Cystatin 10, a Novel Chondrocyte-specific Protein, May Promote the Last Steps of the Chondrocyte Differentiation Pathway*

This study attempts to characterize cystatin 10 (Cst10), which we recently identified as a novel protein implicated in endochondral ossification. Expression of Cst10 was specific to cartilage, localized in the cytosol of prehypertrophic and hypertrophic chondrocytes of the mouse growth plate. In the mouse chondrogenic cell line ATDC5, Cst10 expression preceded type X collagen expression and increased in synchrony with maturation. When we compared ATDC5 cells transfected with Cst10 cDNA with cells transfected with a mock vector, hypertrophic maturation and mineralization of chondrocytes were promoted by Cst10 gene overexpression in that type X collagen expression was observed earlier, and alizarin red staining was stronger. On the other hand, type II collagen expression and Alcian blue staining, both of which are markers of the early stage of chondrocyte differentiation, were similar in both cells. Overexpression of the Cst10 gene also caused fragmentation of nuclei, the appearance of annexin V, a change in the mitochondrial membrane potential, and activation of caspases. These results strongly suggest that Cst10 may play an important role in the last steps of the chondrocyte differentiation pathway as an inducer of maturation, followed by apoptosis of chondrocytes.

Endochondral ossification is an essential process for skeletal development, bone growth, and fracture healing and is implicated in pathological conditions such as osteoarthritis and ectopic ossification. During this process, chondrocytes first proliferate and then progressively differentiate into mature hypertrophic chondrocytes. Once fully matured, these hypertrophic cells mineralize the surrounding matrix and undergo apoptosis. This is followed by a local recruitment of blood vessels and osteoclasts, leading to progressive replacement of cartilage by bone. Thus, in this process of endochondral bone formation, proliferation, maturation, mineralization, and apoptosis of chondrocytes must be properly coordinated. To eluci-date the molecular mechanisms of endochondral ossification, we have been attempting to isolate novel genes implicated in this process (1)(2)(3)(4). For this study, we took advantage of the naturally occurring mouse mutant ttw (tiptoe walking), which exhibits ectopic ossification in various soft tissues such as tendons, cartilage, and ligaments of the extremities and the spine (5). We previously found that ttw is caused by a nonsense mutation of the nucleotide pyrophosphatase gene encoding an ectoenzyme generating phosphate and pyrophosphate (4). Based on the fact that a high phosphate diet accelerates ectopic ossification of ttw, using a differential display method, we identified nine mouse genes whose expression is regulated by a high phosphate diet (1). Six of the nine genes were novel; and among them, we isolated one, termed cystatin 10 (Cst10), 1 that is up-regulated by a high phosphate diet and is expressed exclusively in cartilage, suggesting its specific role in endochondral bone formation.
In this study, we first characterized temporal and spatial expression patterns of Cst10, a novel member of the cystatin superfamily. The cystatin superfamily is known to inhibit the papain-like cysteine proteinases cathepsins B, H, and L by the formation of a tight reversible complex (6). These cysteine proteinases are thought to be associated with terminal degradation of proteins in lysosomes, so the cystatin superfamily is ubiquitously expressed and exhibits various biological functions (7). However, the present study reveals that Cst10 is expressed exclusively in mature chondrocytes. In addition, overexpression of the Cst10 gene accelerates hypertrophic maturation, mineralization, and apoptosis of chondrocytes. These data suggest a crucial and specific role of Cst10 in the later stage of endochondral ossification, implying a physiological role distinct from those other members of the cystatin superfamily.

EXPERIMENTAL PROCEDURES
Determination of the Genomic Structure of the Mouse Cst10 Gene-Bacterial artificial chromosome (BAC) clones containing the mouse Cst10 gene were isolated using a BAC PCR screening system (Genome Systems, St. Louis, MO) according to the manufacturer's protocol. The set of primers used for screening was Cst10/BAC/F (5Ј-TCCTGAG-GATATATGTCAGGC-3Ј) and Cst10/BAC/R (5Ј-ATCTCTGTCTGAG-GAAAGGAG-3Ј). To determine the size of introns of the Cst10 gene, interexon PCRs were carried out with primers designed according to the cDNA sequence we determined in this study. The BAC clones and PCR products were sequenced directly, and the exon-intron junctions were determined by comparing the genomic sequences obtained with the corresponding cDNA sequences.
Chromosomal Localization-To determine the chromosomal localization of the mouse Cst10 gene, we performed fluorescence in situ hybridization as described previously (8). A BAC clone containing the mouse Cst10 gene was labeled and hybridized to the mouse metaphase chromosome. Hybridization signals were rendered visible with fluorescein isothiocyanate (FITC)-avidin. Precise assignments of the signals were determined by visualization of the replicated G-bands.
Animals-The ddY and ttw mice were purchased from Shizuoka Laboratories Animal Center (Shizuoka, Japan) and the Central Institute for Experimental Animals (Kanagawa, Japan), respectively. All animal experiments were performed according to the guidelines of the International Association for the Study of Pain (9).
Cell Culture-Primary mesenchymal cells (osteoblasts, chondrocytes, and fibroblasts) were extracted from the calvariae, costal cartilage, and skin, respectively, of neonatal ddY mice as described previously (10,11). Cells were cultured in ␣-modified minimal essential medium (Invitrogen) containing 5% fetal bovine serum (Invitrogen) at 37°C. Mouse chondrogenic ATDC5 cells were obtained from the RIKEN Cell Bank (Saitama, Japan). The cells were cultured in medium consisting of a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 medium (Invitrogen) containing 5% fetal bovine serum, 10 g/ml human transferrin (Roche Applied Science, Mannheim, Germany), and 3 ϫ 10 Ϫ8 M sodium selenite (Sigma) as described previously (12). The inoculum density of the cells was 4 ϫ 10 4 cells/well in 12multiwell plates (Corning Inc., New York). For induction of chondrogenesis, the cells were cultured in medium supplemented with 10 g/ml bovine insulin (Wako Pure Chemicals, Osaka, Japan). Cells were maintained at 37°C in a humidified atmosphere of 5% CO 2 in air. The medium was replaced every other day.
Expression of the Cst10 Transcript-Expression of the Cst10 mRNA was examined by semiquantitative reverse transcription (RT)-PCR, followed by Southern blotting using auricular cartilage from ttw mice, mesenchymal cells from neonatal ddY mice, and cultured mouse chondrogenic ATDC5 cells. For the experiments with ttw mice, the mice were divided into two groups according to the content of phosphate in the diet, i.e. high (0.87%) and low (0%) phosphate groups after weaning at 3 weeks of age. The animals were killed 0, 1, 3, 5, 7, 10, and 14 days after the start of the diet, and the auricular cartilage was resected en bloc. Total RNAs were extracted using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. RT-PCR was done using the following set of primers: Cst10/3Ј/F (5Ј-TCC TGA GGA TAT ATG TCA AT-3Ј) and Cst10/3Ј/R (5Ј-GAA CAG TGG GCC TTT GAA AA-3Ј). The following amplification cycle was used: 2 min of initial denaturation at 94°C, followed by 35 cycles at 94, 60, and 72°C for 30 s each plus extension at 72°C for 4 min. The primers and RT-PCR conditions used for type II and X collagens were as described previously (13). PCR products were electrophoresed and detected by Southern hybridization. For quantification of type X collagen mRNA levels, the density of each band was measured by NIH Image Version 1.62 2 and is expressed as the ratio to the density of glyceraldehyde-3-phosphate dehydrogenase.
Immunoblot Analysis-Polyclonal antibody against the full-length Cst10 protein was raised in rabbits using a synthetic peptide of Cst10. For preparation of the whole cell lysate, adherent and detached cells were collected and resuspended in chilled lysis buffer (10 mM Tris-HCl (pH 7.4), 1% Nonidet P-40, 0.1% SDS, 1 mM EDTA, 2 mM sodium orthovanadate, 10 mM sodium fluoride, and 10 mg/ml aprotinin). Collected cells were allowed to lyse by sonication on ice. The homogenate was centrifuged for 5 min in a microcentrifuge at 4°C, and the supernatants were collected and boiled in SDS sample buffer. The culture medium was collected and centrifuged for 5 min in a microcentrifuge at 4°C, and the supernatants were collected. The pellet obtained from the supernatants by centrifugation at 19,000 ϫ g for 20 min was resuspended in SDS sample buffer. Fifty-g portions of SDS sample buffer were loaded onto SDS-polyacrylamide gels and blotted onto polyvinylidene difluoride membrane (Amersham Biosciences). Mouse recombinant cystatin C (CstC) protein was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Protein bands on Western blots were visualized by chemiluminescent detection (ECL, Amersham Biosciences).
Immunohistochemistry-The samples harvested from embryonic mice (18 days postcoitus) were demineralized in 10% EDTA for 1 week at 4°C. The specimens were dehydrated with increasing concentrations of ethanol and then embedded in paraffin. Cst10 immunolocalization was examined in 4-m-thick dewaxed paraffin sections. The sections were treated with phosphate-buffered saline (PBS) containing 0.3% hydrogen peroxide for 30 min at room temperature and then with PBS containing 1% bovine serum albumin (Sigma) for 60 min at room temperature. They were then incubated with polyclonal antibody against mouse Cst10 for 24 h at 4°C and with horseradish peroxidase-conjugated goat anti-rabbit IgG antibody (Dakopatts, Glostrup, Denmark) at a dilution of 1:500 for 1 h at room temperature. After washing with PBS, the sections were immersed in diaminobenzidine solution for 10 min at room temperature for visualization and counterstained with hematoxylin. Nonimmune rabbit serum at the same concentration was used as a negative control. For ultrastructural analysis, these sections were observed under a transmission electron microscope (H-7100, Hitachi, Tokyo) following the pre-embedding method described previously (14). Briefly, embryonic mice (18 days postcoitus) were perfused through the left ventricle with a 2-ml dose of a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.08 M cacodylate buffer (pH 7.4). The tibiae were then removed and immersed in the same fixative for 2 h at 4°C. Specimens were decalcified with 4.13% EDTA at 4°C for 1 day and cryosectioned at a thickness of 10 m by cryomicrotome. The cryosections were treated following methods similar to those used for immunohistochemical investigations, being post-fixed in 1% OsO 4 in 0.1 M cacodylate buffer at 4°C for 1 h. The specimens were dehydrated in a graded ethanol series and embedded in Poly/Bed 812 resin (Polysciences, Warrington, PA). Ultrathin sections stained with lead citrate were used for transmission electron microscopic observation.
Establishment of ATDC5 Cells Stably Transfected with the Cst10 Gene-The entire sequence of Cst10 cDNA was amplified by PCR using Pfu DNA polymerase (Stratagene, La Jolla, CA) and inserted into the mammalian expression vector pcDNA3.1 (Invitrogen) including the cytomegalovirus promoter with a hemagglutinin epitope tag at the C terminus (pCMV-Cst10). The subcloned cDNA fragment was confirmed by sequencing. ATDC5 cells (2 ϫ 10 5 ) were plated in a 6-cm culture dish 24 h before transfection. pCMV-Cst10 (4 mg/6-cm culture dish) or the mock vector (pCMV) was transfected into ATDC5 cells by lipofection using SuperFect transfection reagent (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions. Two days later, cells were diluted 10-fold and incubated in a maintenance medium containing 400 g/ml Geneticin (Invitrogen). After 2 weeks, we isolated drug-resistant colonies, each of which was derived from a single clone, and cultured them separately. To confirm the reproducibility of the effects of Cst10 overexpression, we performed two experiments with independent transfection procedures. For the first experiment, we established 18 stable clones from 18 different colonies of ATDC5 cells transfected with pCMV-Cst10 and selected three clones (clones 3, 8, and 12) with the highest expression of Cst10 (pCMV-Cst10/ATDC5) by RT-PCR analysis. Three mock vector-transfected clones (clones 1-3) that were confirmed by RT-PCR not to express Cst10 were also selected as negative controls (pCMV/ATDC5). For the second experiment, we independently transfected pCMV-Cst10 into ATDC5 cells as described above, randomly established four stable clones, and examined the relationship of expressions between Cst10 and collagens by RT-PCR.
Alcian Blue and Alizarin Red Staining-pCMV-Cst10/ATDC5 and pCMV/ATDC5 cells were placed in 12-multiwell plates and cultured. Twenty-one days after induction by insulin, cells were rinsed with PBS and fixed with 95% methanol for 20 min. They were then stained overnight with 0.1% Alcian blue 8GS (Fluka, Buchs, Switzerland) in 0.1 M HCl. Twenty-eight days after induction by insulin, cultures were stained with 1% alizarin red S (pH 4.0) (Sigma) after fixation with 95% ethanol.
Detection of Apoptosis and Activities of Caspases-Apoptosis of pCMV-Cst10/ATDC5 and pCMV/ATDC5 cells was examined by nuclear staining with Hoechst 33342, externalization of phosphatidylserine residues using FITC-labeled annexin V, mitochondrial membrane potential, and flow cytometric analysis. Annexin V binding assay was performed using an FITC-labeled annexin V apoptosis detection kit (Medical and Biological Laboratories, Nagoya, Japan) according to the manufacturer's protocol (15)(16)(17). Briefly, ATDC5 cells were harvested 7 days after induction with insulin and washed with PBS. Cells were then incubated with binding buffer (10 mM HEPES (pH 7.4), 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , and 1.8 mM CaCl 2 ) containing 2 l of FITClabeled annexin V and 5 g of propidium iodide (PI) for 15 min at room temperature in the dark. After incubation, they were viewed under a fluorescent microscope. PI was added to distinguish cells with membrane permeability due to the loss of membrane integrity, which is characteristic of necrotic cell death. To assess the mitochondrial membrane potential, a MitoCapture apoptosis detection kit (Medical and Biological Laboratories) was used (18). ATDC5 cells harvested 7 days 2 Available at rsb.info.nih.gov/nih-image/download.html.
after induction were incubated with MitoCapture solution for 15 min at 37°C and viewed under a fluorescent microscope using a band-pass filter (detects FITC and rhodamine). In healthy cells, MitoCapture accumulates and aggregates in the mitochondria, giving off a bright red fluorescence. In apoptotic cells, MitoCapture cannot aggregate in the mitochondria due to the altered mitochondrial membrane potential and thus remains in the cytoplasm in its monomer form, fluorescing green. For flow cytometric analysis, ATDC5 cells were harvested 0, 1, 3, 5, 7, 10, and 14 days after induction by insulin and fixed with 75% ethanol and PBS at 4°C for 1 h. After rinsing twice with PBS, cells were incubated for 30 min with 1 ml of PBS containing 1 mg of boiled RNase at 37°C and then stained with 1 ml of PBS containing 10 g of PI. A total of 2 ϫ 10 4 cells were analyzed with a flow cytometer (FACSCaliber, BD Biosciences). To determine whether caspase-3 is activated in pCMV-Cst10/ATDC5, a PhiLux kit (Medical and Biological Laboratories) was used according to the manufacturer's protocol (19). Briefly, ATDC5 cells harvested 7 days after induction were incubated with 10 mM GDEVDGI and labeled with two molecules of rhodamine, which was selectively cut by caspase-3. After incubation, cells were viewed under a fluorescent microscope. Caspase-3 activity was determined with a caspase-3/CPP32 colorimetric protease assay kit (Medical and Biological Laboratories) according to the manufacturer's protocol. In brief, ATDC5 cells were harvested 7 days after induction with insulin and lysed in lysis buffer. Cell lysate (150 g of protein in 50 l of lysis buffer) was incubated with DEVD-p-nitroanilide (pNA) as a substrate for 1 h at 37°C, and the amount of pNA generated was determined spectrophotometrically at 405 nm. Caspase-8 and caspase-9 activities were determined with caspase-8/FLICE and caspase-9/Mch6 colorimetric protease assay kits, respectively (Medical and Biological Laboratories), according to the manufacturer's protocols. IETD-p-nitroanilide for caspase-8 and LEHD-p-nitroanilide for caspase-9 were used as substrates. After incubation for 1 h at 37°C, the amount of p-nitroanilide generated was determined spectrophotometrically at 405 nm (20 -22).
Statistical Analysis-Means of groups were compared by analysis of variance, and significance of differences was determined by post-hoc testing using Bonferroni's method.

RESULTS
Characterization of the Cst10 Gene-In a previous study, using differential display analysis, we identified nine genes, including Cst10, whose expression is regulated in the auricular cartilage of ttw mice fed a high phosphate diet (1). The fulllength cDNA sequence of the mouse Cst10 gene determined by cating a novel member of the type II cystatin superfamily with 40.5 and 39.0% homologies to mouse and human CstC (Cst3), respectively, the closest cystatin family member (Fig. 1A). Its homologies to other mouse cystatins were around or less than 30%: 31.5% to Cst9, 31.4% to Cst7, 27.0% to CstEM, 28.6% to CstSC, and 26.7% to CstTE. To investigate the localization of the Cst10 gene in the mouse chromosome, Ͼ50 metaphase cells were examined by fluorescence in situ hybridization using a BAC clone containing the mouse Cst10 gene as a probe. Specific hybridization signals were identified on chromosome 2 in almost all cells, and no significant background was observed at any other chromosomal sites (Fig. 1B).
Temporal and Spatial Expression of Cst10 in Vivo and in Vitro-We first examined the temporal expression pattern of Cst10 mRNA levels in the auricular cartilage of ttw mice whose endochondral ossification was enhanced with a high phosphate diet. Expression appeared 3 days after weaning and was upregulated by a high phosphate diet at 5 days and thereafter ( Fig. 2A). Our previous study on the tissue distribution of Cst10 expression in a variety of mouse tissues showed that this gene is expressed exclusively in cartilage (1). We therefore examined the expression pattern of Cst10 using cell cultures. Among three cultured mesenchymal cells from neonatal ddY mice (primary osteoblasts from calvariae, chondrocytes from costal cartilage, and fibroblasts from skin), Cst10 expression was confirmed to be specific to chondrocytes (Fig. 2B). To characterize the expression pattern during differentiation of chondrocytes, we used the mouse chondrogenic cell line ATDC5, which can be induced to differentiate into mature chondrocytes in the presence of insulin (12). During induction of differentiation with insulin, expression of type II collagen remained unchanged throughout the culture period up to 14 days, whereas that of type X collagen, a marker for hypertrophic chondrocytes, appeared 7 days after induction (Fig. 2C). Expression of the Cst10 gene appeared at 3 days and increased thereafter, indicating that Cst10 expression is in synchrony with the maturation of chondrocytes.
To examine the localization of Cst10 in cartilage, we first confirmed by Western blot analysis the specificity of a polyclonal antibody against Cst10 without cross-reactivity with CstC, the closest member of the cystatin superfamily (Fig. 2D). Using this antibody, we performed immunohistochemical analysis on the growth plates of embryonic ddY mice and found that Cst10 was expressed mainly in mature chondrocytes, including FIG. 2. Temporal and spatial expression of Cst10. A, time course of Cst10 mRNA expression in the auricular cartilage of ttw mice fed high and low phosphate diets. Mice were serially killed at the indicated days after weaning and the start of the diet at 3 weeks of age, and total RNAs were extracted from resected auricular cartilage. Expression was examined by semiquantitative RT-PCR, followed by Southern blotting. B, expression of Cst10 mRNA in cultured primary osteoblasts, chondrocytes, and fibroblasts. Cells were isolated from the calvariae, costal cartilage, and skin of neonatal ddY mice, respectively. Expression was examined by semiquantitative RT-PCR, followed by Southern blotting. C, temporal expression patterns of Cst10, type II collagen (Col II), and type X collagen (Col X) during differentiation of mouse chondrogenic ATDC5 cells. At the indicated days after induction with insulin, cells were harvested, and mRNA levels were examined by semiquantitative RT-PCR. D, specific reaction of the polyclonal antibody against Cst10 used in this study. The antibody was confirmed to bind to mouse recombinant Cst10, but not to mouse recombinant CstC, by Western blot analysis. E, light microscopic immunohistochemistry of Cst10 in the growth plate of an embryonic ddY mouse (18 days postcoitus). Blue, red, and green lines indicate proliferative, prehypertrophic, and hypertrophic layers, respectively. Immunopositive cells are stained brown. F, electron microscopic immunohistochemistry of Cst10 in a hypertrophic chondrocyte in the growth plate of an embryonic ddY mouse (18 days postcoitus). The lower panel shows an enlargement of the boxed area in the upper panel. Immunochemical localizations shown as electron-dense particles (asterisks) were observed in the cytosolic areas, but not in the nucleus (N) or the rough endoplasmic reticulum (rER). Bars, 10 m (upper panel) and 1 m (lower panel). GAPDH, glyceraldehyde-3-phosphate dehydrogenase. prehypertrophic and hypertrophic cells (Fig. 2E). Electron microscopic examination of a hypertrophic chondrocyte revealed that Cst10 was immunolocalized in the cytosolic areas, but was not found in the nucleus or within the lumen of the rough endoplasmic reticulum (Fig. 2F). These findings suggest that Cst10 is not transported into the Golgi-endoplasmic reticulum system, but acts as an intracellular protein in the cytosol.
Overexpression of the Cst10 Gene Accelerates Maturation of ATDC5 Cells-To elucidate the function of Cst10 in chondrocytes, we established stable clones of ATDC5 cells overexpressing the Cst10 gene (pCMV-Cst10/ATDC5). We first compared by RT-PCR the differentiation of pCMV-Cst10/ATDC5 cells with that of control clones of ATDC5 cells transfected with the mock vector (pCMV/ATDC5) upon induction with insulin (Fig.  3A). In pCMV-Cst10/ATDC5 cells, Cst10 mRNA expression was clearly seen not only after, but also before induction (time 0), whereas in pCMV/ATDC5 cells, expression was faintly seen 7 days after induction and increased moderately thereafter. Western blot analysis revealed that the Cst10 protein was localized in the cell lysate, but not in the culture medium of pCMV-Cst10/ATDC5 cells (Fig. 3B), indicating that Cst10 is not a secreted protein. Expression of type II collagen, which is known to be produced by chondrocytes from their early phase of differentiation, was constitutively seen before and after induction, and expression was not different between pCMV-Cst10/ ATDC5 and pCMV/ATDC5 cells (Fig. 3A). However, expression of type X collagen, a marker of hypertrophic chondrocytes, was observed earlier and was stronger in pCMV-Cst10/ATDC5 cells than in pCMV/ATDC5 cells. We also compared the cartilage nodule formation and mineralization between cultured pCMV-Cst10/ATDC5 and pCMV/ATDC5 cells using Alcian blue and alizarin red staining, respectively. No difference was seen in the Alcian blue staining between the two cells; however, alizarin red staining was stronger in cultured pCMV-Cst10/ ATDC5 cells than in pCMV/ATDC5 cells (Fig. 3C).
To confirm the reproducibility of the effects of Cst10 overexpression, we performed another experiment using ATDC5 cell clones that were independently transfected with pCMV-Cst10 and isolated. In this experiment, we randomly established four stable clones and compared the mRNA levels of Cst10 and type II and X collagens by RT-PCR before (time 0) and 5 days after induction (Fig. 3D). The Cst10 mRNA levels were not changed before or after induction in all clones. There was a good correlation between the Cst10 and type X collagen levels at 5 days, although the type II collagen levels were similar among the clones (Fig. 3D). These results indicate that overexpression of FIG. 3. Effect of Cst10 overexpression on maturation of ATDC5 cells. We performed two experiments with independent transfection of the Cst10 gene. For the first experiment (A-C), 18 stable clones of ATDC5 cells transfected with pCMV-Cst10 were established as described under "Experimental Procedures," and three clones (clones 3,8,12) with the highest expression of Cst10 (pCMV-Cst10/ATDC5) were selected by RT-PCR analysis. Three clones (clones 1-3) transfected with the mock vector (pCMV/ATDC5) were also selected as negative controls. Although the data shown in A-C are for representative clones (clone 3 for pCMV-Cst10/ATDC5 and clone 1 for pCMV/ATDC5), the results were reproducible when other clones in each group (clones 8 and 12 for pCMV-Cst10/ATDC5 and clones 2 and 3 for pCMV/ATDC5) were used. A, shown are the temporal expression patterns of Cst10, type II collagen (Col II), and type X collagen (Col X) during differentiation of pCMV-Cst10/ATDC5 and pCMV/ATDC5 cells cultured in the presence of insulin. At the indicated days after induction with insulin, cells were harvested, and the mRNA levels were examined by semiquantitative RT-PCR. B, Cst10 protein levels in the cell lysate and medium of pCMV-Cst10/ATDC5 cells cultured for 5 days in the presence of insulin were determined by Western blot analysis. The Cst10 protein could not be detected in the culture medium even if it was lyophilized and condensed (data not shown). C, pCMV-Cst10/ATDC5 and pCMV/ATDC5 cells were stained with Alcian blue and alizarin red at 21 and 28 days of culture, respectively, in the presence of insulin. D, for the second experiment, ATDC5 cells were transfected with pCMV-Cst10, independent of the first experiment, and four stable clones (clones 1-4) were randomly established. Expression of Cst10 and type II and X collagens was compared among the clones by semiquantitative RT-PCR before (time 0) and 5 days after induction with insulin. Gapdh, glyceraldehyde-3phosphate dehydrogenase.
the Cst10 gene accelerates the later (but not earlier) stage of chondrocyte differentiation and mineralization.
Overexpression of Cst10 Leads to Apoptosis of ATDC5 Cells-Staining with Hoechst 33342 revealed the existence of cells with fragmented and condensed nuclei with increased fluorescence, suggesting apoptotic cell death in pCMV-Cst10/ATDC5 cells, but not in pCMV/ATDC5 cells (Fig. 4A). To distinguish between apoptosis and necrosis in these cells, the kinetics of loss of membrane integrity was examined by double staining with annexin V and PI. Annexin V is known to stain positive in the early stage of apoptotic cells that retain the ability to exclude vital dyes, whereas PI becomes positive in necrotic cells that have lost membrane integrity and that have undergone rapid swelling and lysis (15)(16)(17). Most pCMV-Cst10/ATDC5 cells were stained green, indicating that annexin V was positive; however, pCMV/ATDC5 cells were stained red, indicating that PI was positive (Fig. 4B). In the analysis of apoptosis through the mitochondrial pathway, many pCMV-Cst10/ ATDC5 cells were stained green, indicating the change in the mitochondrial membrane potential, whereas there were no positively stained cells in the pCMV/ATDC5 cell culture (Fig. 4C). We further examined the involvement of caspases in the stimulation of chondrocyte apoptosis by Cst10. Staining of substrates specific to caspase-3 revealed that there were many cells with a high activity of this caspase in pCMV-Cst10/ ATDC5 cells, but not in pCMV/ATDC5 cells (Fig. 4D). Furthermore, we examined the activities of caspase-3, -8, and -9 by measuring pNA free of each substrate. pCMV-Cst10/ATDC5 cells exhibited significantly higher activities of all these caspases compared with pCMV/ATDC5 cells (Fig. 4E). Caspase-3 and caspase-9, which are known to be associated mainly with the mitochondrial pathway of apoptosis, were strongly activated by Cst10 gene overexpression. These results imply the importance of the mitochondrial pathway in the induction of chondrocyte apoptosis by Cst10.
To determine whether the induction of apoptosis by Cst10 overexpression is a direct action or secondary to that of maturation, we examined the time course of hypertrophic matura- FIG. 4. Effect of Cst10 overexpression on apoptosis of ATDC5 cells. A, fragmented nuclei immunostained with Hoechst 33342 were seen in pCMV-Cst10/ATDC5 cells. B, annexin V-positive cells stained green were predominantly seen in pCMV-Cst10/ATDC5 cells, whereas PI-positive cells stained red were seen in pCMV/ATDC5 cells. C, green fluorescence indicating a change in the mitochondrial membrane potential was seen only in pCMV-Cst10/ATDC5 cells, whereas most pCMV/ATDC5 cells gave off red fluorescence, indicating healthy mitochondrial membranes. D, green fluorescence indicating positive caspase-3 activity was predominantly seen in pCMV-Cst10/ATDC5 cells. Although A-D are images of representative clones of the first experiment (clone 3 for pCMV-Cst10/ATDC5 and clone 1 for pCMV/ATDC5) 7 days after induction with insulin, similar findings were obtained when two other clones in each group (clones 8 and 12 for pCMV-Cst10/ATDC5 and clones 2 and 3 for pCMV/ATDC5) were used. E, shown are the averages of the activities of caspase-3, -8, and -9 in cultured pCMV-Cst10/ATDC5 (clones 3, 8, and 12) and pCMV/ATDC5 (clones 1-3) cells in the first experiment 7 days after induction. The cell lysate was incubated with substrates DEVD-pNA, IETD-pNA, and LEHD-pNA, which were selectively cut by caspase-3, -8, and -9, respectively. The amount of pNA free of substrates was measured. Data are expressed as means (bars) Ϯ S.E. (error bars) for (four cultures/clone) ϫ (three clones/group). *, p Ͻ 0.01, and **, p Ͻ 0.001, significant difference between pCMV-Cst10 and pCMV. F, shown is the time course of hypertrophic maturation (OO) and apoptosis (---) of cultured pCMV-Cst10/ATDC5 (E) and pCMV/ATDC5 (q) cells after induction with insulin. Hypertrophic maturation was determined by the ratio of the band densities of type X collagen (Col X) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) upon RT-PCR shown in Fig. 3A. Apoptosis was determined by the average of the percentage of cells in the sub-G 1 population upon flow cytometric analysis of three clones of each pCMV-Cst10/ ATDC5 (clones 3, 8, and 12) and pCMV/ATDC5 (clones 1-3) in the first experiment. Data are expressed as means (symbols) Ϯ S.E. (error bars) for (four cultures/clone) ϫ (three clones/group). *, p Ͻ 0.01, significant increase compared with day 0. tion and apoptosis determined by the type X collagen mRNA level shown in Fig. 3A and the percentage of cells in the sub-G 1 population by flow cytometric analysis, respectively (Fig. 4F). In pCMV-Cst10/ATDC5 cells, significant induction of hypertrophy was seen at 3 days and that of apoptosis at 10 days, whereas in pCMV/ATDC5 cells, hypertrophy was at 7 days and apoptosis at 14 days. Hence, the time gaps between hypertrophy and apoptosis were similar (ϳ7 days) in both cells, suggesting that induction of apoptosis by overexpression of the Cst10 gene may not be direct, but is secondary to that of hypertrophic maturation. DISCUSSION Based on our previous study in which we identified a novel gene (Cst10) whose expression is up-regulated during ossification of auricular cartilage by a high phosphate diet in ttw mice (1), in the present study, we investigated the possible role of Cst10 in endochondral ossification. Expression was seen exclusively in differentiated chondrocytes in both in vivo and in vitro mouse models. Overexpression of the Cst10 gene in ATDC5 cells induced their maturation, followed by apoptosis, suggesting an important role of Cst10 in the last steps of chondrocyte differentiation.
Cst10 is a novel member of the cystatin superfamily and shows ϳ40% homology to both mouse and human CstC. CstC is an abundant extracellular inhibitor of all cysteine proteinases of the papain superfamily (23) and is related to several human disorders such as atherosclerosis and aortic aneurysms (24) and hereditary amyloid angiopathy of the brain (25,26). CstC also down-regulates bone resorption by inhibiting osteoclastic proteolytic enzymes released into the resorption lacunae in skeletal tissues (27)(28)(29). Unlike other members of the cystatin superfamily such as CstC, expression of Cst10 is limited to chondrocytes, implying some specific role of Cst10 in cartilage homeostasis. In addition, electron microscopic examination (Fig. 2F) and Western blot analysis (Fig. 3B) of cultured pCMV-Cst10/ATDC5 cells indicated that Cst10 is not a secreted protein, but is an intracellular enzyme localized in the cytosol. It is therefore possible that Cst10 has a function distinct from those of other cystatin members that act as extracellular inhibitors of proteinases. Investigation of the physiological role of Cst10 in cartilage in vivo is now under way by creating a Cst10 genedeficient mouse line using homologous recombination in embryonic stem cells.
Regarding the regulation of chondrocyte differentiation and maturation, several secreted molecules have been identified. Bone morphogenetic proteins (BMPs) are reported to be positive regulators, whereas parathyroid hormone-related protein (PTHrP) acts to slow the rate of chondrocyte maturation and to maintain chondrocytes in a proliferative state. Analyses of PTHrP knockout (30,31) and transgenic (32,33) mice point to this molecule as a major factor that diminishes the rate of chondrocyte hypertrophy and the subsequent endochondral ossification. Ihh (Indian hedgehog), which co-localizes with BMP-6 in the region of post-proliferative chondrocytes, is known to be another negative regulator of chondrocyte hypertrophy (32,34). Ihh induces PTHrP expression, and PTHrP reciprocally inhibits Ihh expression, thus forming a negative feedback loop to regulate the rate of chondrocyte hypertrophy. Ihh also blocks BMP-6 expression, suggesting a tightly controlled progression of maturation by secreted molecules within the cartilage (35). Signaling related to cystatins may possibly be involved in the regulatory network of these secreted molecules. In fact, CstC produced by osteoblasts inhibits bone resorption by PTHrP in malignancy (27,28). In this study as well, ATDC5 cells overexpressing Cst10 showed phenotypes similar to those of chondrocytes from PTHrP-deficient mice: enhanced hypertrophic maturation, mineralization, and apoptosis (30,31). Studies on the regulation of expression and/or activity of Cst10 by BMPs, PTHrP, and Ihh will help elucidate the molecular mechanism of endochondral ossification in greater detail.
There are many reports demonstrating the association of cysteine proteinases and their inhibitors with apoptotic cell death. Disruption of lysosomal membranes and release of lysosomal enzymes, including proteinases such as cathepsins B (36 -38), D (39,40), and L (41,42), are known to cause apoptosis (43,44). This pathway is thought to be mediated mainly by cleavage of the Bcl-2 family member Bid (45). Caspases, which play an integral part in the apoptosis pathway, also belong to a family of cysteine proteinases with aspartate specificity. Some cysteine proteinase inhibitors are reported to block apoptosis (41,46); and among them, leupeptin and E-64 specifically suppress activation of caspase-3-like proteinases (47). Hence, cystatins could also down-regulate apoptosis by inhibiting lysosomal enzymes or caspases. In fact, loss-of-function mutations in the cystatin B gene cause a severe neurological disorder known as Unverricht-Lundoborg disease in humans (48,49) and myoclonic seizures and ataxia in mice (50,51), in which cerebellar granule cells appear to undergo apoptosis. In contrast, CstC has been reported to induce apoptosis in cultured rat neurons (52) and during mouse embryo implantation and placentation (53). The present study has also shown that Cst10 overexpression led to apoptosis in chondrocytes. A direct contribution of proteinases to chondrocyte apoptosis was actually implied by the finding that matrix metalloproteinase-9-deficient mice exhibit normal hypertrophic maturation, but delayed apoptosis (54). However, the action of Cst10 on chondrocyte apoptosis is not likely to be direct, but secondary to its stimulation of maturation, because the time gaps between hypertrophy and apoptosis were similar between cells with and without Cst10 overexpression (Fig. 4F). Furthermore, in our preliminary experiments, Cst10 gene overexpression failed to promote apoptosis in the non-chondrogenic cell line COS-7 (data not shown). Recent studies on various knockout mice also suggest that there is a direct coupling between maturation and apoptosis in chondrocytes. PTHrP-deficient mice exhibit acceleration of both hypertrophy and apoptosis in growth plate chondrocytes (30,31). Mice deficient in Bcl-2, an inhibitor of apoptosis, exhibit premature chondrocyte differentiation (33). Although changes in the mitochondrial membrane potential and related caspases were seen in Cst10-expressing chondrocytes, this also may not be a specific action of Cst10 because normal chondrocyte apoptosis is known to involve Bcl-2, mitochondrial integrity, and caspases (33,55). Further studies on intracellular interactions between Cst10 and signaling molecules of apoptosis such as Bcl-2 will elucidate more precise mechanisms by which Cst10 overexpression leads to chondrocyte apoptosis.
Chondrocyte maturation, mineralization, and apoptosis in the endochondral ossification process are observed in physiological development and growth as well as under pathological conditions such as osteoarthritis and ectopic ossification. Understanding the molecular mechanisms of endochondral ossification through the Cst10 signaling pathway may therefore help elucidate not only the mechanism of skeletal development and growth, but also the pathophysiology of these diseases.