SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Objective

The membrane-anchored metalloproteinase disintegrin ADAM15 is up-regulated in osteoarthritis and has been implicated in proteolysis and cell–matrix interactions. To address its role in cartilage metabolism, we performed an analysis of joint morphology in aging mice with a targeted inactivation of the ADAM15 gene (ADAM15−/−). In addition, a human chondrocyte cell line overexpressing ADAM15 was used to investigate the role of ADAM15 in an in vitro model of chondrocyte–matrix interactions.

Methods

Knee joint sections from 3-, 6-, and 12–14-month-old ADAM15−/− and wild-type (WT) 129/SvJ mice were examined for synovial hyperplasia, cartilage degradation, and osteophyte formation. Stable transfection of the human T/C28a4 chondrocyte cell line with full-length human ADAM15 complementary DNA led to the establishment of ADAM15-overexpressing chondrocytes that were further analyzed for their capability to adhere to and to survive on cartilage matrix molecules (fibronectin and types II and VI collagen) under conditions of serum starvation. ADAM15 expression was investigated by reverse transcription–polymerase chain reaction and Western blotting.

Results

Aging ADAM15−/− mice exhibited accelerated development of osteoarthritic lesions compared with WT mice, and the difference was statistically significant at age 12 months. The osteoarthritic changes preferentially affected male ADAM15−/− mice. ADAM15 overexpression in T/C28a4 cells led to the specific reinforcement of chondrocyte adhesion to cartilage types II and VI collagen, and this was associated with enhanced cell viability under conditions of serum starvation.

Conclusion

The accelerated development of murine osteoarthritis in ADAM15 deficiency as well as the proadhesive and cell survival–promoting in vitro effect of ADAM15 overexpression suggest a homeostatic rather than a destructive role of ADAM15 in cartilage remodeling.

ADAM15 belongs to a family of adamalysin (ADAM) metalloproteinase disintegrins (MDCs) that are membrane-anchored glycoproteins containing modular metalloproteinase, disintegrin, cysteine-rich, and epidermal growth factor–like domains, followed in most cases by a transmembrane region and a cytoplasmic tail (1). ADAMs have been implicated in fertilization, myogenesis, neurogenesis, and protein ectodomain shedding (for recent reviews, see refs. 1 and2) and are also thought to play roles in cell–cell or cell–matrix adhesion through interactions with integrins (3, 4) or syndecans (5, 6).

For ADAM15, a variety of potential functions have been postulated. Thus, cell–cell interaction studies using its recombinant extracellular domains provide experimental evidence for a role in integrin ligation (αvβ3, α5β1, α9β1) (3,7,8). The localization to adherens junctions in endothelial cell cultures and an enhancement of cell–cell interactions in ADAM15-overexpressing fibroblasts (9) indicate a role in cell adhesion. In addition, ADAM15 contains a catalytic site consensus sequence for zinc-dependent metalloproteinases, and the purified recombinant protein is catalytically active (10). The cytoplasmic domain of ADAM15 harbors potential signaling motifs, such as src homology 3 (SH3) ligand domains, and has been shown to interact with Src kinase family members (11, 12).

More recently, the generation of lines of mice with a targeted disruption of ADAM15 allowed an evaluation of its role in mouse development and adult homeostasis (13). ADAM15 deficiency did not affect fertility and viability and was not associated with any evident pathologic phenotype. However, reduced responses to experimental hypoxia and tumor-induced angiogenesis in the knockout mice indicated a role of ADAM15 in pathologic neovascularization.

Another pathologic condition that has been associated with ADAM15 expression is degenerative joint disease (14, 15). The hypothesis of a role of this MDC in promoting pathologic extracellular matrix remodeling in the joints had been proposed on the basis of a strong up-regulation by chondrocytes in human osteoarthritic and neoplastic cartilage specimens (14). This hypothetical catabolic effect of ADAM15 on joint integrity was further investigated in the present comparative study of aging ADAM15-deficient (ADAM15−/−) and wild-type (WT) mice representing a spontaneous age-dependent murine model of progressive joint degeneration. The histopathologic scores used for quantification of degenerative joint disease were significantly increased (up to 3-fold) in the ADAM15−/− mice compared with the respective age- and sex-matched WT controls, although the severity of osteoarthritic lesions was more pronounced in male than in female ADAM15−/− mice.

These results render unlikely the originally proposed catabolic effect of ADAM15 on cartilage metabolism, and they instead suggest a long-term protective role of ADAM15 in maintaining joint integrity. Accordingly, overexpression of ADAM15 in the human chondrocytic cell line T/C28a4 (16) results in increased adhesion and cell survival on matrices of types II and VI collagen, which represent the major constituents of the perichondrocytic collagenous meshwork in cartilage. This newly unraveled potential of ADAM15 to enhance extracellular survival signals might be a relevant compensatory mechanism to counteract chondrocytic stress factors in joint degeneration.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Materials.

The chondrocyte cell line T/C28a4 (16) was kindly provided by Dr. M. B. Goldring (Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA). Polyclonal and monoclonal antibodies against human ADAM15 were from R&D Systems (Wiesbaden, Germany). The adhesion-blocking antibodies against α1 integrin (clone FB12), α2 integrin (clone BHA2.1), and β1 integrin (clone P4C10) were from Chemicon (Hofheim, Germany). The GRGDTP blocking peptide and control peptide GRADSP were from Calbiochem (Bad Soden, Germany). The rabbit anti-actin antibody was from Sigma-Aldrich (Taufkirchen, Germany). Type II collagen from rat swarm chondrosarcoma was a gift from Dr. K. von der Mark (Center for Molecular Medicine, University of Erlangen–Nuremberg, Erlangen, Germany). Bovine type VI collagen was purchased from Biomol (Hamburg, Germany). Human plasma fibronectin was from Chemicon.

ADAM15-knockout mice.

The generation and initial analysis of ADAM15−/− mice has been described previously (13). For the histologic study of the knee joints, groups of ADAM15−/− mice (129/SvJ) of different age ranges and corresponding litter mates were analyzed. These included 3-month-old ADAM15−/− mice (5 males and 5 females) and WT mice (5 males and 5 females), 6-month-old ADAM15−/− mice (5 males and 5 females) and WT mice (5 males and 5 females), and 12–14-month-old mice (13 WT and 11 ADAM15−/− females and 8 WT and 12 ADAM15−/− males).

Preparation of mouse joints.

After mice were euthanized and skin was removed, both knee joints were excised and fixed in 4% paraformaldehyde and decalcified in 0.4M EDTA (pH 8.0) for several weeks. In order to get cross-sections of various depths of 1 whole joint, the joint was cut once frontally in the middle, and the resulting 2 pieces (∼3-mm thick) were embedded together in paraffin. A minimum of 20 serial tissue sections (2 μm) were cut on a microtome and stained with hematoxylin and eosin or toluidine blue in order to visualize proteoglycan loss.

Scoring of mouse joints.

A total of 10 parameters were applied to evaluate degenerative changes of the knees (generally, 0 points were given for normal, nondegenerated appearance of the joint), as follows: 1) for degeneration, 1 point for an uneven appearance of the entire cartilage surface; 2) for erosion, 1 point for a small superficial lesion confined to an area of <10% of the cartilage surface, 2 points for a wider and deeper lesion of about half of the cartilage surface, 3 points for deep fissures and clefts that reach down to the bone, and 4 points for a complete loss of the cartilage layer leading to the exposure of the subchondral bone to the joint space; 3) for necrosis (assessed by toluidine blue staining), 1 point for smaller necrotic lesions, and 2 points for larger necrotic lesions; 4) for tendon/ligament metaplasia, 1 point for the presence of metaplastic cells; 5) for meniscus degeneration, 1 point; 6) for synovial hyperplasia, 1 point for small areas of hyperplastic lining, and 2 points for extensive areas of hyperplastic lining; 7) for synovial detritus, 1 point for 1–2 pieces of cartilage incorporated into the synovial membrane, and 2 points for >4 pieces; 8) for formation of osteophytes, 1 point; 9) for chondrocyte clusters, 1 point; and 10) for chondrocyte clusters in the ligaments, 1 point.

The resulting scores of the left and right knee joints of each mouse were added together to calculate the individual score. In order to obtain the average score of all mice in 1 group, the sum of the individual scores was divided by the number of mice in the group. The scoring of the knee joints was performed in a blinded manner by 2 independent investigators.

Transfection of the chondrocyte cell line T/C28a4.

A full-length complementary DNA (cDNA) of human ADAM15 was cloned into pcDNA3 vector as described previously (15) and was transfected into subconfluently growing T/C28a4 cells using FuGENE according to the manufacturer's instructions (Roche Diagnostics, Mannheim, Germany). Briefly, T/C28a4 cells were transfected in 6-well tissue culture plates and subsequently exposed to permanent high concentrations of G418 (15 mg/ml; Calbiochem) to ensure stable transfection with the pcDNA3 vector that confers resistance to G418. The G418 concentration resulting in 100% lethality to nontransfected T/C28a4 cells had been determined previously (10 mg/ml). All stably transfected cells were pooled and maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen, Karlsruhe, Germany) containing G418 (10 mg/ml). Cells that were transfected with the empty pcDNA3 vector and kept under identical G418 exposure and cell culture conditions served as the control.

Western blotting.

ADAM15-transfected and vector-transfected cells were washed with ice-cold phosphate buffered saline (PBS), scraped off the tissue culture flask, and lysed in radioimmunoprecipitation assay solution (150 mM NaCl, 50 mM Tris HCl, 1% Nonidet P40, 1% Triton X-100, 0.25% sodium desoxycholate, 5 mM EDTA, pH 7.6) containing a proteinase inhibitor cocktail (Roche Diagnostics). Cell lysates were electrophoresed on a 10% sodium dodecyl sulfate (SDS)–polyacrylamide gel and transferred to a nitrocellulose filter that was first incubated with the antibody against ADAM15 (1:1,000) followed by incubation with a horseradish peroxidase–conjugated secondary antibody (Dako, Hamburg, Germany). The protein bands were visualized using the ECL Plus Western Blotting Detection System (Amersham Biosciences, Freiburg, Germany).

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

In order to detect ADAM15 messenger RNA (mRNA) in the transfected, overexpressing chondrocyte cell line, RT-PCR using ADAM15-specific primers on the DNase I–treated RNA was performed as described previously (14). The reverse-transcribed cDNA was prepared using oligo(dT) priming with Superscript II reverse transcriptase according to the supplier's instructions (Invitrogen).

Cell adhesion assays.

Ninety-six–well Maxisorp plates (Nunc, Wiesbaden, Germany) were coated with 40 μg/ml type II collagen, 20 μg/ml type VI collagen, and 10 μg/ml fibronectin, incubated overnight at 4°C, and blocked with 1% heat-denatured bovine serum albumin (BSA) (Fraction V; Sigma-Aldrich) for 1 hour at 37°C. The cells were harvested from subconfluent cultures using trypsin cell dissociation buffer, washed with soybean trypsin inhibitor (Sigma-Aldrich), counted, and diluted to 1 × 106 cells/ml in serum-free medium. After serum starvation for 30 minutes at 37°C, 1 × 105 cells (100-μl volume) were added to the wells and allowed to adhere for 3 hours at 37°C. After washing with prewarmed (37°C) PBS, the adherent cells were fixed in 0.1% glutaraldehyde solution for 15 minutes and stained with a filtered 1% crystal violet solution for 25 minutes. The wells were washed with water, and the cells were lysed with 1% SDS in PBS. The optical density (OD) was measured at 595 nm. The adhesion assays were performed in octuple wells. Results shown are representative of at least 5 independent experiments performed.

Inhibition of cell attachment.

The assessment of cell adhesion blocking compounds was performed under the same culture and assay conditions described above. Nunc 96-well Maxisorp plates were coated with 2 μg/ml type II collagen, and the cells (1 × 105 cells in a 100-μl volume) were preincubated for 45 minutes at 37°C with monoclonal antibodies specific for α1 integrin, α2 integrin, or β1 integrin at concentrations of 10 μg/ml for the vector-transfected cells and 50 μg/ml for the ADAM15-transfected cells. Nonspecific mouse IgG was applied at the respective concentrations for negative control. The GRGDTP blocking peptide and the control peptide GRADSP were used at a concentration range of 100–300 μM. Inhibition assays of cell attachment were performed in quadruple wells and reproduced in at least 5 independent experiments.

Cell viability assessment.

Three different assays were performed to analyze viability of T/C28a4 cells (1 × 105 cells in a 100-μl volume) upon incubation in collagen-coated wells under serum-free conditions. Each experiment was performed at least 5 times in quadruple tests. Results were analyzed using unpaired t-tests.

Quantification of propidium iodide (PI)–Triton–stained degraded apoptotic nuclei.

Two hundred microliters of PI–Triton staining solution (0.1% sodium citrate, 0.1% Triton X-100, and 1 μg/ml PI) was added to the cells and incubated for 2 hours to release adherent cells. Upon transfer to polypropylene tubes, the cell suspension was adjusted to a total volume of 1 ml by adding another 700 μl of the PI–Triton staining solution and incubated for 24 hours at 4°C in the dark. The cells with PI-stained nuclei were then analyzed by flow cytometry (Epics XL; Coulter, Miami, FL). Flow cytometric quantification of cells with degraded nuclei characterized by their sub-G1 DNA content served as a parameter of apoptosis (17, 18).

Quantification of the release of lactate dehydrogenase (LDH) as a marker for cell membrane integrity.

For quantification of LDH release, the CytoTox 96 Non-Radioactive Cytotoxicity Assay was used according to the supplier's instructions (Promega, Madison, WI). After serum withdrawal and incubation in type II collagen–coated wells, the cells were spun at 250g for 10 minutes. The culture supernatants and lysed cell pellets were separately processed for photometric assessment of LDH activity at an OD of 490 nm. The percentage of LDH release was calculated as the ratio of LDH activity in the supernatant to the total LDH activity (cell pellet + supernatant) × 100.

MTS assay for monitoring metabolic activity.

The CellTiter 96 Aqueous One Solution Cell Proliferation Assay (MTS assay) from Promega was used according to the supplier's instructions to measure the metabolic conversion of MTS tetrazolium to a chromogenic formazan product that is recordable at an OD of 490 nm.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Increased severity of degenerative joint disease in aging ADAM15-deficient mice.

Systematic histopathologic examination of serial sections through the entire joints of both knees revealed the accelerated development of osteoarthritic lesions in ADAM15−/− mice compared with WT control animals. The application of a histopathologic grading system allowed for the comparison of the quantitative extent of osteoarthritic lesions in the different cohorts of mice (129/SvJ background), and the results are shown in Figure 1.

thumbnail image

Figure 1. Histologic scores in ADAM15-deficient and wild-type (WT) mice. Individual scores and median values are indicated for male and female cohorts of 12–14-month-old WT and ADAM15−/− mice. P values were calculated by Wilcoxon's rank sum test.

Download figure to PowerPoint

At ages 3 and 6 months, no significant differences in the osteoarthritis scores were detectable in cohorts of male and female ADAM15−/− mice (n = 10) compared with WT controls (n = 10) (data not shown). The only noteworthy difference was the occurrence of synovial hyperplasia (Figure 2b) in 6 of 10 ADAM15−/− mice (3 females and 3 males) in the 6-month-old cohort, resulting in a total score that was marginally elevated in comparison with the age-matched WT controls. The prevalence of osteoarthritic lesions increased in the cohort of 12–14-month-old ADAM15−/− mice, preferentially affecting the males. Thus, the median osteoarthritis scores in the group of male ADAM15−/− mice (12.9) was 3-fold greater than that in the age- and sex-matched WT controls (4.3) (P < 0.002 by Wilcoxon's rank sum test). Although the median histopathologic score also exhibited a slight increase with age in the 12–14-month-old female ADAM15−/− mice (3.8), the difference from the score in the WT mice (1.0) did not reach the threshold of statistical significance. Figure 1 shows the scoring of individual mice and the median score for each cohort of mice to provide information on interindividual variations.

thumbnail image

Figure 2. Histopathology of the synovial membrane in ADAM15−/− 129/SvJ mice. In contrast to the thin synovial lining (arrowheads) in the normal joint of a 6-month-old male wild-type mouse (toluidine blue staining in a), some of the age-matched ADAM15−/− mice develop a marked synovial hyperplasia (arrowheads in b; hematoxylin and eosin [H&E] staining). At later stages of degenerative joint disease in 12-month-old ADAM15−/− mice, tissue destruction is associated with the liberation of cartilage debris (arrows in c; toluidine blue staining) into the joint space. The detritus becomes integrated into the synovial tissue (arrows in d; H&E staining) and gives rise to inflammatory responses. (Original magnification × 100.)

Download figure to PowerPoint

Typical histomorphologic changes that were detected preferentially in the knee joints of male ADAM15-knockout mice are shown in Figures 2 and 3. Mild signs of osteoarthritic changes were hyperplasia of the synovial membrane (Figure 2b), as shown in comparison with its monolayer appearance in a normal control joint (Figure 2a), and loss of proteoglycans (Figure 3d) from the cartilage matrix (normal morphology is shown in Figures 3a and b). More pronounced lesions consisted of cartilage erosions of various degrees of severity, from superficial changes (Figure 3c) to deep fissures (Figures 3c, e, and f), areas of necrotic cartilage (Figure 3h), and finally, to the complete loss of cartilage, with exposure of the bone cavity to the joint space (Figure 3g). Detritus derived from the cartilage became incorporated into the synovial membrane, which gave rise to subsequent inflammatory responses (detritus synovitis) (Figures 2c and d). Remarkably, this hallmark of irreversible cartilage damage was detectable in 7 of 12 12–14-month-old male ADAM15−/− mice, but in none of the age- and sex-matched WT controls.

thumbnail image

Figure 3. Histopathology of articular cartilage in ADAM15−/− 129/SvJ mice. a and b, Intact cartilage of 13-month-old wild-type mice (hematoxylin and eosin [H&E] staining and toluidine blue staining, respectively). ck, Characteristic histologic features of cartilage degeneration in 12–14-month-old ADAM15−/− mice. c, Fibrillation and fissuring (H&E staining). d, Loss of proteoglycans as detected by toluidine blue staining (arrows indicate cartilage surface). e and f, Progressive cartilage destruction, from deep fibrillation to complete loss of the entire cartilage, as shown in the boxed area in e (H&E staining), an enlarged, toluidine blue–stained parallel section is shown in f. g, Complete loss of the entire cartilage layer (eburnation) (H&E staining). h, Necrotic cartilage, as indicated by sharply demarcated areas of proteoglycan loss (arrows) affecting the entire depth, from the surface to the tidemark (toluidine blue staining). i, Degeneration of the fibrocartilaginous meniscus as indicated by the arrow (H&E staining). j, Chondrocyte clusters (H&E staining). k, Cellular metaplasia of ligaments (toluidine blue staining). l, Osteophyte formation (toluidine blue staining). m = meniscus. (Original magnification × 100 in a and b, ei, k, and l; × 200 in c, d, and j.)

Download figure to PowerPoint

Cartilage degeneration also affected the menisci (Figure 3i), and cells in tendons and ligaments exhibited signs of metaplasia (Figure 3k). In addition, regenerative changes leading to osteophyte formation were detectable in the bone (Figure 3l). Thus, the entire spectrum of osteoarthritic lesions was already present in male ADAM15-knockout mice at ages 12–14 months, a time at which sex-matched WT controls, on average, exhibited only initial signs of cartilage degeneration. Accordingly, it can be concluded that an accelerated development of osteoarthritic lesions in male, but not female, mice occurs as a result of the targeted gene disruption of ADAM15.

Concordant findings of a disease-promoting effect of ADAM15 deficiency were obtained in smaller and somewhat older cohorts of WT and ADAM15−/− mice (ages 18–20 months; n = 5 per group) on a mixed 129/SvJ and C57BL/6 background. The osteoarthritic lesions were more pronounced in the males than in the females, and the scores in the ADAM15−/− mice exceeded those in the respective WT cohort by a factor of ∼3, on average (data not shown). These results indicate that the effect of ADAM15 deficiency on the aggravation of aging-dependent osteoarthritis development is not confined to an inbred 129/SvJ genetic background.

Overexpression of ADAM15 in the chondrocyte cell line T/C28a4.

Stable transfection of T/C28a4 cells with the full-length cDNA of human ADAM15 resulted in strong expression of ADAM15 mRNA, as evidenced by the results of RT-PCR shown in Figure 4A, compared with negative results in vector-transfected T/C28a4 control cells. Overexpression of ADAM15 mRNA was accompanied by a respective increase in protein levels, as demonstrated by the intense staining of an electroblotted protein band of ∼90 kd with a specific anti-ADAM15 monoclonal antibody (Figure 4B).

thumbnail image

Figure 4. ADAM15 expression in stably transfected T/C28a4 chondrocytes. Upon transfection of the human cell line (see Materials and Methods), expression of ADAM15 in cDNA-transfected cells (+) is shown in comparison with vector-transfected controls (−). A, Two examples of specific detection of ADAM15 mRNA by reverse transcription–polymerase chain reaction. B, Western blot analysis of whole cell lysates (20 μg and 50 μg) developed with a specific monoclonal antibody reveals ADAM15 overexpression at the protein level in transfected cells. Development of the same blot with an actin-specific antibody is shown as a control for the protein transfer.

Download figure to PowerPoint

Enhanced adhesion to types II and VI collagen resulting from overexpression of ADAM15 in the chondrocyte cell line T/C28a4.

The hypothesis that ADAM15 might alter cell–matrix interactions was analyzed by comparing ADAM15-overexpressing T/C28a4 chondrocytes with vector-transfected control cells. The results of cell adhesion assays using microtiter wells coated with different extracellular components of the cartilage matrix (e.g., fibronectin and types II and VI collagen) and BSA as a control are shown in Figure 5A.

thumbnail image

Figure 5. A, ADAM15-overexpressing T/C28a4 chondrocytes exhibit reinforced binding to cartilage type II collagen (Coll II) and type VI collagen (Coll VI). Shown are results of representative adhesion experiments with ADAM15-transfected cells (solid bars) and vector-transfected control cells (open bars). Values are the mean and SD of 8 determinations. Although ADAM15 overexpression did not affect cell adhesion to fibronectin (Fn), it clearly increased the ability of T/C28a4 cells to adhere to types II and VI collagen. Adhesion to type II collagen was completely inhibited by 5 mM EDTA. B, The attachment of ADAM15-transfected cells (solid bars) and vector-transfected control cells (open bars) to type II collagen could be completely blocked by incubation of the cells with 300 μM GRGDTP peptide, whereas the control GRADSP peptide did not exhibit any effect (not shown). C, The adhesion of ADAM15-transfected cells (solid bars) and vector-transfected control cells (open bars) to type II collagen could be inhibited by blocking monoclonal antibodies (10 μg/ml for vector-transfected control cells; 50 μg/ml for ADAM15-transfected cells) directed toward the α1, α2, or β1 integrin chains. Equivalent amounts of nonspecific mouse IgG did not influence collagen binding of the cells. Results in B and C are the percentage of adherent cells in the treated relative to the untreated T/C28a4 populations (mean and SD of 4 determinations in a representative experiment). OD = optical density; BSA = bovine serum albumin.

Download figure to PowerPoint

Although ADAM15- and vector-transfected T/C28a4 cells did not bind to BSA, they attached equally well to fibronectin. However, significant differences between ADAM15-overexpressing and control T/C28a4 cells became apparent with respect to their capacity to adhere to types II and VI cartilage collagen (Figure 5A). While the vector-transfected chondrocytic cells exhibited a rather weak ability to bind to types II and VI collagen, the ADAM15 transfection led to a significant gain in collagen attachment. Collagen binding of ADAM15- and vector-transfected T/C28a4 cells was completely inhibited in the presence of 5 mM EDTA in the medium, indicating the dependence of adherence on divalent cations. EDTA sensitivity is a characteristic of cell attachment mediated by integrins; accordingly, low-level expression of the respective α1, α2, and β1 integrin chains involved in collagen binding (19) was detected on T/C28a4 cells without any effect on ADAM15 overexpression (Western blot results not shown).

Further evidence for the conservation of integrin dependency in the reinforced collagen binding of the ADAM15-transfected cells was provided by the inhibitory effect of the RGD peptide (GRGDTP) (Figure 5B). The blocking capacity of the GRGDTP peptide on collagen attachment was quite similar for the vector- and ADAM15-transfected T/C28a4 cells, whereas the control peptide GRADSP did not exhibit any effect (results not shown).

Additional experiments with α1, α2, and β1 integrin–blocking antibodies, the results of which are shown in Figure 5C, revealed the capacity of the respective integrin-blocking antibodies, but not control IgG, to inhibit cell adhesion to type II collagen. Collagen adherence of both vector- and ADAM15-transfected T/C28a4 cells exhibited comparable sensitivity to integrin-blocking antibodies, although somewhat higher antibody concentrations (10 μg/ml versus 50 μg/ml, respectively) were required to achieve the same inhibitory effect in the latter cell population. Although the blocking antibodies were low to moderately dosed compared with earlier investigations (100 μg/ml) (see ref. 19), an effective inhibition of both vector- and ADAM15-transfected cell adhesion was obtained (up to 70%).

Thus, the results of experiments on the adhesion of T/C28a4 cells to extracellular matrix components (in which we tested the dependence of adherence on divalent cations, RGD peptide, and α1, α2, and β1 integrin–blocking antibodies) suggest that ADAM15 overexpression leads to a specific reinforcement of integrin-mediated collagen attachment.

Association of reinforced collagen adhesion in ADAM15-overexpressing T/C28a4 cells with prolonged survival under conditions of serum starvation.

Since cell binding to extracellular matrix components provides the transmission of outside-in signals that are critical for cell survival, we performed an extensive analysis of cell viability under conditions of serum starvation. Following varying periods of serum withdrawal (0–4 hours) and subsequent serum-free incubation in type II collagen–coated microtiter wells for 3 hours, vector-transfected and ADAM15-overexpressing T/C28a4 chondrocytes were analyzed for the occurrence of apoptotic chromatin disintegration by fluorescence-activated cell sorting analysis of PI-stained cells (Figure 6A) and for the development of cell membrane leakiness by LDH release (Figure 6B). In addition, metabolic activity was monitored by the conversion of MTS tetrazolium to formazan as a correlate of cell viability (Figure 6C).

thumbnail image

Figure 6. Prolonged survival of ADAM15-overexpressing T/C28a4 cells under conditions of serum starvation. Depending on the duration of initial serum withdrawal (0–3 hours), vector-transfected (broken lines) and ADAM15-overexpressing (solid lines) T/C28a4 chondrocytes exhibited differences in their capacity to survive on type II collagen–coated microtiter wells. Results for parameters of cell viability from representative experiments are shown as the mean ± SD of quadruplicate determinations. ∗ = P < 0.0001 for vector-transfected versus ADAM15-overexpressing T/C28a4 chondrocytes. A, Flow cytometric assessment of propidium iodide–Triton–stained T/C28a4 cells allowed for the quantification of nuclei with degraded chromatin. Prolongation of serum withdrawal to ≥2 hours before starting incubation on collagen-coated wells caused a rise in apoptotic cell death that was significantly increased in vector-transfected cells. B, Release of lactate dehydrogenase (LDH) activity as a parameter of cell membrane leakiness was significantly increased in vector-transfected cells (by 60%, compared with 30% in ADAM15-transfected cells) when the initial time of serum withdrawal was extended beyond 2 hours. C, The percentage of metabolic activity of vector-transfected and ADAM15-transfected cells, as assessed by the conversion of MTS tetrazolium to a formazan product, is shown relative to the 100% value at the start of serum starvation (0 hours).

Download figure to PowerPoint

As shown in Figures 6A and B, overexpression of ADAM15 in T/C28a4 chondrocytes reduced the percentage of cells with degraded nuclei and the development of cell membrane leakiness in response to serum starvation to one-third to one-half that of vector-transfected control cells. Accordingly, the highly significant enforcement of cell survival on type II collagen was reflected by an increased metabolic rate of the ADAM15-transfected T/C28a4 chondrocytes in the MTS assay (Figure 6C). Thus, our investigations of cell viability suggest that overexpression of ADAM15 in T/C28a4 chondrocytes confers an increased resistance to the proapoptotic effect of serum starvation, which was independently corroborated by the morphology of 4′,6-diamidino-2-phenylindole–stained ADAM15- and vector-transfected cells in the respective experiments (results not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The purpose of the present study was to gain insight into the role of ADAM15 in the development of osteoarthritis by a histopathologic study of the age-dependent development of degenerative cartilage and bone lesions in ADAM15-deficient mice (13) compared with WT controls. While an analysis of ADAM15−/− mice did not uncover any evident defects during development or adult homeostasis, this study did not include a thorough analysis of potential defects in cartilage or bone in adult mice.

A potential role of ADAM15 as a factor that promotes tissue remodeling in degenerative joint disease, most likely associated with a proteolytic effect of its catalytic domain on extracellular matrix molecules, was originally suggested by its strong up-regulation in synovial and cartilage specimens derived from human osteoarthritic lesions (14, 15), although alternative interpretations could not be formally excluded. In this respect, the monitoring of age-dependent morphologic changes in ADAM15−/− mice offered a unique opportunity to evaluate potential relationships between the expression of ADAM15 and the development of osteoarthritic lesions. However, instead of the expected amelioration of joint pathology in ADAM15−/− mice, we observed more severe signs of cartilage and bone degradation than in the respective age- and sex-matched WT controls. Accordingly, the results in the knockout model suggest a protective role of ADAM15 in the maintenance of joint integrity.

Thus, the targeted gene disruption seems to weaken the compensatory mechanisms of articular cartilage that are required to preserve long-term structural integrity and to withstand life-long mechanical stress in the aging knockout mice. While these results shed new light on the role of ADAM15 in the metabolism of aging cartilage, they also raise new intriguing questions concerning the underlying mechanisms of action.

The multidomain structure of ADAM15 harbors several functional domains that could hypothetically be involved in mediating protective effects on cartilage. In an initial attempt to establish an in vitro system for the investigation of potential mechanisms contributing to protective effects, full-length cDNA of human ADAM15 was stably transfected into the human chondrocytic cell line T/C28a4 (16). To investigate whether cell–matrix interactions might be modulated by ADAM15 overexpression, we analyzed the binding properties of ADAM15- and vector-transfected T/C28a4 cells to different extracellular matrix molecules. Interestingly, we noted a specific enhancement of in vitro adhesion to types II and VI collagen as physiologic components of the pericellular collagen meshwork in normal cartilage (20), but we noted no enhancement of in vitro adhesion to fibronectin. The improved collagen adhesion of the ADAM15-overexpressing T/C28a4 cells compared with the vector-transfected controls was accompanied by a prolongation of viability under conditions of serum starvation in vitro, indicating the reinforcement of matrix contact–dependent survival signals by ADAM15.

Chondrocyte attachment to collagen is known to involve the integrin-mediated transduction of signals that are critical for cell survival (21, 22), and we could confirm reported data on the expression of the respective collagen-binding integrins α1β1 and α2β1 (19) in T/C28a4 cells. Accordingly, a modulating interaction with outside-in signaling of integrins had to be considered as a hypothesis to explain the effect of ADAM15 on the reinforcement of cell adhesion. Additional circumstantial evidence supporting this assumption is derived from the known capability of ADAM15 to function as an integrin ligand (αvβ3 and α9β1) (3, 7, 8). However, a direct interaction with collagen-binding integrins is rather unlikely, since it has been shown that the recombinant extracellular domains of ADAM15 do not bind to α1β1 and α2β1 integrins (7). Likewise, a direct cis-acting effect of ADAM15 on the extracellular domains of the fibronectin-binding integrins αvβ3 and α5β1 can also be excluded as an explanation for the modulated cell–matrix attachment. Although both integrins have been shown to interact with the recombinant disintegrin domain of human (but not mouse) ADAM15 (3, 7, 8), fibronectin binding of T/C28a4 cells remained unaffected by overexpression of the human ortholog.

Therefore, the molecular mechanism that is critical for the enhancement of collagen binding in ADAM15-overexpressing chondrocytes remains to be elucidated in future studies. Nevertheless, the present in vitro studies revealed new facets of ADAM15 action that could be relevant as compensatory mechanisms to sustain chondrocyte–matrix adhesion and cell survival in cartilage degeneration. This hypothesis about the homeostatic effects of ADAM15 is entirely consistent with the observed aggravation of cartilage degeneration in aging ADAM15−/− mice compared with WT controls.

Moreover, the consequences of ADAM15 gene disruption on cartilage–matrix interactions suggested by the results of the transfection experiments with the T/C28a4 cells in the present investigation are consistent with more recent findings on accelerated, aging-dependent osteoarthritis development in α1 integrin–deficient mice (23). This knockout model of degenerative joint disease shares a surprisingly high degree of phenotypic similarity with the ADAM15-deficient mice. Whereas both α1 integrin and ADAM15 are physiologically expressed in hypertrophic cartilage of the growth plate, their roles do not seem to be critical for normal skeletal development in the respective knockout mice. ADAM15 and α1 integrin are both up-regulated in remodeling cartilage (14, 23), and the targeted disruption of the respective genes in mice leads to accelerated development of osteoarthritic lesions. For α1 integrin, the phenotypic consequences of the gene knockout occur somewhat earlier (23) and have been associated with the disturbance of its function in promoting adhesion-dependent cell survival and progression through the G1 phase of the cell cycle in response to mitogenic growth factors (24).

Interestingly, the transfection experiments in the present study indicate a reinforcing role of ADAM15 on the same effector mechanism, namely, adhesion-dependent cell survival. Moreover, for the intracellular SH3 ligand homology domains of ADAM15, interactions with adaptor molecules and Src kinase family members have been described (11, 12) that are also essential in the α1 signaling cascades. In light of the data described here, it will now be interesting to determine whether ADAM15-mediated signal transduction might lead to a convergence with the α1 integrin pathway to provide reinforced survival signals for chondrocytes in the remodeling osteoarthritic cartilage.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The authors are grateful to Dr. M. B. Goldring (Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA) for providing the chondrocyte cell line T/C28a4. We thank Prof. Martin Herrmann for his expert advice on the flow cytometric determination of apoptotic nuclei.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES