The transcription factor SOX9 regulates cell cycle and differentiation genes in chondrocytic CFK2 cells.

SOX9 is a transcription factor that is essential for chondrocyte differentiation and cartilage formation. We stably overexpressed SOX9 cDNA in the rat chondrocytic cell line CFK2. Compared with the vector control, a greater proportion of SOX9-transfected cells accumulated in the G0/G1 phase. This was associated with an increase in mRNA and protein expression of p21(cip1), an inhibitor of cyclin-dependent kinase activity. SOX9 enhanced p21(cip1) promoter activity in a luciferase reporter assay. CFK2 cells overexpressing SOX9 became more elongated and adhesive and demonstrated a shift in cytoplasmic F-actin distribution. N-cadherin mRNA levels were elevated in the SOX9-transfected cells, and SOX9 enhanced N-cadherin promoter activity. By electrophoretic mobility shift assay, nuclear extracts of SOX9-transfected CFK2 cells specifically bound an oligonucleotide comprising an N-cadherin promoter region containing a consensus SOX9-binding motif. The transcriptional activity of SOX9 depended upon nuclear localization signals required for SOX9 nuclear entry. Differentiation of transfected CFK2 cells was accelerated as evidenced by more rapid accumulation of alkaline phosphatase activity, increased production of proteoglycans, and increased calcium accumulation, and this was associated with decreased ERK1 expression. These studies demonstrate that SOX9 alters the rate of cell cycle progression of chondrocytes and their differentiation by enhancing or inhibiting the expression of selected genes, including p21(cip1) and ERK1, and that N-cadherin is an additional direct target of this transcriptional regulator.

Skeletal morphogenesis begins with the induction of undifferentiated mesenchyme to form regions of high cell density or condensations that represent the outlines of future components of the skeleton. Cells in the center of these condensations change their phenotype to form pre-chondrocytes and then chondrocytes capable of producing cartilaginous matrix and forming the cartilage anlage of developing bone (1). A cassette of cytokines and growth factors is involved in the stepwise development and maturation of the cartilaginous growth plate and includes, among others, bone morphogenetic protein, fibroblast growth factor, parathyroid hormone-related peptide, and Indian hedgehog (2,3). Several transcription factors such as HOX and PAX family members are involved in the patterning of skeletal primordia (4 -6). Recently, SOX9, a transcription factor containing a high mobility group DNA-binding domain, has been implicated in commitment of mesenchymal cells to the chondrogenic lineage and to chondrocyte differentiation. Its expression in all pre-chondrocytic and chondrocytic cells during embryonic development is consistent with this role (7)(8)(9). SOX9 null embryonic stem cells are inhibited in their differentiation to chondrocytes and persist as mesenchymal cells (10). SOX9 binding and activation of chondrocyte-specific enhancer elements in the chondrocyte-specific Col2a1 and Col11a2 genes has been demonstrated (11)(12)(13). The former appears to be enhanced by cAMP-dependent phosphorylation (14). SOX9 is also expressed in testis, brain, and heart and is a relative of the mammalian Y-linked sex-determining gene, SRY. In humans, inactivating mutations in the SOX9 gene cause campomelic dysplasia, a malformation syndrome associated with sex reversal and severe abnormalities in skeletal development resulting in neonatal death (15)(16)(17). To explore the mechanism of SOX9 function in chondrocyte development, we overexpressed SOX9 in the chondrocytic cell line CFK2 and examined its effect on cell proliferation and differentiation. Our results demonstrate an effect of SOX9 on cell cycle regulation and identify several key genes that may be modulated by SOX9 activity and contribute to its capacity to enhance chondrocyte differentiation.

unless otherwise indicated.
Construction of Plasmids-A 2-kb HindIII-NotI fragment encoding full-length human SOX9 cDNA (12) was subcloned into the expression vector pcDNA3. This fragment was devoid of most of the untranslated regions and therefore differed in size from the endogenous 4.5-kb SOX9 transcript. The human SOX9 protein has 509 amino acids. A SOX9 deletion mutant (SOX9-Myc⌬NLS) lacking the first 176 amino acids (within which there are putative nuclear localization signals (NLSs)) and encoding a Myc epitope tag at the COOH terminus was generated by polymerase chain reaction using the above plasmid as template. The forward primer was 5Ј-GGATCCGGTCGCCACCATGAAGAACGGGC-AGGCG-3Ј, which encodes an ATG initiation codon (boldface) flanked upstream by a 5Ј-untranslated region Kozak consensus sequence and downstream by codon 177-181 of SOX9. The reverse primer, 5Ј-TCAG-TTGTTCAGGTCCTCTTCGCTAATCAGCTTTTGTTCCATAGAAGGT-CGAGTGAGC-3Ј, encodes the Myc epitope tag (italic) placed downstream of the COOH terminus (amino acid 509). Following polymerase chain reaction amplification with high fidelity Vent polymerase (New England Biolabs Inc.), the product was cloned into the pCRII vector, and the insert was excised with HindIII and NotI and cloned into the polylinker of pcDNA3 to generate SOX9-Myc⌬NLS. A construct (SOX9-Myc) encoding the full-length SOX9 followed by a Myc tag was reconstituted by cloning the full-length SOX9 cDNA into the SOX9-Myc⌬NLS plasmid. The correctness of all recombinants was confirmed by nucleotide sequencing. p21 and N-cadherin Gene Promoter-Reporter Constructs-A p21 construct in which 2.4 kb of the human p21 gene promoter lies upstream of the luciferase reporter gene in the pGL2 vector (20) was kindly provided by Dr. J.-J. Lebrun (McGill University). A 3.1-kb DNA fragment of the chicken cadherin gene promoter was excised from pNDCAT (21) with HindIII and XbaI and cloned into the polylinker of the pBluescript II vector. This was excised as a KpnI-SacI fragment and cloned into the pGL3 vector upstream of the luciferase reporter gene.
For transient transfections, COS-7 cells were maintained in Dulbecco's modified Eagle's medium with high glucose, 10% FBS, penicillin, and streptomycin. Cells were plated at a cell density of 3 ϫ 10 5 cells in six-well Falcon plates 24 h prior to transfection. They were washed with phosphate-buffered saline (PBS) and overlaid with 2 g of SOX9-Myc or SOX9-Myc⌬NLS in 500 l of Dulbecco's modified Eagle's medium with 4 l of LipofectAMINE reagent. After 16 h, the transfection mixture was aspirated, and cells were cultured in complete medium for 18 h.
Analysis of Promoter Activity-Transfected cells were solubilized by mixing in extraction buffer containing 15 mM MgSO 4 , 4 mM EGTA, 1 mM dithiothreitol, 25 mM glycylglycine, and 1% Triton X-100 and kept on ice for 30 min. After centrifugation at 12,000 rpm for 5 min at 4°C, the supernatants were aspirated. For luciferase assay, 45 l of cell lysate buffered in 3 M ATP, 10 mM KH 2 PO 4 (pH 7.8), and 10 mM MgCl 2 were mixed briefly, and the luciferase activity was measured in a chemiluminometer with a 100-l injection of 25 mM luciferin per reaction. For ␤-galactosidase assay, a 1.5 mg/ml solution of o-nitrophenyl-␤-D-galactopyranoside was prepared in buffer containing 60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgCl 2 , and 50 mM ␤-mercaptoethanol. 5 l of cell extract were mixed with 100 l of o-nitrophenyl-␤-D-galactopyranoside solution and incubated at 37°C until the yellow color appeared. Spectrophotometric measurement was made at 420 nm.
Electrophoretic Mobility Shift Assays-CFK2 cells grown in 150-mm plates to ϳ70% confluency were washed with PBS and 1 mM EDTA; suspended in buffer containing 10 mM Tris (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, and protease inhibitors; and slowly swelled. Nonidet P-40 was added to 0.2%, and the suspension was mixed vigorously for 30 s and centrifuged for 30 s at 12,000 rpm. The supernatant was discarded, and the pellet was resuspended in buffer containing 20 mM Tris (pH 7.9), 400 mM NaCl, 1.0 mM EDTA, 1.0 mM EGTA, and protease inhibitors. The mixture was vigorously shaken at 4°C for 1 h and centrifuged at 4°C for 15 min at 12,000 rpm. 10-l aliquots of the supernatant were stored at Ϫ70°C.
DNA-protein interactions were assayed by electrophoretic mobility shift assay. A double-stranded oligonucleotide that corresponds to part of the N-cadherin promoter (uppercase letters) was generated by annealing oligomers 5Ј-ggCCTCATTTACATTGTTGTAACCAAAAGT and 5Ј-ggACTTTTGGTTACAACAATGTAAATGAGG. (The consensus binding sites for SOX9 are in boldface.) G residues (lowercase letters) were added at the 5Ј-ends for labeling with 32 P. 100,000 cpm of 32 P-labeled DNA and 5 g of nuclear extract were used per reaction. The binding buffer was 20 mM HEPES (pH 7.9), 50 mM KCl, 10% (v/v) glycerol, 100 g/ml poly(dI-dC), 0.5 mM EDTA, 0.5 mM dithiothreitol, 1.0 mM phenylmethylsulfonyl fluoride, 0.05% Nonidet P-40, and 300 g/ml bovine serum albumin.
Microscopic Analyses-For phase-contrast microscopy, stably transfected CFK2 cells (1 ϫ 10 5 ) were seeded in 100-mm plates and grown for 5 days, with changes of medium on alternate days. Cells were photographed in a digital image processing center (Northern Eclipse Empix Imaging Inc., Toronto, Ontario, Canada). For immunofluorescent labeling of the actin filaments, stably transfected CFK2 cells were cultured and then washed with PBS and fixed in PLP solution (2% formaldehyde, 0.075 M L-lysine, and 2 mg/ml sodium periodate (pH 7.4)) for 20 min. After several PBS washings, cells were permeabilized with 0.1% Triton X-100 in PBS and again washed with PBS. Cells were stained with 50 g/ml fluorescein isothiocyanate-conjugated phalloidin in PBS for 1 h at room temperature. Cells were again washed several times with PBS to remove the unbound phalloidin conjugate. Mounted slides were photographed in a fluorescent microscope at a wavelength of 513 nm.
Alkaline phosphatase activity was determined in the stably transfected CFK2 cells by histochemical staining. Briefly, cells were grown in six-well plates in RPMI 1640 medium containing 10% FBS for 14 days with changes of medium every 3 days. Cells were fixed in PLP solution for 20 min. Reagent A (20 mg of naphthol AS-MX phosphate in 1 ml of ethylene glycol monoethyl ether) and reagent B (40 mg of fast red in 1 ml of H 2 O) were added to 100 ml of 100 mM Tris malate buffer (pH 9.2) just before staining. Staining was carried out for 30 min at room temperature. Cells were washed and dried and photographed.
For staining of glycosaminoglycans, ϳ8 ϫ 10 4 stably transfected CFK2 cells were seeded in six-well multiwell plates in RPMI 1640 medium containing 10% FBS. Cells were grown for 10 days with changes of medium every 3 days. On day 14, cultures were fixed with 95% ethanol and then stained for glycosaminoglycans with Alcian blue and counterstained with periodic acid-Schiff reagent as described previously (18).
Accumulation of calcium in the CFK2 cells were determined by histochemical staining with alizarin red S. Briefly, stably transfected cells were grown in six-well plates in RPMI 1640 medium containing 10% FBS. After 10 days in culture, cells were fixed in PLP solution for 20 min. Following washing with water, cells were incubated in 2 mg/ml alizarin red S (pH 4.2) for 30 min. Cells were washed in water and dried.
For immunofluorescence, to determine the cellular localization of SOX9, transiently transfected COS-7 cells were trypsinized, seeded in Falcon slide chambers (Becton Dickinson), and incubated at 37°C in a carbon dioxide chamber for 36 h. Cells were then washed twice with PBS, fixed in PLP solution for 20 min, washed with PBS, and permeabilized with 0.5% Triton X-100. Cells were blocked with 10% FBS in Dulbecco's modified Eagle's medium for 30 min, washed, and incubated for 1 h at 37°C with anti-Myc monoclonal antibody 9E10 (1:200 dilution). Cells were subsequently washed and incubated in turn with fluorescein-conjugated sheep anti-mouse secondary antibody (1:300 dilution). Slides were rinsed five times with PBS, mounted in 70% glycerol, examined by fluorescent microscopy, and photographed using a green filter.

Expression of SOX Transcripts in Chondrocytic CFK2
Cells-In initial studies, Northern blot analysis of endogenous SOX9 revealed a hybridization band of the expected size (4.5 kb), indicating that wild-type cells (data not shown) and CFK2 cells stably transfected with empty vector express this transcription factor, although the expression is relatively weak (Fig. 1A). CFK2 cells stably transfected with the 2-kb SOX9 cDNA expressed this transgene at a 10-fold higher level (Fig.  1A). In view of the fact that SOX5 and/or SOX6 may interact with SOX9 to promote chondrogenesis, we assessed the expression levels of these transcription factors. SOX5 (but not SOX6) transcripts were also detected in both wild-type and SOX9- overexpressing cells (Fig. 1B). The levels of endogenous SOX9 transcripts (Fig. 1A), but not of L-SOX5 (Fig. 1B), increased in the SOX9-transfected cells, suggesting positive feedback regulation of SOX9.
Alteration of Cell Cycle Distribution Is Induced by SOX9 -Inamuch as SOX9 is known to enhance chondrocyte differentiation, we assessed whether SOX9 might also inhibit cell cycle progression to facilitate its pro-differentiating function. In CFK2 cells stably transfected with SOX9, an increased proportion of cells (49%) was indeed observed in the G 0 /G 1 phase compared with empty vector-transfected cells (33%) (Fig. 2). Almost 2-fold more SOX9-transfected cells were found in the S phase, but fewer SOX9-overexpressing cells were observed in the G 2 /M phase. We therefore examined the protein levels of several known regulators of the G 0 /G 1 phase by Western blotting (Fig. 3). Although retinoblastoma protein (Rb) levels were unchanged, a doubling in expression of the cell cycle-dependent kinase inhibitor p21 cip1 (p21) was observed, whereas the level of p53 protein was down-regulated in the SOX9-transfected cells.
Northern blotting demonstrated increased mRNA levels of p21 in the SOX9-transfected cells (Fig. 4A). To determine whether this increase occurred, at least in part, as a result of increased gene transcription, we examined the effect of SOX9 on p21 promoter activity. Cotransfection of SOX9 cDNA and a construct in which 2.5 kb of the p21 promoter drives a luciferase reporter gene (20) induced the reporter expression 2-fold compared with transfection of the p21 promoter construct alone and 10-fold compared with cotransfection of SOX9 cDNA and the pGL2 empty vector (Fig. 4B).
We identified two putative NLSs situated at positions 104 -120 and 163-176 of the SOX9 protein (Table I). We examined the requirement of these sequences for entry of SOX9 into the nuclear compartment to exert transcriptional activation. A SOX9-Myc⌬NLS cDNA, in which both NLSs were removed and which was tagged with a Myc epitope, and intact Myc-tagged SOX9 were transiently transfected separately into COS-7 cells. Immunofluorescence revealed nuclear localization of intact SOX9 (Fig. 5A); however, deletion of the NLS abrogated nuclear transport of the protein, which remained in the cytoplasmic compartment (Fig. 5B).
Alteration of Cell Morphology by SOX9 -In view of the fact that SOX9 enhances the condensation of pre-chondrocytic mesenchymal cells, we assessed whether overexpression of SOX9 might alter CFK2 cell morphology and cell adhesiveness. After  (2)). Northern blotting was performed as described under "Experimental Procedures." B, luciferase reporter activity in COS-7 cells transiently transfected with plasmid pGL2 without (pGL2Basic) or with (p21::GL2) 2.4 kb of the p21 promoter linked to the luciferase reporter gene and cotransfected with empty vector expressing SOX9 (SOX9). Transfection and luciferase measurements were performed as described under "Experimental Procedures." Each bar represents the mean Ϯ S.E. of triplicate determinations. This is representative of results obtained from three independent experiments. 7 days in culture, CFK2 cells overexpressing SOX9 became more adhesive to each other and morphologically elongated relative to empty vector-transfected cells (Fig. 6, A-D). The pattern of distribution of the cytoskeletal protein actin was altered by SOX9 overexpression as shown by staining with phalloidin (Fig. 7, A-D), with actin fibers clustered more toward the plasma membrane.  8. SOX9 induces N-cadherin protein and mRNA levels. A, Western blots of N-cadherin (N-Cad) and ␤-catenin (␤-cat) in protein extracts from empty vector-transfected CFK2 cells (Vector) and from two representative clones of CFK2 cells transfected with SOX9 cDNA (sox9 (1) and sox9 (2)). Western blotting was performed as described under "Experimental Procedures." B, Northern blots of N-cadherin and 18 S RNA on total RNA extracted from empty vector-transfected CFK2 cells and from two representative clones of CFK2 cells transfected with SOX9 cDNA.

SOX9 Induces N-cadherin Gene Transcription in CFK2
Cells-Increased levels of the adhesion molecule N-cadherin, but not of ␤-catenin, were observed as CFK2 cells overexpressing SOX9 became more adhesive (Fig. 8A). Northern blotting demonstrated increased N-cadherin gene expression in the SOX9-overexpressing cells (Fig. 8B).
We assessed whether this increase in N-cadherin was due, at least in part, to increased transcription. A 3.1-kb N-cadherin promoter fragment was cloned upstream of the luciferase reporter gene in plasmid pGL3. When this recombinant was transiently transfected into COS-7 cells together with SOX9 or SOX9-Myc cDNA, a 4-fold increase in reporter activity was observed (Fig. 9A). No increase was observed when the Ncadherin promoter-luciferase reporter construct was cotransfected with the SOX9-Myc⌬NLS construct. To determine whether SOX9 transcriptional activation of N-cadherin promoter activity was due to direct SOX9 binding, an electrophoretic mobility shift assay was performed. A double-stranded radiolabeled oligonucleotide encoding a consensus SOX9-bind-ing motif in the N-cadherin promoter bound to nuclear protein from SOX9-transfected CFK2 cells (Fig. 9B, lane 2) and to nuclear protein from SOX9-Myc-transfected CFK2 cells (Fig.  9C, lane 2). Binding was reduced by increasing concentrations of unlabeled oligonucleotide, indicating the specificity of the DNA-protein interaction (Fig. 9B, lanes 3 and 4). Binding to nuclear protein from vector-transfected cells was significantly reduced (Fig. 9B, lane 5). Electrophoretic mobility of the oligonucleotide bound to nuclear protein from SOX9-Myc-transfected cells was retarded in the presence of anti-Myc antibody (Fig. 9C, lane 3), demonstrating the presence of SOX9 in the bound complex.
Stimulation of Differentiation by SOX9 -After 10 days in culture, alkaline phosphatase levels and proteoglycan accumulation were higher in SOX9-overexpressing CFK2 cells than in empty vector-transfected cells (Fig. 10A). Calcium accumulation was also increased in the SOX9-overexpressing cells.
Inasmuch as ERK1 down-regulation has been associated with increased chondrocyte differentiation (27, 28), we exam- ined the effect of SOX9 on ERK1 expression. CFK2 cells overexpressing SOX9 displayed reduced ERK1 protein levels (Fig.  10B). No difference was observed in ERK2. Northern blotting showed that the decrease in ERK1 protein in SOX9-transfected cells was associated with a decrease in ERK1 gene expression (Fig. 10C), whereas no change was observed in the level of the ERK2 transcript. DISCUSSION SOX genes are evolutionarily conserved and play critical roles in decisions of cell fate in a variety of developmental processes. SOX9 is a master regulator of cartilage formation and may act cooperatively with SOX5 and SOX6 to regulate chondrogenesis (29). To assess effects of SOX9 on chondrocyte development, we overexpressed this transcription factor in a rat chondrocytic cell line (CFK2) that normally expresses low levels of SOX9 as well as SOX5. SOX9-overexpressing CFK2 cells exhibited increased levels of the endogenous transcript, suggesting that SOX9 may feed back to enhance its own gene transcription.
During initial stages of chondrocyte differentiation, prechondrocytic mesenchymal cells aggregate and turn on cartilage-specific genes, but presumably must also undergo growth arrest. SOX9 is known to control the initial steps of chondrocyte differentiation, and recent studies have emphasized the role that cell cycle regulators play in the coordination of chondrocyte proliferation and differentiation (30). We therefore examined the effect of SOX9 on the cell cycle distribution of transfected CFK2 cells. We demonstrate here for the first time the role of SOX9 as a cell cycle regulator and provide a molecular mechanism for this action.
In SOX9-overexpressing cells, a greater proportion of the cells accumulated in the G 0 /G 1 phase, although some increase in the S phase was also observed. Although the retinoblastoma protein level remained unchanged between the vector-and SOX9-transfected cells, the level of p53 was significantly reduced (Fig. 3), suggesting that p53 is negatively regulated by SOX9 in chondrocytes. The tumor suppressor gene p53 plays a major role in cell cycle regulation and apoptosis, stalling cells at the G 1 phase when DNA is damaged and allowing it to be repaired before progression through the cell cycle resumes (31)(32)(33). Consequently, the decreased p53 level may have contributed to the enhanced proportion of SOX9-overexpressing cells in S phase. However, the major alteration observed was the increase in cells in G 0 /G 1 together with increased expression of the cell cycle-dependent kinase inhibitor p21 cip1 . Similar to other SOX and high mobility group proteins (34,35), SOX9 binds to a consensus DNA sequence, C(A/T)TTG(A/T)(A/T) (7,12,36). Using a luciferase reporter construct, we showed that SOX9 can activate transcription via a p21 cip1 promoter sequence. The 2.5-kb promoter region of p21 that we used contains three consensus sequences of the type described above, one or more of which may bind the high mobility group box of the SOX9 protein, facilitating the observed transactivation. The action of increased p21 cip1 expression may then promote G 0 /G 1 arrest of increased numbers of the chondrocytic cells and, as with transforming growth factor-␤ (37), facilitate their enhanced differentiation.
Human SOX9, which is 96% identical to the mouse protein, has 509 amino acids. Functional domains include a high mobility group DNA-binding motif at positions 104 -182 and a COOH-terminal transactivation domain at positions 402-509 that is SPQ-rich (15,16,22). There is a PQA domain at positions 339 -379, the function of which is unknown. Signal-dependent nuclear transport of proteins (38,39) can occur by multiple pathways mediated by different cellular factors (40). We noted that two putative bipartite nuclear localization sequences at positions 104 -120 and 163-176 are present in the human SOX9 protein (Table I). Our studies demonstrate the importance of these sequences for nuclear import of the protein (Fig. 5), which is a prerequisite for SOX9 to exert its role as a transactivator (Fig. 9).
CFK2 cells overexpressing SOX9 altered their cell shape, becoming more adhesive; and the distribution of the cytoskeletal protein actin changed. Previous studies have shown the importance of a number of adhesion molecules during condensation of chondroprogenitor mesenchymal cells, including Ncadherin, neural cell adhesion molecule, and fibronectin (41-44). We therefore examined the possible role of N-cadherin as a downstream target of SOX9. N-cadherin is one of a family of FIG. 10. SOX9 stimulates differentiation. A, photomicrograph of CFK2 cultures transfected with empty vector (lower row) or with SOX9 cDNA (upper row) and stained for alkaline phosphatase (ALP), proteoglycans (PG), or Ca 2ϩ . The alkaline phosphatase staining, the Alcian blue/periodic acid-Schiff staining for glycosaminoglycans, and the alizarin red S staining for calcium is described under "Experimental Procedures." B, Western blots of ERK1 and ERK2 in protein extracts from CFK2 cells transfected with empty vector or from two representative clones of CFK2 cells transfected with SOX9 cDNA (SOX9(1) and SOX9 (2)). Western blotting was performed as described under "Experimental Procedures." C, Northern blots of ERK2 (upper panel), ERK1 (middle panel), and 18 S RNA (lower panel) on total RNA extracted from CFK2 cells transfected with empty vector (Vector) or from two representative clones of CFK2 cells transfected with SOX9 cDNA. Northern blotting was performed as described under "Experimental Procedures." calcium-dependent intracellular cell adhesion molecules that, through their interaction with the cytoskeleton, physically link cells with each other and with the extracellular matrix and play a critical role in development (45). N-cadherin protein and mRNA levels were elevated in the SOX9-transfected CFK2 cells. In contrast, the levels of the adhesion molecule ␤-catenin were unchanged. We noted several putative SOX9-binding motifs in the chicken N-cadherin gene promoter, and we examined the capacity of SOX9 to transactivate the N-cadherin gene. SOX9 markedly increased promoter activity, and nuclear extracts from SOX9-transfected CFK2 cells specifically bound an oligonucleotide encoding a SOX9-binding motif derived from the N-cadherin promoter sequence. Consequently, N-cadherin may be an important downstream target of SOX9 in altering chondrocytic cell adhesion.
SOX9 accelerated the differentiation of the CFK2 cells along the chondrocyte pathway (Fig. 10). Previous studies have implicated the modulation of ERK1 in chondrogenesis (27), and we assessed the effect of SOX9 on this protein. Both transcript and protein levels of ERK1 were reduced in SOX9-overexpressing cells, but this was not observed with ERK2. ERK1 downregulation may lie in the pathway of SOX9-induced chondrocyte differentiation. Interestingly, fibroblast growth factorinduced SOX9 expression has been reported to be associated with stimulation of the mitogen-activated protein kinase pathway (46). Reciprocal modulation of the activity of this pathway may therefore be critical in sustaining the levels of SOX9 necessary for differentiation and the capacity of SOX9 to act as a differentiating agent.
Our results show that SOX9 can act both as a transcriptional activator and a repressor of gene expression in chondrocytic cells. This is fully consistent with the recent demonstration that SOX9 regulates target genes by pairing with specific partners (47). Thus, a transcription factor (␣A-crystallin-binding protein-1 with a zinc finger domain) has been shown to interact with SOX9 to negatively regulate a chondrocyte-specific enhancer of the type II collagen gene (48), whereas the interaction between SOX9 and steroidogenic factor-1 leads to the activation of the AMH gene (49). It will be important to investigate whether any of the co-regulators identified to date interact with SOX9 to modulate the reduction we identified in ERK1 or the increase in p21 or N-cadherin expression.
Our studies have identified several additional mechanisms and downstream modulators that are important for the function of SOX9 as a master regulator of chondrogenesis. They emphasize some of the pleiotropic effects that SOX9 exerts to coordinate the complex series of events inherent in chondrocyte differentiation.