The Transcription Factor Protein Sox11 Enhances Early Osteoblast Differentiation by Facilitating Proliferation and the Survival of Mesenchymal and Osteoblast Progenitors*

Background: Sox11 deletion mice exhibit developmental defects, yet the detailed roles of Sox11 in osteogenesis are unknown. Results: Sox11 stimulated proliferation and survival of mesenchymal and osteoblast precursor cells through up-regulation of Tead2, as well as stimulated osteogenesis via Runx2 and Osterix. Conclusion: Sox11 positively regulates osteogenesis. Significance: Novel functions of Sox11 and its target genes during osteogenesis were established. Sox11 deletion mice are known to exhibit developmental defects of craniofacial skeletal malformations, asplenia, and hypoplasia of the lung, stomach, and pancreas. Despite the importance of Sox11 in the developing skeleton, the role of Sox11 in osteogenesis has not been studied yet. In this study, we identified that Sox11 is an important transcription factor for regulating the proliferation and survival of osteoblast precursor cells as well as the self-renewal potency of mesenchymal progenitor cells via up-regulation of Tead2. Furthermore, Sox11 also plays an important role in the segregation of functional osteoblast lineage progenitors from osteochondrogenic progenitors. Facilitation of osteoblast differentiation from mesenchymal cells was achieved by enhanced expression of the osteoblast lineage specific transcription factors Runx2 and Osterix. Morpholino-targeted disruption of Sox11 in zebrafish impaired organogenesis, including the bones, which were under mineralized. These results indicated that Sox11 plays a crucial role in the proliferation and survival of mesenchymal and osteoblast precursors by Tead2, and osteogenic differentiation by regulating Runx2 and Osterix.

The formation of vertebrate skeletal elements exclusively depends on the differentiation of two cell types, osteoblasts and chondrocytes, which are derived from common mesenchymal precursor cells (1). Several transcription factors contribute to the commitment of mesenchymal progenitors to osteogenesis.
The crucial transcription factors shown to be involved in osteogenesis include runt-related transcription factor 2 (Runx2), 2 Sp7/Osterix, Dlx3, and Msx2 (2)(3)(4)(5)(6). In particular, Runx2 has been shown to be an essential transcription factor for early osteoblast differentiation (7). In a previous study, Runx2 knockout mice exhibited a complete lack of mineralized bone due to defects in osteoblast differentiation (8 -10). For the commitment of mesenchymal stem cells to osteochondrogenic progenitors and osteoblast differentiation, another transcription factor Osterix is absolutely required (11)(12)(13). Osterix has been shown to specifically induce osteoblast differentiation and bone formation in vivo. Like Runx2-null mice, Osterix-null mice also exhibited a complete absence of bone matrix and osteoblasts. Osterix-deficient mice demonstrated a lack of mineralized matrix in only the bones formed by intramembranous ossification (11)(12)(13). Furthermore, Osterix was shown to be regulated by Runx2 in a Runx2-dependent and independent manner during osteoblast differentiation (14). However, the upstream regulation thereof, except for Runx2, and the molecular mechanisms underlying the action of Osterix during osteogenesis are less understood.
Sox genes encode transcription factors that contain a high mobility group (HMG), which interacts with DNA binding domains. Based on their protein specificity, Sox proteins are divided into eight different groups (SoxA-SoxH) (15)(16)(17). Sox genes have also been identified as the master genes for the fate determination and cell survival of specific cell types (18). Mes-enchymal progenitor cells give rise to many cell lineages including osteoblasts, chondrocytes, fibroblasts, adipocytes, and myocytes during organogenesis. Only a few studies have shown the molecular requirements for the self-renewality and lineagespecific decisions of mesenchymal and osteoblasts cells. The best known examples of genes involved in the fate decision of mesenchymal progenitor cells are Sox5, Sox6, and Sox9, which function together to specify the fate and differentiation of chondrocytes (19). The transcription factors Sox4, Sox11, and Sox12 belong to the SoxC gene family (20), and share a high degree of identity in both the high-mobility-group domain and a group specific transactivation domain. In previous mouse embryonic studies, these three genes were shown to be highly and concomitantly expressed in multipotent neural and mesenchymal cells. Furthermore, specific inactivation of SoxC family genes demonstrated the essentiality of these genes in neural and mesenchymal progenitor survival (18,21). However, the detailed role of SoxC factors in specifying the fate and differentiation of osteoblasts has not yet been demonstrated.
In a previous study, Sox11-deficient mice exhibited various craniofacial and skeletal malformations; Sox11 (ϩ/Ϫ) mice were small in body size and asplenic, and 40% of these mice had a cleft palate or cleft lips, L4,5 vertebrae that were duplicated, and had a tail that was kinked (22). In these mice, the last two or three sternebrae as well as the xiphoid process were defective and irregularly mineralized. It was also shown that Sox4 (ϩ/Ϫ) and Sox11 (Ϫ/Ϫ) embryos died around E10.5 (18,22). Also, limbs failed to bud, somites were rudimentary, and embryo growth was arrested at the development stage of E8.5 (18). Even though phenotypic evidence of skeletal malformations in Sox11 knock-out mice has already been shown, there are no detailed findings for the precise role of Sox11 in osteogenesis.
In the present study, we demonstrated that Sox11 works as an important regulator of the proliferation and apoptosis of osteoblast precursor and early osteoblast lineage cells. Knockdown of Sox11 in primary calvaria cells reduced the expression of osteoblast fate specifying transcription factors, mainly Runx2 and Osterix. In vivo experiments with antisense morpholino oligonucleotide (MO)-mediated knock down of Sox11 in zebrafish revealed a significant decrease in bone formation and osteogenesis shown by alizarin red staining.
Isolation and Culture of Primary Calvaria Cells-Primary calvaria cells used in this study were isolated from the parietal bones of 3-4 day neonatal mice calvaria (ICR) after serial diges-tion in 1ϫ HBSS digestion medium containing 0.02% collagenase (Invitrogen), 0.05% trypsin (Invitrogen), and 0.53 mM EDTA (Invitrogen). The fractions containing primary cells were collected after the 3rd, 4th, and 5th digestions and these cells were passed through a cell strainer to separate traces of cell debris. Finally, these cells were cultured in ␣-MEM and incubated at 37°C in a humidified atmosphere with 5% CO 2 and 95% air, and the medium was changed every 2-3 days until the cells reached full confluence.
Animals-Adult ICR mice were housed in a temperature controlled condition (22°C) under artificial illumination and standard humidity with access to food and water ad libitum. Embryos and postnatal pups were obtained from time-mated pregnant mice. The presence of a vaginal plug was confirmed as embryonic day 0 (E0). Embryos and tissue samples were harvested at different developmental stages. All animal procedures were approved by the institutional animal research committee.
Western Blot Analysis-Protein samples were extracted from MC3T3-E1, primary calvaria cells, and bone tissues using protein cell lysis buffer (Intron Biotechnology, Korea), containing a protease inhibitor mixture (Roche Diagnostics Korea Co. Ltd.). After sonication of the cells in lysis buffer for 30 -60 s, the protein extracts were centrifuged at 13000 rpm at 4°C for 10 min. The protein extracts were then separated by electrophoresis on a 12% acrylamide gel and then transferred to an activated PVDF membrane (Millipore, MA). The membrane was blocked in TBST buffer (TBSϩ0.1% Tween-20) containing 5% skim milk powder or bovine serum albumin (BSA) for 1 h at room temperature and incubated overnight at 4°C with anti-Sox11 polyclonal antibody and anti-Runx2 antibody (Santa Cruz Biotechnology), Tead2 (Millipore), Osterix (Abcam), p-p53, caspase-3-cleaved (Cell Signaling Technology, Inc.) Bax, Bcl2 and anti-␤-Actin monoclonal antibody (Santa Cruz Biotechnology) separately. After washing, the membranes were transferred to secondary antibody, HRP-conjugated anti-goat, anti-rabbit, and antimouse immunoglobulin (Santa Cruz Biotechnology) for 1 h at room temperature. After washing, the resultant antibody binding to the membrane was detected on x-ray film using an ECL chemiluminescence detection kit (Santa Cruz Biotechnology) following manufacturer's instructions.
Alkaline Phosphatase Staining-Alkaline phosphatase (ALP) staining was performed in osteoblast differentiated cells after the cells were fixed in a citrate-acetone-formaldehyde solution (Sigma) for 30 s and washed once with deionized water at room temperature. These cells were then stained with an alkaline dye mixture prepared according to the manufacturer's instructions with an alkaline phosphatase staining kit (Sigma). The cells were incubated in the alkaline dye mixture for 1 h to overnight at 18 -26°C, and examined microscopically.
Cell Survival and Apoptosis Assays-Cell survival assay was performed after 72 h of siSox11 and control siRNA transfection in MC3T3-E1 and primary calvaria cells using a CCK-8 assay kit following the manufacturer's protocol (Dojindo Molecular Technologies). This assay was performed to determine the number of viable cells during cell proliferation. The number of viable cells was determined after siRNA transfection in a 12-well dish and washing of the cells with 1ϫ PBS and trypsin-EDTA treatment. Finally, the cells were collected in 1 ml of growth medium and counted with a C-Chip DHC-NO1 disposable hemocytometer. An apoptotic assay was performed after knockdown of Sox11 followed by propidium iodide (PI) staining. Stained cells were visualized and photographed with a fluorescent microscope. For absolute quantification of the apoptotic cells, the PI-stained cells were analyzed by flow cytometry (FACS-Calibur flow cytometry systems, Dickinson and Company).
Cytotoxicity Assay-Lactate dehydrogenase (LDH) cytotoxicity assay was performed after 72 h of Sox11 siRNA transfection in MC3T3-E1 and primary calvaria cells. The cytotoxicity effects after knockdown of Sox11 were evaluated by measuring LDH activity with a nonradioactive cytotoxicity assay kit following the manufacturer's protocol (Promega). Stable cytosolic LDH enzyme was released after cell lysis converted a tetrazolium salt into a red color formazan. The color formation was directly proportional to the number of lysed cells.
Colony Formation Assay-Bone marrow cells were flushed from the femurs and tibias of 2-month-old mice, filtered, and seeded in two 35-mm dishes containing a minimum essential medium (␣-MEM) supplemented with 10% fetal bovine serum (FBS) and a final concentration of 1ϫ penicillin/streptomycin (Welgene Inc., Daegu, South Korea). The medium was replaced with fresh medium every 2 days. After 1 week in culture, bone marrow mesenchymal stromal cells were trypsinized and seeded in a 6-well or 12-well dish for siSox11 and siTead2 transfection along with the control siRNA. After 48 h of transfection, the cells were seeded in a 35-mm dish for colony-forming unit fibroblast (CFU-F) assay, and after 7-10 days in cell culture, the cells were fixed and stained with crystal violet. The number and diameter of colonies (colony Ն15 number of cells) were counted and measured microscopically. The colony number showed as mean values from the three replicates of a representative experiment.
Chromatin Immunoprecipitation (ChIP)-PCR-MC3T3-E1 pre-osteoblast and C3H10T1/2 mesenchymal cells were used for ChIP assay. The protocol was slightly modified from the user manual of the ChIP assay kit (Upstate Biotechnology). Crosslinking was performed on ϳ80 -100% confluent cells in a 10-cm dish using 1% formaldehyde in the culture medium mentioned above for specific cell types. Crosslinking was stopped by the addition of 2.5 M glycine for 2 min at room temperature, followed by washing of the cells with cold PBS and resuspension of the cells in 1 ml of ChIP lysis buffer. The resuspended cells were sonicated for 15 min in cycles of 10 s on and off at 4°C. The sonicated product (ϳ500 bp) was then used as the input after reverse cross linking. The precleared lysates were immunoprecipitated with anti-Sox11 antibody (H-290 Sc-20096) and normal rabbit IgG at 4°C overnight on a rocker. Immunoprecipitated DNA was collected and purified with pre-blocked A/ G-agarose beads and subsequently eluted. This DNA was then analyzed by PCR with specific primers to identify the putative Sox11 binding element containing regions in the Osterix intron and Runx2 promoter. The chip primers used for Sox11 binding element regions on Runx2 promoter were M1; F.P: 5Ј-TCCAA-GAGGGCAAAAGAAGA-3Ј, R.P: 5Ј-TCTTTCATCAGCGTG-GACAG, M2; F.P: 5Ј-AGTCACTGTCCACGCTGATG-3Ј, R.P: 5Ј-TTTTATTACGTGGCGGCTCT-3Ј, M3:5Ј-TCCCCAGGC-TAACACTTTTG-3Ј, R.P: 5Ј-ACTGAGTGTGTGGCGTT-CTG-3Ј.The PCR conditions were optimized with each of the primer sets before Chip PCR was performed.
Luciferase Reporter Activity Assay-Osterix intron DNA fragments were amplified from mouse genomic DNA, which were sequenced and cloned into TA cloning vectors (RBC Life Sciences). After sequencing, the fragments were sub cloned into pGL3-promoter vectors after digestion with the restriction enzymes KpnI/BglII. The Runx2 promoter construct used in this study was previously described by our group (23). Point mutations at putative Sox11 binding sites were introduced with an Expand Long Range dNTPack kit (Roche Applied Science, Mannheim, Germany) and point mutations were confirmed by DNA sequencing before analyzing reporter activity. For luciferase reporter activity, MC3T3-E1 cells (3 ϫ 10 4 cells/well) were seeded in 24-well plates before transfection. Transfection was performed with a mixture of 200 ng/well of each plasmid containing regulatory region and the Sox11 expression plasmid (a gift from Prof. M. Wegner) or a control vector, and 10 ng/well of pRL-TK plasmid as a transfection control were transfected with Lipofectamine plus reagent (Invitrogen) according to the manufacturer's protocol. Luciferase activity was measured in the cell lysates after 48 h of transfection using a Promega Dual Luciferase assay kit with a GLOMAX 20/20 luminometer.
Fish Maintenance, Morpholino, and Microinjection-Zebrafish were raised under standard conditions at 28.5°C with a light-dark cycle of 14/10 h. Zebrafish embryos were obtained from a cross between AB wild-type adults. These were cultured in embryonic water with methylene blue to inhibit fungal growth. Antisense morpholino oligonucleotides directed against zebrafish sox11 (sox11a Mo; 5Ј-CG CTGTTGTCCG-TTTGCTGCACCAT-3Ј, sox11b Mo; 5Ј-CTGTGCTCCGTC-TGCTGCACCA TGT-3Ј) and control Mo; 5Ј-CCTCTTACC-TCAGTTACAATTTATA 3Ј were obtained from Gene Tools, Inc. Morpholinos were solubilized in Danieau buffer to a final concentration of 1 mM, and ϳ330 pg of morpholinos was injected into each zebrafish embryo at the 1-or 2-cell stage. The zebrafish embryos were then microinjected with control or a Sox11 specific morpholino oligomer. Microinjection of Sox11 morpholino was carried out under a dissection microscope (Leica) using a WPI microinjector and a picopump controller.
Treatment of Primary Calvaria Cells with Bone Anabolic Agents-To investigate the upstream signal of Sox11, primary calvaria cells were treated with BMP-2, FGF-2 and PTH in a 12-well dish, and cells were harvested after 48 h for RT-PCR analysis. On the other hand, a constitutively active form of BMPR1A plasmid DNA was transfected along with empty vector and samples were harvested after 48 h of transfection for RT-PCR and Western blot analysis.
Statistical Analysis-All data were compared and analyzed with the respective controls using Student's t test and expressed as mean Ϯ S.D. All the experiments were performed at least three times in triplicates. The t test was performed using Sigma Plot software. All p values Յ 0.05 were considered statistically significant.

Endogenous Sox11 Expression in Different Cell Lines and
during Osteoblast Differentiation-Initially, we analyzed the endogenous expression pattern of Sox11 mRNA and protein in different bone cells, ATDC5 cells, calvaria, and long bone tissues of mice by RT-PCR and Western blot analysis (Fig. 1, A and  B). The expression of Sox11 was slightly higher in MC3T3-E1 pre-osteoblast cells and primary calvaria cells than MLOY4 cells and ATDC5 cells (Fig. 1, A and B). The expression of Sox11 was also detected in calvaria and long bone tissues of adult mice. The expression of Sox11 mRNA and protein were relatively high in calvaria tissue compared with the long bone tissue samples (Fig. 1, A and B). Furthermore, immunohistochemical analysis also confirmed expression of Sox11 in bone tissues during embryonic and postnatal stages of development. A detectable level of Sox11 was seen in different bone tissue sections (data not shown). We also studied the expression of Sox11 during osteoblast differentiation in MC3T3E1 cells after treatment with osteoblast differentiation medium. In these cells, Sox11 was expressed in the early days of differentiation, but a gradual decrease in the expression was observed as osteoblast differentiation progressed (Fig. 1C, upper panel); the differentiation of osteoblast cells was confirmed by alkaline phosphatase staining (Fig. 1C, lower panel).
Sox11 Regulates the Cell Viability, Survival, and Proliferation of Primary Osteoblast Cells-To investigate the function of Sox11 in osteoblast cells, we performed cell viability and apoptosis assays in MC3T3-E1 cells and primary mouse calvaria cells after knockdown of Sox11. The cell numbers and viability were remarkably decreased in Sox11 siRNA transfected cells compared with control siRNA (Fig. 2, A and B). Furthermore, LDH enzymatic activity was significantly increased after transfection of Sox11 siRNA, which indicated an increase in cell death due to enhancement of cytotoxicity (Fig. 2C). In addition, Sox11 siRNA induced a significant increase in apoptotic cells (Fig. 2D). Flow cytometry analysis further confirmed a significant difference in apoptotic cells between the control siRNA and Sox11 siRNA transfected cells (Fig. 2E). The mechanism of the apoptotic signaling pathway was studied by Western blotting analysis of Bcl2 family proteins (Fig. 2F), and showed that knockdown of Sox11 in primary calvaria cells led to increased expression of p-p53, cleaved caspase-3 and the pro-apoptotic protein Bax, as well as decreased expression of the anti-apoptotic protein Bcl2 (Fig. 2F). To overrule the concerns of the off target effects of siRNA, we used three different Sox11 siRNAs from different regions of the same gene sequence and performed the cell viability assay. This result shows significant decrease of cell viability in all three Sox11 siRNA-treated cells (Fig. 2G). These results suggested that Sox11 is essential for the proliferation, viability, and anti-apoptosis of osteoblast lineage cells.
Role of Sox11 in Self Renewal and Cell Proliferation and Survival through the Hippo Signaling Pathway-Next, we analyzed the self-renewal capacity of primary bone marrow mesenchymal cells after knockdown of Sox11 and Tead2. There was a significant decrease in the self-renewal capacity of primary bone marrow mesenchymal cells, showed by decreased colony formation after knockdown of Sox11 and Tead2 (Fig. 3, A-C). Altogether, these results suggested that Sox11 is essential for the self renewal of mesenchymal progenitors, as well as for the proliferation and cell survival of osteoblast precursor cells and mesenchymal progenitors.
The transcription factor Tead2, which plays a key role in the Hippo signaling pathway, was identified as an interacting partner with Yap and Taz in vivo and was shown to prevent cell death upon SoxC inactivation (18). Since Sox11 belongs among the SoxC transcription factors, we analyzed Tead2 and other hippo signaling genes after knockdown of Sox11 in primary calvarial cells. These results showed that Sox11 knockdown decreased the mRNA expression of Tead2 and Birc5 (Fig. 3D).
On the other hand, overexpression of Sox11 induced the expression of Hippo signaling pathway genes (Fig. 3F) and furthermore Western blotting analysis showed that knockdown of Sox11 decreased the expression of Tead2 (Fig. 3E) and overexpression of Sox11 stimulates the expression of Tead2 (Fig. 3G). Furthermore, overexpression of Sox11 and Tead2 after knockdown of Sox11 in MC3T3-E1 cells rescued cell proliferation to a significant extent in Sox11 siRNA pre-treated cells, compared with the Sox11 siRNA pre-treated cells transfected with empty vector (Fig. 3H). These results suggested that Tead2 is a downstream target of Sox11 in osteoblasts, and that the cell survival and proliferation of osteoblast cells might be regulated through Tead2 of the Hippo signaling pathway.
Sox11 Regulates the Expression of Osteoblast Lineage Marker Genes-Since there was a progressive decrease of Sox11 expression during differentiation of osteoblasts, we evaluated the expression of Sox11 and other well known osteoblast differentiation markers after treatment with osteoblast induction media in C3H10T1/2 and bone marrow mesenchymal cells (BMScs). These cells were able to differentiate into osteoblasts in the presence of osteogenic media as demonstrated by ALP staining (Fig. 4C). Both Sox11 and Osterix revealed a strong expression during 7 days of osteogenic induction of C3H10T1/2 cells and during 3 days of osteogenic induction of BMScs (Fig. 4B), osterix expression was abruptly decreased at 14 days to normal mRNA levels in C3H10T1/2 cells (Fig. 4A) and detectable protein expression levels in C3H10T1/2 and BMScs (Fig. 4B). Meanwhile, Runx2 mRNA and protein levels were found to be increased on day 7, which decreased day by day at the protein level as differentiation progressed in C3H10T1/2 cells and BMScs (Fig. 4, A and B). Furthermore, knockdown of Sox11 followed by osteogenic differentiation in C3H10T1/2 and BMScs inhibited the differentiation of these cells into ALP positive osteoblasts (Fig. 4D). Since we found a decrease in ALP positive cells after knockdown of Sox11, we further analyzed osteoblast specific markers in primary calvaria cells after knockdown or overexpression of Sox11. As a result, the expressions of the osteoblast specific marker genes Runx2, Col1a1, Osterix, and Ocn were increased after overexpression of Sox11 (Fig. 4, E and F). On the other hand, knockdown of Sox11 decreased the expressions of these osteoblast marker genes, as demonstrated by RT-PCR and qRT-PCR data (Fig. 4, G and H). Taken together, these results suggested that Sox11 plays an important role in osteoblast lineage development and early osteoblast differentiation, and that the segregation of osteoblast lineage cells from mesenchymal progenitor cells is achieved through the expression of Osterix and Runx2.
Osterix and Runx2 Were Found to Be the Direct Target Genes of Sox11 Protein-To study the direct involvement of Sox11 in the induction of Osterix expression, we analyzed the promoter and intron sequences of mouse Osterix. Interestingly, the first intron of Osterix contained two putative SoxC group-binding elements (SBE), one was located at the ϳ750 bp region of the 1st intron (intron1a) and the another one was located at the ϳ3600 bp region of the 1 st intron (intron1b) (Fig. 5A). Chromatin immunoprecipitation (ChiP) assays confirmed that Sox11 interacted directly with the putative binding site located at intron1b region of the Osterix intron. Immunoprecipitates were prepared from C3H10T1/2 cells using a Sox11 antibody. PCR amplification of the immunoprecipitates revealed that Sox11 was strongly bound to the Osterix intron1b region (Fig. 5B; lower panel). To further clarify whether the SBE is involved in Osterix regulation, we generated a wild-type (CTTTGTT) and a mutant (CTGTGGT) Osterix intron1b enhancer (240 bp) construct containing a putative SBE in the pGL3-promoter vector. In addition to the ChiP assay, luciferase reporter assay revealed that the activity of the wild-type Osterix intron1b enhancer element was significantly increased 2-3 fold after overexpression of Sox11 in MC3T3-E1 cells, whereas the construct bearing mutations at the putative SBE site did not show any significant increase in luciferase activity compared with the wild type after Sox11 overexpression (Fig. 5C). We also identified three putative SBEs in the Runx2 promoter region located Ϫ1657 bps upstream from the 5Ј-UTR start site (Fig. 5D). For chromatin immunoprecipitation, the immunoprecipitates were prepared from MC3T3-E1 and C3H10T1/2 cells along with Sox11 antibody. PCR amplification of the immunoprecipitates and Runx2 promoter mutation (TGCAGTG) analysis revealed that Sox11 was bound to the 3rd putative binding element (TACAATG) region of the Runx2 promoter (Fig. 5, E and F). Luciferase assay with pGL3-Runx2-pro showed a significant increase in luciferase reporter activity after overexpression of Sox11 compared with pGL-3 basic and pGL-3-M3 -Runx2-pro mutant (Fig. 5F). Whereas the other mutant Runx2 promoter vectors (M1 and M2) showed wild type like activity suggesting the specificity of Sox11 binding site located on pGL-3-M3-Runx2 promoter (Fig. 5, E and F). Furthermore, the endogenous expression patterns of Sox11, Runx2 and Osterix in calvaria sections revealed that they are co-expressed during E16 and E18 stages of development (Fig. 5G). Altogether, our findings suggested that Sox11 regulates osteoblastogenesis by binding to the SBEs at both the first intron of Osterix and the Runx2 promoter in a sequence-specific manner, and contributes to the commitment and differentiation of osteoblast cells.

Severe Decrease of Ossification and Osteogenic Marker Expression in Sox11
Morpholinos Targeted Zebrafish-To further evaluate our in vitro data regarding the role of Sox11 in osteogenesis, in vivo experiments were conducted with knockdown of Sox11 by applying morpholinos against zsox11a and zsox11b in a zebrafish model. The effect of morpholino and the knockdown efficiency of sox11a and sox11b genes were determined initially by blocking zsox11a-GFP and zsox11b-GFP expression vectors; decreased green fluorescence signals in morpholino-treated embryos confirmed the knock down efficiency and target specificity of each sox11 morpholinos (Fig. 6,  A and B). After injecting the zsox11a, zsox11b, and zsox11a/b morpholinos into the zebrafish embryos, we found that body size and osteogenesis were severely impaired. The decreased osteogenesis in zsox11-targeted zebrafish was shown by alizarin red skeletal staining at the development age of 5.5 days (Fig.  6C). Treatment with the combined zsox11a/b morpholinos demonstrated more severe impairment of bone formation and mineralization, compared with the individually treated ones, zsox11a morpholino or zsox11b morpholino (Fig. 6C). However, increased concentrations of the zsox11a/b morpholinos also impaired cartilage formation. Therefore, we thought that impairment of both bone and cartilage formation was caused by a decrease in the number of osteoblast precursor cells due to enhanced apoptosis and suppression of proliferation after knockdown of Sox11. Furthermore, to demonstrate that the craniofacial bone defects in the zebrafish after sox11 targeted knockdown was due to the effect on osteoblast marker genes, we performed whole-mount in situ hybridization analysis in 3-4 dpf morpholino-treated zebrafish embryos using runx2a, runx2b, and osterix probes. Therein, the expression of runx2a and runx2b was detected in the operculum, brachiostegal ray, dentary, and maxilla of control embryos and control morpholino treated embryos at 3-4 dpf (Fig. 6D). In contrast, runx2a and runx2b expression was found to be decreased in craniofacial bones including the brachiostegal ray, dentary, and maxilla of sox11a, sox11b, and sox11a/b morpholino treated embryos, it was also noticed that the decreased expression of runx2a was greater than runx2b in sox11a/b morpholinos-treated embryos (Fig. 6D). The expression of osterix in the maxilla, ectopterygoid, brachiostegal ray and operculum was found to be decreased in the sox11a, sox11b and sox11a/b morphants at 3 dpf compared with control and control morpholino treated embryos (Fig. 6D). The effects of sox11 on osteoblast marker genes were found to be more severe in sox11a/b morpholino treated zebrafish embryos (Fig. 6D). Taken together, the in vivo evidences strongly supported our in vitro data that Sox11 plays an important role in osteoblastogenesis by promoting the survival and proliferation of mesenchymal cells and osteoblast precursor cells along with further facilitation of early osteoblast differentiation.
BMP-2 Is the Upstream Regulator of Sox11 in Osteoblast Cells-In this study, we showed that Sox11 expression was necessary for the commitment and early differentiation of osteoblasts from mesenchymal progenitors. To investigate the upstream regulatory factors that stimulate the expression of Sox11, we treated primary calvaria cells with different bone anabolic agents, rhBMP-2, FGF-2, and hPTH. After 48 h of treatment, Sox11 expression was stimulated by BMP-2 treatment and overexpression of a constitutive active form of BMPRIA (Fig. 7,  A and B). Furthermore, knockdown and overexpression of Bmpr1a showed correlative mRNA expression patterns for Smad4 and Sox11 (Fig. 7B), indicating the possible involvement of Smad pathway, whereas FGF-2 and PTH did not show any significant change in Sox11 expression (Fig. 7, C and D). Accordingly, these results suggested that BMP-2 stimulates the expression Sox11, and acts as an upstream regulator of Sox11. Based on our data, we proposed a model in which Sox11 plays an important role in the survival and proliferation of mesenchymal progenitors and osteoblast precursor cells via the activation of Tead2, prior to further osteoblast differentiation. Furthermore, we showed that Sox11 participates in the fate decision and lineage specification of osteoprogenitor cells through the binding and up-regulation of master transcription factors of osteogenesis, Runx2 and Osterix (Fig. 7E).

DISCUSSION
In mammals, the SoxC group proteins Sox4, Sox11 and Sox12 are co-expressed in embryonic neuronal progenitors and in mesenchymal cells throughout the development of many organs (15,20). While Sox12 null mice were shown to be viable without any obvious malformation, Sox4-null embryos and Sox11-null mice died from heart malformations and widespread defects (18). Sox11 knock-out mice displayed microphthalmia, open eyelids, cleft palates, cleft lips, hypoplastic lungs, and asplenia (22). Bone phenotypes exhibited severe cranial and noncranial skeletal malformations, as the skull was defectively mineralized and a split in the lumbar vertebrae was identified (18,22). Even though phenotypic evidence for Sox11 knock-out mice has been shown, the precise role of Sox11 in osteogenesis has not been completely investigated, and there is no detailed information on the target genes affected in these functional knock-out mice.
In this study, the expression of Sox11 in E16, E18, and PN14 sections of calvaria was shown by immunohistochemistry, and was localized in the primary spongiosa and periosteum of the long bone sections of E16, E18, and PN14 pups (data not shown). In vitro study also revealed the expression FIGURE 5. Sox11 promotes the transcriptional activity of Osterix and Runx2. A, nucleotide sequences of putative Sox11 binding elements located in Intron1 (Intron1a and Intron1b) of Osterix. The putative Sox11 binding sequences are shown in red color. B, chromatin immunoprecipitation assays were performed using a Sox11 specific antibody or control IgG in cell lysates of C3H10T1/2 cells. Sox11 bound specifically to intron 1b region only (lower panel). C, Luciferase reporter assays were performed to investigate the effect of Sox11 on Osterix intron1b enhancer activity in MC3T3-E1 cells. Sox11 overexpression significantly increased the activity of the wild-type enhancer element. The mutant Sox11 binding element attenuated the reporter activity increased by Sox11 overexpression. Luciferase reporter activity was measured and normalized to pRL-TK activity. D, nucleotide sequence of the Runx2 promoter region upstream to the 5Ј-UTR sequence. The putative Sox11 binding elements are indicated in red color (ttcaaag (M1, M2), tacaatg (M3)). E, chromatin immunoprecipitation assays were performed using Sox11 specific antibody or control IgG in cell lysates of MC3T3-E1 and C3H10T1/2 cells. Sox11 specifically bound to the region containing the tacaatg sequence (M3), not to the ttcaaag sequences (M1 and M2), of chip DNA in MC3T3-E1 cells. F, luciferase reporter assay was performed to investigate the effect of Sox11 on Runx2 promoter (Ϫ1663 Ϫ5ЈUTR) activity in MC3T3-E1 cells. Sox11 overexpression significantly increased wild type Runx2 promoter activity compared with pGL-3 basic and mutant Runx2 promoter (TGCAGTG; pGL-3-M3-Runx2-pro), as well as the other two Runx2 promoter mutants (TGCAGAG (pGL-3-M1-Runx2-pro, TGCAGAG (pGL-3-M2-Runx2-pro). The activity of M1 and M2 mutants was slightly higher than that of wild type (pGL-3-Runx2-pro). Luciferase reporter activity was measured and normalized to pRL-TK activity. Data shown here represent the mean Ϯ S.D. from transfection in four different wells. G, endogenous expression patterns of Sox11, Osterix, and Runx2 in E16 and E18 sections of calvarial bone. Arrows indicate the positive immunostaining signals for Sox11, Osterix, and Runx2. Arrows indicate the immunostaining signals for Sox11, Osterix and Runx2. * indicates p Յ 0.05.
of Sox11 in primary cultured mouse calvarial cells and MC3T3-E1 preosteoblast cells. Induction of osteogenesis in C3H10T1/2 mesenchymal cells stimulated osteoblast differentiation, and we also observed a remarkable increase in the expressions of Sox11, Osterix, and Runx2 at day 7 with detectable expression of alkaline phosphatase. However, the expression of Sox11 decreased to the basal level at day 14, where the expression of ALP reached a maximum (Fig. 4, A-C). In another experiment on the differentiation of MC3T3-E1 cells in the osteogenic induction medium, Sox11 expression was observed between day 1 and day 7, where there was a low amount of ALP expression. However, progression of osteoblast differentiation inhibited the expression of Sox11 abruptly (Fig. 1C). This suggested that Sox11 may be essential during the early stage of osteoblast differentiation.
The expression of Sox11 was shown to be crucial to cell survival and inhibition of apoptosis in several tissues (18). Knockdown of Sox11 in primary cultured calvarial cells and MC3T3-E1 cells induced apoptosis and decreased the cell number and proliferating capacity of both cells (Fig. 2, A-E). Knockdown of Sox11 and Tead2 inhibited the self-renewal capacity of bone marrow mesenchymal cells (Fig. 3, A-C). Furthermore, the number of ALP positive-stained cells decreased with Sox11 knockdown followed by osteogenic induction of mesenchymal cells (Fig. 4D). In a previous study, Tead2 was identified as a direct target of SoxC genes (18). Tead2 is a transcriptional mediator of the Hippo signaling pathway and functions in cell survival of mesenchymal progenitors. Similarly, we also observed a significant decrease in Tead2 expression in primary cultured calvarial cells after knockdown of Sox11 (Fig. 3, D and  E), while overexpression of Sox11 stimulated Tead2 expression (Fig. 3, F and G). This result suggested that Tead2 might be a direct target of Sox11 in mesenchymal and osteoblast progenitor cells and functions in cell survival and proliferation. The Hippo signaling pathway is known to play a central role in the regulation of organ size and tissue homeostasis in different species (24). The transcription factor Tead2 forms a complex with the transcriptional co-activators Yki, Yap, and Taz and bind to target DNA to mediate tissue overgrowth as well as other functions (18). Furthermore, direct manipulation of Tead2 in bone marrow mesenchymal cells also showed decreased colony formation (Fig. 3B, lower panel). Therefore, we speculated that Sox11 ensures the crucial function of proliferation and the survival of mesenchymal and osteoblast progenitors via activation of the Tead2 in the Hippo signaling pathway prior to osteoblast differentiation.
Runx2 is a master transcription factor involved in the commitment of osteoblast and chondroblast lineage cells (8 -10), and Osterix is known to be critically involved in the fate determination of osteochondroprogenitor cells into osteoprogenitor cells. To further elucidate the involvement of Sox11 in osteoblastogenesis, we studied the endogenous expression patterns of Sox11, Runx2, and Osterix after osteogenic induction in C3H10T1/2 mesenchymal cells and also in bone marrow mesenchymal stem cells. The induced expression of Sox11, Runx2, and Osterix was observed at day 7 in C3H10T1/2 cells (Fig. 4, A  and B) and day 3 in bone marrow mesenchymal stem cells (Fig.  4B). Knockdown of Sox11 suppressed the expression of osteoblast marker genes Runx2, Col1a1, and Osterix (Fig. 4, G and H). Since, Sox11 influenced the expression patterns of Osterix and Runx2, we analyzed the promoter and intron regions of Runx2 and Osterix for SoxC group consensus binding elements. Interestingly, we identified that the 1st intron of Osterix contained two SoxC consensus elements, and one of these consensus elements was similar to the one which was previously reported as a Sox11 consensus element in a Tead2 intron (18). The Runx2 promoter had three SoxC consensus elements. Furthermore, chromatin immunoprecipitation analysis confirmed that Sox11 could bind to one of the SoxC consensus sequences present in the 1st intron of Osterix and the Runx2 promoter (Fig. 5). Luciferase reporter activity with these regulatory sequences after overexpression of Sox11 further confirmed the direct regulation of Osterix and Runx2 by Sox11. Altogether these data indicated that the involvement of Sox11 in the commitment of osteoblasts and early osteoblast differentiation is mediated through Osterix and Runx2, respectively. In contrast, the expression of Sox11 was mainly observed before significant expression of chondrocyte specific marker genes after induction of chondrogenic differentiation by ATDC5 cells (data not shown). Furthermore, Sox11 overexpression was unable to stimulate the markers of chondrocyte differentiation (data not shown). This suggested that Sox11 plays an important role in the segregation of osteochondrogenic progenitor cells and facilitates early osteoblast differentiation via up-regulation of Runx2 and Osterix. Along with the changes in osteoblast differentiation markers during knockdown and overexpression of Sox11, we have been investigating the co-operative expression of Sox11 and Sox4 during osteoblast differentiation. Preliminary data suggested that the SoxC group factors Sox11 and Sox4 were expressed in pre-osteoblasts. However, during osteoblast differentiation, Sox4 was remarkably increased, whereas the levels of Sox11 were decreased (data not shown). Our findings also supported a previous report on the involvement of Sox4 in osteoblast differentiation (25). In addition, the co-expression patterns of Sox4, Sox11, and Sox12 in mesenchymal and neuronal tissues were previously reported (20). All of the above results indicated that the SoxC family plays an important role in osteogenesis: early osteoblast differentiation is ensured by Sox11 and late osteoblast differentiation might be regulated by Sox4 co-operatively. However, the co-operative action of SoxC group genes in osteoblast differentiation remains to be investigated in future studies.
Zebrafish were previously shown to be a common and useful model for studying vertebrate development and gene function (23). Morpholino-targeted disruption of Sox11 in zebrafish further supported our in vitro data of the role of Sox11 in osteogenesis. Unlike in mice, Sox11 has two isoforms in zebrafish. Even though the two isoforms of Sox11 in zebrafish have distinct expression patterns, their combined expression is known to be similar to those of chick and mouse embryos (26). In this study, morpholino-targeted knockdown of Sox11a and Sox11b in zebrafish severely impaired bone and cartilage development (Fig. 6C). Compared with the control morpholino treated zebrafish, the size of bone as well as the whole body was smaller in sox11a, sox11b, and sox11a/b combined sox11 knockdown. In addition, the bones were defectively mineralized, and brain development and organogenesis were also impaired. The impairment of bone formation in zebrafish was more evident in sox11a/b combined morpholino treatment (Fig. 6, C and D), and the majority of the fish died when treated with sox11a/bcombined morpholino. We speculated that impairment of both bone and cartilage was due to the loss of precursor cells after knockdown of Sox11.
In this study, BMP-2, not PTH or FGF2, was identified as the upstream regulator of Sox11 (Fig. 7, A-D). Overexpression of the constitutive active form of BMPRIA, which is the activator of BMP-2, strongly stimulated the expression of Sox11. It is well known that BMP-2 plays an important role in osteoblast differentiation by stimulating Osterix expression (14,27). We showed that Sox11 could regulate Osterix expression directly during osteoblast differentiation of mesenchymal cells. From these studies, we discerned that the upstream regulatory factor may be common for both factors during osteogenesis. In a previous study, BMP-2 not only promoted cartilage formation in a mouse cartilage defect model, but also increased the formation of osteophytes (ectopic bone) at a defect site (28). Nell-1 has been reported to exert stimulatory effects on rabbit chondrocyte proliferation and cartilage specific extracellular matrix deposition via activation of Runx2 and inhibition of Osterix (29). BMP-2 and osteogenic medium could influence the expression of Sox11 by acting as stimulatory factors. Altogether, these data showed that Sox11 could stimulate osteoblast lineage specification from mesenchymal cells via up-regulation of Osterix and Runx2. We also speculated that the involvement of Sox11 in proliferation and survival of pre-osteoblast cells and mesenchymal cells might be mediated through the Hippo signaling pathway (Fig. 7E).
In conclusion, we showed that Sox11, a transiently expressed transcriptional factor, increases the pool of mesenchymal and osteoblast progenitors via Hippo signaling pathway to reach an optimal cell mass before proceeding to further the differentiation of osteoblasts. Furthermore, Sox11 plays an important role in the lineage specification of osteochondrogenic mesenchymal progenitor cells via the up-regulation of Runx2 and Osterix by binding to their regulatory regions. Targeted deletion of Sox11 in osteoblast precursors will strengthen our understanding of molecular regulation of osteogenesis and will help to develop new therapeutic targets for tissue engineering in bone metabolic diseases.