TGFβ-mediated FGF signaling is crucial for regulating cranial neural crest cell proliferation during frontal bone development

Skull vault malformations are a major group of congenital birth defects in humans. The cranial neural crest (CNC), a population of multipotent embryonic progenitor cells, plays an integral role in craniofacial morphogenesis. Cranial neural crest cells migrate ventrolaterally from the neural tube,populating the frontonasal prominence and the branchial arches, and contribute extensively to the formation of mesenchymal structures in the head and neck. In the mouse model, it has been elegantly demonstrated that the calvarial bones are of both CNC and mesoderm origin. Specifically, in the case of frontal bone development, the CNC-derived ectomesenchyme is the sole source for populating the frontal bone primordium and its underlying dura mater(Jiang et al., 2002).

Members of the transforming growth factor β superfamily mediate a wide range of biological activities, including cell proliferation, differentiation,extracellular matrix formation and induction of homeobox genes, suggesting that TGFβ signaling is important in pattern formation during embryogenesis. TGFβ ligands signal through their receptors, which are members of a transmembrane serine/threonine kinase family(Massague, 1992; Massague, 1998; Eickelberg et al., 2002). The TGFβ ligand binds to the TGFβ type II receptor and triggers heterodimerization with a TGFβ type I receptor(Massague, 1998). Following heterodimerization, the TGFβ IIR transphosphorylates the type I receptor,resulting in the propagation of a phosphorylation signal to downstream substrates SMAD2 and SMAD3 (Wrana et al.,1994; Derynck and Zhang,1996). To date, overwhelming evidence supports the notion that both TGFβ IR and TGFβ IIR are indispensable in eliciting the biological response of TGFβ (Wrana et al., 1992; Attisano et al.,1993).

Tgfbr2 is expressed in the CNC-derived ectomesenchyme and cranial sutures, and has been recently demonstrated to play an important role during human skull development (Pelton et al.,1990; Lawler et al.,1994; Loeys et al.,2005). We have previously shown that conditional inactivation of the Tgfbr2 gene in neural crest cells result in cleft palate, small mandible, calvarial and cardiovascular defects (such as PTA and aorticopulmonary septum malformations, which are detailed elsewhere) with complete phenotype penetrance. Furthermore, it is clear that the inductive signaling within the CNC-derived dura mater is crucial for both the CNC- and non-CNC-derived calvarial bone development(Ito et al., 2003). Despite this information, we still do not have a comprehensive understanding of the molecular and cellular mechanisms of the TGFβ signaling-mediated frontal bone development.

In this study, we show that TGFβ is required for CNC cell proliferation in both the orbital and the calvarial (lateral) aspects of the frontal bone primordium. However, osteoblast differentiation is dependent on TGFβ signaling only in the calvarial aspect, not the orbital aspect, of the developing frontal bone. In searching for members of TGFβ signaling network that are involved in regulating CNC cell proliferation and differentiation, we show a compromised FGF signaling in the Tgfbr2mutant. FGF signaling plays a crucial role in regulating calvarial development. Fgfr2-deficient mice show compromised osteoblast proliferation (Yu et al.,2003). In Tgfbr2 mutant mice, Fgf2 and Fgfr2 expression are significantly reduced. Exogenous FGF2 can rescue the proliferation defect in the frontal bone primordium of Tgfbr2mutant, thus demonstrating the biological importance of this signaling cascade in regulating CNC cell proliferation. Our study provides a useful animal model towards a comprehensive understanding of normal craniofacial development, as well as CNC-related congenital malformations.

Both Wnt1-Cre transgenic line and R26R conditional reporter allele have been described previously(Danielian et al., 1998; Soriano, 1999). Mating Wnt1-Cre and R26R mice generated transgenic mice with progenies of neural crest cells labeled with indelible β-galactosidase(Chai et al., 2000).

Whole embryos (E9.5 and E10.5) were stained for β-galactosidase activity according to standard procedures, as previously described(Chai et al., 2000).

Mouse embryonic tissue was frozen, sectioned and then stained according to standard procedures. Specifically, mouse tissue was dissected in PBS and fixed by immersion in 0.2% glutaraldehyde solution for 30 minutes at room temperature. Tissue was soaked in 10% sucrose in PBS for 30 minutes at 4°C, incubated in PBS plus 2 mM MgCl₂, 30% sucrose and 50% OCT at 4°C for 2 hours, then frozen in OCT. Sections were cut at 10-20 μm and mounted on polylysine-coated slides. The mounted tissue sections were fixed in 0.2% glutaraldehyde for 10 minutes on ice, rinsed briefly in PBS and rinsed in detergent solution (0.005% NP-40 and 0.01% sodium deoxycholate in PBS) for 10 minutes at 4°C. Thereafter, the slides were washed in PBS for 10 minutes and stained in X-gal staining solution overnight at room temperature in the dark. Sections were counterstained with Nuclear Fast Red and Eosin.

Mating Tgfbr2^fl/+;Wnt1-Cre with Tgfbr2^fl/fl mice generated Tgfbr2^fl/fl;Wnt1-Cre null alleles that were genotyped using PCR primers as previously described(Chytil et al., 2002). All samples were fixed in 10% buffered formalin and processed into serial paraffin sections using routine procedures. For general morphology, deparaffinized sections were stained with Hematoxylin and Eosin.

DNA synthesis activity within the frontal bone primordium was monitored by intraperitoneal BrdU (5-bromo-2′-deoxy-uridine, Sigma) injection (100μg/g body weight) at E12.5, 13.5 and 14.5. One hour after the injection,mice were sacrificed and embryos were fixed in Carnoy's fixative solution and processed. Serial sections of the specimen were cut at 5 μm intervals. Detection of BrdU labeled cells was carried out by using a BrdU Labeling and Detection Kit and following manufacturer's protocol (Boehringer Mannheim). BrdU-positive and total number of cells within the frontal bone primordium were counted from five randomly selected sections per sample. Five samples were evaluated from each experimental group. Student's t-test was applied for statistical analysis. A P value of less than 0.05 was considered statistically significant. TUNEL assay was performed using the In Situ Cell Death Detection (fluorescein) Kit (Roche Molecular Biochemicals) by following the manufacturer's protocol.

In situ hybridization was performed according to the methods of Wilkinson et al. (Wilkinson et al., 1998). Several negative controls (e.g. sense probe and no probe) were run in parallel with the experimental reaction. Details of the experimental procedures are available upon request.

Timed-pregnant mice were sacrificed on postcoital day 14.5 or 16.5 and staged according to external developmental characteristics. Frontal bone primordium explants (six per treatment group) were cultured in serumless,chemically defined BGJB medium according to standard methods (Chai et al.,1994).

For delivery of TGFβ1 or TGFβ2, affi-gel blue beads (BioRad),diameter 50-80 μm, were used. The beads were washed in phosphate-buffered saline (PBS) and then incubated for 1 hour at room temperature in 10 μg/ml TGFβ1 or TGFβ2 (R&D). Heparin-acrylic beads (Sigma, 200 μm diameter) were soaked in FGF2 (100 μg/ml, R&D) for 1 hour and applied into the calvaria explant. Control beads were incubated with in 0.1% BSA. TGFβ- or BSA-containing beads were placed adjacent to the margin of frontal bone primordium. In both wild-type and Tgfbr2^fl/fl;Wnt1-Cre mutant sample, one side of the frontal bone primordium was treated with BSA bead and the other side treated with TGFβ or FGF2 bead.

During skull development, TGFβ ligand and its type II receptor are colocalized within the frontal bone primordium, implying an autocrine function(Fitzpatrick et al., 1990; Pelton et al., 1990; Lawler et al., 1994). The Wnt1 transgene drives Cre expression specifically in the neural crest lineage and was used to generate Tgfbr2 conditional null allele (Chai et al., 2000; Ito et al., 2003). The anatomical features of a frontal bone include the calvarial aspect, which covers the frontal lobe of brain, and the orbital aspect, which forms the roof of bony orbit. At E16.5, all Tgfbr2^fl/fl;Wnt1-cre mutants had identical frontal bone defects, including failed development of the calvarial aspect of skull and a smaller than normal bony rudiment above the orbit (Fig. 1B). We performed histological analyses of the calvaria development at multiple time points in order to characterize the frontal bone defect in the Tgfbr2^fl/fl;Wnt1-cre mutant sample. Apparent condensation of the CNC-derived mesenchyme was observed in the frontal bone primordium of the wild-type embryo at E12.5 (Fig. 1C). There was comparable development of the frontal primordium in the Tgfbr2^fl/fl;Wnt1-cre mutant sample at E12.5(Fig. 1D). At E13.5, the frontal primordium of the wild type was more defined and extended superiorly and laterally (Fig. 1E). In the Tgfbr2^fl/fl;Wnt1-cre mutant, however, the frontal primordium development was retarded (Fig. 1F). Frontal bone matrix was clearly visible in the wild-type sample at E14.5 (Fig. 1G). In the Tgfbr2^fl/fl;Wnt1-cre mutant sample, the calvarial aspect of the frontal bone primordium failed to form bone matrix(Fig. 1H, arrow), while the orbital aspect did so (Fig. 1H,arrowhead). As embryonic development continued, a well-formed frontal bone was present at both the calvarial and orbital aspects of the frontal bone in the wild type sample (Fig. 1I,K). In the Tgfbr2^fl/fl;Wnt1-cre mutant sample, only the orbital aspect of the frontal bone formed, while the calvarial aspect did not(Fig. 1J,L).

In order to test whether a CNC migration defect might have contributed to the deficiency of the CNC-derived frontal primordium, and ultimately the failure of frontal bone development, we crossed the Tgfbr2conditional allele with R26R transgenic mice and generated embryos with Tgfbr2^fl/fl;R26R;Wnt1-Cre genotype. All of these embryos had frontal bone hypoplasia identical to that seen in the Tgfbr2^fl/fl;Wnt1-Cre mutant mice. Whole-mount and sectioned β-gal staining showed no difference in migration and distribution of CNC cells within the frontonasal prominence and the first branchial arch between Tgfbr2^fl/fl;R26R;Wnt1-Cre mutant and the wild-type control embryos from E9.5 to E11.5(Ito et al., 2003). Following the beginning of telencephalon development at E10.5, the CNC-derived mesenchyme was focused above the eye in both the wild-type and the Tgfbr2^fl/fl;R26R;Wnt1-Cre mutant. Histological sections showed that the CNC-derived mesenchyme began to form the frontal bone primordium at E11.5 (Fig. 2A). There was no apparent difference in the contribution of CNC cells within the frontal primordium between the Tgfbr2^fl/fl;R26R;Wnt1-Cremutant and wild-type sample (Fig. 2A,B). Our data suggest that there was no CNC migration defect that might have resulted in a deficient number of CNC-derived mesenchymal cells within the forming frontal primordium and in failure of frontal bone development in Tgfbr2^fl/fl;Wnt1-Cre mutant mice. At E13.5,loss of TGFβ signaling affected the expansion of the CNC-derived frontal primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant when compared with the wild-type control (Fig. 2C,D). Taken together, our data are consistent with a specific requirement for TGFβ signaling in the CNC-derived frontal primordium during calvarial morphogenesis.

To explore the cellular mechanism responsible for the failure of frontal bone development in Tgfbr2^fl/fl;Wnt1-Cre mutant embryos,we investigated whether there was a decrease in cell proliferation, an increase in apoptosis or a combination of both in the CNC-derived frontal bone primordium. Cell proliferation within the CNC-derived frontal primordium, as measured by BrdU incorporation, appeared to be identical between control and Tgfbr2^fl/fl;Wnt1-Cre mutant embryos at E12.5(Fig. 3A,B, Table 1). Because the calvarial aspect of the frontal bone primordium failed to form bone matrix, whereas the orbital aspect did form bone in the Tgfbr2^fl/fl;Wnt1-Cremutant, we decided to examine the cell proliferation activity separately within these two aspects of the frontal bone primordium. At E13.5, there was no apparent reduction in cell proliferation activity within the calvarial or the orbital aspects of the frontal bone in the Tgfbr2^fl/fl;Wnt1-Cre mutant when compared with the wild-type control (Fig. 3C,D,G, Table 1). At E14.5, there was significantly reduced cell proliferation activity (P<0.05) within the calvarial aspect of the frontal primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant[Fig. 3F (arrow), 3G, Table 1]. Interestingly, a cell proliferation defect was also detected in the orbital aspect of the frontal bone in the Tgfbr2^fl/fl;Wnt1-Cre mutant [Fig. 3F (arrowhead), 3G, Table 1]. Taken together, our study suggests that the cell proliferation defect is responsible for the failure of frontal primordium bone formation and reduction in the size of the orbital aspect of the frontal bone (Fig. 3G). We conclude that TGFβ signaling specifically controls CNC cell proliferation during frontal bone development.

To investigate whether an increase in apoptosis might have contributed to the failure of frontal bone development in the Tgfbr2^fl/fl;Wnt1-Cre mutant, we performed an apoptosis analysis. At E13.5, no apparent apoptotic activity was detected in the frontal bone primordium of either the Tgfbr2^fl/fl;Wnt1-Cre mutant or the wild-type samples (Fig. 4A,B). Sporadic apoptotic signals were present within the frontal bone primordium of both the wild type and the Tgfbr2^fl/fl;Wnt1-Cre mutant at E14.5, suggesting that loss of TGFβ signaling did not affect the survival of the CNC-derived mesenchyme during frontal bone development(Fig. 4C,D). In summary, our data suggest that the primary function of TGFβ signaling is to control cell proliferation within the CNC-derived frontal primordium during calvarial morphogenesis.

To test for an osteoprogenitor differentiation defect that might be responsible for the failure of frontal bone development in the Tgfbr2^fl/fl;Wnt1-Cre mutant, we performed an osteogenic marker analysis. As shown in Fig. 1, the frontal bone primordium first appeared to be smaller than the wild-type control at E13.5. We investigated the differentiation status of the CNC-derived osteoprogenitors starting at this stage. There was comparable Runx2, type I collagen, osterix and Ibsp (previously known as BSP) expression in the developing frontal primordium between the wild-type and Tgfbr2^fl/fl;Wnt1-Cre mutant samples(Fig. 5A-H), suggesting normal osteoblast differentiation in Tgfbr2^fl/fl;Wnt1-Cre mutant at E13.5. At E14.5, the number of cells expressing Runx2 reduced in the calvarial aspect of the frontal bone primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant when compared with the wild-type control (Fig. 5I,J). These data suggest that there was a defect in osteoprogenitor cell differentiation within this cell population in the absence of TGFβsignaling. Interestingly, analysis of two of the early osteogenic lineage markers revealed comparable expression of type I collagen and osteonectin in the frontal primordium of both the wild-type control and the Tgfbr2^fl/fl;Wnt1-Cre mutant(Fig. 5K-N), suggesting the presence of early osteoprogenitor cells in the frontal primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant mice.

Osterix is a zinc-finger-containing transcription factor that is required for osteoblast differentiation and acts downstream of RUNX2 during osteoblast differentiation (Nakashima et al.,2002). The expression of osterix was detected in the calvarial aspect of the frontal primordium in the wild-type sample(Fig. 5O, arrow). No osterix expression was detected within the calvarial aspect of the frontal primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample, suggesting a defect in osteogenic progenitor differentiation(Fig. 5P, arrow). Although reduced, osterix expression was evident in the orbital aspect of the frontal bone (Fig. 5P, arrowhead),suggesting that the osteoprogenitor differentiation within this portion of the frontal bone was not dependent on TGFβ signaling. To evaluate osteoprogenitor terminal differentiation and bone matrix maturation, we examined the expression of late osteogenic markers. There was virtually no detectable expression of osteopontin, Ibsp, osteocalcin and ALP in the calvarial aspect of the frontal primordium (arrow) in the Tgfbr2^fl/fl;Wnt1-Cre mutant, suggesting a bone matrix maturation defect in the mutant mice (Fig. 5Q-X). Overall, our osteogenic marker analysis indicated that the CNC-derived osteogenic progenitor cells were unable to complete the differentiation process (e.g. matrix maturation and mineralization) in the calvarial aspect of the frontal primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant.

As shown in this study, loss of Tgfbr2 affects osteoprogenitor cell proliferation and differentiation in the Tgfbr2^fl/fl;Wnt1-Cre mutant. The identity remains unclear of the molecular signaling cascade involved in regulating osteoprogenitor cell proliferation and differentiation during frontal bone development. To investigate this molecular mechanism, we examined the expression of genes known to be crucial in regulating skull development. Recent studies suggest that FGFR2 is crucial for osteoprogenitor cell proliferation during skull development and conditional inactivation of Fgfr2 results in a decrease in osteoblast proliferation and bone formation (Iseki et al., 1999; Yu et al., 2003). To test whether TGFβ and FGF signaling cascade may work together to regulate cell proliferation, we compared the expression of Fgfr2 between the wild-type and the Tgfbr2^fl/fl;Wnt1-Cre mutant samples. At E13.5, the expression of Fgfr2 in the frontal primordium was reduced in the Tgfbr2^fl/fl;Wnt1-Cre mutant when compared with the wild type (Fig. 6A,B). At E14.5, the defect of the frontal primordium was more evident and there was no detectable Fgfr2 expression in the calvarial aspect of the frontal primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample(Fig. 6D). In the wild-type sample, however, active Fgfr2 expression was detected within the frontal primordium (Fig. 6C). Interestingly, the expression of Fgfr2 was also reduced in the orbital aspect of the frontal bone in the Tgfbr2^fl/fl;Wnt1-Cre mutant, supporting the argument that TGFβ mediated Fgfr2 expression was crucial for osteoprogenitor proliferation (Fig. 6D,arrowhead). Because there was no change in apoptotic activity within the frontal primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant, our data suggest that TGFβ-mediated FGF signaling is involved in controlling osteoprogenitor proliferation during calvarial development.

Dlx5 is known as a crucial mediator of calvarial osteoblast differentiation. Loss of Dlx5 results in multiple craniofacial defects including agenesis of frontal bones(Depew et al., 1999; Tadic et al., 2002; Holleville et al., 2003). The similarity in frontal bone development defect between the Tgfbr2^fl/fl;Wnt1-Cre and Dlx5 mutant mice suggests that TGFβ and DLX signaling network may function together to regulate osteoblast differentiation during calvarial morphogenesis. To test this hypothesis, we examined the expression of Dlx5 in the frontal primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant samples. At E13.5, well-defined expression of Dlx5 was detected within the CNC-derived frontal primordium of the control sample(Fig. 6E). Dlx5expression was also detected in the frontal bone primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant(Fig. 6F). As bone matrix began to form, Dlx5 was continuously expressed in the frontal primordium of the wild-type sample at E14.5 (Fig. 6G). In the Tgfbr2^fl/fl;Wnt1-Cre mutant,however, Dlx5 expression was not detectable on the calvarial aspect of the frontal primordium (Fig. 6H). Interestingly, the osteogenic differentiation defect became apparent in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample at this stage (Fig. 5), suggesting that TGFβ-mediated Dlx5 expression may play a crucial role in controlling osteoprogenitor cell differentiation during calvarial morphogenesis. However, Dlx5 expression was unaffected in the orbital aspect of the frontal bone in the Tgfbr2^fl/fl;Wnt1-Cremutant (Fig. 6H, arrowhead). The fact that osteoblast differentiation occurred successfully in this region(Fig. 5) suggests that TGFβ signaling is dispensable for osteogenesis at the orbital aspect of the frontal bone.

Twist1 (previously known as Twist) is a crucial regulator of osteoblast differentiation in mice and humans(Jabs, 2001). Loss of Twist1 function results in defects in the head mesenchyme, failure of neural tube closure in the cranial region, and other abnormalities(Chen and Behringer, 1995). An important function of TWIST1 is to suppress Fgfr gene expression, thereby inhibiting osteogenic differentiation of the cranial suture mesenchyme, thus maintaining the potency of cranial sutures during craniofacial development. During frontal bone development, Twist1 is expressed in the undifferentiated CNC-derived mesenchyme prior to osteoblast differentiation(Rice et al., 2000). In the Tgfbr2^fl/fl;Wnt1-Cre mutant, elevated Twist1expression was detected in the frontal bone primordium where bone matrix failed to form at E14.5 (Fig. 6J, arrow). In the wild-type sample, bone matrix was formed in the frontal bone primordium and there was no detectable Twist1 expression at E14.5 (Fig. 6I), suggesting downregulation of the Twist1 gene following osteoblast differentiation. In the orbital aspect of the frontal bone where bone matrix was detected in both the Tgfbr2^fl/fl;Wnt1-Cre mutant and the wild type, there was no detectable Twist1 expression(Fig. 6I,J, arrowhead). These data suggests that TGFβ signaling is required to suppress Twist1expression in order to achieve normal CNC proliferation and differentiation during frontal bone development.

Based on the analysis of CNC cell defects in the frontal bone primordium of Tgfbr2^fl/fl;Wnt1-Cre mutant, TGFβ signaling is crucial for the CNC-derived osteoblast proliferation. To further test this hypothesis, we treated frontal bone primordium explant with either TGFβbeads or BSA control beads and evaluated CNC cell proliferation activity. E13.5 (n=6), E14.5 (n=10) or E16.5 (n=14) frontal bone primordium explants treated with BSA beads for 1 day showed normal frontal bone primordium development with few proliferating cells, as indicated by BrdU staining (Fig. 7A,D,E16.5 sample was shown for better morphology). By contrast, both TGFβ1(n=10 explants) and TGFβ2 (n=7 explants) beads were able to dramatically expand the frontal bone primordium by increasing cell proliferation activity (three- to fourfold increase in BrdU incorporation as compared to the control) (Fig. 7B,C,E,F). Based on both our in vivo and in vitro analyses, we conclude that TGFβ signaling is important for expansion of frontal bone primordium by regulating CNC proliferation.

We hypothesized that TGFβ is crucial for regulating the expression of FGF ligand that itself induces Fgfr2 expression. E14.5 frontal bone primordium explant treated with BSA beads (n=6) showed normal levels of Fgfr2 and Fgf2 expression(Fig. 7G, data not shown). TGFβ1 beads treated explants (n=6) showed a dramatic increase in Fgfr2 expression, suggesting that TGFβ signaling can induce FGF signaling in the developing calvaria (Fig. 7H). Finally, in order to test whether FGF signals downstream of TGFβ in regulating CNC proliferation in the frontal primordium, we performed experiments using calvarial explants from the Tgfbr2^fl/fl;Wnt1-Cre mutant sample. BSA beads(n=4) were unable to restore cell proliferation when compared with the wild-type control (Fig. 7I). FGF2 beads treated explants (n=6) showed dramatic increase of Fgfr2 expression, restored cell proliferative activity back to normal, and greatly expanded the cell mass in the Tgfbr2^fl/fl;Wnt1-Cre mutant frontal bone primordium explant (Fig. 7J). These results demonstrate that FGF signaling is a crucial downstream component of the TGFβ pathway/network in regulating CNC proliferation during frontal bone development.

Finally, in order to test whether TGFβ mediated FGF signaling is dependent on the presence of TGFβ IIR, we compared TGFβ induced Fgfr2 expression in the cultured frontal and the parietal primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant sample. We used the parietal bone primordium because it is mesoderm derived. The conditional inactivation of Tgfbr2 targeted the neural crest derived cells,therefore, the function of TGFβ IIR was intact in the parietal bone primordium. E14.5 or E16.5 frontal bone primordium explants (n=4)treated with TGFβ beads failed to show an induction of Fgfr2expression (data not shown), while parietal bone primordium explants of the Tgfbr2^fl/fl;Wnt1-Cre mutant treated with TGFβ showed a dramatic increase in Fgfr2 expression(Fig. 7K). Our data suggest that TGFβ IIR is indispensable in mediating TGFβ regulated FGF signaling and that the CNC-derived dura mater provides inductive signaling in regulating mesoderm-derived parietal bone development.

The development of the skull depends upon the coordinated actions of multiple growth and transcription factors that regulate the fate of the CNC-or mesoderm-derived cells (Hall and Ekanayake, 1991; Francis-West et al., 1998; Wilkie and Morris-Kay, 2001). The important findings of this study include that TGFβ signaling is specifically required in the CNC-derived frontal bone primordium during skull development. TGFβ IIR appears to be a crucial regulator for proliferation and differentiation of osteoblasts, but is not required for the survival of osteoblast progenitors. Most importantly,TGFβ-mediated FGF signaling controls CNC proliferation in the skull. This discovery provides a link between two important families of signaling molecules and begins to provide a cellular mechanism for how different growth factors may work together to exert their specific regulatory function during frontal bone development.

Tgfbr2 is specifically expressed during the condensation,proliferation and differentiation stages of the frontal bone development,suggesting that it has a crucial role in regulating osteogenesis(Lawler et al., 1994; Iseki et al., 1995). Previous studies showed that TGFβ signaling plays a pivotal role in regulating postnatal calvarial development and repair. In particular, TGFβ signaling stimulates osteogenic progenitor cell proliferation and can induce premature suture obliteration in cultured fetal rat calvaria(Opperman et al., 2000). In addition, TGFβ signaling in the immature dura mater (in newborn and immature animals) can induce calvarial bone repair, while diminished TGFβsignaling in the mature dura mater fails to repair calvarial defects,suggesting that TGFβ signaling is a crucial regulator for calvarial ossification (Greenwald et al.,2000).

Deletion of any one Tgfb isoform is probably compensated for by the presence of the other, as the three TGFβ isoforms are expressed in overlapping patterns (Pelton et al.,1991; Dickson et al.,1995; Sanford et al.,1997). Inactivation of both Tgfb2 and Tgfb3results in severe developmental defects that include craniofacial malformations and early embryonic lethality, suggesting that the concerted function of TGFβ isoforms in regulating craniofacial development(Dunker and Krieglstein,2002). More recently, an animal model with a compound mutation of SMAD2 and SMAD3 has demonstrated a critical function of TGFβ signaling in regulating craniofacial development (Liu et al., 2004). As there is no CNC migration defect in the Tgfbr2^fl/fl;Wnt1-Cre mutant, our observations are consistent with a general model in which local signals regulate the fate and function of neural crest cells as they migrate to their final destinations. TGFβ is one such signal and is required for the proper development of the CNC-derived frontal primordium during calvarial morphogenesis.

TGFβ is known to regulate the fate of multipotential progenitor cells instructively by regulating the expression or function of tissue-specific growth and transcription factors (Moses and Serra, 1996). We provide evidence that TGFβ signaling is specifically required in the CNC-derived frontal bone primordium during initial calvarial morphogenesis. Loss of Tgfbr2 affects osteoprogenitor cell proliferation. Although osteoprogenitor cells produce bone matrix proteins, their main function is to maintain the progenitor pool through cell proliferation (Yu et al.,2003). It is important to point out that, although both the calvarial and orbital aspects of the frontal primordium show a significant defect in cell proliferation, only the calvarial aspect of the frontal bone primordium fails to form, while the orbital aspect is developed (although smaller than the one of the wild type) in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample. This discrepancy may reflect differential regulatory mechanisms for osteoprogenitor differentiation between different regions of the frontal bone.

Fgfr2 gene expression is clearly downregulated in the frontal bone primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant. Fgfr2 is one of the earliest genes expressed in the condensing mesenchyme and its expression can be induced by FGF2(Delezoide et al., 1998; Orr-Urtreger et al., 1991; Ornitz and Marie, 2002; Yu et al., 2003). At the functional level, FGFR2 signaling regulates proliferation of osteoprogenitor cells and the anabolic function of mature osteoblasts, which apparently has also been compromised in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample. In addition, the expression of Fgfr1 is also dramatically reduced in the frontal bone primordium (data not shown). As both Fgf2and Fgfr2 expression are dramatically reduced and exogenous FGF2 restores Fgfr2 expression as well as rescues CNC proliferation in the cultured frontal bone primordium in Tgfbr2^fl/fl;Wnt1-Cremutant, our study suggests that TGFβ-regulated FGF signaling is crucial for CNC proliferation and provides the first in vivo model to demonstrate the crucial role of this signaling cascade in regulating the fate of CNC cells during calvarial morphogenesis. Besides using FGF2, we have also attempted to rescue cell proliferation defect by introducing other morphogens (such as BMP4) into the frontal primordium of Tgfbr2^fl/fl;Wnt1-Cremutant sample. To date, only exogenous FGF signaling was able to restore cell proliferation activity in the mutant sample, demonstrating the functional specificity of TGFβ-mediated FGF signaling in regulating osteoprogenitor proliferation.

TGFβ is deposited at high level in the forming skeletal tissue and has important effects on bone cell function during development and repair(Filvaroff et al., 1999). Loss of Tgfbr2 in neural crest derivatives affects the differentiation of osteoblast cells during frontal bone development. The presence of early osteogenic differentiation markers (such as type I collagen and osteonectin)in the frontal primordium suggests that the CNC-derived frontal mesenchyme is present but fails to differentiate into an osteogenic lineage in the Tgfbr2^fl/fl;Wnt1-Cre mutant at E14.5(Fig. 5). The failure to express mature osteogenic differentiation makers (such as osterix, osteopontin or osteocalcin) in the CNC-derived frontal bone primordium suggests that TGFβ signaling is required for osteoprogenitor cell differentiation. Conversely, the orbital aspect of the frontal bone anlagen expresses all osteogenic differentiation markers and can form bone in the Tgfbr2^fl/fl;Wnt1-Cre mutant, demonstrating a differential regulatory mechanism during osteogenesis.

Of the possible downstream transcription factors that may be controlled by TGFβ signaling and are crucial for osteoblast differentiation, Dlx5 and Twist1 expression are clearly affected in the frontal primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant. Dlx5 plays a crucial role in skull formation(Depew et al., 1999; Acampora et al., 1999). Targeted inactivation of Dlx5 results in multiple craniofacial defects, including frontal bone malformations(Depew et al., 1999). Several lines of evidence suggest that Dlx genes are crucial for regulating osteoblast differentiation. First, Dlx single (e.g. Dlx5) or double(Dlx5/Dlx6) null mutation results in abnormal skull development(Depew et al., 1999; Depew et al., 2002). Second, Dlx5 is crucial for regulating osteocalcin expression and promoting the maturation of bone cell phenotypes(Ryoo et al., 1997; Newberry et al., 1998; Depew et al., 1999). Third, a recent study suggests that Dlx5 binds directly to the alkaline phosphatase promoter to stimulate Alp expression in response to BMP2 treatment. Msx2 counteracts the osteogenic-inducing property of Dlx5 during osteoblast differentiation(Kim et al., 2004). Our study shows that Dlx5 expression is significantly suppressed in the frontal bone primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant. Taken together, we propose that TGFβ-mediated Dlx5 expression plays a crucial role in regulating the terminal differentiation of osteoblast during calvarial morphogenesis.

TWIST proteins are basic helix-loop-helix (bHLH)-containing transcription factors. Haploinsufficiency of Twist1 results in craniosynostosis and suggests that Twist1 generally functions as an osteoblast differentiation inhibitor (el Ghouzzi et al., 1997; Howard et al.,1997). The increased Twist1 expression in the frontal bone primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant has significant implications. In terms of osteoblast progenitor proliferation, Fgfr2 and Twist1 have overlapping expression patterns and FGFRs may mediate a signal downstream of Twist1 because murine Twist1expression precedes that of Fgfr genes during skull development(Jabs, 2001; Holleville et al., 2003). It is conceivable that elevated Twist1 expression inhibits Fgfr gene expression in the frontal primordium, which results in a compromised osteoprogenitor cell proliferation in the Tgfbr2^fl/fl;Wnt1-Cre mutant. In parallel, TWIST1 protein has recently been shown to inhibit RUNX2 function during skeletogenesis(Bialek et al., 2004). Elevated Twist1 expression may inhibit Dlx5 expression to inhibit the terminal differentiation of osteoprogenitor cells within the CNC-derived frontal bone primordium in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample and to prevent normal bone formation.

This study clearly demonstrates the important inductive function of dura mater in regulating calvarial bone development as conditional inactivation of TGFβ signaling in the CNC-derived dura mater resulted in a defect in the mesoderm-derived parietal bone primordium (T.S., Y.I., P.B. and Y.C.,unpublished). The successful induction of Fgfr2 expression by implanted TGFβ beads in the parietal bone primordium of the Tgfbr2^fl/fl;Wnt1-Cre mutant sample shows that the TGFβ signaling cascade is intact in this mesoderm derived tissue and the parietal bone defect may be the result of a defect in another CNC-derived tissue. This finding is consistent with previously published results in which the ossification of parietal bone requires interaction with the underlying neural crest derived meninges (Jiang et al., 2002). However, it is not clear whether the TGFβsignaling mediated ossification of the CNC-derived frontal bone is cell autonomous or is completely dependent on the inductive signals from dura mater, which can only be tested by a specific inactivation of TGFβsignaling in the meninges.

In summary, our study shows that loss of Tgfbr2 in the neural crest does not affect the migration of CNC cells during frontal bone development. TGFβ IIR is specifically required for proliferation and terminal differentiation of the osteoprogenitor cells. TGFβ-mediated FGF signaling is crucial for the CNC-derived osteoprogenitor proliferation. As exogenous FGF signaling can rescue only the defect in cell proliferation but is unable to promote differentiation of osteoprogenitors in the frontal bone primordium of Tgfbr2^fl/fl;Wnt1-Cre mutant sample, it is apparent that there are additional signaling pathways that rely on TGFβsignaling to induce osteoblast differentiation in the calvarial aspect of the frontal bone primordium. Specifically, TGFβ-mediated Twist1 and Dlx5 expression may play a crucial role for osteoblast differentiation and osteoid mineralization. In the orbital aspect of the frontal bone primordium, however, TGFβ signaling is apparently dispensable for differentiation of osteoprogenitor cells. It is tempting to speculate that signals from the orbit may compensate the loss of TGFβsignaling in the orbital aspect of frontal bone and allow the successful differentiation of osteoprogenitor cells in the Tgfbr2^fl/fl;Wnt1-Cre mutant sample. Future studies using this animal model will provide useful information on the mechanism of TGFβ signaling in regulating human skull development.

We thank Rob Maxson, Henry Sucov, Brauch Frenkel and Julie Mayo for critical reading of the manuscript; Xun Xu for his assistance in the laboratory; D. Ornitz and J. L. R. Rubenstein for reagents; and H. Moses for the Tgfbr2^fl/fl mice. This study was supported by grants from the National Institute of Dental and Craniofacial Research, NIH(DE012711, DE014078 and DE017007) to Yang Chai.