 1275 This prompted an examination of whether peptide V could influence MMP induction by HBFN-f in human articular cartilage (Figure 5). Incubation of cartilage Figure 4. Effects of CS-1 peptide on HBFN-f–induced MMP production in explant culture. CS-1 derived from the IIICS domain of HBFN-f was added at 10 g/ml from day 0 in the presence or absence of HBFN-f at 100 nM. Control cultures (C) contained no additives. Conditioned media were collected on days 4, 8, and 12. Secreted levels of MMPs 1 and 13 in conditioned media were analyzed by immunoblotting. Secreted levels of MMPs 2 and 9 in conditioned media were evaluated by gelatin zymography. Three separate experiments were performed, with similar results. Each lane represents a 4-day accumulation of material and not the total accumulation from day 0. See Figure 1 for other definitions. Figure 3. Heparin-binding fibronectin fragment (HBFN-f)–induced cleavage of type II collagen by collagenase in explant culture. The collagenase-generated cleavage epitope in type II collagen was measured by enzyme-linked immunosorbent assay in media and in cartilage (the latter following proteolysis of collagen to release epitope). HBFN-f at 100 nM was added from day 0. Control cultures contained no additives. Cartilage explants and conditioned media with (open bars) or without (solid bars) treatment with HBFN-f were harvested on days 4 and 8. Two separate experiments were performed, with similar results. Values are the mean and SD of 4 determinations. ⴱ ⫽ P ⬍ 0.05 versus control. with 10 M peptide V alone had no effect on MMP production, while HBFN-f induced MMPs 1, 2, 9, and 13. Peptide V at 10 M completely suppressed HBFNf–stimulated production of MMPs 2, 9, and 13 to control levels, and suppressed HBFN-induced MMP-1 production by 81 ⫾ 4%, 79 ⫾ 7%, and 58 ⫾ 4% during days 0–4, 4–8, and 8–12, respectively (n ⫽ 3). A significant difference was found between the secreted levels of the individual MMPs from HBFN-f–treated cultures in the presence and absence of peptide V (P ⬍ 0.05 by t-test). In contrast to peptide V, the scrambled peptide V at 10 M had no significant effect on HBFN-f–stimulated MMP production (Figure 6A). In addition, lower concentrations of peptide V (1 M) failed to suppress HBFN-f–induced MMP (Figure 6B). Effects of heparan sulfate on HBFN-f–stimulated MMP production. Cell surface heparan sulfate has been implicated in cell adhesion to several sites in the 1276 YASUDA ET AL Involvement of CD44 in HBFN-f–induced MMP production in human articular cartilage. CD44 is expressed on a variety of cell types, including chondrocytes. While the principal ligand for CD44 is hyaluronan (12), HBFN-f is also known to bind CD44 (11). Because antibody OS/37 has been shown to block CD44 binding to its ligand (34), an attempt was made to clarify the Figure 5. Suppression of HBFN-f–stimulated production of MMPs 1, 2, 9, and 13 by peptide V in explant culture. Peptide V derived from III14 domain of HBFN-f was added at 10 M from day 0 in the presence or absence of HBFN-f at 100 nM. Control cultures (C) contained no additives. Conditioned media were collected on days 4, 8, and 12. Secreted levels of MMPs 1 and 13 in conditioned media were analyzed by immunoblotting. MMPs 2 and 9 secreted into media were evaluated by gelatin zymography. Three separate experiments were performed, with similar results. Each lane represents a 4-day accumulation of material and not the total accumulation from day 0. See Figure 1 for other definitions. III12–14 domains of HBFN-f, and peptide V can bind heparan sulfate (6,7,31–33). Thus, 1–100 g/ml of heparan sulfate proteoglycan (HSPG) was added to cartilage explant cultures to ascertain whether HSPG ligation with the HSPG-binding sites of HBFN-f including peptide V could block HBFN-f–induced MMP. Coincubation of articular cartilage with 100 g/ml HSPG in the presence of 100 nM HBFN-f resulted in complete reduction of HBFN-f–stimulated secretion of MMPs 2, 9, and 13, with decreased secretion of MMP-1 by 80 ⫾ 9% (n ⫽ 3) (Figure 7). There was a significant difference between the secreted levels of the individual MMPs from HBFNf–treated cultures in the presence and absence of HSPG (P ⬍ 0.05 by t-test). Based on the experiments using peptide V (Figures 5 and 6) and HSPG (Figure 7), we concluded that HBFN-f–stimulated MMP production in human articular cartilage involves the peptide V domain of HBFN-f. Figure 6. A, Effects of scrambled peptide V on HBFN-f–stimulated MMP production. Human articular cartilage was incubated with HBFN-f at 100 nM from day 0 in the presence or absence of peptide V or the scrambled peptide V at 10 M. Secreted levels of MMPs 1, 2, 9, and 13 during days 4–8 in the conditioned media were analyzed by immunoblotting. B, Dose-dependent effects of peptide V on HBFN-f– stimulated MMP production. HBFN-f at 100 nM was added from day 0 in the presence or absence of peptide V at 1 or 10 M. Secreted levels of MMPs 1 and 13 during days 4–8 in the conditioned media were analyzed by immunoblotting. Three separate experiments were performed, with similar results. See Figure 1 for definitions. CD44 IN MMP INDUCTION BY FIBRONECTIN FRAGMENT 1277 binding of HBFN-f mediated via the same binding site for HSPG in HBFN-f (11). In addition, HSPG and CSPG interact with the same site within the III14 repeat of HBFN-f, which contains the peptide V sequence (37). Thus, ligation of peptide V with GAGs on CD44 is likely the reason why the peptide blocked the access of antibody OS/37 to CD44, which could result in the inhibition of HBFN-f–stimulated MMP production in articular cartilage. DISCUSSION Figure 7. Suppression of HBFN-f–stimulated production of MMPs 1, 2, 9, and 13 by heparan sulfate proteoglycan (HSPG) in explant culture. HBFN-f at 100 nM was added from day 0 in the presence or absence of HSPG at 100 g/ml. Control cultures contained no additives. Secreted levels of MMPs 1, 2, 9, and 13 during days 0–4 in conditioned media were analyzed by immunoblotting. Three separate experiments were performed, with similar results. See Figure 1 for other definitions. involvement of CD44 in MMP induction stimulated by HBFN-f in articular cartilage by use of the antibody. In contrast to control IgG1, antibody OS/37 blocked HBFN-f–stimulated production of MMPs 2, 9, and 13 by 99 ⫾ 2%, 99 ⫾ 2%, and 95 ⫾ 5%, respectively (n ⫽ 3) (Figure 8). In addition, the antibody partially inhibited HBFN-f–induced MMP-1 by 50 ⫾ 8% (n ⫽ 3) (Figure 8). A significant difference was found between the secreted levels of the individual MMPs from HBFN-f– treated cultures in the presence and absence of the anti-CD44 antibody (P ⬍ 0.05 by t-test). Analysis by fluorescence microscopy revealed that FITC-conjugated antibody OS/37 localized CD44 in association with chondrocytes (Figure 9A), indicating that occupancy of CD44 by antibody OS/37 on chondrocytes can block HBFN-f– induced MMP production. The inhibitory effect of antibody OS/37 on HBFN-f–induced MMP production (Figure 8) appeared to be similar to that of peptide V (Figure 6) and HSPG (Figure 7). When articular cartilage was incubated with FITC-conjugated OS/37 following preincubation with excessive amounts of peptide V, antibody OS/37 failed to bind CD44 on chondrocytes (Figure 9B). Glycosaminoglycans (GAGs) can attach to the membrane proximal portion of the extracellular domain of CD44 (35). Chondroitin sulfate proteoglycan (CSPG) has been found on the standard isoform of CD44 (36), and it is required for the Degradation products of fibronectin are of interest as amplifiers or catalysts in diseased joints, including those in RA and OA (38), because of their ability to stimulate MMP induction (17,19,20) and cartilage destruction (17,19). In this study, we demonstrated that HBFN-f can stimulate type II collagen cleavage by collagenase in human articular cartilage explant culture under serum-free conditions, in association with enhanced production of collagenases (MMPs 1 and 13) and gelatinases (MMPs 2 and 9). These findings extend the results of our previous study, which showed that Figure 8. Suppression of HBFN-f–stimulated production of MMPs 1, 2, 9, and 13 by anti-CD44 antibody OS/37 in explant culture. Articular cartilage was incubated with HBFN-f at 100 nM from day 0, with or without pretreatment with antibody OS/37 or control IgG1. Secreted levels of MMPs 1, 2, 9, and 13 during days 0–4 in conditioned media were analyzed by immunoblotting. Three separate experiments were performed, with similar results. See Figure 1 for definitions. 1278 Figure 9. Suppression of binding of OS/37 to CD44 on chondrocytes by peptide V. A, Following blocking with bovine serum albumin (BSA), articular cartilage was incubated with fluorescein isothiocyanate (FITC)–conjugated antibody OS/37 or subclass-matched FITCconjugated mouse IgG1. B, After blocking with BSA, articular cartilage was preincubated with or without peptide V and then incubated with FITC-conjugated antibody OS/37. Bar ⫽ 50 m; in A and 200 m in B. HBFN-f can induce type II collagen degradation caused mainly by MMP-13 in bovine articular cartilage under serum-free conditions (19). The observation that injection of fibronectin fragments into rabbit knee joints induces depletion of cartilage proteoglycan is consistent with the pathophysiologic significance of the fragments (39). Elevated levels of these fragments are present in OA cartilage (15) and in the synovial fluid of OA and RA patients (15,16). In OA synovial fluids, ⬃1 M of 100–200-kd fibronectin fragments have been found (16). Since the levels of fibronectin fragments in OA cartilage have been suggested to be similar to those in OA synovial fluids (15), the contents of fibronectin fragments may reach 100 nM in OA cartilage, YASUDA ET AL comparable to the concentration used in the present study. Thus, fibronectin fragments could play an important role in cartilage destruction in arthritis. Although fibronectin isoforms containing III12–14 and IIICS domains are present in human femoral head cartilage (40), the presence of HBFN-f in diseased articular cartilage or synovial fluid in the current study remains unclear. The CD44 gene has 19 exons, 12 of which may be alternatively spliced to produce a number of different isoforms (41). Restricted expression of CD44 isoforms and posttranslational glycosylation of the parent protein provide diverse functions of CD44. Of the CD44 isoforms, CD44H is commonly expressed in human articular chondrocytes (42). Although CD44H is predominant, messenger RNA containing the V3 exon of CD44 is also found in chondrocytes (42). The diversity of CD44 is further amplified by the differential use of GAG attachment sites on its extracellular domain. While CSPG is attached to the membrane proximal portion of the external domain of CD44H (35), HSPG can bind CD44 at V3 in the membrane proximal extracellular domain of CD44v (43). CSPG and HSPG use identical or overlapping binding sites in the repeats III13 and III14 of HBFN-f (11,37). In our study, the observed suppression of HBFN-f–stimulated MMP production by peptide V suggests that the peptide V domain, a binding site of HBFN-f for cell surface HSPG, is required for HBFN-f– activated MMP induction in human articular cartilage. The results from our inhibition studies using HSPG are consistent with this idea. Thus, the results using the antiCD44 antibody indicate that HBFN-f may directly bind GAGs on CD44 through the peptide V sequence. The characterization of GAGs on CD44 on chondrocytes that can interact with peptide V is under investigation. The pathologic roles of CD44 in arthritis are not fully understood. CD44 is expressed by different cells in the RA synovium (44). Anti-CD44 treatment in other studies has been shown to result in a reduction of tissue swelling and leukocyte infiltration in a murine arthritis model (45), as well as in the inhibition of cartilage invasion by RA synovial fibroblasts (46). Immunohistochemical studies have demonstrated up-regulation of CD44 on chondrocytes in articular cartilage from patients with OA (47) and RA (48). Recent studies suggest that up-regulated CD44 on chondrocytes plays a role in increased internalization and degradation of hyaluronan (49). However, this is the first evidence that MMP induction by fragments of fibronectin may be mediated by CD44. In addition to IL-1␣, fibronectin fragments have been shown to enhance CD44 expression in bovine articular chondrocytes (50). Thus, increased generation of CD44 IN MMP INDUCTION BY FIBRONECTIN FRAGMENT 1279 HBFN-f in OA and RA may up-regulate CD44 on OA and RA chondrocytes, further supporting activation of MMP production by HBFN-f. Based on the results of the present study, CD44-directed therapy may therefore help prevent cartilage destruction by HBFN-f in OA and RA. Whether other fragments of fibronectin can stimulate MMP production through a similar mechanism remains to be clarified. The present study cannot exclude the possibility that the action of HBFN-f on chondrocytes may involve other receptors besides CD44. While fibronectin can bind several integrins and other cell surface protein ligands (51), fragments of fibronectin could interact with more than one receptor on chondrocytes. Because receptors such as integrins can work together cooperatively, blocking any one of the receptors may inactivate fibronectin fragments. Although the results of the present study indicate the minor role of the ␣4␤1 integrin in HBFN-f–induced MMP, CS-1 did stimulate MMP-1 production when used with HBFN-f (Figure 4). The interaction between CS-1 and chondrocytes may alter signaling events by blocking the binding of the natural ligand to its receptor, because one of the proposed mechanisms for the action of fibronectin fragments is that their binding to a fibronectin receptor or to other matrix components near the receptor perturbs the binding of native fibronectin to ␣5␤1 integrin (21). When rabbit synovial fibroblasts bind to the central RGD-containing fragments of fibronectin via ␣5␤1 integrin, expression of MMPs 1 and 3 is induced (20). In addition, antisense oligonucleotides to ␣5 integrin inhibit chondrolysis induced by 29-kd NH2terminal heparin-binding and 50-kd NH2-terminal gelatin-binding fragments of fibronectin (21). Thus, both cell-binding and non–cell-binding fragments of fibronectin could individually operate through ␣5␤1 integrin. Integrin ␣5␤1 is the primary receptor involved in the assembly of dimeric fibronectin into the extracellular matrix (52). The I1–5 repeats of NH2-terminal heparinbinding fragment of fibronectin block the assembly of fibronectin into fibrils, and fibronectin dimers lacking these domains fail to be incorporated into fibrils (53– 56). Since the III12–14 repeats of HBFN-f have also been shown to contribute to fibronectin fibrillogenesis (57), HBFN-f may interfere with fibronectin assembly and indirectly alter ␣5␤1 signaling. Vol. 1. 14th ed. Baltimore: Lippincott Williams & Wilkins; 2001. p. 226–84. Arend WP, Dayer J-M. Inhibition of the production and effects of interleukin-1 and tumor necrosis factor ␣ in rheumatoid arthritis. Arthritis Rheum 1995;38:151–60. Burton-Wurster N, Lust G. Fibronectin in cartilage. In: Carson S, editor. Fibronectin in health and disease. Boca Raton (FL): CRC Press; 1989. p. 243–55. Yamada KM. Adhesive recognition sequences. J Biol Chem 1991;266:12809–12. McCarthy JB, Chelberg MK, Mickelson DJ, Furcht LT. Localization and chemical synthesis of fibronectin peptides with melanoma adhesion and heparin binding activities. Biochemistry 1988;27: 1380–8. Drake SL, Klein DJ, Mickelson DJ, Oegema TR, Furcht LT, McCarthy JB. Cell surface phosphatidylinositol-anchored heparan sulfate proteoglycan initiates mouse melanoma cell adhesion to a fibronectin-derived, heparin-binding synthetic peptide. J Cell Biol 1992;117:1331–41. Iida J, Skubitz AP, Furcht LT, Wayner EA, McCarthy JB. Coordinate role for cell surface chondroitin sulfate proteoglycan and ␣4␤1 integrin in mediating melanoma cell adhesion to fibronectin. J Cell Biol 1992;118:431–44. Wayner EA, Garcia-Pardo A, Humphries MJ, McDonald JA, Carter WG. Identification and characterization of the T lymphocyte adhesion receptor for an alternative cell attachment domain (CS-1) in plasma fibronectin. J Cell Biol 1989;109:1321–30. Mould AP, Humphries MJ. Identification of a novel recognition sequence for the ␣4␤1 in the COOH-terminal heparin-binding domain of fibronectin. EMBO J 1991;10:4089–95. Mould AP, Komoriya A, Yamada KM, Humphries MJ. The CS5 peptide is a second site in the IIICS region of fibronectin recognized by the integrin ␣4␤1: inhibition of ␣4␤1 function by RGD peptide homologues. J Biol Chem 1991;266:3579–85. Jalkanen S, Jalkanen M. Lymphocyte CD44 binds the COOHterminal heparin-binding domain of fibronectin. J Cell Biol 1992; 116:817–25. Aruffo A, Stamenkovic I, Melnick M, Underhill CB, Seed B. CD44 is the principal cell surface receptor for hyaluronate. Cell 1990;61: 1303–13. Miller DR, Mankin HJ, Shoji H, D’Ambrosia RD. Identification of fibronectin in preparations of osteoarthritic human cartilage. Connect Tissue Res 1984;12:267–75. Jones KL, Brown M, Ali SY, Brown RA. An immunohistochemical study of fibronectin in human osteoarthritic and disease free articular cartilage. Ann Rheum Dis 1987;46:809–15. Homandberg GA, Wen C, Hui F. Cartilage damaging activities of fibronectin fragments derived from cartilage and synovial fluid. Osteoarthritis Cartilage 1998;6:231–44. Xie DL, Meyers R, Homandberg GA. Fibronectin fragments in osteoarthritic synovial fluid. J Rheumatol 1992;19:1448–52. Homandberg GA, Meyers R, Xie DL. Fibronectin fragments cause chondrolysis of bovine articular cartilage slices in culture. J Biol Chem 1992;267:3597–604. Homandberg GA, Hui F, Wen C, Purple C, Bewsey K, Koepp H, et al. Fibronectin-fragment-induced cartilage chondrolysis is associated with release of catabolic cytokines. Biochem J 1997;321: 751–7. Yasuda T, Poole AR. A fibronectin fragment induces type II collagen degradation by collagenase through an interleukin1–mediated pathway. Arthritis Rheum 2002;46:138–48. Werb Z, Tremble PM, Behrendtsen O, Crowley E, Damsky CH. Signal transduction through the fibronectin receptor induces collagenase and stromelysin gene expression. J Cell Biol 1989;109: 877–89. Homandberg GA, Costa V, Ummadi V, Pichika R. Antisense oligonucleotides to the integrin receptor subunit ␣5 decrease 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. REFERENCES 1. Poole AR. Cartilage in health and disease. In: Koopman WJ, editor. Arthritis and allied conditions: a textbook of rheumatology. 21. 1280 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. fibronectin fragment mediated cartilage chondrolysis. Osteoarthritis Cartilage 2002;10:381–93. Huebsch JC, McCarthy JB, Diglio CA, Mooradian DL. Endothelial cell interactions with synthetic peptides from the carboxylterminal heparin-binding domains of fibronectin. Circ Res 1995; 77:43–53. Chon JH, Chaikof EL. Soluble heparin-binding peptides regulate chemokinesis and cell adhesive forces. Am J Physiol Cell Physiol 2001;280:C1394–402. Xie DL, Hui F, Homandberg GA. Fibronectin fragments alter matrix protein synthesis in cartilage tissue cultured in vitro. Arch Biochem Biophys 1993;307:110–8. Morales TI, Wahl LM, Hascall VC. The effect of bacterial lipopolysaccharides on the biosynthesis and release of proteoglycans from calf articular cartilage cultures. J Biol Chem 1984;259: 6720–9. Billinghurst RC, Dahlberg L, Ionescu M, Reiner A, Bourne R, Rorabeck C, et al. Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage. J Clin Invest 1997;99:1534–45. Dahlberg L, Billinghurst RC, Manner P, Nelson F, Webb G, Ionescu M, et al. Selective enhancement of collagenase-mediated cleavage of resident type II collagen in cultured osteoarthritic cartilage and arrest with a synthetic inhibitor that spares collagenase 1 (matrix metalloproteinase 1). Arthritis Rheum 2000;43: 673–82. Billinghurst RC, Wu W, Ionescu M, Reiner A, Dahlberg L, Chen J, et al. Comparison of the degradation of type II collagen and proteoglycan in nasal and articular cartilages induced by interleukin-1 and the selective inhibition of type II collagen cleavage by collagenase. Arthritis Rheum 2000;43:664–72. Labarca C, Paigen K. A single, rapid, and sensitive DNA assay procedure. Anal Biochem 1980;102:344–52. Ostergaard K, Salter DM, Petersen J, Bendtzen K, Hvolris J, Andersen CB. Expression of ␣ and ␤ subunits of the integrin superfamily in articular cartilage from macroscopically normal and osteoarthritic human femoral heads. Ann Rheum Dis 1998;57: 303–8. Woods A, McCarthy JB, Furcht LT, Couchman JR. A synthetic peptide from the COOH-terminal heparin-binding domain of fibronectin promotes focal adhesion formation. Mol Biol Cell 1993;4:605–13. Lories V, Cassiman JJ, Van den Berghe H, David G. Differential expression of cell surface heparan sulfate proteoglycans in human mammary epithelial cell and fibroblasts. J Biol Chem 1992;267: 1116–22. Giuseppetti JM, McCarthy JB, Letourneau PC. Isolation and partial characterization of a cell-surface heparan sulfate proteoglycan from embryonic rat spinal cord. J Neurosci Res 1994;37: 584–95. Murakami S, Shimabukuro Y, Miki Y, Saho T, Hino E, Kasai D, et al. Inducible binding of human lymphocytes to hyaluronate via CD44 does not require cytoskeleton association but does require new protein synthesis. J Immunol 1994;152:467–77. Lesley J, Hyman R. CD44 structure and function. Front Biosci 1998;3:D616–30. Henke CA, Roongta U, Mickelson DJ, Knutson JR, McCarthy JB. CD44-related chondroitin sulfate proteoglycan, a cell surface receptor implicated with tumor cell invasion, mediates endothelial cell migration on fibrinogen and invasion into a fibrin matrix. J Clin Invest 1996;97:2541–52. Barkalow FJ, Schwarzbauer JE. Interactions between fibronectin and chondroitin sulfate are modulated by molecular context. J Biol Chem 1994;269:3957–62. Homandberg GA. Potential regulation of cartilage metabolism in osteoarthritis by fibronectin fragments. Front Biosci 1999;4: D713–30. YASUDA ET AL 39. Homandberg GA, Meyers R, Williams JM. Intraarticular injection of fibronectin fragments causes severe depletion of cartilage proteoglycans in vivo. J Rheumatol 1993;20:1378–82. 40. Parker AE, Boutell J, Carr A, Maciewicz RA. Novel cartilagespecific splice variants of fibronectin. Osteoarthritis Cartilage 2002;10:528–34. 41. Screaton GR, Bell MV, Jackson DG, Cornelis FB, Gerth U, Bell JI. Genomic structure of DNA encoding the lymphocyte homing receptor CD44 reveals at least 12 alternatively spliced exons. Proc Natl Acad Sci U S A 1992;89:12160–4. 42. Salter DM, Godolphin JL, Gourlay MS, Lawson MF, Hughs DE, Dunne E. Analysis of human articular chondrocyte CD44 isoform expression and function in health and disease. J Pathol 1996;179: 396–402. 43. Bennett KL, Jackson DG, Simon JC, Tanczos E, Peach R, Modrell B, et al. CD44 isoforms containing exon V3 are responsible for the presentation of heparin-binding growth factor. J Cell Biol 1995; 128:687–98. 44. Henderson KJ, Edwards JC, Worrall JG. Expression of CD44 in normal and rheumatoid synovium and cultured synovial fibroblasts. Ann Rheum Dis 1994;53:729–34. 45. Mikecz K, Brennan FR, Kim JH, Glant TT. Anti-CD44 treatment abrogates tissue oedema and leukocyte infiltration in murine arthritis. Nat Med 1995;1:558–63. 46. Neidhart M, Gay RE, Gay S. Anti–interleukin-1 and antiCD44 interventions producing significant inhibition of cartilage destruction in an in vitro model of cartilage invasion by rheumatoid arthritis synovial fibroblasts. Arthritis Rheum 2000;43: 1719–28. 47. Ostergaard K, Salter DM, Andersen CB, Petersen J, Bendtzen K. CD44 expression is up-regulated in the deep zone of osteoarthritic cartilage from human femoral heads. Histopathology 1997;31: 451–9. 48. Takagi T, Okamoto R, Suzuki K, Hayashi T, Sato M, Sato M, et al. Up-regulation of CD44 in rheumatoid chondrocytes. Scand J Rheumatol 2001;30:110–3. 49. Aguiar DJ, Knudson W, Knudson CB. Internalization of the hyaluronan receptor CD44 by chondrocytes. Exp Cell Res 1999; 252:292–302. 50. Chow G, Knudson CB, Homandberg G, Knudson W. Increased expression of CD44 in bovine articular chondrocytes by catabolic cellular mediators. J Biol Chem 1995;270:27734–41. 51. Johansson S, Svineng G, Wennerberg K, Armulik A, Lohikangas L. Fibronectin-integrin interactions. Front Biosci 1997;2: D126–46. 52. Dzamba BJ, Bultmann H, Akiyama SK, Peters DM. Substratespecific binding of the amino terminus of fibronectin to an integrin complex in focal adhesions. J Biol Chem 1994;269: 19646–52. 53. McKeown-Longo PJ, Mosher DF. Interaction of the 70,000mol-wt amino-terminal fragment of fibronectin with the matrix-assembly receptor of fibroblasts. J Cell Biol 1985;100: 364–74. 54. Quade BJ, McDonald JA. Fibronectin’s amino-terminal matrix assembly site is located within the 29-kDa amino-terminal domain containing five type I repeats. J Biol Chem 1988;263: 19602–9. 55. Schwarzbauer JE. Identification of the fibronectin sequences required for assembly of a fibrillar matrix. J Cell Biol 1991;113: 1463–73. 56. Sottile J, Schwarzbauer J, Selegue J, Mosher DF. Five type I modules of fibronectin form a functional unit that binds to fibroblasts and Staphylococcus aureus. J Biol Chem 1991;266: 12840–3. 57. Bultmann H, Santas AJ, Peters DM. Fibronectin fibrillogenesis involves the heparin II binding domain of fibronectin. J Biol Chem 1998;273:2601–9.