ADAM15 exerts an antiapoptotic effect on osteoarthritic chondrocytes via up-regulation of the X-linked inhibitor of apoptosis




To investigate the capacity of ADAM15, a disintegrin metalloproteinase that is up-regulated in osteoarthritic (OA) cartilage, to protect chondrocytes against apoptosis induced by growth factor deprivation and genotoxic stress.


Caspase 3/7 activity was determined in primary OA and ADAM15-transfected T/C28a4 chondrocytes upon exposure to the DNA-damaging agent camptothecin or serum withdrawal. Camptothecin-induced cytotoxicity was determined by measuring cellular ATP content. (Anti-)apoptotic proteins were analyzed by immunoblotting, and levels of messenger RNA (mRNA) for X-linked inhibitor of apoptosis (XIAP) were determined using real-time polymerase chain reaction. RNA interference was applied for down-regulation of ADAM15 and XIAP expression. Immunohistochemistry analysis of normal and OA cartilage samples was performed using XIAP- and ADAM15-specific antibodies.


ADAM15-transfected chondrocytes cultured on a collagen matrix displayed significantly reduced caspase 3/7 activity upon serum or intermittent matrix withdrawal, compared with vector-transfected control cells. Apoptosis induction by camptothecin exposure also led to significantly elevated caspase 3/7 activity and reduced cell viability of the vector-transfected compared with ADAM15-transfected chondrocytes. Increased levels of activated caspase 3 and cleaved poly(ADP-ribose) polymerase were detected in the vector controls. XIAP, an inhibitor of activated caspase 3, was significantly up-regulated (∼3-fold) at the protein and mRNA levels in ADAM15-transfected chondrocytes upon camptothecin treatment. Specific down-regulation of either ADAM15 or XIAP in OA chondrocytes led to significant sensitization to camptothecin-induced caspase 3/7 activity. Immunohistochemical analysis revealed low to moderate XIAP expression in normal specimens and markedly increased XIAP staining, colocalizing with ADAM15, in OA cartilage.


ADAM15 conveys antiapoptotic properties to OA chondrocytes that might sustain their potential to better resist the influence of death-inducing stimuli under pathophysiologic conditions.

ADAM15 is a transmembrane-anchored multidomain glycoprotein that belongs to the large family of the disintegrin metalloproteinases (for review, see ref. 1). ADAM15 has been implicated in several conditions involving inflammatory (2) and neoplastic extracellular matrix (ECM) remodeling (3, 4) and neovascularization (5), and its overexpression could be closely associated with the progression of aggressive forms of prostate and breast cancer (for review, see ref. 4). However, ADAM15 up-regulation is not confined to tumorigenesis, but is also detectable in non-neoplastic conditions involving tissue remodeling, such as degenerative joint disease (6–8). It is markedly up-regulated even at early stages of osteoarthritis (OA), whereas its expression remains below the threshold of detectability in healthy cartilage (6). Catalytic activities of ADAM15 have been demonstrated in conjunction with ectodomain shedding (9–11) and type IV collagen cleavage (12), thereby implying at least a theoretical possibility of its contribution to chondrocyte-mediated ECM breakdown in OA cartilage. However, there is currently no direct experimental evidence for such a role of ADAM15 in proteolytic cartilage remodeling in vivo. Moreover, our previous studies, which revealed accelerated development of OA lesions in ADAM15-deficient mice compared with wild-type mice (7), suggest a homeostatic rather than a destructive role of ADAM15 in cartilage remodeling.

More recent data further suggest that ADAM15 can exert additional functions that are independent of its potential proteolytic activity and highly relevant to the cell–matrix interactions and outside-in signaling in OA chondrocytes ( 8). Thus, the pro domain of ADAM15 confers binding to cartilage-specific types II and VI collagens (7, 8). Moreover, the cytoplasmic tail of ADAM15 has a modulatory effect on the phosphorylation of focal adhesion kinase (FAK) during chondrocyte–collagen interaction (8). Since FAK plays a central role in the integration of growth factor receptor and ECM signals by forming a signaling scaffold involved in growth, differentiation, and survival pathways (13, 14), its modulation by up-regulation of ADAM15 in OA chondrocytes is likely to affect further downstream cascades that control cell fate and thus the integrity of the entire cartilage. As the replenishment of ECM molecules is accomplished exclusively by chondrocytes, any significant impairment of their vitality would inevitably result in the loss of functional cartilage matrix. Accordingly, substantial DNA damage (15–18) and other alterations indicative of cellular degeneration are indeed detectable in OA chondrocytes and are likely involved in the activation of effector pathways of programmed cell death. Thus, apoptosis of chondrocytes is a feature of OA cartilage, although its quantitative contribution to matrix degeneration is still a matter of debate (for review, see ref. 19).

Irrespective of these uncertainties, there appears to be an obvious discrepancy between the extent of detectable severe cellular damage and the relatively small fraction of apoptotic cells (<1%) traceable at any time in OA cartilage ( 15). It seems that instead of dying, a considerable proportion of OA chondrocytes remain preserved in a pre- or para-apoptotic phenotype (20) with a discoordinated gene expression pattern, for reasons that are thus far unknown. In the present investigation we obtained clear experimental evidence of the capacity of ADAM15 to exert antiapoptotic effects in stressed human chondrocytes in vitro, thereby suggesting its contribution to the preservation of a preapoptotic chondrocytic phenotype in OA cartilage. Our results with an ADAM15-transfected chondrocyte line as well as with primary human OA chondrocytes revealed a striking role of ADAM15 as a potent suppressor of apoptosis upon induction of cell death by various stimuli. This action of ADAM15 is accomplished by up-regulation of X-linked inhibitor of apoptosis (XIAP).



An ADAM15-specific mouse monoclonal antibody (mAb) recognizing the pro domain ( 8) was obtained from R&D Systems, and mouse anti-XIAP mAb was obtained from BD Biosciences. Rabbit anti–cleaved caspase 3 and cleaved poly(ADP-ribose) polymerase (PARP) were from Cell Signaling Technology. Antitubulin antibody was from Epitomics (Biomol), and anti-pTyr99 antibody was from Santa Cruz Biotechnology. Camptothecin was from Calbiochem. Bovine type II collagen (CII) was from MD BioSciences; bovine type VI collagen was from Biomol.

Cell culture.

T/C-28a4 chondrocytic cells ( 21) (kindly provided by Dr. M. B. Goldring, Hospital for Special Surgery, New York, NY) were grown in Dulbecco's modified Eagle's medium (DMEM) as previously described (8). For all subsequent tests, cells were grown to subconfluence (4 × 106 cells/ 75-cm2 culture flask). Human OA chondrocytes were kept in DMEM/Hams F-12 medium containing 10% fetal calf serum (FCS).

Isolation of human OA chondrocytes.

OA chondrocytes were isolated, as previously described ( 22), from hip or knee joints at the time of endoprosthetic joint replacement for severe OA. All patients had provided written informed consent, and approval was obtained from the Ethics Committee of University Hospital Frankfurt am Main.

Generation of permanent cell lines transfected with ADAM15.

The transfection and selection procedures were performed as previously described ( 8). Cells transfected with empty pexchange-1 vector served as control.

Determination of caspase 3/7 activity.

MaxiSorp plates (96-well; Nunc) were coated with CII (5 μg/ml) and CVI (2.5 μg/ml) as previously described ( 7). ADAM15-transfected T/C-28a4 cells (1 × 104) and OA chondrocytes (2 × 104) were grown for 24 hours in DMEM containing 10% FCS. After removal of the medium, cells were treated with DMEM alone for up to 36 hours or with 20 μM camptothecin for up to 8 hours at 37°C. Caspase 3/7 activity was measured with the CaspaseGlo 3/7 Assay (Promega), using the Mithras LB 940 plate reader (Berthold Biotechnologies).

Matrix withdrawal.

ADAM15- and vector-transfected T/C-28a4 chondrocytes were harvested by gentle trypsinization, and 1 × 104 cells seeded into a 96-well tissue culture plate that had been blocked with 1% bovine serum albumin for 1 hour at 37°C in order to prevent cells from attaching to the well. After 1–4 hours, cells were transferred into CII-coated wells and allowed to adhere for 1 hour at 37°C, and caspase 3/7 activity was measured using the CaspaseGlo Assay.

Cell viability assay.

ADAM15- and vector-transfected T/C-28a4 chondrocytes (1 × 104) were grown in a 96-well plate for 24 hours at 37°C and treated with camptothecin in various concentrations (0–100 μM) for 18 hours. The number of viable cells was determined by quantification of the ATP present in the well using the CellTiterGlo Luminescent Cell Viability Assay (Promega), in which the amount of ATP is directly proportional to the number of viable cells.

Preparation of cell lysates and Western blotting.

Cell lysates were prepared and immunoblotted as previously described ( 8). Exposed x-ray films were scanned, and signal density was determined using ImageJ ( Results of the densitometric analysis were used for calculation of changes in protein levels.

Silencing of XIAP and ADAM15 in cell culture cells by RNA interference.

T/C-28a4 chondrocytes (1 × 104 cells/ well in a 96-well plate; 1.5 × 105 cells/well in a 24-well plate) and human OA chondrocytes from 7 different donors (3 × 104 cells/well in a 96-well plate; 3 × 105 cells/well in a 24-well plate) were seeded into a culture plate and grown for 24 hours. Cells were treated with 5 nM Silencer Select Predesigned and Validated small interfering RNA (siRNA) for XIAP (siRNA XIAP I [5′-GCAGAUUUAUCAACGGCUUtt-3′] and siRNA XIAP II [5′-GGAUAUACUCAGUUAACAAtt-3′]) and ADAM15 (siRNA ADAM I [5′-GAUCUACUCUGGGAGACAAtt-3′] and siRNA ADAM II [5′-CAUUAUUUCGCGAAUCCAAtt-3′]) (Ambion/Applied Biosystems) for 48 hours, using Saint Red according to the instructions of the supplier (Synvolux). The nonsilencing siRNA control #1 (Ambion) was used as a negative control.

Real-time quantitative reverse transcription–polymerase chain reaction (PCR).

Total RNA was isolated using RNeasy Mini spin columns combined with on-column DNase treatment (Qiagen). RNA (2 μg) was reverse-transcribed using Moloney murine leukemia virus reverse transcriptase according to the protocol recommended by the supplier (Promega). PCR amplification was performed on an MJ Research/Bio-Rad cycler using QuantiFast SYBR Green Mastermix according to the protocol of the manufacturer (Qiagen), with the following intron- spanning primer pairs: for XIAP, 5′-ACGGATCTTTACTTTTGGGAC-3′ and 5′-CACCCTGGATACCATTTAGCAT-3′, and for GAPDH, 5′-GAAGGTGAAGGTCGGAGTC-3′ and 5′-GAAGATGGTGATGGGATTTC-3′. Amplification was performed with initial denaturation at 95°C for 5 minutes followed by 40 cycles of 95°C for 10 seconds and 60°C for 30 seconds. Quadruplicates of each sample were measured for calculation of the mean and standard deviation, and each sample was normalized for GAPDH messenger RNA (mRNA) content. Relative quantification of XIAP mRNA levels in camptothecin-treated ADAM15- and vector control–transfected cells was performed using the comparative threshold cycle (ΔΔCt) method (Applied Biosystems User Bulletin).

Conventional immunohistochemistry.

Normal cartilage specimens (n = 4) were obtained from knee joints within 24 hours postmortem, and OA cartilage (n = 8) was obtained at the time of joint replacement surgery of the hip or knee. All patients met American College of Rheumatology criteria for the classification of OA ( 23, 24). Serial sections of normal and OA cartilage (Mankin grades 2–7 [25]) were stained with mouse anti-ADAM15 antibodies (1:100), rabbit anti-XIAP (1:1,000; Abcam), and a streptavidin–biotin complex (BioGenex), as previously described (22). Laser scanning microscopy was performed upon double-staining with the above-mentioned specific and fluorescence-labeled secondary antibodies.

Statistical analysis.

Data are presented as the mean ± SD of quadruplicates from at least 3 independently performed assays. Statistical significance was determined using Student's unpaired t-test. P values less than 0.05 were considered significant.


ADAM15 expression is associated with reduced caspase 3/7 activity upon induction of apoptosis by matrix or serum withdrawal.

We have shown earlier that ADAM15-overexpressing chondrocytic T/C-28a4 cells (ADAM15-transfected cells) have a significantly enhanced ability to attach to cartilage collagens (CII and CVI) ( 7). Furthermore, compared with vector-transfected control cells, ADAM15-transfected cells exhibited significantly increased viability after removal of matrix contact intermittently for several hours, presumably due to reinforced survival signals perceived in conjunction with the increased capacity of ADAM15-transfected cells to adhere to cartilage collagens (7).

To analyze whether ADAM15 expression improves cell viability under these conditions via an effect on caspase-dependent apoptosis, trypsinized cells that had been kept in suspension on 1% bovine serum albumin for 1–4 hours to prevent adherence were then allowed to attach to CII, and the activity of caspase 3/7, an apoptotic agent, was determined. Significantly reduced caspase 3/7 activity (mean ± SD 2.6 ± 0.4–fold reduction) was observed in ADAM15-transfected cells as compared with vector-transfected controls (Figure 1A). Subsequently, we investigated whether the increased resistance to apoptosis induction conferred by ADAM15 was also detectable with the use of other cell death inducers, e.g., growth factor deprivation. For that purpose, ADAM15- and vector-transfected cells were grown for 24 hours on CII and CVI, and apoptosis was induced by serum withdrawal for up to 36 hours. Caspase 3/7 activity was significantly reduced (2.2 ± 0.7–fold) by serum withdrawal in ADAM15 cells compared with vector-transfected cells, indicating that the antiapoptotic effect of ADAM15 is not restricted to apoptosis induced by loss of ECM contact.

Figure 1.

Association of ADAM15 expression in T/C-28a4 chondrocytes with reduced caspase 3/7 activity after induction of apoptosis by matrix or growth factor deprivation. A, ADAM15- or vector-transfected cells were initially kept in suspension for 1–4 hours and subsequently plated onto type II collagen (CII) for 1 hour to allow adherence, and caspase 3/7 activity was measured. B, ADAM15- or vector-transfected cells were grown on CII (upper panel) or CVI (lower panel) for 24 hours, and induction of caspase 3/7 activity by serum withdrawal was monitored for up to 36 hours. Values are the mean ± SD. ∗ = P < 0.009 versus vector control.

ADAM15-transfected cells have reduced sensitivity to camptothecin-induced caspase 3/7 activity.

Camptothecin is a chemotherapeutic agent that acts as an inhibitor of DNA topoisomerase I and causes damage to DNA via induction of strand breaks ( 26). Thus, we used chondrocyte treatment with camptothecin as an in vitro model of DNA damage–induced caspase 3/7 activity. In this model, caspase 3/7 activity remained significantly lower (by a mean ± SD of 2.8 ± 0.7–fold) in ADAM15-transfected cells compared with vector-transfected controls (Figure 2A), thereby confirming that the protective effect of ADAM15 also applies to conditions of genotoxic stress. In accordance with this, cell viability after exposure to increasing amounts of camptothecin (0–100 μM) for 18 hours remained significantly higher in ADAM15-transfected cells (>80% of initial value at maximum concentration) compared with vector-transfected cells (<20%) (Figure 2B). Figure 2C depicts the altered cell morphology induced by treatment with 20 μM camptothecin for 8 hours: whereas the majority of the vector-transfected cells were already completely detached and floating in the medium, the ADAM15-cells remained attached to the bottom, and only a minority of cells had started to round.

Figure 2.

Reduced apoptosis of ADAM15-expressing T/C-28a4 chondrocytes after camptothecin treatment. A, Caspase 3/7 activity after exposure to 20 μM camptothecin for 0–8 hours. B, ATP content (which correlates with cell viability) after exposure to 0–100 μM camptothecin for 18 hours. Values in A and B are the mean ± SD. ∗ = P < 0.05; ∗∗ = P < 0.0005, versus vector control. C, Photomicrographs of ADAM15- and vector- transfected cells incubated with 20 μM camptothecin for 8 hours. The majority of vector-transfected cells are already detached and floating, whereas the ADAM15-transfected cells have started to round (arrows), with only a few cells detached (magnification × 10).

ADAM15 expression is associated with up-regulation of XIAP in response to camptothecin treatment.

ADAM15- and vector-transfected cells were incubated with 20 μM camptothecin for 0–8 hours, lysed, and analyzed by Western blotting for the presence of various pro- and antiapoptotic proteins. Cleaved caspase 3 was markedly induced in vector-transfected cells compared with ADAM15-transfected cells as was cleaved PARP, the substrate protein for caspase 3, (Figure 3A) thereby providing further evidence of the above-described antiapoptotic effects of ADAM15. XIAP, known as an inhibitor of activated caspase 3 and a contributor to the prosurvival effect of cartilage oligomeric protein on chondrocytes ( 27), was markedly up-regulated (∼2.8 fold, averaged over all measured time points in 5 independently performed experiments) in ADAM15-transfected cells compared with the respective vector control (Figure 3A), with tubulin used as a loading control. The ADAM15 expression level remained unchanged throughout all time points analyzed, and no proteolytic cleavage of the pro domain, a prerequisite for the catalytic activation of metalloproteinases, could be detected.

Figure 3.

ADAM15-expressing cells counteract camptothecin-induced apoptosis by up-regulating X-linked inhibitor of apoptosis (XIAP). A, Vector-transfected cells (−) and ADAM15-transfected cells (+) were incubated with 20 μM camptothecin for 0–8 hours, lysed, and subjected to immunoblotting with antibodies against proteins involved in the apoptotic pathway. Expression of activated caspase 3 and concomitantly cleaved poly(ADP-ribose) polymerase (PARP) was increased in the vector controls, whereas the antiapoptotic protein XIAP was up-regulated in ADAM15-expressing cells relative to controls. Tubulin was used as loading control, and ADAM15 expression was determined using anti-ADAM15 antibodies. Optical density of the XIAP bands was analyzed using tubulin for normalization, and the fold change relative to controls was calculated from 5 independently performed experiments. Values are the mean and SD. B, XIAP mRNA in ADAM15- and vector-transfected cells upon camptothecin treatment was assessed by real-time reverse transcription–polymerase chain reaction. The ΔCt value was normalized to GAPDH, and the fold change of XIAP gene expression in ADAM15-tranfected cells relative to vector-transfected cells was calculated using the ΔΔCt method; ΔCt values are the mean ± SD.

The respective XIAP mRNA levels determined by real-time PCR revealed an overall significant increase (2.2–3.5-fold) in the ADAM15-transfected compared with vector-transfected cells, after incubation with camptothecin for 0–8 hours. However, this relative increase was due mainly to reduced XIAP mRNA expression in the vector control cells, indicated by an increased number of amplification cycles (shown as ΔCt normalized to GAPDH) (Figure 3B), whereas the mRNA levels remained constant in the ADAM15-transfected cells. Accordingly, our results suggest that posttranscriptional control mechanisms ( 28, 29) might also contribute to the elevation of XIAP at the protein level detected in ADAM15-transfected cells.

Down-regulation of XIAP by silencing of RNA results in increased camptothecin-induced caspase 3/7 activity in ADAM15-transfected cells.

To elucidate the role of XIAP in ADAM15-mediated antiapoptotic pathways, ADAM15-transfected T/C-28a4 chondrocytes were transfected with 2 different siRNA for XIAP: XIAP I and XIAP II. Treatment with either siRNA at a concentration of 5 nM resulted in efficient down-regulation of XIAP protein expression (by ∼95% after 48 hours), as evidenced by comparison with the whole cell lysate of untreated ADAM15 control cells (Figure 4A). No effect was observed with the nonsilencing control. Subsequently, the effect of XIAP silencing on induction of caspase 3/7 activity by exposure to 20 μM camptothecin for 0–8 hours was determined. Transient knockdown of XIAP by either XIAP I or XIAP II led to significantly increased caspase 3 activity (∼2-fold compared with nonsilencing control).

Figure 4.

Knockdown of X-linked inhibitor of apoptosis (XIAP) in ADAM15-expressing chondrocytes leads to enhanced camptothecin-induced caspase 3/7 activity. A, Immunoblotting with anti-XIAP antibodies showed that XIAP expression was efficiently down-regulated by XIAP-specific small interfering RNA (siRNA) XIAP I and XIAP II compared with nonsilencing control (Neg). Down-regulation of XIAP by siRNA XIAP I and XIAP II led to increased caspase activity compared with the levels observed with nonsilencing control, after apoptosis induction by camptothecin. Values are the mean and SD. ∗ = 0.05; ∗∗ = 0.0002, versus nonsilencing control. Lys = whole cell lysate of ADAM15-transfected cells. B, Analogous to the findings with XIAP siRNA, ADAM15 gene expression was silenced in ADAM15-transfected T/C-28a4 chondrocytes by treatment with ADAM15 siRNA ADAM I and ADAM II compared with nonsilencing control, as shown by immunoblotting using ADAM15 antibodies. C, Camptothecin treatment led to significantly increased caspase activity upon ADAM15 silencing by siRNA ADAM I and ADAM II compared with nonsilencing control. Values are the mean and SD. ∗ = 0.02; ∗∗ = 0.0002, versus nonsilencing control. D, Concomitantly decreased expression of XIAP was observed only in ADAM15-silenced cells after camptothecin exposure.

Analogous knockdown experiments applying siRNA for ADAM15 (ADAM I and ADAM II) and a nonsilencing control were performed in ADAM15-transfected T/C-28a4 chondrocytes. ADAM15 protein expression was reduced by 90% with the ADAM I probe and by ∼80% with the ADAM II probe, relative to the nonsilencing control (Figure 4B). As was observed with silencing of XIAP, siRNA-mediated knockdown of ADAM15 also resulted in a significant (2–3-fold) rise in caspase 3 activity upon camptothecin challenge (Figure 4C), thereby confirming the above-described evidence of a protective role of ADAM15 in genotoxic cell stress. Furthermore, down-regulation of ADAM15 by siRNA ADAM I and ADAM II resulted in a 1.8–2.5-fold decrease in XIAP protein expression compared with that observed with the nonsilencing control (Figure 4D).

XIAP and ADAM15 expression in human normal and OA cartilage.

In moderately damaged OA cartilage, which shows a loss of the superficial zone, clefts, and cell clustering (Mankin grades 2–7), strongly enhanced expression of XIAP was observed in all samples throughout all cartilage layers from the superficial zone, middle zone, and upper deep zone (Figure 5e). No differences in XIAP expression were found between single cells and cell clusters (Figure 5f). In normal cartilage, however, overall XIAP protein expression was reduced and was detectable only in the superficial zone and the upper region of the middle zone (Figure 5d). Analogous with these findings, ADAM15 expression in OA cartilage also was found in all cartilage zones (Figure 5b), whereas in normal cartilage no expression was detectable, and only very sparse positive chondrocytes were found in the superficial zone. Strong expression was seen both in cell clusters and in single cells in the OA samples (Figure 5c). Laser scanning microscopy revealed coexpression of ADAM15 and XIAP in the same chondrocytes (Figures 5g–i).

Figure 5.

Immunostaining of ADAM15 and X-linked inhibitor of apoptosis (XIAP) in normal and osteoarthritic (OA) cartilage. a–f, Immunohistochemistry results. Only very low levels of ADAM15 expression were seen in normal cartilage (a) (arrow), but strong up-regulation was observed in all zones of OA cartilage (b). Expression of XIAP was reduced and was observed only in the superficial zone and the upper layer of the middle zone in normal cartilage (d), whereas elevated expression of XIAP was seen throughout all layers of OA cartilage (e). Higher-magnification views of ADAM15 (c) and XIAP (f) protein expression in chondrocyte clusters and single cells in OA cartilage are also shown (magnification × 10 in a, b, d, and e; × 40 in c and f). g–i, Confocal laser microscopy of an OA chondrocyte cluster stained with ADAM15 (g) and XIAP (h) and an overlay view (i), showing coexpression of both molecules, but no subcellular colocalization.

Silencing of ADAM15 or XIAP in OA chondrocytes leads to enhanced camptothecin-induced caspase 3/7 activity.

To analyze whether OA chondrocytes, in which ADAM15 is up-regulated in vivo during the disease course ( 6) (Figures 5a and b), can be sensitized to camptothecin-induced caspase 3 activity in vitro by ADAM15 knockdown, silencing studies were performed on 7 different donors using 2 different siRNA (ADAM I and ADAM II) as well as a nonsilencing control probe. Both siRNA repressed ADAM15 protein expression markedly (by ∼85%) relative to the nonsilencing siRNA control, as shown by immunoblotting with an ADAM15-specific antibody (Figure 6A). The transient down-regulation of ADAM15 expression by both siRNA clearly sensitized the OA chondrocytes to camptothecin-induced apoptosis, as evidenced by the highly significant 2.2-fold (ADAM I)–2.7-fold (ADAM II) increase of caspase 3/7 activity in comparison with that observed with the nonsilencing siRNA control (Figure 6B). Additionally, the down-regulation of ADAM15 resulted in a reduced XIAP protein level (decrease ∼2-fold compared with the nonsilencing control) (Figure 6C).

Figure 6.

Knockdown of ADAM15 in osteoarthritis (OA) chondrocytes enhances camptothecin-induced caspase 3/7 activity. A, OA chondrocytes were treated with siRNA ADAM I, ADAM II, or nonsilencing control for 48 hours. Immunoblotting with ADAM15-specific antibodies showed that ADAM15 expression was down-regulated by ADAM15-specific siRNA compared with nonsilencing control. B, Down-regulation of ADAM15 by siRNA ADAM I and ADAM II led to increased caspase 3/7 activity compared with the levels observed with nonsilencing control, after apoptosis induction with camptothecin. Values are the mean and SD. ∗ = P < 0.05; ∗∗ = P < 0.005; ∗∗∗ = P < 0.0005, versus nonsilencing control. C, This was accompanied by lower XIAP expression in ADAM15-silenced chondrocytes only, as shown by Western blotting. D, Analogously, XIAP gene expression was silenced in OA chondrocytes by siRNA XIAP I and XIAP II compared with nonsilencing control, as shown by Western blotting. Camptothecin treatment resulted in significantly increased caspase activity upon XIAP silencing by XIAP I and XIAP II compared with nonsilencing control. ∗∗ = P < 0.003; ∗∗∗ = P < 0.0008, versus nonsilencing control. See Figure 4 for other definitions.

An analogous knockdown of XIAP expression was performed, and the degree of XIAP silencing by the siRNA XIAP I and XIAP II in OA chondrocytes was assessed by immunoblotting of cell lysates using an anti-XIAP–specific antibody. XIAP is synthesized by OA chondrocytes (Figures 5e and f and 6). An 85% reduction of XIAP protein expression was achieved with both XIAP siRNA (Figure 6D). The subsequent induction of apoptosis by camptothecin led to a significant 2.3–2.5 fold increase in caspase activity in chondrocytes transfected with either of the XIAP siRNA compared with chondrocytes that received the nonsilencing probe (Figure 6D).


Results of the present study provide substantial evidence of a novel function of ADAM15 as a protein with antiapoptotic properties in OA cartilage. OA is a degenerative joint disease characterized by the progressive loss of intact cartilage matrix and impairment of the phenotypic stability and cellular integrity of chondrocytes, which, in combination, lead to failure to maintain tissue homeostasis (for review, see ref. 30). Thus, the chondrocytes in OA cartilage are exposed to enhanced stress, e.g., by mechanical forces or by cytokine exposure that is associated with increased production of reactive oxygen (31, 32) which, in turn, has been incriminated in accelerated aging processes due to the accumulation of oxidatively damaged molecules (33).

In this regard, genomic DNA is an extraordinarily delicate target of oxidative damage that can cause significant changes in the transcription pattern of genes required for normal chondrocyte function and viability. Many studies show that OA chondrocytes exhibit significant DNA damage ( 15–18), which should normally render these cells susceptible to programmed cell death. Accordingly, experimental evidence also clearly suggests that apoptosis occurs in OA cartilage at an increased frequency compared with that in normal cartilage (for review, see ref. 19). However, considering the extent of detectable DNA damage in OA chondrocytes, the occurrence of apoptosis rather appears to be a relatively rare event in OA cartilage, affecting <1% of the total cell population (15), whereas a considerable proportion of the chondrocytes seem to survive in a pre- or para-apoptotic phenotype for reasons that are thus far unknown (20). In this respect, the present results revealing previously unrecognized antiapoptotic properties of ADAM15 suggest a mechanism that could contribute to the rescue of chondrocytes from genotoxic-stress–induced apoptosis. Thus, we have shown that ADAM15-transfected chondrocytes exhibit significantly lower caspase 3/7 activity and subsequent PARP cleavage compared with vector-transfected cells, following induction of apoptosis by camptothecin exposure as a model of genotoxic stress or other cell death stimuli, such as growth factor deprivation or loss of matrix contact.

Caspase 3 is crucial in virtually every model of apoptosis ( 34) and belongs to a subfamily of effector caspases, activated far downstream of apoptosis initiation. Among the natural substrates of caspase 3 are many proteins involved in cell maintenance and/or repair (35), and its selective pharmacologic inhibition has been shown to inhibit chondrocyte apoptosis and to maintain cell functionality (36). It is therefore intriguing that our results suggest that the up-regulation of ADAM15 in OA chondrocytes (2) is capable of activating an antiapoptotic pathway that finally interferes at the level of effector caspases. Our study provides unequivocal evidence of the ADAM15-dependent suppression of DNA damage–induced caspase 3 activity by up-regulation of XIAP, which belongs to the family of inhibitor of apoptosis proteins (37).

The pathways underlying ADAM15-induced XIAP up-regulation upon genotoxic stress seem to be rather complex, involving transcriptional and posttranscriptional mechanisms. The increase of XIAP protein in ADAM15-transfected chondrocytes is not mirrored by an equivalent rise in mRNA levels, in contrast to findings in control chondrocytes, which exhibit a concomitant decrease of XIAP protein and mRNA levels under identical conditions. Accordingly, it appears that ADAM15-mediated XIAP up-regulation is at least partially due to stabilizing protein interactions and/or interference with XIAP mRNA decay, representing already described paradigmatic posttranscriptional control mechanisms of XIAP expression ( 28, 29). In this context, it is of note that silencing of XIAP in ADAM15-expressing OA chondrocytes did not affect ADAM15 expression (data not shown), indicating that a potential reciprocal regulatory connection between XIAP and ADAM15 is rather unlikely.

XIAP is probably the most potent member of this family in terms of its unique ability among the mammalian IAPs to inhibit activated caspases and thereby to suppress both the mitochondrial and the death receptor–mediated pathways of apoptosis ( 37). XIAP directly inhibits both initiator caspase 9 and effector caspases 3 and 7. Accordingly, our investigations using RNA interference technology show that the specific knockdown of XIAP protein expression in ADAM15-transfected T/C-28a4 chondrocytes as well as in primary OA chondrocytes leads to an increase in caspase 3 activity upon exposure to camptothecin as an apoptosis-inducing stimulus. Corresponding results have been obtained in studies aimed at the development of therapeutic strategies to sensitize neoplastic cells to apoptosis induction, which have shown that XIAP-specific RNA interference renders breast cancer and chondrosarcoma cells more susceptible to chemotherapeutic agents (38, 39) and enhances radiosensitivity of grade II chondrosarcoma (40).

In light of the pivotal role of XIAP in apoptosis regulation, the literature on its expression and cellular distribution in cartilage is surprisingly sparse. To our knowledge, the present immunohistochemical investigation is the first to demonstrate intense and widely distributed cellular XIAP staining in all layers of OA cartilage, in contrast to its moderate expression in chondrocytes of normal cartilage, with limited detectability only in the superficial zone and the uppermost parts of the middle zone. Not only is XIAP expression in OA cartilage markedly up-regulated and widespread, but it seems to be coexpressed with ADAM15 in OA. Accordingly, the ex vivo detectable up-regulation of XIAP in ADAM15-positive chondrocytes might at least partially reflect the newly elucidated functional role of ADAM15 in an in vitro model of genotoxic stress–induced caspase 3 activation. Clearly such an interpretation remains somewhat restricted by the inherent limited transferability of in vitro studies to in vivo conditions. However, since the up-regulation of ADAM15 is already detectable in early OA ( 2), this mechanism might contribute to the escape of the stressed OA chondrocytes from apoptosis and could at least partially explain why aging ADAM15-deficient mice exhibit accelerated OA development compared with wild-type controls (7).

The precise sequence of events responsible for the up-regulation of XIAP in the stressed chondrocytes remains to be elucidated in future studies. However, a few possibilities can be envisaged. It might be hypothesized that ADAM15-mediated shedding of growth factor receptor ligands plays a role, analogous to scenarios that have been proposed for its role in tumorigenesis; e.g., E-cadherin shedding with subsequent transactivation of the ErbB receptor by soluble cadherin ( 11). However, such a hypothetical mechanism is critically dependent on the assumption that ADAM15 is proteolytically active, and is rather unlikely to explain our results in chondrocytes. The protease domain of ADAM15 is kept inactive by the pro domain, which needs to be cleaved by furin-type proconvertases for activation, as has been demonstrated for the homologous ADAMs 10 and 17 (41, 42). In our experiments, however, no convertase activity became detectable, since transfection of the T/C-28a4 chondrocytes with full-length ADAM15 always produced the intact molecule, without the appearance of any fragment corresponding to a furin-cleaved ADAM15 in Western blot analysis.

Thus, we suggest another scenario as potentially relevant to the antiapoptotic properties of ADAM15, relating to its recently elucidated function as a proadhesive molecule. ADAM15 binds via its pro domain to cartilage-specific collagens, thereby leading to reinforcement of integrin-mediated cartilage–matrix adhesion ( 7, 8). Moreover, during chondrocyte–collagen interaction ADAM15 can modulate the phosphorylation of FAK (8), which is a central signaling scaffold that integrates growth factor receptor and ECM signals into various connected downstream cascades, including crucial survival pathways. It remains an intriguing possibility that ADAM15 might exert its protective role for the stressed chondrocyte in OA cartilage by tuning outside- in signals via the modulation of FAK, which is also part of the mechanotransduction pathway (43).

However, the protective effect provided by ADAM15 to stressed OA chondrocytes could be a double-edged sword, since the advantage of rescue from cell death is likely to be at least partially counterbalanced by the preservation of an altered chondrocytic phenotype resulting from accumulated genotoxic damage that would otherwise inevitably be censored by programmed cell death. Accordingly, the ADAM15-mediated chondroprotection could help to delay the time frame until chondrocytes fail to maintain homeostasis in OA cartilage. However, the capacity of this compensation mechanism to help the postmitotic chondrocytes fulfill their function in the replenishment of ECM molecules for periods of time as prolonged as decades is presumably limited, thereby explaining the low but detectable rate of chondrocyte apoptosis in OA cartilage despite ADAM15 up-regulation. Thus, the novel role of ADAM15 as a potent suppressor of apoptosis, accomplished by up-regulating XIAP in response to different death-inducing stimuli, might be of prime relevance to the understanding of mechanisms leading to the altered preapoptotic phenotype of chondrocytes in OA cartilage. This warrants further elucidation in future studies.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Burkhardt had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Böhm, Burkhardt.

Acquisition of data. Böhm, Hess, Krause, Schirner, Ewald, Aigner.

Analysis and interpretation of data. Böhm, Aigner, Burkhardt.