WISP3-dependent regulation of type II collagen and aggrecan production in chondrocytes

Authors


Abstract

Objective

WISP3 (Wnt-1–inducible secreted protein 3) is a member of the CCN (connective tissue growth factor, cysteine-rich 61, nephroblastoma overexpressed) family of connective tissue growth factors. WISP3 mutations have been linked to progressive pseudorheumatoid dysplasia (PPRD). The present study was conducted to investigate whether WISP3 is responsible for the expression of cartilage-specific molecules.

Methods

WISP3 expression in human cartilage was assessed by immunostaining with anti-WISP3 antibody. The effect of WISP3 on chondrocyte-specific gene regulation was determined by transfecting human chondrocyte lines C-28/I2 and T/C-28a2 with a WISP3 expression vector. Alterations in WISP3-mediated messenger RNA and protein expression of cartilage-specific molecules were assessed by reverse transcriptase–polymerase chain reaction and immunoblotting.

Results

Immunohistochemistry experiments demonstrated that WISP3 protein is expressed in the midzone chondrocytes of normal adult articular cartilage, in chondrocyte clusters of osteoarthritic cartilage, and in the zone of proliferating chondrocytes of fetal growth cartilage. Human chondrocyte lines C-28/I2 and T/C-28a2 transfected with a WISP3 expression vector produced increased amounts of the cartilage-specific matrix molecules type II collagen and aggrecan, in part via activation of the sex-determining region Y–type high mobility group box (SOX) family of transcription factors. In contrast, a mutant WISP3, previously found to be associated with PPRD, had impaired effects on cartilage-specific gene expression.

Conclusion

Our experimental results suggest that WISP3 supports cartilage integrity by regulating the expression of type II collagen and aggrecan, and mutations linked with PPRD can compromise this function and produce cartilage loss.

The WISP3 (Wnt-1–inducible secreted protein 3) gene product is a member of the WISP family of growth modulators. WISP1, the first cloned member of the WISP family, is induced by Wnt-1 in certain cell types (1, 2). However, WISP3 does not appear to be Wnt-1 inducible (1). Due to high sequence homology, the WISP gene products (WISPs 1, 2, and 3) are considered to be members of the CCN (connective tissue growth factor [CTGF], cysteine-rich 61 [Cyr61], nephroblastoma overexpressed [NOV]) family of connective tissue growth factors (3, 4). CCN growth factors are mostly extracellular matrix (ECM)–associated proteins that regulate cell migration and adhesion, cell proliferation and survival, and cell differentiation in connective tissues (3–6). The CCN members CTGF and Cyr61 not only are involved in important developmental processes, such as chondrogenesis, osteogenesis, and angiogenesis, but also have been implicated in fibrotic disorders and wound healing (3–9). The precise functions of the WISP proteins in connective tissues have not yet been established at the molecular level.

WISP3 maps to chromosome 6q21–22 and encodes a 354–amino acid protein (10), with an estimated molecular weight of ∼40 kd. WISP3 messenger RNA (mRNA) has been detected mainly in mesenchymal cells and tissues (1, 10). Low-to-moderate levels of WISP3 mRNA have also been detected in other tissues, such as kidney, testis, placenta, ovary, prostate, and small intestine (1, 10). Like other CCN family members, WISP3 encodes several potentially functional domains, each domain corresponding roughly to 1 exon (3, 4, 10). Exon 1 encodes a peptide signal sequence, exon 2 encodes a domain that bears homology to the amino-terminal domain of the insulin-like growth factor binding proteins (IGFBPs), exon 3 encodes a von Willebrand factor type C repeat domain that may participate in peptide oligomerization, exon 4 encodes a thrombospondin type 1 domain that may bind to sulfated glycosaminoglycans, and exon 5 encodes a cysteine knot domain that may be involved in dimerization and receptor binding. The domain organization of WISP3 suggests that it encodes a scaffold protein regulating the biologic functions of other proteins.

WISP3 mutations have been linked to an autosomal-recessive skeletal disorder, progressive pseudorheumatoid dysplasia (PPRD), that is marked by abnormal skeletal growth and progressive cartilage loss (10, 11). The identification of 9 independent WISP3 mutations in unrelated PPRD patients suggests that WISP3 mutations might play an important role in the development of PPRD. Because extraskeletal manifestations have not been reported in PPRD and cartilage appears to be the tissue primarily affected (10), it is possible that WISP3 regulates cartilage homeostasis. However, the mechanisms of action of WISP3 in skeletal development and cartilage maintenance have not been elucidated. Because mutations in different WISP3 domains are all associated with similar clinical features, i.e., skeletal malformation and cartilage loss (10), it appears that all the WISP3 domains are required for proper skeletal development and maintenance of cartilage integrity.

Cartilage is a specialized connective tissue containing a large amount of ECM composed of a dense network of collagen fibers and a high concentration of proteoglycans, which provide the tissue with its ability to withstand compression and shear stress (12, 13). Cartilage ECM is formed and maintained by chondrocytes (12–15). Although the mechanisms of ECM turnover and homeostasis have not been studied in great detail, research in several laboratories has substantiated the importance of cartilage-specific collagens and the large aggregating proteoglycan aggrecan in the maintenance of cartilage integrity and homeostasis (12–15). Extensive research has also confirmed the importance of the sex-determining region Y–type high mobility group box (SOX) family of high mobility group transcription factors (L-SOX5, SOX6, and SOX9) in the expression of cartilage-specific proteins (13).

The present study was undertaken to investigate the role of WISP3 in cartilage-specific gene expression. The results suggest that WISP3 up-regulates the expression of the cartilage-specific gene encoding type II collagen (COL2A1) and aggrecan in chondrocytes, at least in part through the activation of SOX family transcription factors, and may be an important molecule in the regulation of cartilage repair/homeostasis.

MATERIALS AND METHODS

Cell cultures.

The immortalized chondrocyte cell lines C-28/I2 and T/C-28a2 were derived from human juvenile costal cartilage and generated by infection with a replication-defective retroviral vector expressing SV40 large T antigen (16). Cultures of C-28/I2 and T/C-28a2 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) in a 5% CO2 incubator at 37°C and passaged at subconfluency every 5–6 days.

Wild-type and mutant WISP3 expression constructs and COL2A1-luciferase reporter construct.

Human WISP3 was cloned from synoviocyte complementary DNA (cDNA) by polymerase chain reaction (PCR) amplification with WISP3-specific primers, bearing Eco RI and Eco RV restriction enzyme sites at their flanking ends, for the purpose of subcloning into the pcDNA3.1 expression vector (Invitrogen, Carlsbad, CA). Human synoviocyte cultures (passage 3) were obtained from the University of California, San Diego Specialized Center for Research on Arthritis and processed immediately for RNA extraction. The primers used for cloning WISP3 were 5′-GGAATGAATTCGTCCCAGCGACATGCAGGG-3′ (forward) and 5′-GCAATGATATCTGGTTTTACAGAATCTTGAGCTC-3′ (reverse). The amplified WISP3 gene product (∼1.1 kb) was subcloned into the Eco RI and Eco RV sites of pcDNA3.1(+). This expression vector was used for all functional studies of WISP3. For the purpose of protein expression studies, a hemagglutinin tag (with an engineered stop codon) was introduced in frame with the WISP3 coding sequence into the Eco RV and Xho I sites of a modified WISP3 expression construct, and the stop codon in the WISP3 coding sequence was deleted by PCR mutagenesis. The amino acid sequence of the hemagglutinin tag was KAFSNCYPYDVPDYASLRS (17).

The mutant Cis78–Arg WISP3 was constructed using a PCR mutagenesis kit (Stratagene, La Jolla, CA). The mutant primers (complementary primers with a T–C transversion), which were used with WISP3-pcDNA3.1 as a template in the PCR reaction, had the sequences 5′-TGGATGCTGTAAAATCCGTGCCAAGCAACC-3′ (forward) and 5′-GGTTGCTTGGCACGGATTTTACAGCATCCA-3′ (reverse). A hemagglutinin tag was introduced separately into the mutant WISP3 expression construct as described above, for the purpose of protein expression studies.

The COL2A1–luciferase reporter construct pXP2-COL2/4.0 contained the COL2A1 sequence spanning –577 to +3,428 bp (16), which was cloned into the pXP2 luciferase reporter gene vector (Promega, Madison, WI).

Cell transfection.

Lipofectamine was used for transfecting C-28/I2 and T/C-28a2 chondrocyte cell lines with the WISP3 and empty vector constructs. Briefly, cells (∼2 × 105/ml) were plated 1 day before transfection in 6-well tissue culture plates (2 ml/well) and incubated at 37°C in 5% CO2. A complex of plasmid DNA (∼1 μg) with 6 μl of Lipofectamine in 200 μl of serum-free, antibiotic-free medium was prepared in a sterile microfuge tube for 30 minutes, after which 800 μl of medium containing 5% FBS was added. A similar complex was prepared for each well of a 6-well plate. The cells in each well of the plate were washed with sterile phosphate buffered saline (PBS) and then incubated with 1 ml of the transfection mixture for 7–12 hours, after which 1 ml of culture medium with 5% FBS was added to each well. After 24 hours, the transfection mixture was replaced with fresh culture medium containing 10% FBS. The incubation was continued for an additional 24–26 hours, and the cells were harvested for either RNA or protein extraction.

RNA extraction and mRNA analysis.

The C-28/I2 chondrocytes were transfected with WISP3, mutant WISP3, or empty vector, and total RNA was extracted using TRIzol (Invitrogen). Primers specific for type II collagen, type I collagen, aggrecan, SOX9, fibronectin, and G3PDH were used for estimating the levels of expression of the corresponding mRNA. Reverse transcriptase–PCR (RT-PCR) was performed using a reverse transcription kit according to the instructions of the manufacturer (Invitrogen). During cDNA synthesis, ∼1.5 μg of RNA was used for each specimen, and 30 cycles of PCR were carried out. The G3PDH gene was used as an internal control. Table 1 summarizes the primer pairs and experimental conditions used for RT-PCR analysis.

Table 1. Primer pairs and experimental conditions
Gene product*Forward and reverse primers (5′–3′)Polymerase chain reaction conditionsExpected product size, bp
  • *

    SOX9 = sex-determining region Y–type high mobility group box 9; WISP3/bgh = Wnt-1–inducible secreted protein 3/vector.

COL2A1GAAACCATCAATGGTGGCTTCC30 cycles, 62° annealing301
 CGATAACAGTCTTGCCCCACTT  
COL1A1ATCCAGCTGACCTTCCTGCG30 cycles, 62° annealing322
 TCGAAGCCGAATTCCTGGTCT  
AggrecanTGAGGAGGGCTGGAACAAGTACC30 cycles, 60° annealing349
 GGAGGTGGTAATTGCAGGGAACA  
FibronectinGTGTGACCCTCATGAGGCAAC30 cycles, 62° annealing299
 TACTCTCGGGAATCTTCTCTGT  
SOX9CACACTACAGCCCCTCCTAC30 cycles, 60° annealing258
 CCTCCTCAAGGTCGAGTGAG  
G3PDHACCACAGTCCATGCCCATCAC30 cycles, 60° annealing450
 TCCACCACCCTGTTGCTGTA  
β-galactosidaseGATCTGCCATTGTCAGACATGTA30 cycles, 60° annealing309
 ACACCAGACCAACTGGTAATGG  
WISP3/bghAGCAACTGTGAAATGAGAAAAGAG30 cycles, 60° annealing438
 AACTAGAAGGCACAGTCGAGG  

Preparation of whole cell protein lysates and immunoblotting.

To prepare whole cell lysates, cells were treated for 1 minute with trypsin–EDTA after removal of culture medium, rinsed once with PBS, harvested to sufficient density, and disrupted with lysis buffer (20 mM Tris HCl [pH 7.5], 500 mM NaCl, 1% Triton X-100, 1 mM EDTA, 50 mM dithiothreitol, and 2 mM phenylmethylsulfonyl fluoride) in a microfuge tube. The lysate was spun down, cleared of debris by centrifugation, and assayed for total protein concentration using Bradford reagent (Bio-Rad, Richmond, CA). Approximately 30 μg of each lysate was used for immunoblotting.

Monoclonal anti-human type II collagen (mouse IgG2a; Calbiochem, La Jolla, CA) was used as primary antibody to detect type II collagen, monoclonal antiactin (mouse IgG1; Sigma, St. Louis, MO) was used as primary antibody to detect actin, antihemagglutinin (mouse IgG1; Cell Signaling Technology, Beverly, MA) was used as primary antibody to detect hemagglutinin-tagged WISP3 protein, and antiluciferase antibody (mouse IgG1; Calbiochem) was used for detecting the luciferase reporter. The secondary antibody for each of the above-mentioned primary antibodies was horseradish peroxidase (HRP)–conjugated goat anti-mouse IgG (Transduction Laboratories, Lexington, KY). Anti-WISP3 (goat IgG) obtained from Santa Cruz Biotechnology (Santa Cruz, CA) was used with HRP-conjugated mouse anti-goat IgG (Santa Cruz Biotechnology) for detecting wild-type WISP3 protein. Secondary antibody incubation of protein blots was followed by visualization with a chemiluminescence system (Amersham Pharmacia Biotech, Piscataway, NJ).

Immunohistochemistry studies.

Three types of human articular cartilage were obtained from tissue banks: cartilage from adult donors with osteoarthritis (OA), cartilage from adult donors without any evidence of disease, and growth plate cartilage from fetuses (∼105 days gestation). The tissues were fixed in formalin and embedded in paraffin. Microtome-cut sections (5 μm) were deparaffinized in 3 changes of Hemo-de and rehydrated in graded ethanol and water. Endogenous biotin– or avidin–binding sites were blocked by sequential incubation for 15 minutes with avidin and biotin (Vector, Burlingame, CA). Nonspecific staining was blocked by incubation of sections with 10% normal serum or bovine serum albumin in PBS. Sections were digested in 2 mg/ml hyaluronidase for 30 minutes and permeabilized in 0.2% Triton X-100/PBS for 5 minutes at room temperature.

After blocking, the sections were incubated with goat anti-human WISP3, an antipeptide antibody against an N-terminal segment of human WISP3 (Santa Cruz Biotechnology), at 2 μg/ml for 1–2 hours at room temperature. After washing the sections 3 times in PBS for 5 minutes each, a second blocking was performed for 10 minutes. The sections were incubated for 30 minutes with diluted biotinylated secondary antibody. The slides were washed 3 times in PBS and incubated for 30 minutes with Vectastain ABC–alkaline phosphatase reagent (Vector). They were washed again, and the sections incubated for 15 minutes with alkaline phosphatase substrate solution. Slides were rinsed with water, counterstained with diluted hematoxylin, rehydrated in graded ethanol, deparaffinized with Hemo-de, and mounted with Refrax mounting medium (Anatech). Tissues from 5 healthy donors, 3 donors with OA, and 2 fetuses were processed for immunostaining; within each of these 3 tissue sources, all specimens yielded similar results.

Image analysis and statistical analysis.

All reported results of RT-PCR analyses of RNA and immunoblotting analyses of protein represent at least 4 separate experiments. Differences in mRNA and protein levels were quantified using Kodak Image Analysis Software, version 2.0.2 for Macintosh. Statistical analysis was performed with Student's t-test.

RESULTS

WISP3 expression pattern in human cartilage.

Although WISP3 mRNA has been detected in both chondrocytes and chondroprogenitor cells (10), the pattern of WISP3 protein expression in cartilage tissue has not been described. Immunohistochemical analysis of normal human articular cartilage with an antibody to an N-terminal peptide of human WISP3 (Figure 1) revealed antigen expression (reddish stain) mostly in the midzone chondrocytes. In OA cartilage, WISP3 was expressed mainly in clusters of proliferating chondrocytes. Immunostaining of fetal cartilage growth plates with the same anti-WISP3 antibody showed maximal WISP3 expression in areas of chondrocyte proliferation, with decreased expression in the hypertrophic zone. This expression pattern suggests that WISP3 is involved in cartilage repair and homeostasis and the maturation of growth plate chondrocytes. The WISP3 staining pattern was mostly intracellular or membrane associated in all tissue sections examined. In OA cartilage sections, however, some ECM staining was clearly visible. No immunoreactivity was detected with an antibody to a C-terminal WISP3 peptide (results not shown) or with control antibody.

Figure 1.

Expression of WISP3 (Wnt-1–inducible secreted protein 3) in human cartilage. Top panels, Immunohistochemical analysis of WISP3 protein in a superficial (Superf)/midzone section and a midzone section of normal adult articular cartilage, a superficial/midzone section of osteoarthritic (OA) cartilage, and a section of fetal growth plate cartilage, using goat anti-human WISP3 antibody. Bottom panels, Analysis of similar sections as in the top panels, using normal goat antibody. Results are representative of 4 separate experiments.

WISP3 regulation of the expression of cartilage-specific proteins.

Because WISP3 mutations have been linked to defective cartilage formation (10), we hypothesized that WISP3 helps to maintain cartilage integrity by promoting the expression of cartilage-specific matrix proteins. To investigate if WISP3 promotes the expression of the cartilage-specific molecules type II collagen and aggrecan, we transfected the chondrocyte cell lines C-28/I2 and T/C-28a2 with a WISP3 expression plasmid. Because the C-28/I2 and T/C-28a2 chondrocyte lines express low levels of WISP3 (Figure 2C) and cartilage-specific collagens (16) in DMEM supplemented with 10% FBS, they provided a suitable system to investigate WISP3-mediated regulation of chondrocyte-specific gene expression.

Figure 2.

WISP3 (Wnt-1–inducible secreted protein 3)–mediated regulation of cartilage-specific molecules in chondrocytes. A, WISP3-transfected C-28/I2 or T/C-28a2 cells (lane 1) expressed 4-fold higher levels of type II collagen (coll type II) protein than cells transfected with empty vector (lane 2) when normalized to actin levels (P ≤ 0.05). B, WISP3-transfected (W) C-28/I2 cells (lanes 1 and 7) expressed 5-fold higher levels of COL2A1 (Coll II) and aggrecan (AG) mRNA than cells transfected with empty vector (E) (lanes 2 and 8), as detected by reverse transcriptase–polymerase chain reaction with COL2A1- and AG-specific primers. The levels of COL1A1 (Coll I), fibronectin (FN), and G3PDH mRNA (lanes 3 versus 4, 5 versus 6, and 9 versus 10, respectively) did not change in response to WISP3 transfection (P ≤ 0.05). WISP3 mRNA transcribed off the WISP3 expression vector was detected in WISP3-transfected C-28/I2 cells, using WISP3-specific and vector-specific (bgh) primers (lane 11) (see Table 1), but not in C-28/I2 cells transfected with empty vector (lane 12). C, Immunoblot of lysates of 293 cells transfected with either WISP3-hemagglutinin (WISP3-HA) or empty vector, using anti-HA antibody (top), and immunoblot of lysates of C-28/I2–T/C-28a2 cells transfected with either WISP3 or empty vector, using anti-WISP3 antibody (bottom). These immunoblots demonstrate that both antibodies recognize the 40-kd WISP3 protein in WISP3-HA/WISP3-transfected cells. Both the anti-N and anti-C terminal antipeptide antibodies against WISP3 worked in immunoblots. All results represent 4 separate experiments.

Figure 2A demonstrates that transient transfection of C-28/I2 and T/C-28a2 cells with a WISP3 expression plasmid (WISP3-pcDNA3) up-regulates type II collagen protein compared with the housekeeping protein actin (P ≤ 0.05). To determine if WISP3 influences regulation of ECM molecules at the transcription level, we compared the relative levels of COL2A1 and aggrecan mRNA in WISP3-transfected versus empty vector–transfected C-28/I2 cells (Figure 2B). The WISP3-transfected C-28/I2 cells expressed 5-fold higher levels of both COL2A1 and aggrecan mRNA than the cells transfected with the empty vector. In contrast, minimal changes were observed in the levels of mRNA encoding nonspecific ECM proteins type I collagen and fibronectin in response to WISP3 transfection. The levels of G3PDH mRNA also did not change. The recombinant WISP3 gene transcribed off the WISP3 expression vector was expressed in WISP3-transfected cells, but not in empty vector–transfected cells. Figure 2C shows that the recombinant WISP3 protein was expressed in both 293 cell transfectants and C-28/I2–T/C-28a2 transfectants, and was recognized by anti-WISP3 antibody.

WISP3 overexpression increases SOX9 expression and COL2A1–luciferase activity.

The SOX family of transcription factors has been reported to be essential for chondrocyte differentiation and proper skeletal development (13). SOX9, SOX6, and L-SOX5 are all known to regulate the transcription of type II collagen and aggrecan by interacting specifically with elements in the promoters and enhancers of the respective genes (13, 18, 19). To investigate if WISP3 could up-regulate the expression of cartilage-specific transcription factor SOX9, we first assessed SOX9 mRNA expression in WISP3-transfected C-28/I2 chondrocytes (Figure 3A). The level of SOX9 mRNA was up-regulated in the WISP3-transfected cells, compared with cells transfected with empty vector. Moreover, the 4-fold increase in SOX9 mRNA correlated with a similar increase in the level of COL2A1 mRNA but not COL1A1 mRNA. To control for transfection efficiency, all batches of the transfectants were cotransfected with a pCMV-LacZ control vector, and were shown to express equivalent levels of the β-galactosidase reporter gene (Figure 3A).

Figure 3.

WISP3 overexpression increases sex-determining region Y–type high mobility group box 9 (SOX9) mRNA levels and COL2A1-luciferase in chondrocytes. A, C-28/I2 chondrocytes were transfected, as described in Figure 2. Overexpression of WISP3 induced 4-fold higher levels of COL2A1 and SOX9 mRNA (lanes 1 and 3) than transfections with empty vector (lanes 2 and 4). Levels of COL1A1 mRNA did not change significantly (lanes 5 and 6), nor did levels of β-galactosidase (β gal) mRNA in cells cotransfected with the reporter vector (lanes 7 and 8). G3PDH levels in the WISP3 and empty vector transfectants (lanes 9 and 10) were also similar (P ≤ 0.05). B, C-28/I2 cells were cotransfected with the COL2A1–luciferase reporter and WISP3 expression vector (W/R) or with the COL2A1–luciferase reporter and empty vector (E/R). WISP3 overexpression (lane 2) stimulated 2-fold higher levels of luciferase reporter protein compared with the empty vector–transfected cells (lane 1), when normalized to actin protein and β-galactosidase reporter levels (P ≤ 0.05). C, Fold increases in SOX9 mRNA and COL2A1-driven luciferase reporter protein levels in cells transfected with WISP3 compared with cells transfected with empty vector. Values are the mean ± SD from 4 experiments. See Figure 2 for other definitions.

To investigate further if WISP3 could activate SOX transcription factors and thereby increase COL2A1 gene transcription in chondrocytes, we cotransfected the C-28/I2 cells with COL2A1–luciferase reporter and either WISP3 expression vector or empty vector (Figure 3B). The WISP3 expression vector up-regulated COL2A1-driven luciferase protein ∼2-fold compared with the empty vector. Similar levels of the housekeeping protein actin were expressed in both WISP3/COL2A1–luciferase reporter and empty vector/COL2A1–luciferase reporter cotransfected cells. Figure 3C summarizes the fold differences in SOX9 mRNA levels between C-28/I2 cells transfected separately with WISP3 and empty vector; it also shows the fold differences in COL2A1–luciferase reporter protein levels between C-28/I2 cells cotransfected separately with WISP3/COL2A1–luciferase and empty vector/COL2A1–luciferase. Since SOX9 is considered to be the major transcription factor that controls cartilage-specific gene expression (18–20), it is possible that WISP3 mediates activation of the cartilage-specific genes type II collagen and aggrecan mainly through SOX9 activation.

A mutant WISP3 (Cys78–Arg) does not activate chondrocyte-specific genes.

Among the various WISP3 mutations that have been linked to PPRD, the Cys78– Arg mutation alters a conserved cysteine residue in a putative IGFBP motif (Figure 4A). The high conservation of this cysteine residue in other CCN proteins suggests that it may be important for WISP3 function. To investigate the effects of the Cys78–Arg mutation on WISP3 function in chondrocytes, we compared collagen and aggrecan mRNA expression levels in C-28/I2 cells transfected separately with wild-type and mutant WISP3 (Figure 4B). Similar to the results shown in Figure 2B, transfection with wild-type WISP3 increased the levels of COL2A1 mRNA and aggrecan mRNA, whereas transfection with the Cys78–Arg mutant produced no pronounced change when compared with the empty vector (Figure 4B). Moreover, similar levels of COL1A1 and G3PDH mRNA in response to all 3 vectors were observed. In accordance with the mRNA expression profile, the Cys78–Arg mutant did not up-regulate type II collagen protein levels in C-28/I2 cells to the same extent as did wild-type WISP3, when compared with empty vector (Figure 4C). The wild-type– and mutant WISP3–transfected cells expressed similar levels of exogenous WISP3 mRNA (Figure 4B), thus demonstrating similar transfection efficiencies.

Figure 4.

The Cys78–Arg mutation partially abolishes WISP3-mediated effects. A, Sequence alignment of the insulin-like growth factor binding protein (IGFBP) domain of WISP3, which is homologous with the N-terminal domain of IGFBPs, with the Cys78–Arg mutation indicated (∗). B, C-28/I2 cells transfected with wild-type WISP3 expressed 4-fold higher levels of COL2A1 mRNA than cells transfected with Cys78–Arg WISP3 mutant (M) or empty vector (lanes 1–3) (P ≤ 0.05). Levels of COL1A1 and G3PDH mRNA were similar in wild-type WISP3–transfected cells, mutant WISP3–transfected cells, and empty vector–transfected cells (lanes 7–9 and 10–12, respectively). AG mRNA was detectable in WISP3-transfected cells (lane 4), but almost undetectable in mutant WISP3–transfected cells (lane 5) or empty vector–transfected cells (lane 6). Similar levels of expression of vector–specific WISP3 mRNA were detected in wild-type WISP3– and mutant WISP3–transfected cells (lanes 13 and 14), but none was detected in empty vector–transfected cells (lane 15). C, Immunoblots demonstrating that C-28/I2 cells transfected with wild-type WISP3 express 3.5-fold higher levels of type II collagen protein (lane 1) than C-28/I2 cells transfected with the Cys78–Arg mutant or empty vector (lanes 2 and 3), when compared with actin expression levels (P ≤ 0.05). D, Immunoblots demonstrating similar expression levels of wild-type WISP3 (lane 2) and mutant WISP3 (lane 1) in C-28/I2 cells transfected with WISP3-HA and Cys78–Arg WISP3-HA, compared with actin expression levels. See Figure 2 for other definitions.

To confirm that the diminished effect of the mutant WISP3 on chondrocyte-specific gene expression was not due to a defect in its expression at the protein level, we analyzed the level of expression of hemagglutinin-tagged Cys78–Arg WISP3 in lysates of transfected C-28/I2 cells (Figure 4D). The hemagglutinin-tagged mutant WISP3 was expressed to almost the same level as hemagglutinin-tagged wild-type WISP3 in C-28/I2 cells, when compared with actin levels. However, it is possible that the mutant WISP3 is unable to undergo the proper conformational changes or intermolecular interactions that are essential for its effects on chondrocyte-specific gene expression.

DISCUSSION

WISP3 (CCN6), along with WISP1 and WISP2, belongs to the CCN family of connective tissue growth factors that include CTGF, Cyr61, and NOV (1–4, 10). CTGF (CCN1) is a multifunctional growth factor that plays a major role in cartilage growth and endochondral ossification (3, 4, 21, 22). It has been reported that CTGF promotes the hypertrophy of growth chondrocytes and proliferation of articular chondrocytes in culture (21, 22). Cyr61 (CCN2) has been shown to promote chondrogenesis in mouse limb bud mesenchymal cells (23). In maturing cartilage of mouse embryos, Cyr61 is detected mostly in the less mature condensing mesenchyme and perichondrium (23). The ability of Cyr61 to promote angiogenesis and fracture repair suggests that it is also required for osteogenesis (3, 4, 24). Recently, NOV (CCN3) has been described as an angiogenic regulator in human umbilical vein endothelial cells (25). WISP1 (CCN4) binds to small, leucine-rich cartilaginous proteoglycans, such as biglycan and decorin (26), but it is not yet clear if such interactions promote connective tissue growth and maintenance. The physiologic role of WISP2 (CCN5) has not been elucidated in detail. Also, exactly how WISP3 functions in bone and cartilage tissues in comparison with other CCN-family proteins is not well understood, although WISP3 mutations have been linked to skeletal malformations and cartilage loss.

Our immunohistochemical analysis of adult articular cartilage sections with an anti-WISP3 antibody revealed expression of the WISP3 protein in chondrocytes mainly in the midzone, which in fact correlates with type II collagen and aggrecan expression patterns (27). Immunostaining of OA cartilage with the same antibody revealed WISP3 expression primarily in clusters of proliferating chondrocytes. Furthermore, immunostaining of growth plate cartilage sections with anti-WISP3 antibody showed expression mainly in the proliferating chondrocytes, with apparently decreased levels when chondrocytes undergo hypertrophy. This expression pattern is consistent with a role of WISP3 in chondrocyte maturation during skeletal development and adult cartilage repair and homeostasis (28).

Although WISP3 expression was detectable in some ECM regions of OA cartilage, it was mostly intracellular or membrane associated in the other tissues examined. However, based on our immunohistochemistry data, one cannot rule out the existence of ECM-bound or extracellular WISP3 protein under normal conditions in vivo. We have, in fact, been able to detect Myc-tagged WISP3 protein with an anti-Myc antibody in the supernatant collected from Myc-tagged WISP3 expression vector–transfected 293 cells in our initial attempts to purify recombinant WISP3 protein (results not shown). It is quite possible that posttranslational modifications in ECM-bound or extracellular WISP3 protein under normal conditions prevented its recognition by the anti–N-terminal WISP3 antipeptide antibody, which we used for our immunohistochemistry experiments. Immunostaining was not detectable with an anti–C-terminal WISP3 antibody, perhaps because of inaccessibility of the C-terminal target in the native folded state of the protein.

Our transfection studies demonstrate that WISP3 promotes 1) the expression of type II collagen and aggrecan in parallel with increased expression of SOX9 and 2) the activation of the COL2A1–luciferase reporter, bearing SOX binding sites, in C-28/I2 cells that are known to express SOX transcription factors (29). SOX9, together with L-SOX5 and SOX6, regulates the expression of cartilage-specific collagens and proteoglycans (13, 19, 20). Various reports have also substantiated the role of SOX9 as a master regulator of such processes (18, 20). Cotransfections of fetal bovine chondrocyte cultures with SOX expression plasmids and the human COL2A1–luciferase reporter have demonstrated that, although all 3 SOX transcription factors are required for maximal stimulation of COL2A1-driven luciferase expression, only SOX9, but not L-SOX5 or SOX6, could increase expression by itself (18). Our experimental results thus suggest that WISP3 mediates regulation of the cartilage-specific molecules type II collagen and aggrecan via activation of SOX transcription factors, and it is possible that SOX9 is the major regulator of transcription.

The modular architecture of WISP3 suggests that it can interact with multiple proteins, perhaps acting as a scaffold. Since the Cys78–Arg mutation in the putative IGFBP domain of WISP3 partially abolishes WISP3 function in chondrocytes, it is conceivable that WISP3 may act, in part, by augmenting the effects of IGFs. IGF-1–mediated up-regulation of cartilage ECM synthesis has already been reported by investigators at several laboratories (30–32). In addition, preliminary experimental results have demonstrated that an anti–IGF-1 neutralizing antibody partially blocks WISP3-mediated up-regulation of COL2A1 mRNA in C-28/I2 cells. Furthermore, specific binding of IGFs to the CCN-family protein CTGF has also been documented (33). Accordingly, it is possible that WISP3 tethers IGF-1 to the IGF-1 receptor by competing with antagonistic IGFBPs (34, 35) (Figure 5). If this is the case, the stimulation of the IGF-1 receptor tyrosine kinase by WISP3 may contribute to the activation of SOX transcription factors and subsequent up-regulation of chondrocyte-specific gene expression, via a MAP kinase signaling cascade, as has been reported (36).

Figure 5.

Potential mechanism(s) of WISP3 (Wnt-1–inducible secreted protein 3)–mediated signaling in chondrocytes. IGFBPs = insulin-like growth factor binding proteins; IGF-1R = insulin-like growth factor 1 receptor; SOX = sex-determining region Y–type high mobility group box.

There may also be other interactions of WISP3 that are capable of regulating chondrocyte-specific gene expression. In light of the fact that the CCN-family proteins CTGF and Cyr61 interact with cell membrane–bound integrins (3, 4), it is conceivable that interactions of WISP3 with chondrocyte-specific integrins could also influence the expression of SOX transcription factors and subsequent up-regulation of type II collagen and aggrecan. The specific WISP3-mediated signaling pathway in chondrocytes, at any given time, is probably determined by the relative stoichiometry of several available signaling components. This hypothesis could explain why several different WISP3 mutations yield a similar phenotype.

Overall, our experimental results suggest that WISP3 regulates the expression of cartilage-specific extracellular matrix proteins in chondrocytes. In the future, it will be important to investigate how the functions of WISP3 in skeletal tissues are temporally regulated in comparison with other CCN-family proteins. Although WISP3 shares a common modular architecture with other CCN proteins, the variable regions of WISP3 may impart differential specificity and affinity for signaling intermediates at different stages of cellular growth and development.

Acknowledgements

We thank Lilo Creighton, Don Vu, Vivian Wong, Yandong Zhao, Brenda Sanders, and Bob Choy for technical assistance.

Ancillary