Tetraspanin CD151 is expressed in osteoarthritic cartilage and is involved in pericellular activation of pro–matrix metalloproteinase 7 in osteoarthritic chondrocytes

Authors


Abstract

Objective

The proenzyme of matrix metalloproteinase 7 (proMMP-7), which can degrade various extracellular matrix (ECM) and non-ECM molecules after being activated, is overexpressed in osteoarthritic (OA) articular cartilage, but the process of its activation in the cartilage remains unknown. The present study was undertaken to investigate the expression of tetraspanin CD151 in OA cartilage and its involvement in proMMP-7 activation.

Methods

The expression of CD151 in articular cartilage was examined by reverse transcription–polymerase chain reaction (RT-PCR), real-time PCR, immunohistochemistry, in situ hybridization, and immunoblotting. Chondrocytes were used to study the interaction between CD151 and proMMP-7, and activation of proMMP-7.

Results

RT-PCR revealed expression of CD151 messenger RNA in all OA cartilage samples, but in only 30% of normal control cartilage samples. Immunohistochemistry and in situ hybridization findings indicated that CD151 was coexpressed with proMMP-7 in chondrocytes, mainly in the superficial and transitional zones of OA cartilage. CD151 immunoreactivity directly correlated with the Mankin score (r = 0.757, P < 0.0001 [n = 30]) and the degree of chondrocyte cloning (r = 0.83, P < 0.0001 [n = 30]) in the cartilage samples. Complexes CD151 and proMMP-7 and their colocalization on the cell membranes were demonstrated by immunoprecipitation and double fluorescence immunostaining of the OA chondrocytes. In situ zymography indicated that chondrocytes exhibit pericellular proteolytic activity, which was abolished by treatment with MMP inhibitors, anti–MMP-7 antibody, or anti-CD151 antibody.

Conclusion

These data demonstrate that CD151 is overexpressed in OA cartilage and suggest that CD151 plays a role in the pericellular activation of proMMP-7, leading to cartilage destruction and/or chondrocyte cloning.

Osteoarthritis (OA) is a common disease affecting various joints, such as the knees, hips, spine, and hands. Loss of proteoglycans from the cartilage extracellular matrix (ECM) and subsequent collagen degradation lead to fibrillation and laceration, and finally, complete loss of the articular cartilage, with exposure of underlying subchondral bone. Elevated proteolytic activity is reported to play a central role in ECM degradation during the destruction of articular cartilage in OA (1, 2). Among the several classes of proteinases, matrix metalloproteinases (MMPs), a gene family of structurally and functionally related zinc endopeptidases (3), are thought to be key enzymes involved in the ECM degradation in cartilage (2). MMPs are synthesized as inactive proenzymes consisting of 3 basic domains, a propeptide, a zinc-binding catalytic site, and a COOH-terminal hemopexin-like domain. The MMP family is broadly classified into secreted-type MMPs and membrane-type MMPs (MT-MMPs), the latter of which possess a transmembrane region or a glycosyl phosphatidylinositol–anchored tail next to the hemopexin-like domain (2, 3).

Many MMPs, including MMP-1 (tissue collagenase) (4), MMP-2 (gelatinase A) (5), MMP-3 (stromelysin 1) (4, 6), MMP-7 (matrilysin 1) (7), MMP-8 (neutrophil collagenase) (8), MMP-9 (gelatinase B) (9), MMP-13 (collagenase 3) (10), and MT1-MMP (5), are expressed in human OA cartilage. Among them, MMP-7 is unique in that it lacks the COOH-terminal hemopexin-like domain (11) and has a broad range of substrate specificity (12–20). Although its substrate specificity was originally reported to be similar to that of MMP-3 (19), it has been shown that among the MMP gene family members, MMP-7 has the highest specific activity against many ECM components (12, 15, 16, 18, 20). Recent studies have indicated that MMP-7 is involved in the shedding of heparin-binding epidermal growth factor precursor (21), E-cadherin (22), and syndecan 1 (23). In addition, biochemical studies performed at our institution (20) have demonstrated that MMP-7 can activate the zymogen of MMP-1 (proMMP-1) and proMMP-9. Thus, MMP-7 is considered to play a dual role in cartilage pathology, through direct digestion of ECM and non-ECM molecules and through activation of proMMP-1 and proMMP-9. However, there is little or no available information on proMMP-7 activation at the cellular level in OA cartilage.

We recently determined that CD151 (PETA-3/SFA-1 [platelet-endothelial cell tetra-span antigen/SF-HT–activated gene 1]) (24, 25), a member of the tetraspanin gene family, binds proMMP-7, and that proMMP-7 is captured and activated on the cell membranes of carcinoma cells through interaction with CD151 (26). This prompted us to study the expression of CD151 in OA articular cartilage and its potential involvement in proMMP-7 activation in chondrocytes. During the time the current manuscript was in preparation, Diaz-Romero et al (27) reported that cultured chondrocytes isolated from normal human articular cartilage expressed several tetraspanin molecules, including CD151. However, they did not investigate expression of CD151 within OA cartilage tissue or its biologic function in chondrocytes. In the present study, we examined the expression and localization of CD151 in OA articular cartilage, and the activation of proMMP-7 through interaction with CD151 in OA chondrocytes.

MATERIALS AND METHODS

Clinical samples and histologic analysis.

Nonosteophytic articular cartilage samples were obtained at arthroplasty from hip joints (n = 5) or knee joints (n = 15) of patients with OA (mean ± SD age 69.8 ± 8.2 years) diagnosed according to the criteria of the American College of Rheumatology (formerly, the American Rheumatism Association) (28). Normal control cartilage samples without macroscopic changes were obtained from hip joints of patients with femoral neck fracture (n = 10) (mean ± SD age 80.4 ± 14.2 years). The cartilage samples were cut into slices (∼3 mm thick), fixed with periodate-lysine-paraformaldehyde (PLP) or 4% paraformaldehyde fixative for ∼24 hours at 4°C, and embedded in paraffin wax after decalcification with 0.5M EDTA (pH 7.4). PLP-fixed paraffin sections (4 μm thick) were stained with hematoxylin and eosin or toluidine blue, and histologic and histochemical characteristics were graded according to the Mankin scale (29). The degree of chondrocyte cloning was determined by calculating the ratio of the number of chondrocytes to the number of lacunae in the cartilage, by a modification of the method reported by Johnson et al (30). Informed consent was obtained from the patients for the experimental use of the surgical samples, according to hospital ethics guidelines.

Reverse transcription–polymerase chain reaction (RT-PCR).

Total RNA was extracted directly from articular cartilage samples (20 OA and 10 normal samples). Articular cartilage was shaved into chips with a scalpel and immediately placed into liquid nitrogen. Frozen cartilage chips (1–2 gm at a time) were pulverized in a Cool Mill (Toyobo, Osaka, Japan) for 60–90 seconds at maximum impact frequency. One gram of the powdered cartilage was added to 5 ml Buffer RLT (Qiagen, Mississauga, Ontario, Canada) containing guanidine thiocyanate, and mixed by gentle rocking for ∼4 hours at room temperature. Subsequent procedures were performed according to the RNeasy Mini Protocol for Plant cells and Tissues (Qiagen). RNA samples were evaluated with an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA), and the tissue samples in which the 28S:18S ribosomal RNA ratios were >1.0 were used for further study. Using a random oligonucleotide hexamer (Takara Bio, Otsu, Japan) and Moloney murine leukemia virus reverse transcriptase (ReverTra Ace; Toyobo), randomly primed complementary DNA (cDNA) was prepared from total RNA (0.5 μg).

A 2.5 μl aliquot of the reaction products was subjected to RT-PCR analysis for the expression of CD151, MMP-7, or GAPDH, for 23 and 25 cycles, 32 and 34 cycles, or 20 and 25 cycles, respectively. PCR was performed in a 20-μl reaction volume containing 500 nM of each primer, 200 μM dNTP, and 0.5 units of Ex Taq DNA polymerase (Takara Bio). The thermal cycle was 1 minute at 94°C, followed by 1 minute at 65.5°C (for human CD151), 67°C (for MMP-7), or 60°C (for GAPDH) and 1 minute at 72°C, followed by 3 minutes at 72°C for the final extension. Nucleotide sequences of the PCR primers were as follows: for human CD151 5′-ACAGCCTACATCCTGGTGGT-3′ (forward), 5′-TTCTCCTTGAGCTCCGTGTT-3′ (reverse); for MMP-7 5′-GGTCACCTACAGGATCGTATCATAT-3′ (forward), 5′-CATCACTGCATTAGGATCAGAGGAA-3′ (reverse); for GAPDH 5′-CCACCCATGGCAAATTCCATGGCA-3′ (forward), 5′-TCTAGACGGCAGGTCAGGTCCACC-3′ (reverse). Expected sizes of the amplified cDNA fragments of CD151, MMP-7, and GAPDH were 197 bp, 373 bp, and 600 bp, respectively. An aliquot of the PCR products was electrophoresed in a 2% agarose gel and stained with ethidium bromide. The nucleotide sequence of the amplified fragments was confirmed by cycle sequencing using a DYEnamic ET dye terminator cycle sequencing kit (MegaBACE; Amersham Pharmacia Biotech, Tokyo, Japan) and MegaBACE 1000 DNA sequencer (Amersham Pharmacia Biotech).

Real-time quantitative PCR.

For quantitative analysis of CD151 expression, cDNA was used as template in a TaqMan real-time PCR assay (ABI Prism7000 Sequence Detection System) according to the protocol recommended by the manufacturer (Applied Biosystems, Foster City, CA). Cycling conditions were as follows: 50°C for 2 minutes and 95°C for 10 minutes, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. The primers (forward 5′-TCCTGCTCCTCATCATCTTTCTG-3′, reverse 5′-CCGTGTTCAGCTGCTGGTAGT-3′) and TaqMan probe (FAM-TCGCTGGTATCCTCGCCTACGCC-TAMRA) for CD151 were chosen using the computer program Primer Express (Applied Biosystems). Sample data were normalized to 18S ribosomal RNA, which was selected as endogenous control using a TaqMan Human Endogenous Control Plate (Applied Biosystems). The total gene specificity of the nucleotide sequences chosen for the primers and probe and the absence of DNA polymorphisms were ascertained via BLASTN and Entrez (http://www.ncbi.nlm.nih.gov/).

In situ hybridization.

Paraffin sections from paraformaldehyde-fixed samples (9 OA and 2 normal cartilage samples) were used for in situ hybridization according to a modification of our previously described method (31). Briefly, single-stranded sense and antisense digoxigenin-labeled RNA probes were generated by in vitro transcription of the cDNA with T3 or T7 RNA polymerase, using the DIG RNA labeling kit according to the protocol recommended by the manufacturer (Boehringer Mannheim, Mannheim, Germany). Template DNA was a cDNA fragment (453 bp) encoding the COOH-terminal extracellular domain of human CD151, which was subcloned in pGEM-3Zf(+) vector from pACT2-CD151EL2 (26). Serial paraffin sections were hybridized with the digoxigenin-labeled antisense or sense probes and then subjected to immunostaining using mouse antidigoxigenin antibody (1:750 dilution; Boehringer Mannheim) followed by application of the avidin–streptavidin complex method using the Catalyzed Signal Amplification system according to the instructions of the manufacturer (Dako, Glostrup, Denmark). After the reactions, the sections were counterstained with hematoxylin.

Immunohistochemistry analysis.

Sections from the PLP-fixed samples (20 OA and 10 normal cartilage samples) were treated with 0.3% H2O2 and 1% bovine serum albumin (BSA) to block endogenous peroxidase and nonspecific binding, respectively. For immunohistochemistry analysis of CD151, the sections were treated with 0.4% pepsin (Dako Japan, Kyoto, Japan) before blocking. They were then incubated with mouse monoclonal antibody against human CD151, which recognizes the extracellular loop of CD151 (11G5a; 10 μg/ml) (Serotec, Oxford, UK), mouse monoclonal antibody against proMMP-7 (141-7B2; 10 μg/ml) (Daiichi Fine Chemical, Takaoka, Japan), or nonimmune mouse IgG (10 μg /ml) (Dako Japan). After reaction with goat antibody against mouse IgG conjugated with peroxidase-labeled Dextran polymer (not diluted) (En Vision+Mouse; Dako Japan), color was developed with 3,3′-diaminobenzidine tetrahydrochloride in 50 mM Tris HCl buffer (pH 7.6) containing 0.006% H2O2. After the reaction, the sections were counterstained with hematoxylin and observed under a light microscope. Immunoreactivity of CD151 (percentage of total chondrocytes that were immunostained) in the OA cartilage samples (n = 20) was measured by observing ∼200 chondrocytes per sample at a magnification of 200×. To examine the colocalization of CD151 and proMMP-7, paired mirror sections of OA cartilage (5 samples) were prepared and each section was subjected to CD151 and proMMP-7 immunohistochemistry analysis as described above.

Chondrocyte cultures and stimulation with proinflammatory cytokines and growth factors.

Chondrocytes were isolated from OA cartilage by incubation with 0.4% (weight/volume) Pronase (Calbiochem, La Jolla, CA) for 1 hour and then with 0.4% (w/v) bacterial collagenase type I (Worthington, Freehold, NJ) for 3 hours at 37°C. Chondrocytes were cultured on culture flasks by plating at a density of 1 × 104 cells/cm2 in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Tissue Culture Biologicals, Tulare, CA) and 25 μg/ml ascorbic acid (Gibco BRL, Grand Island, NY). A chondrocytic phenotype of the cultured cells was confirmed by the positive immunostaining of aggrecan and type II collagen observed using mouse monoclonal antibody against human aggrecan (1:10 dilution; Abcam, Cambridge, UK) and rabbit polyclonal antibody against human type II collagen (1:20 dilution; Monosan, Uden, The Netherlands), respectively (31) (results not shown). In order to examine the stimulatory effects of proinflammatory cytokines and growth factors on CD151 expression, cultured OA chondrocytes were starved for 24 hours in serum-free DMEM/F-12 containing 0.2% lactalbumin hydrolysate and then left untreated or treated for 24 hours with 0.1, 1, or 10 ng/ml interleukin-1α (IL-1α; Dainippon, Osaka, Japan), tumor necrosis factor α (TNFα; Dainippon), transforming growth factor β (TGFβ; R&D Systems, Minneapolis, MN), or vascular endothelial growth factor 165 (VEGF165; R&D Systems). In some experiments, OA chondrocytes were treated for 24 hours with both 1 ng/ml IL-1α and 10 ng/ml TNFα. After the treatments, total RNA was extracted from chondrocytes and expression of CD151 and MMP-7 messenger RNA (mRNA) was examined by RT-PCR as described above.

Immunoprecipitation and immunoblotting.

Since MMP-7 expression in OA chondrocytes is up-regulated by IL-1α and TNFα, as shown in previous studies at our laboratory (7), OA chondrocytes were treated with 1 ng/ml IL-1α and 10 ng/ml TNFα for 48 hours, and then cell lysates were subjected to immunoprecipitation with anti-CD151 antibody (14A2.H1; BD PharMingen, San Diego, CA) or nonimmune mouse IgG1κ (BD PharMingen) and protein G–Sepharose 4 Fast Flow beads (Amersham Pharmacia Biotech), after preclearing with the beads according to previously described methods (32). Immunoprecipitates were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (12.5% total acrylamide) under reducing conditions. After electrophoresis, proteins were electrotransferred onto polyvinylidene difluoride membranes (Atto, Tokyo, Japan), and the membranes were incubated with 125I-labeled anti–MMP-7 antibody (125-20H11; 10 μg/ml) or 125I-labeled anti-CD151 antibody (11G5a; 10 μg/ml) at 4°C for 12 hours; nonspecific reaction was blocked with 3% BSA in phosphate buffered saline (PBS). The bound antibodies were detected using an imaging plate and the BAS-2000 system (Fuji, Tokyo, Japan). Iodination of the antibodies was performed using Iodogen reagent (Pierce, Rockford, IL) as previously described (32).

Double immunohistochemistry analysis of CD151 and MMP-7.

OA chondrocytes cultured on Lab-Tek chamber slides (Nalge Nunc International, Tokyo, Japan) were treated with IL-1α and TNFα as described above, and then unfixed cells were incubated with mouse monoclonal antibody against human CD151 (10 μg/ml) and rabbit polyclonal antibody against MMP-7 (M8683; 10 μg/ml) (Sigma-Aldrich, St. Louis, MO) according to our previously described methods (26). As controls, chondrocytes were reacted with nonimmune mouse IgG (10 μg/ml) and nonimmune rabbit IgG (10 μg/ml; Dako Japan). They were washed 3 times with PBS and reacted with fluorescein isothiocyanate– or tetramethylrhodamine isothiocyanate–conjugated secondary antibodies (1:50 dilution; Dako Japan). After washing, coverslips were mounted with fluorescent mounting medium (Dako Japan). Immunofluorescence was examined using a laser scanning confocal microscope (Fluoview FV300; Olympus, Tokyo, Japan).

In situ zymography using crosslinked carboxymethylated transferrin (CCm-Tf) films.

Chondrocytes with or without IL-1α and TNFα treatment were detached by incubation for 10 minutes with 0.05% trypsin and 0.2% EDTA, and the activity of trypsin was completely blocked with trypsin neutralization solution (Cambrex, East Rutherford, NJ). Cells were suspended with serum-free DMEM/F-12 containing 0.2% lactalbumin hydrolysate and then seeded on CCm-Tf films, which were prepared by coating polyethylene telephthalate support with CCm-Tf (2.5 μm thickness) according to our previously described methods (33). Cells on the films were incubated in a 5% CO2 incubator for 6 hours at 37°C and then stained with amido black 10B. For some experiments, suspensions of cytokine-treated chondrocytes were incubated on CCm-Tf films treated with 0.1Mo-phenanthroline, or suspensions in the presence of 20 μg/ml recombinant tissue inhibitor of metalloproteinases 1 (rTIMP-1) (34), anti–MMP-7 antibody (25 μg/ml), anti-CD151 antibody (25 μg/ml), nonimmune mouse IgG (25 μg/ml), or nonimmune rabbit IgG (25 μg/ml) were seeded on CCm-Tf films, followed by incubation for 6 hours at 37°C.

Statistical analysis.

The Mann-Whitney U-test was used to compare the significance of the difference in results obtained with OA samples and normal samples. Spearman's rank correlation with simple linear regression was used for analysis of the relationships between parameters. P values less than 0.05 were considered significant.

RESULTS

Expression of CD151 and MMP-7 mRNA in OA and normal cartilage.

The expression of mRNA for CD151 and MMP-7 in OA and normal cartilage tissue was examined by RT-PCR. CD151 was expressed in all of the OA cartilage samples (20 of 20), whereas it was detected in only 30% of the normal control cartilage samples (3 of 10) (Figure 1A). MMP-7 was detected in 70% of the OA cartilage samples (14 of 20) and in only 10% of the control cartilage samples (1 of 10). The specific amplification from the target mRNA for CD151 and MMP-7 was confirmed by sequencing the amplified DNA products (results not shown).

Figure 1.

Expression of CD151 and matrix metalloproteinase 7 (MMP-7) mRNA in articular cartilage samples and in situ hybridization of CD151 in osteoarthritic (OA) cartilage. A, Total RNA extracted from OA and normal (NOR) cartilage was reverse transcribed into cDNA, followed by polymerase chain reaction (PCR) using specific primer pairs for CD151, MMP-7, and GAPDH as described in Materials and Methods. Ten representative OA samples and 5 representative normal samples are shown. PCR cycle numbers are shown in parentheses. B, Relative level of CD151 expression in OA and normal cartilage samples. ∗ = P < 0.005. C, Photomicrographs showing the results of application of digoxigenin-labeled antisense and sense probes to paraffin sections as described in Materials and Methods. A positive signal in OA chondrocytes was obtained with the antisense probe (arrows), but not with the sense probe. Bar = 25 μm. Inset shows positive chondrocytes at a higher magnification.

Real-time quantitative PCR analysis of CD151 in OA and normal cartilage.

Since CD151 was predominantly expressed in OA cartilage samples, we further examined the mRNA expression levels by real-time quantitative PCR. As shown in Figure 1B, the relative level of CD151 expression was significantly higher in OA cartilage (mean ± SD 0.86 ± 0.37) than in control cartilage (0.48 ± 0.21) (P < 0.005).

In situ hybridization findings.

CD151 mRNA–expressing cells in the articular cartilage were identified by in situ hybridization. Chondrocytes in the superficial and transitional zones of OA cartilage were frequently labeled with the anti-sense RNA probe (Figure 1C), whereas specimens from normal cartilage were not labeled (results not shown). The sense probe showed negligible background signal in OA cartilage (Figure 1C) and normal cartilage (results not shown).

Immunolocalization of CD151 in OA and normal cartilage.

The cartilage specimens from nonosteophytic areas of OA cartilage had typical OA changes, such as surface irregularities, fibrillation, and fissuring. Mankin scores in the samples ranged from 3 to 12 (mean ± SD 6.7 ± 2.4; n = 20). Control samples from normal articular cartilage showed few or no microscopic changes, with Mankin scores ranging from 0 to 2 (mean ± SD 1.2 ± 0.6; n = 10). Immunohistochemistry analysis demonstrated that in 85% of the samples (17 of 20), CD151 localized to the OA chondrocytes mainly in the superficial and transitional zones (Figures 2B and C). The chondrocytes located in the radial zone were stained when the cartilage had deep fissures reaching the transitional zone. Clustered chondrocytes were frequently labeled (Figures 2B and C).

Figure 2.

Immunolocalization of CD151 in A, normal and B–D, osteoarthritic (OA) cartilage tissue, and its correlations with E, Mankin score and F, degree of chondrocyte cloning. Normal and OA cartilage samples were immunostained with A–C, anti-CD151 antibody or D, nonimmune mouse IgG as described in Materials and Methods. A, Normal cartilage (Mankin grade 1), showing negligible staining. B, Cartilage from a patient with moderate OA (Mankin grade 8), showing staining of chondrocytes, including clustered cells (arrow), in the superficial zone. C, Cartilage from a patient with severe OA (Mankin grade 11), showing staining of clustered chondrocytes (arrows) in the transitional and radial zones. D, Cartilage from a patient with severe OA (Mankin grade 11) after incubation with nonimmune IgG. No staining is seen. Bar = 100 μm. E and F, Positive correlations between CD151 immunoreactivity and Mankin score (r = 0.757, P < 0.0001; n = 30) and between CD151 immunoreactivity and degree of chondrocyte cloning (r = 0.83, P < 0.0001; n = 30), respectively. Open circles represent normal cartilage samples; solid circles represent OA cartilage samples.

When the immunoreactivity (percentage of immunostained chondrocytes) in OA cartilage was calculated, it was found that ∼30% of the total chondrocytes (mean ± SD 30.8 ± 24.8%) immunostained positively for CD151 in the OA samples. The immunoreactivity was 8.4 ± 18.4% in samples from patients with mild OA (Mankin scores 3–6; n = 9), 34.0 ± 24.0% in samples from patients with moderate OA (Mankin scores 7–10; n = 9), and 67.0 ± 23.1% in samples from patients with severe OA (Mankin scores 11–14; n = 2). CD151 staining was found in 10% of the normal cartilage samples (1 of 10), within a few chondrocytes in the superficial zone (0.6 ± 2.0%) (Figure 2A). OA and normal cartilage samples immunostained with nonimmune mouse IgG showed no reactivity (Figure 2D and results not shown). CD151 immunoreactivity was significantly higher in OA cartilage than in normal cartilage (P < 0.001). A linear correlation between immunoreactivity and Mankin score (r = 0.757, P < 0.0001; n = 30) (Figure 2E) was observed. In addition, there was a similar direct correlation between immunoreactivity and the degree of chondrocyte cloning (r = 0.83, P < 0.0001; n = 30) (Figure 2F).

Since it was previously demonstrated that MMP-7 immunolocalization in OA chondrocytes is directly correlated with the Mankin score (7), we further examined the colocalization of CD151 and MMP-7 in paired mirror sections of OA cartilage (n = 5). As shown in Figure 3, CD151 and MMP-7 were colocalized in the chondrocytes in the superficial and transitional zones of OA cartilage.

Figure 3.

Colocalization of CD151 and MMP-7 in OA cartilage. Mirror sections of OA cartilage from a patient with mild OA (Mankin grade 3 [A and B]) and a patient with moderate OA (Mankin grade 9 [C and D]) were immunostained with anti-CD151 antibody (A and C) or anti–MMP-7 antibody (B and D) and counterstained with hematoxylin as described in Materials and Methods. Bar = 50 μm. Insets show higher magnification of the boxed areas indicated. See Figure 1 for definitions.

Effects of proinflammatory cytokines and growth factors on CD151 expression in OA chondrocytes.

To study the effects of IL-1α, TNFα, TGFβ, and VEGF165 on CD151 expression levels, mRNA expression was examined in OA chondrocytes that were left untreated or were treated with IL-1α, TNFα, TGFβ, or VEGF165 (all at 0.1, 1, or 10 ng/ml). No changes in levels of CD151 expression associated with these treatments were found (data not shown). When chondrocytes were treated with both IL-1α (1 ng/ml) and TNFα (10 ng/ml), MMP-7 expression was induced to a high level as previously reported (7), but CD151 expression was not altered (data not shown).

Coimmunoprecipitation of CD151 and proMMP-7 from OA chondrocytes.

Since OA chondrocytes treated with IL-1α and TNFα produced both CD151 and proMMP-7, lysates of the chondrocytes were subjected to immunoprecipitation with anti-CD151 antibody to examine the interaction between CD151 and proMMP-7. As shown in Figure 4, anti-CD151 antibody immunoprecipitated proMMP-7 of 29 kd but not active MMP-7 of 19 kd, both of which can be recognized by the anti–MMP-7 antibody used. Immunoprecipitation of CD151 (30 kd) was confirmed by immunoblotting of the precipitates with anti-CD151 antibody (Figure 4). No immunoprecipitation was noted with nonimmune mouse IgG.

Figure 4.

Demonstration of the CD151–proMMP-7 complex formation by immunoprecipitation. Lysates of OA chondrocytes treated with both interleukin-1α and tumor necrosis factor α were subjected to immunoprecipitation (IP) with anti-CD151 antibody (lanes 4 and 6) or nonimmune mouse IgG (NI) (lanes 3 and 5), and then immunoblotted (IB) with 125I-labeled anti–MMP-7 antibody (lanes 3 and 4) or anti-CD151 antibody (lanes 5 and 6) as described in Materials and Methods. Note that proMMP-7 of 29 kd (arrowhead) is coimmunoprecipitated with CD151 of 30 kd (arrow). Lanes 1 and 2, Authentic purified proMMP-7 and active MMP-7, respectively. See Figure 1 for other definitions.

Colocalization of CD151 and MMP-7 on cell membranes of OA chondrocytes.

In order to show colocalization of CD151 and MMP-7 in chondrocytes, double immunostaining was performed on unfixed OA chondrocytes stimulated with IL-1α and TNFα. Both CD151 and MMP-7 were localized on the cell membranes of chondrocytes in a dotted pattern (Figures 5A and B), and a merged image of the double immunostaining showed yellow dots on the membranes (Figure 5C), indicating that the 2 molecules coexist on the cell membranes. Immunostaining using nonimmune IgG revealed no signal (Figure 5D).

Figure 5.

Colocalization of MMP-7 and CD151 on the cell membranes of OA chondrocytes. OA chondrocytes treated with both interleukin-1α and tumor necrosis factor α without fixation were reacted with anti-CD151 monoclonal antibody and anti–MMP-7 rabbit polyclonal antibody, incubated with fluorescein isothiocyanate– or tetramethylrhodamine isothiocyanate–conjugated secondary antibodies, and observed under a laser scanning confocal microscope as described in Materials and Methods. A, Localization of CD151. B, Localization of MMP-7. C, Merged image of CD151 and MMP-7. D, Immunostaining with nonimmune IgG. Bar = 10 μm. See Figure 1 for definitions.

Detection of MMP-7 activity on cell membranes of OA chondrocytes.

To detect proteolytic activity of MMP-7 in OA chondrocytes, in situ zymography using CCm-Tf films was performed. When untreated OA chondrocytes were incubated on CCm-Tf films, they showed negligible or no digestion because of negligible proMMP-7 expression (results not shown). However, OA chondrocytes treated with IL-1α and TNFα exhibited pericellular digestion of CCm-Tf (Figure 6A). The proteolytic activity of stimulated chondrocytes was almost completely abolished when the cells were incubated on films coated with o-phenanthroline (Figure 6B). The activity was also blocked with rTIMP-1, a specific MMP inhibitor (Figure 6C). These results indicate that the lysis was due to MMP activity. Negligible activity was detected when the chondrocytes were treated with anti–MMP-7 antibody (Figure 6D) or anti-CD151 antibody (Figure 6E), although nonimmune rabbit and mouse IgG had no effect (Figure 6F and results not shown).

Figure 6.

Demonstration of MMP-7 activity by in situ zymography of OA chondrocytes. OA chondrocytes treated with both interleukin-1α and tumor necrosis factor α were incubated on crosslinked carboxymethylated transferrin (CCm-Tf) films and stained with amido black 10B as described in Materials and Methods. A, Control OA chondrocytes. B, OA chondrocytes incubated on o-phenanthroline–coated CCm-Tf films. C, OA chondrocytes in the presence of recombinant tissue inhibitor of metalloproteinases 1 (rTIMP-1). D, OA chondrocytes in the presence of anti–MMP-7 antibody. E, OA chondrocytes in the presence of anti-CD151 antibody. F, OA chondrocytes in the presence of nonimmune rabbit IgG. Note that pericellular activity (arrowheads) is abolished by treatment with o-phenanthroline, rTIMP-1, anti–MMP-7 antibody, or anti-CD151 antibody. Chondrocytes have a round shape because they did not completely spread on the CCm-Tf films during the 6-hour incubation. Bar = 10 μm. See Figure 1 for other definitions.

DISCUSSION

In the present study, we have demonstrated that CD151, a member of the tetraspanin family, is overexpressed in human OA articular cartilage. This was shown by RT-PCR, real-time PCR, in situ hybridization, and immunohistochemistry experiments with cartilage tissue. The expression of CD151 was further confirmed by RT-PCR, immunofluorescence staining, and immunoprecipitation in studies of cultured OA chondrocytes.

A recent flow cytometric study of cell surface markers demonstrated that cultured chondrocytes isolated from normal human articular cartilage express several tetraspanin molecules including CD9, CD63, CD81, CD82, and CD151, as well as integrins and adhesion molecules, receptors, ectoenzymes, and surface molecules (27), although the report did not provide information about the relative expression of these molecules in normal cartilage compared with OA cartilage. In our present work, CD151 was expressed by chondrocytes in almost all OA cartilage samples, whereas expression was confined to <30% of normal cartilage samples. The preferential expression in OA cartilage suggested the presence of a system of regulation of CD151 expression in joint tissue such as cartilage, but our study showed that neither IL-1α, TNFα, TGFβ, nor VEGF165 affected CD151 expression levels on cultured chondrocytes. Thus, factors regulating CD151 expression in chondrocytes remain to be elucidated.

Findings of previous studies have suggested that CD151 is involved in adhesion-dependent signaling in epithelial, endothelial, and muscle cells, as well as platelets and megakaryocytes, through association with integrins (35–39) and that it also promotes the invasion and metastasis of tumor cells (40). We have recently reported that CD151 can interact with propeptide of proMMP-7 and activate proMMP-7 on the cell membranes of carcinoma cells, and we suggested that this activation process could be important for cancer invasion and metastasis (26). Our findings in the present study provide the first evidence that CD151 and proMMP-7 are coexpressed in chondrocytes from human OA articular cartilage and are coimmunoprecipitated from the chondrocytes. In addition, we have demonstrated colocalization of CD151 and proMMP-7 on the cell membranes of OA chondrocytes, by confocal microscopy. Most importantly, we detected MMP-7 activity on the cell surface of OA chondrocytes, by in situ zymography. The finding that the activity was specifically abolished by treatment with anti–MMP-7 antibody and anti-CD151 antibody as well as with rTIMP-1 and o-phenanthroline suggests that proMMP-7 is activated on the cell surface of OA chondrocytes through interaction with CD151.

Despite the detection of MMP-7 activity in OA chondrocytes, the mechanism of CD151-mediated proMMP-7 activation in OA chondrocytes is not known. Because CD151 has no recognized proteolytic activity, another activator of proMMP-7 may be necessary for this to occur. Findings in a biochemical study conducted at our institution (20) indicated that proMMP-7 is fully activated by MMP-3. MMP-3 is commonly expressed in OA chondrocytes (4, 6) and readily activated by several serine proteinases, such as plasmin (41). Thus, it is tempting to speculate that MMP-3 acts as an activator of proMMP-7 that has been captured on the cell membranes of OA chondrocytes through interaction with CD151.

Another possible activation mechanism may be autoactivation of proMMP-7, since proMMP-7 in vitro has a propensity to autoactivate and to degrade itself in a concentration-dependent manner (42, 43). In addition, the finding that proMMP-7 activation occurs only on OA chondrocytes plated on films coated with CCm-Tf, an MMP-7 substrate, suggests that this activation requires a substrate of MMP-7. Increased concentrations of proMMP-7 captured on cell membranes via the action of CD151 may lead to spontaneous activation through conformational changes of the proenzyme, which are dependent on the presence of its substrate. Understanding of the mechanism of CD151-mediated proMMP-7 activation on OA chondrocytes will require further study.

CD151 immunoreactivity (the percentage of immunoreactive cells in relation to the total number of chondrocytes) was directly correlated with the Mankin score, a marker of cartilage destruction (29). Since the expression of MMP-7 is also known to correlate with the Mankin score (7), the correlation between CD151 expression and Mankin score suggests that CD151-mediated proMMP-7 activation might be implicated in OA cartilage destruction. The present study showed that the level of CD151 immunoreactivity also correlates with the degree of chondrocyte cloning in the cartilage.

Accumulated evidence from functional analyses of MMP-7 has demonstrated that MMP-7 is capable of shedding various membrane proteins, including heparin-binding epidermal growth factor (21), E-cadherin (22), and syndecan 1 (23). In a parallel study, we have demonstrated that heparin-binding epidermal growth factor and its receptor are overexpressed in OA cartilage (Okada A, et al: unpublished observations). Therefore, it can be speculated that CD151 plays a role in chondrocyte proliferation by facilitating the availability of heparin-binding epidermal growth factor shed by MMP-7. A recent study showed that active aggrecanase 1 (also called ADAMTS-4) binds to syndecan 1 to localize its aggrecan-degrading activity on the cell surface (44). Thus, shedding of syndecan 1 from chondrocytes via MMP-7 might reduce pericellular aggrecanase activity and prevent chondrocyte migration after proliferation, promoting chondrocytes to form cell clusters. This hypothesis could be studied with existing methodology.

In summary, we have demonstrated that CD151 is overexpressed and colocalized with MMP-7 in the chondrocytes of OA cartilage. Immunohistochemistry analysis showed that the immunoreactivity of CD151 is directly correlated with the Mankin score and the degree of chondrocyte cloning in cartilage tissue. The interaction of these molecules results in pericellular activation of proMMP-7 on OA chondrocytes. Our data suggest that this activation of proMMP-7 mediated by CD151 is implicated in the pathologic changes that occur in OA cartilage, such as ECM degradation and/or chondrocyte cloning.

Acknowledgements

We thank Dr. Edward D. Harris, Jr., (Stanford University, School of Medicine, Palo Alto, CA) for reviewing the manuscript. We are also grateful to Drs. Y. Horiuchi, M. Kihara, T. Ohtani, E. Nomura, Y. Suda, H. Maruiwa, M. Kurimura, H. Hotta, M. Jinnouchi, and A. Okada for providing us with cartilage samples, and to Ms M. Uchiyama and Mr. H. Abe for skillful technical assistance.

Ancillary