Coordinated regulation of the activities of bone morphogenetic protein (BMP) and its inhibitors is essential for skeletal development since loss-of-function experiments show that both BMPs and BMP inhibitory signals, such as noggin, are required to establish proper formation of skeletal tissues. In this paper, we asked how and when noggin would be functional to interact with BMPs during skeletogenesis in mammals. For this purpose, we first analyzed the spatial and temporal patterns of noggin, BMP-2, BMP-4, and BMP-7 expression during early skeletogenesis in mouse embryos. In situ hybridization study revealed that noggin expression was detected at a low level in limb mesenchyme, whereas BMP-7 was expressed at a high level throughout limb mesenchyme 10.5 days postcoitum (dpc) in mouse embryos. One day later, noggin mRNA was expressed at a high level in the prechondrogenic condensations in appendicular and axial skeletal primordia, where sox9 transcripts were also expressed. At this stage, noggin-expressing cells were surrounded by those expressing BMP-7. The chondrogenic cell condensation continued to express noggin transcripts in 12.5 dpc and 13.5 dpc embryos, and again the noggin-expressing cells within the cartilaginous tissue were surrounded by those expressing BMP-7. We further examined interaction of noggin and BMPs by using organ cultures of 11.5 dpc mouse forelimbs and found that implantation of carriers containing BMP-7 protein into the forelimb explants induced noggin expression in the limb mesenchyme. BMP-7 also induced type II collagen and sox9 mRNAs in the same cell population, indicating that noggin induction occurred in the chondrogenic precursor cells. BMP-7 effects on noggin expression were observed in a dose-dependent manner within a dose range of 10–100 ng/μl. These results suggest that BMP-7 induced expression of noggin transcripts within skeletal cell condensation and that this noggin expression in turn could act antagonistically to attenuate BMP action in the early skeletogenesis.
During embryonic development, primordia of skeletal tissues appear at appropriate time points and positions as cellular aggregates in the mesenchymal tissue. The patterning of the skeletal primordia is likely to be governed by secreted signals and transcriptional regulators. For instance, in the appendicular skeleton, implantation of sonic hedgehog (SHH) protein into the anterior margin of the limb bud induces duplication of the digits.(1) Loss of function mutants have demonstrated regulatory roles of Hox genes in the limb skeletal patterning. One of the Hox d cluster genes, Hoxd12, is shown to affect chondrogenic aggregation through the regulation of SHH expression.(2) Fibroblast growth factors are expressed from the apical ectodermal ridge of the limb bud and mimic the functions of the apical ectodermal ridge. Fibroblast growth factors are also required for the proximal-distal axis determination and growth of the limb.(3) During the development of axial skeletons, bone morphogenetic protein (BMP)-4 is expressed in the dorsal mesoderm and dorsal neural tube, and exogenous application of BMP-2 or BMP-4 before the sclerotomal condensation stage results in skeletal malformation.(4)
This positional information is transmitted to the mesenchymal cells and interpreted by them to form cellular aggregation. The transforming growth factor-β superfamily of growth factors has been suggested to be involved in this step. Application of exogenous transforming growth factor-β3 on precondensation limb mesenchymal cells promotes synthesis of type II collagen mRNA in micromass cultures.(5) growth/differentiation factor (GDF)-5/BMP-14 is expressed in the prechondrogenic cell condensations in the limb and its mutation results in brachypodism.(6) In short ear mice, BMP-5 mutation is implicated to be the cause of the defects of cell condensation.(7) Other BMPs are also expressed during the cell condensation stage; however, precise functions of each BMP in early skeletogenesis have yet to be determined.(8–10)
Recently, several important secreted signaling molecules have been shown to perform their functions as antagonists against BMP actions by directly binding to BMP ligands.(11) Among them, noggin is originally isolated as an organizer signal in Xenopus laevis development.(12) Dorsally expressed noggin inactivates the action of BMP-4 by interfering with the binding of BMP ligand to its receptor. In vitro, noggin is able to bind to BMP-2, BMP-4, and BMP-7 protein.(13) In mammals, noggin is expressed in the neural tube, notochord, somites, and later cartilage.(14,15) Noggin-null mutant mice show severe cartilaginous dysplasia and joint defects.(15) In the limb bud of noggin mutant mice, Hoxd13 expression is not affected, showing that noggin may function after the skeletal pattering is determined. In chick embryos, ectopic expression of noggin alters the limb skeletal morphology but does not affect anteroposterior axis pattern.(16,17) In a pluripotent mesenchymal cell line, C1, noggin is expressed at a low level and its expression is up-regulated by BMP.(18) Therefore, noggin may exert its function during early skeletogenesis; however, it has not yet been determined when and where noggin would be functional in embryonic skeletal development.
In this paper, we analyzed the expression pattern of noggin in conjunction with that of BMP-2, BMP-4, and BMP-7 and the early chondrogenic marker, sox9, in mouse development. Based on the endogenous expression patterns of noggin in early cell condensation, we performed organ cultures of 11.5 days postcoitum (dpc) limb buds to search for signals that may regulate noggin expression during the early skeletogenesis. Our results indicated that noggin started to be expressed at the stage of cell condensation and this was closely associated with BMP-7 expression. Moreover, we found that noggin expression in skeletal precursor cells was triggered by BMP-7.
MATERIALS AND METHODS
In situ hybridization and organ culture studies were performed using embryos collected from mating between ICR outbred mice.
Whole mount in situ hybridization
Whole mount in situ hybridization was performed as previously described.(4) Briefly, embryos were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight. The embryos were washed twice in PBS containing 0.1% Tween 20 (PBT), followed by dehydration through a series of methanol/PBT at 4°C (25, 50, 75, and twice with 100% methanol). Dehydrated embryos were stored at –30°C until use.
The following procedures were performed at room temperature with and incubation time of 5 minutes unless noted. The embryos were rehydrated through a graded series of methanol/PBT (75, 50, and 25% methanol) and were bleached in 6% H2O2 in PBT for 1 h. Following three washes with PBT, embryos were permeabilized with proteinase K (10 μg/ml) (Sigma Chemical Co., St. Louis, MO, U.S.A.) for 15 minutes, washed three times with 2 mg/ml glycine in PBT, and refixed with 4% paraformaldehyde + 0.2% glutaraldehyde in PBS for 20 minutes. Following three PBT washes, embryos were incubated with prehybridization buffer (50% formamide, 5× SSC, pH 4.5, 1% SDS, 50 μg/ml yeast tRNA, 50 μg/ml heparin, 0.1% chaps) for 1 h at 70°C. The prehybridization buffer was then replaced with hybridization buffer (prehybridization buffer plus digoxigenin-labeled RNA probe at 300 ng/ml) and incubated overnight at 70°C. Labeling of RNA probes with digoxigenin-11–UTP was performed as described.(4) Probes used were cRNAs for noggin,(15) BMP-2,(8) BMP-4,(8) BMP-7,(19) Sox9(20) and type II collagen.(21)
Hybridized embryos were washed twice for 30 minutes at 70°C with solution I (50% formamide, 5× SSC, pH 4.5, 1% SDS), once for 20 minutes at 70°C with 1:1 solution I/solution II (solution II: 0.5 M NaCl, 10 mM Tris-HCl, pH 7.5, 0.1% Tween 20), twice with solution II, treated for 30 minutes at 37°C with RNAse A (100 μg/ml) (Sigma) in solution II, and washed twice for 30 minutes at 65°C with solution III (50% formamide, 2× SSC, pH 4.5). Following three washes with Tris-buffered saline (TBS) containing 0.1% Tween 20 (TBST), the embryos were blocked with 10% heat-inactivated sheep serum in TBST for 3 h at 4°C. Nonspecific binding was prevented by preincubating anti-digoxigenin Fab/alkaline phosphatase conjugate (Roche, Indianapolis, IN, U.S.A.) in TBST containing 1% heat-inactivated sheep serum and 0.2% heat-inactivated 14.5 dpc mouse embryonic powder for 3 h at 4°C. After overnight incubation at 4°C with preadsorbed antibody which was diluted 1:1500 in TBST with 1% heat-inactivated sheep serum, embryos were washed three times for 30 minutes and overnight at 4°C in TBST + 2 mM levamisole. The buffer was changed three times for 10 minutes with NTMT (100 mM NaCl, 100 mM Tris-HCl pH 9.5, 50 mM MgCl2, 0.1% Tween 20) containing 2 mM levamisole and color reactions were performed in BM Purple AP solution (Boehringer Mannheim) in the dark without rocking. The color reactions were stopped by two washes in PBT + 1 mM EDTA. Stained embryos were kept in the PBT solution at 4°C.
In situ hybridization
In situ hybridization of paraffin sections were carried out based on the protocol kindly provided by Dr. Ebensperger (University of California–San Francisco, San Francisco, CA, U.S.A.). Briefly, embryos were fixed overnight in 4% paraformaldehyde, pH 7.4, in PBS and embedded in paraffin wax. Sections were made 6 μm thick and placed on the 3-aminopropyltriethoxysilan–coated slides. Each section was consecutively placed on four or five independent slides to examine spacial relationship among different probes.
The sections were dewaxed in xylene and rehydrated in PBS and treated with 10 μg/ml proteinase K (Sigma) for 5 minutes. The sections were then refixed in 4% paraformaldehyde in PBS for 20 minutes and washed twice in PBS, followed by incubation with Tris-glycine buffer for 30 minutes. Hybridization was performed at 65°C overnight in hybridization buffer (40% formamide, 5× SSC, 1× Denhardt's solution, 100 μg/ml herring sperm DNA, 100 μg/ml tRNA) containing 1 ng/μl of the digoxigenin-labeled mouse noggin. The slides were washed three times in 5× SSC, twice in 20% formamide/0.5× SSC, and treated with 10 mg/ml RNAse A in NTE (0.5 M NaCl, 10mM Tris-HCl, pH 7.0, 5 mM EDTA). After washing twice in 2× SSC and once in Blocking Solution (100 mM Maleic acid, 150 mM NaCl, 1% Blocking reagent), the sections were incubated with Blocking Solution + 1/3000 antidigoxigenin-antibody (Roche) overnight at 4°C. Then the sections were washed five times with TBS. To visualize, the sections were placed in BM purple AP substrate. After stopping the reaction, they were dehydrated and mounted in Glycerogel (Dako Corp., Carpinteria, CA, U.S.A.).
Organ culture study of forelimb buds
Forelimb buds were dissected out from 11.5 dpc mouse embryos. The limb buds were placed on membrane filters supported by metal grids according to the Trowel technique(22) and cultured in BGJb medium (GIBCO, Rockville, MD, U.S.A.). Then application of BMP proteins with fibrous glass matrix (FGM) was performed. FGM, composed of mainly silica, was used as carrier matrix in this experiment since it has been shown to be more effective delivery substance for BMP than the guanidine HCl-insoluble bone matrix.(23) Small pieces of FGM were soaked in 0.1 μl of PBS supplemented with or without recombinant BMP protein, and a chip of the matrix 50–100 μm average diameter was picked with a no. 5 forceps under a stereomicroscope. Then, the FGM was microsurgically implanted into the anterior, posterior, or middle part of 11.5 dpc limb buds. Induction of noggin by BMP-7 in the limb bud occurred irrespective of the site of implantation.
Human BMP-2 and human BMP-7 (OP1) were kindly provided by Drs. John Wozney (Genetics Institute, Andover, MA, U.S.A.) and Kuber Sampath (Creative BioMolecules, Boston, MA, U.S.A.), respectively. Both BMP-2 and BMP-7 were expressed in CHO cells, and the final purification was performed by reversed phase high-performance liquid chromatography. Recombinant human BMP-2 and BMP-7 were diluted in PBS with 0.1% bovine serum albumin.
We first determined when and where noggin transcripts start to be expressed during skeletal development and compared the noggin expression profile with that of BMP-2, BMP-4, and BMP-7 to elucidate spatial and temporal relationship among these molecules. Whole mount in situ hybridization (WISH) and in situ hybridization on sections (ISH) were performed in 10.5, 11.5, 12.5, and 13.5 dpc mouse embryos by using digoxigenin-labeled noggin, BMP-2, BMP-4, and BMP-7 cRNAs as probes. Expression of sox9, which is known to be expressed in the primordia of cartilaginous tissue, was also monitored as a marker for prechondrogenic cells.
Noggin mRNA was weakly expressed in the limb bud in the 10.5 dpc embryos
In 10.5 dpc embryos, WISH study revealed that weak expression of noggin was observed in the posterior part of the limb bud, where transcripts for BMP-2, BMP-4, and BMP-7 were expressed (Figs. 1A–1D). Noggin transcripts were also expressed in the notochord, medial part of somites, and neural tube (data not shown). In the overlapping area of limb bud, BMP-7 transcripts were expressed in an area larger than that of noggin (Fig. 1B). Both BMP-2 and BMP-4 mRNAs were highly localized in the posterolateral part of limb mesoderm (double arrowhead in Figs. 1C and 1D). BMP-4 transcripts were also expressed in the anterior part of the limb, where noggin expression was not observed (single arrowhead in Fig. 1D).
In 11.5 dpc embryos, noggin mRNA was expressed in the developing skeletal primordia
In 11.5 dpc embryos, cellular condensation starts to form primordia of skeletal tissues. WISH study revealed that noggin mRNA was localized in the appendicular and axial skeletal elements in 11.5 dpc embryos (Fig. 2A). In the sections, expression of noggin transcripts was observed in the core regions of developing skeletal blastemas in limbs and sclerotomal mesenchyme. Many of these regions were also positive for sox9 expression (arrowhead in Figs. 2A–2C, compared with Figs. 2G–2I). We performed ISH on five consecutive sections so that we could compare expression of noggin with that of the other four probes. The area where noggin transcripts were expressed at high levels was indicated by the white dot lines (Figs. 2F, 2I, 2L, and 2O). The noggin expression was also observed at a low level in a more proximal region, where sox9 mRNA was expressed (white arrowhead in Fig. 2C, compared with Fig. 2I). In each skeletal blastema, sox9 transcripts were expressed in larger areas than those of noggin. This observation suggests that noggin is expressed in a certain stage of cells in the same skeletal primordia. Expression of BMP-7 mRNA was observed in the trunk and limb mesenchyme as was the case for 10.5 dpc embryos (Figs. 2D and 2E). In contrast to the expression of noggin and sox9, however, the center of the cellular condensation, marked by the white dot lines, was negative for BMP-7 expression (Fig. 2F). Rather, BMP-7 expression was observed in the peripheral regions which demarcated the skeletal blastemas (arrowhead in Fig. 2F). The cells positive for BMP-2 expression were condensed in the proximal limb mesenchyme; however, its localization was more proximal to that of noggin (Figs. 2K and 2L). In hands and feet, BMP-2 transcripts were expressed in the interdigital areas where noggin expression was not observed (arrowhead in Fig. 2J, compared with Fig. 2A). BMP-4 transcripts were expressed in the limb and axial mesenchyme, though their localization looked more distal to the cellular condensation in limb compared with BMP-7 (arrowhead in Figs. 2N and 2O, compared with Fig. 2F). Thus, in 11.5 dpc embryos, expression of noggin and sox9 transcripts overlapped in the precartilaginous cellular condensations, whereas BMP-7 transcripts were expressed in the mesenchyme surrounding the noggin-expressing cellular condensations.
Noggin transcripts were expressed in 12.5 dpc skeletal blastemas
In the transversal sections of 12.5 dpc embryos, the precartilaginous tissues of ribs and vertebrae were positive for noggin expression(Fig. 3A). Other precartilaginous tissues such as scapulae and limb skeleton also express noggin mRNA, and the noggin expression was overlapped with that of sox9 (data not shown). BMP-7 expression was observed in the peripheral regions which surrounded the noggin-expressing cell condensation, marked by white dot lines (arrowhead in Fig. 3B compared with Fig. 3C). Expression of BMP-4 was observed in the more external connective tissues (data not shown).
Cells expressing noggin transcripts were surrounded by those expressing BMP-7 in the cartilaginous tissue in 13.5 dpc mice embryos
In situ hybridization on saggital sections of 13.5 dpc embryos revealed that both transcripts for noggin (Figs. 4A and 4B) and BMP-7 (Figs. 4C and 4D) were expressed in the vertebral primordia and ribs. At higher magnification, cells expressing noggin mRNA were observed in the core of developing cartilage in vertebrae (arrowhead in Fig. 4B), which was surrounded by the cells expressing BMP-7 (arrowhead in Fig. 4D). In the extraskeletal sites, BMP-7 transcripts were expressed at high levels in the cells along the ventral surface of the spinal cord (arrow in Fig. 4D), presumably the primordia of dura, where noggin expression was not observed. Thus, in 13.5 dpc embryos, BMP-7 expression flanked noggin expression only in cartilaginous tissue similar to the observation in 11.5 dpc and 12.5 dpc embryos.
BMP-7 induced noggin expression in the 11.5 dpc limb primordia
In situ hybridization study revealed that intense noggin expression started in the cellular condensation of skeletal primordia at 11.5 dpc, BMP-7 expressing cells surrounded those of noggin in the skeletal primordia of 11.5, 12.5, and 13.5 dpc embryos, and expression of BMP-7 preceded that of noggin at 10.5 dpc limb bud. These observations on the spatial and temporal relationship between noggin and BMP-7 expression raised the possibility that noggin expression is regulated by BMP-7. To address this point directly, we performed organ culture experiments using 11.5 dpc forelimbs and investigated the effects of BMP-7 on noggin expression (Fig. 5).
Implantation of the carrier FGM with BMP-7 recombinant protein into the explant of forelimb buds in cultures induced noggin expression 24 h after the implantation (Figs. 6A–6C). The area of induction of noggin expression by BMP-7 expanded in a dose-dependent manner up to 100 ng/μl. Furthermore, a low dosage of 10 ng/μl was effective to induce noggin expression in this system (Fig. 6A). As described in the Materials and Methods, one tip of FGM for implantation experiments was a 50–100 μm average diameter; therefore, the volume would be ∼0.125–1 × 10−3 mm3 (see also Fig. 6B). Based on this calculation, even if the carrier contained all corresponding volume of the protein solution, each implantation would deliver < 10 pg of the protein in quantity, suggesting that a small quantity of BMP-7 is enough to induce noggin mRNA expression. BMP-2 at higher concentrations, such as 1700 ng/μl, also induced weak noggin expression (Fig. 6E); however, 100 ng/μl BMP-2 was ineffective (Fig. 6D), indicating that BMP-7 is more potent than BMP-2. Induction of noggin by BMP-7 occurred irrespective of the sites of implantation in the limb bud of this stage (data not shown). Implantation of FGM soaked in PBS alone did not induce noggin expression and served as a control (Fig. 6F).
We then asked if the cell population responding to BMP-7 has characteristics of skeletal precursors. Twenty-four hours after implantation, we fixed the tissue and prepared serial section to examine whether noggin-expressing cells also express prechondroblastic markers (Fig. 7). High levels of ectopic noggin expression were observed in the mesenchymal tissue near the carrier of FGM containing BMP-7 protein (Figs. 7A and 7D). In the adjacent sections, expression of both transcripts for sox9 (Figs. 7B and 7E) and type II collagen (Figs. 7C and 7F) was induced in the regions almost identical to those of the noggin-expressing cells, indicating that BMP-7 implantation induced ectopic noggin expression in the early prechondroblastic cell condensation.
In this paper, we demonstrated that noggin was expressed during the onset of skeletal development in the core regions of cell condensation in limb and trunk as early as 11.5 dpc in the embryos and that the noggin expression overlapped with that of sox9. The noggin expression in the core of skeletal primordia was observed in the 12.5 dpc (Fig. 3) and 13.5 dpc (Fig. 4) embryos and continued up to at least prehypertrophic cartilage of 15.5 dpc embryos (data not shown). Moreover, the cell population where BMP-7 induced noggin expression was also positive for the expression of sox9 and type II collagen. These expression profiles, based on in situ data, strongly suggest the important roles of noggin in early skeletal development. Indeed, noggin mutant mice did not show any defects in the early limb axis patterning and the later cartilage maturation of long bone, suggesting that noggin may be functional during the early stage of skeletal development.(15)
By comparing expression patterns of noggin mRNA with those of BMP-2, BMP-4, and BMP-7, we found that only BMP-7 expression, but not BMP-2 and BMP-4, is associated with noggin expression in the early skeletogenesis. A close spatial relationship between noggin and BMP-7 expression in the 11.5, 12.5, and 13.5 dpc embryos suggests that noggin functions to suppress BMP-7 activity during early skeletogenesis. A similar spatial relationship between BMP-4 and noggin was reported in Xenopus embryos. In the Xenopus gastrula embryos, BMP-4 is expressed in the entire marginal zone mesoderm except for the dorsal region where noggin is expressed.(24) Dorsally expressed noggin suppress BMP-4 activity to establish dorsoventral patterning of the mesoderm in Xenopus.(25) In the skeletal primordia, noggin expression was observed in the cells central and adjacent to those expressing BMP-7 but not in the BMP-7– expressing cells themselves. Possibly BMP-7 expressing cells may not have competence to respond to BMP-7 in an autocrine manner to express noggin. Alternatively, BMP-7–expressing cells may have a feedback mechanism in response to locally high levels of BMP-7 to shut off noggin expression and keep BMP-7 expression on.
The cell population where noggin expression is induced by the implantation with BMP-7 protein showed early chondroblastic characteristics. In the mesoblastic cell line C1 in vitro, we observed the induction of noggin mRNA expression and up-regulation of type II collagen transcripts by the addition of BMP-4/7 heterodimer protein or BMP-7 protein.(18) Moreover, induction of noggin by BMP-7 occurred irrespective of the site of implantation in the 11.5 dpc limb buds. We speculate that noggin induction by BMP-7 is intrinsically programmed in the mesodermal cells and that this noggin expression may be required to attenuate the BMP action during early skeletal cell differentiation.
Using small pieces of FGM we tried to address if a small quantity of BMP is effective in our experimental system of organ cultures. We showed that a minute amount of BMP-7 (< 10 pg) induces noggin expression, whereas a higher dose (more than 170-fold in quantity) was required to exert similar activity when BMP-2 protein was used. Based on these dose-response experiments, we speculated that noggin induction is more specific to BMP-7 than BMP-2 during early skeletogenesis. In accordance with this notion, in mesoblastic C1 cells, 100 ng/ml of BMP-7 protein induced noggin expression, but 100 ng/ml of BMP-2 protein did not.(18) In early skeletogenesis, specific components of BMP receptors, such as ALK2, which transmits BMP-7 signals but not BMP-2 signals,(26) may function to turn on noggin expression. Recently, Merino et al.(17) reported that implantation of affigel blue beads with 1 μg/μl of BMP-2 expanded noggin expression when implanted at the tip of the digits, but not when implanted in the interdigital region in the stage 28 chick embryos. Others also showed that BMP-2 induced noggin expression in rat osteoblasts derived from 22 day fetal rat calvariae.(27) Therefore, later in chondrogenesis and osteogenesis, BMP-2 may be active to promote noggin expression.
To date, several other BMP regulators, including chordin,(28) cerberus,(29) gremlin,(30) and DAN,(31) have been shown to modulate BMP activity as antagonists. We speculate that some of these antagonists may block specific BMP signals, for example, those transduced by BMP-2 but not those transduced by BMP-7. So far, at least 15 family members of BMPs have been reported. To elucidate BMP action in early skeletogenesis, therefore, it should be determined which BMPs and BMP antagonists are expressed during early skeletogenesis.
Noggin null mice showed hyperplasia of cartilage, whereas BMP-7 null mice did not show such a severe phenotype. One possible explanation for this discrepancy is that other BMP isoforms may be functional to compensate for the absence of BMP-7 in the BMP-7 null mice.(10,32) Since the regulation of early skeletogenesis is crucial for the vertebrate, the pivotal function of BMP-7 could be compensated for by the other BMPs such as BMP-5 or BMP-6. Another possibility is that the main player could exist other than BMP-7 to induce noggin expression in the early skeletal condensation.
Based on our observation, we postulated one model to explain how the interaction between BMP and noggin may be functional during the early skeletogenesis in mammals. BMP-7 and/or other BMPs determine the location of noggin and sox9 expression in the site of primary condensation of skeletal blastemas. Then, noggin in turn attenuates the activity of BMP derived from surrounding tissues. This noggin expression persists until the prehypertrophic stage of chondrogenesis, and BMP–noggin interaction may continue during these periods. In conclusion, the appropriate interaction between BMP and an anti-BMP signal noggin may be required to establish the proper skeletal formation.
We are grateful to Dr. Richard Harland for communication of unpublished data and the materials including noggin cDNA. We thank Drs. H. Oppermann of Creative Molecules, Inc., B.L. Hogan, and P. Koopmann for providing OP-1 (BMP-7), BMP-2, and BMP-4, and Sox9 probes, respectively. We also thank Drs. J. Wozney and K. Sampath for providing of recombinant BMP-2 and BMP-7 protein, respectively. This work was supported by the grants from the Ministry of Education, Science and Culture of Japan (#09771531, #11152209, #10044246, #09307034), JSPS, CREST, and NASDA).