Fuz Regulates Craniofacial Development through Tissue Specific Responses to Signaling Factors

The planar cell polarity effector gene Fuz regulates ciliogenesis and Fuz loss of function studies reveal an array of embryonic phenotypes. However, cilia defects can affect many signaling pathways and, in humans, cilia defects underlie several craniofacial anomalies. To address this, we analyzed the craniofacial phenotype and signaling responses of the Fuz−/− mice. We demonstrate a unique role for Fuz in regulating both Hedgehog (Hh) and Wnt/β-catenin signaling during craniofacial development. Fuz expression first appears in the dorsal tissues and later in ventral tissues and craniofacial regions during embryonic development coincident with cilia development. The Fuz−/− mice exhibit severe craniofacial deformities including anophthalmia, agenesis of the tongue and incisors, a hypoplastic mandible, cleft palate, ossification/skeletal defects and hyperplastic malformed Meckel's cartilage. Hh signaling is down-regulated in the Fuz null mice, while canonical Wnt signaling is up-regulated revealing the antagonistic relationship of these two pathways. Meckel's cartilage is expanded in the Fuz−/− mice due to increased cell proliferation associated with the up-regulation of Wnt canonical target genes and decreased non-canonical pathway genes. Interestingly, cilia development was decreased in the mandible mesenchyme of Fuz null mice, suggesting that cilia may antagonize Wnt signaling in this tissue. Furthermore, expression of Fuz decreased expression of Wnt pathway genes as well as a Wnt-dependent reporter. Finally, chromatin IP experiments demonstrate that β-catenin/TCF-binding directly regulates Fuz expression. These data demonstrate a new model for coordination of Hh and Wnt signaling and reveal a Fuz-dependent negative feedback loop controlling Wnt/β-catenin signaling.


Introduction
Vertebrate craniofacial development is a complicated process requiring the coordination of multiple signaling pathways and tissue interactions among the three germ layers and neural crest cells in three dimensions. A number of signaling pathways have been implicated in craniofacial development including Hedgehog (Hh), Wnt/b-catenin, TGF-b, Fibroblast Growth Factor (Fgf), Notch, and Planar Cell Polarity (PCP) signaling [1,2,3,4,5]. However, the interactions and components shared among different signaling pathways are not well understood. The recent identification of the PCP effector gene Fuzzy (Fuz) as an important regulator of Hedgehog (Hh) signaling suggests that there may be substantial crosstalk between the different molecular cues. Furthermore, Fuz can coordinate ciliogenesis and secretion, two processes that affect a variety of signaling pathways [6,7]. Our data suggest Fuz plays a pivotal role in the Wnt and Hedgehog pathways [6,7,8]. Furthermore, loss of Fuz leads to dramatic defects in craniofacial development.
PCP signaling, initially discovered in Drosophila, controls a diverse range of polarized cellular behaviors [9]. Recent studies have shown that PCP signaling is also important in regulating cell interactions and tissue movements [10,11,12]. In mice, loss of PCP genes leads to disruption of polarized structures such as the stereociliary bundles in the cochlea [13,14,15,16]. In addition, mutants in PCP genes display an array of morphogenetic defects affecting the neural tube, heart, kidney and other tissues [1]. Thus, the PCP pathway affects a wide range of cell-cell interactions via coordination of morphogenesis, tissue polarity, and potentially growth.
The PCP pathway makes use of the Wnt signaling components Frizzled and Dishevelled, but is b-catenin-independent [17]. Studies of vertebrate planar cell polarity have primarily focused on the ''core PCP'' pathway. Thus, the function of vertebrate PCP ''effectors'', which act downstream of Dishevelled, is unclear. This report focuses on one of the PCP effector proteins, Fuzzy or Fuz, which encodes a transmembrane protein required for tissue polarity. Elegant genetic studies in Drosophila have shown that fuzzy, along with inturned, plays a role in maintaining cytoskeletal integrity in the wing hairs; this role is genetically downstream of dishevelled and frizzled [18,19,20,21].
Recent studies suggest that mutations affecting ciliogenesis lead to congenital human anomalies. Cilia appear to act as mechanical and chemical sensors that interpret extracellular cues via signaling pathways such as Hh, PDGF and Wnt [22]. Because there is no protein synthesis in the cilium, assembly and maintenance of the cilium requires intraflagellar transport (IFT). Thus, loss of IFT proteins leads to catastrophic effects on ciliogenesis and cilia function. Human conditions resulting from defects in ciliogenesis or IFT include situs inversus, polycystic kidney disease, ear defects and craniofacial anomalies [23,24].
Several proteins involved in PCP signaling, including Dishevelled, localize to the base of cilia [25,26,27,28]. This suggests that PCP signaling may be important for cilia function and possibly for coordination of other signaling pathways. In this context, we were particularly interested in the canonical Wnt/b-catenin pathway, as the absence of cilia results in enhanced b-catenin nuclear localization and downstream gene transcription [29,30,31].
In this study, we analyzed the expression, regulation and functional requirements of Fuz during vertebrate craniofacial development. We found that Fuz 2/2 mice display massive craniofacial defects including anophthalmia, agenesis of the tongue and incisors and a hypoplastic mandible leading to cleft palate. In contrast, we found that Meckel's cartilage, an embryonic structure that develops into the mandible and portions of the inner ear, was hyperplastic and malformed. Some aspects of the craniofacial phenotype, such as missing teeth and cleft palate, could be attributed to decreased Hh signaling. However, the increased growth of Meckel's cartilage was associated with increased Wnt signaling.
To further examine the molecular consequences of Fuz loss of function, we analyzed expression of Wnt target genes in Fuz 2/2 mice. We found that a number of b-catenin/TCF target genes were up-regulated in our mutants, suggesting that Fuz plays a dual role in Wnt signaling, in both the canonical and non-canonical pathways. Furthermore, we found that the Fuz protein could repress expression of Wnt pathway genes as well as a Wntdependent reporter, suggesting a direct role within the canonical pathway. Finally, most interesting, we identified b-catenin/TCFbinding sites in the Fuz gene, which we confirmed by chromatin immunoprecipitation. Together, these data imply that Fuz may be a critical factor linking the Hh, PCP and Wnt/b-catenin signaling pathways and may function as a switch to balance the activities of these pathways during craniofacial development.

Results
Fuz expression pattern during mouse development X-gal staining of E9.5 Fuz LacZ/+ mouse embryos revealed Fuz LacZ expression was restricted to the dorsal tissue and brain ( Fig. 1A). At E11.5, Fuz LacZ expression was concentrated in the brain, spinal cord and eyes. However, relatively weak expression was also detected in the heart, limbs and craniofacial region (Fig. 1B). At E12.5, Fuz LacZ expression in the craniofacial region began to expand from dorsal to the ventral regions ( Fig. 1C). At E14.5, Fuz LacZ expression became stronger and widespread throughout the craniofacial region (Fig. 1C). Fuz expression at E12.5 in the oral epithelium, mesenchyme and Meckel's Cartilage is shown in sagittal sections (Fig. 1D). Fuz LacZ expression at E14.5 in the oral epithelium, mesenchyme, palate (PL), and tongue (TE) has increased from E12.5 (Fig. 1E). Fuz expression in Meckel's cartilage (MC) and the perichondrium (PC) are shown in high magnification sections (Fig. 1F). In addition, we assessed Fuz expression in different cell lines by RT-PCR. Fuz is highly expressed in LS-8 cells (mouse oral epithelium), C3H10T1/2 cells (mouse embryonic fibroblast), HEK 293 FT cells (human embryonic kidney fibroblast) and SW1353 cells (human chondrocyte). It had relatively weak expression in MDPC-23 cells (mouse dental mesenchyme), and no expression in CHO cells (hamster ovary) (Fig. 1G).
Severe craniofacial defects in the Fuz 2/2 mice To study Fuz function during craniofacial development, we used the Fuz null mouse created with a gene trap cassette inserted in the second exon of the Fuz gene [7]. The lack of both copies of the Fuz gene in homozygotes (Fuz 2/2 ) was lethal for mice immediately after birth. At E18.5, all null embryos have a hypoplastic mandible and maxilla, and anophthalmia ( Fig. 2A). An abnormal bulge in the mandible and a secondary cleft palate are observed in mutant embryos ( Fig. 2A,B). Histological analysis on E18.5 embryos revealed further craniofacial defects including a malformed tongue and missing incisors. The ventral bulge of the mutant mandible is due to a hyperplastic and malformed Meckel's cartilage (Fig. 2C,D). The Fuz 2/2 mice do not form incisor tooth buds and lower and upper incisor tooth initiation did not begin and the tongue muscles appear to be fused with the mandible. The hyperplastic and malformed Meckel's cartilage suggested an increased proliferation of Meckel's cartilage cells.
To examine the developmental progress of craniofacial formation, embryos were harvested at different developmental time points and analyzed. At E12.5, the mandible of the Fuz 2/2 embryos appeared normal compared with those of their heterozygous littermates. Meckel's cartilage of the Fuz 2/2 embryos had a normal shape and similar size with heterozygous embryos (Fig. 3A). The normal mouse craniofacial structures such as the choroid plexus (CP) differentiating from the roof of fourth ventricle, the choroid plexus extending into the lateral ventricle (CPL), the corpus striatum mediale (STM), the optic recess of the diencephalon (OR) and the cochlea (CO) are present in the E12.5 Fuz +/2 heterozygous mouse (Fig. 3A). These structures are lost in the Fuz null mice at this stage, only the cochlea and corpus striatum are observed and these structures are abnormal (Fig. 3A). In contrast, at E14.5 the Fuz 2/2 entire mandible is thickened in the dorsal-ventral axis and shortened in the anterior-posterior axis compared with those of heterozygous littermates (Fig. 3B). The trigeminal nerve (TG) appears to replace the pituitary primordium or Rathke's pouch (RP) and many of the brain structures are not developed (Fig. 3B). Meckel's cartilage has begun to elongate in the dorsal-ventral direction instead of the anterior-posterior direction (Fig. 3B). The upper and lower incisor buds have not developed while the tongue had a rudimentary root structure and failed to elongate in the anterior-posterior axis (Fig. 3B). The malformed Meckel's Cartilage not only elongated dorsoventrally but also the dorsal end expanded in the median-lateral axis. An abnormal growth of the palate tissue (PLT) or palate shelf was observed in the E14.5 Fuz null mice compared to the normal palate (PL) seen in the heterozygous mice (Fig. 3B). Coronal sections of E16.5 Fuz 2/2 mice revealed descending palate shelves (PL) suggesting a delay of elevation, a lack of eyes, however the upper and lower molar (ML) tooth buds have formed (Fig. 3C). Meckel's cartilage has developed in the midline and extends dorsally (Fig. 3C). There were many other defects observed in the nasal cavity, facial bones, cartilage and brain that will not be examined in this report. However, the loss of Fuz affects patterning of much of the craniofacial region with a loss of some structures (incisors, tongue, eyes and bone) and an expansion of other structures (Meckel's cartilage).

Defective craniofacial bone development in Fuz null mice
To determine if bone formation was affected in the Fuz null mice, E16.0 embryos were stained with Alcian Blue/Alizarin Red and compared to their heterozygous littermates ( Fig. 4A-D). The blue cartilage stain revealed defective Meckel's cartilage in the Fuz 2/2 mice and malformed cartilage of the craniofacial region (Fig. 4B,D). The anterior Meckel's cartilage is deformed with an ascending branch and reduced ossification (red stain) around Meckel's cartilage. Specifically, membranous ossification of the premaxilla (pmx), maxilla (mx), mandible (md), frontal (fnt) and parietal (par) bones are missing at this stage. Ossification of the sphenoid (sb) and basioccipital (bb) bones are also defective in the Fuz null mice (Fig. 4D). To determine if craniofacial bone development was delayed or reduced in the Fuz null mice we analyzed E18.5 Fuz null mice and found a delayed ossification of the craniofacial region compared to heterozygous littermates ( Fig. 4E-H). Further research is ongoing to understand the delayed bone development and in this report we focused on the regulatory mechanisms directing Meckel's cartilage formation.

Increased proliferation of Meckel's cartilage cells in Fuz null embryos
Cell proliferation of E14.5 Meckel's cartilage was measured by immunofluorescence with a Ki67 antibody (Fig. 5A). The Ki67 positive cell number to total cell number (DAPI) within Meckel's cartilage was calculated to estimate the proliferation ratio. Compared with a 75% ratio observed in heterozygous Meckel's cartilage, the ratio in null samples was significantly increased to 88% (Fig. 5B). These cells are proliferating at a higher rate compared to Fuz heterozygous Meckel's cartilage.

Fuz regulates cilia development in the mandibular mesenchyme
We asked if cilia formation was affected in the mandible by immunofluorescence with an Arl13b antibody in mandible mesenchyme at E14.5. The amount of primary cilium was significantly decreased in the Fuz 2/2 mandible mesenchyme (Fig. 6). The average cilium number per 5000 mm 2 in the sagittal sections of the mutant mandible mesenchyme was 2.9, compared to 16.5 in the wild type. Because Hh signaling is regulated by the primary cilium components, we hypothesized that the Hh signaling was altered due to the cilium defect in the Fuz null mouse.
Down-regulation of Hedgehog signaling in Fuz 2/2 embryos A loss of primary cilia correlates with an absence of membrane associated Smoothened and failure to activate downstream transcription factors responding to the Hh signal [4,32,33]. We asked if sonic hedgehog (Shh) signaling was altered in the craniofacial region of Fuz mutant embryos. A whole-mount in situ hybridization assay was performed at E9.5 and showed an overall reduction of Patched 1 (Ptch1), and Gli1 expression levels while Shh levels were unchanged (Fig. 7). Given that Ptch1 and Gli1are indicative of Hh signaling [34,35], these data revealed a general down-regulation of Hh signaling in the Fuz null mouse.
Immunohistochemistry experiments revealed that Ptch1 was highly expressed in the tongue, oral and dental epithelium and mesenchyme, but weakly expressed in Meckel's cartilage in E14.5 heterozygous embryos (Fig. 8A). In the Fuz null embryos, Ptch1 expression was decreased overall in the oral cavity as well as in Meckel's cartilage (Fig. 8B). Real-time PCR with mRNA of dissected Meckel's cartilage and surrounding mesenchyme confirmed the down-regulation of Ptch1 (Fig. 8E). Real-time PCR also revealed a significant decrease in Gli1 transcripts in the Fuz null embryos (Fig. 8E). At E14.5, Gli2 had a similar expression pattern with Ptch1 and Gli2 protein was decreased in the null embryos compared to heterozygotes (Fig. 8C,D), though its transcript level was not significantly changed (Fig. 8E). Shh transcript levels were also reduced in the null embryos suggesting that Fuz was required for the maintenance of Hh signaling. Hh signaling is essential for establishing and maintaining dorsal-ventral patterning and required for incisor development. The disrupted Hh signaling provides a possible explanation for the malformed Meckel's cartilage and the missing incisors. Hh signaling is also able to stimulate cell proliferation by activating Cyclin D1 (Ccnd1) and Cyclin D2 (Ccnd2) expression [36]. However, the decreased Hh signaling is contrary to the increased proliferation of Meckel's cartilage. We hypothesized that another mechanism must be involved in the enhanced proliferation of Meckel's cartilage.
Increased Wnt/b-catenin signaling in Fuz 2/2 embryos Wnt/b-catenin signaling is regulated by the basal component of the primary cilium [29,30] and Hh signaling interacts with Wnt/ b-catenin signaling during multiple developmental programs [37,38,39]. We asked if Wnt/b-catenin signaling was altered in the Fuz null embryos, which could affect cell proliferation. Immunofluorescence with a b-catenin antibody was performed using sagittal sections of E14.5 embryos. These experiments revealed that bcatenin was increased in the Fuz 2/2 oral epithelium, mesenchyme ( Fig. 9A) and Meckel's cartilage (Fig. 9B), compared to those of heterozygous littermates. Because Lef-1 is a transcription factor whose activity and expression are positively regulated by Wnt/bcatenin signaling, we asked if Lef-1 expression was modulated in the Fuz null mutants. Lef-1 is normally expressed in the oral and dental epithelium and mesenchyme at E14.5 (Fig. 9C,D, upper panels). In the null embryos Lef-1 expression in the oral mesenchyme was increased (Fig. 9C,D, lower panels), confirmed by Real-time PCR (Fig. 9F). Tcf4 (Tcf7l2) expression is also regulated by Wnt/b-catenin signaling [40], and immunofluorescence assays revealed increased Tcf4 in Fuz null Meckel's cartilage at E14.5, and further confirmed by Real-time PCR (Fig. 9E,F). Inspection of a group of Wnt/b -catenin target genes revealed increased expression of Axin2, Cyclin D1, Cyclin D2 and Runx2 in Fuz 2/2 Meckel's cartilage (Fig. 9F), [41,42,43,44]. On the contrary, non-canonical Wnt signaling including Wnt5a, Ror2, and their downstream target Pcdh8 (known as Papc in Xenopus) [45] were down regulated in Fuz 2/2 Meckel's cartilage (Fig. 9F). As a   control real-time PCR demonstrated a lack of Fuz transcripts in the null embryos (Fig. 9F). These data indicate that Fuz acts as a repressor of Wnt/b-catenin signaling during craniofacial development. The increased Cyclin D1 and D2 provide a possible explanation of the enhanced cell proliferation in the Fuz 2/2 Meckel's cartilage.
To confirm the repressor role of Fuz on Wnt/b-catenin signaling, Fuz expression plasmid was co-expressed with the 7xTopflash reporter in both HEK 293 FT cells and CHO cells. Co-transfection of Fuz with the 7xTopflash reporter resulted in a 50% decrease in Topflash reporter activity, compared to that with the control vector in both cell lines (Fig. 9G). The result was consistent between 293 cells, which has high endogenous Fuz expression and CHO cells which do not endogenously express Fuz.
These results indicate a repressor role of Fuz on Wnt/b-catenin signaling.
b-catenin directly activated the Fuz promoter Sequence analyses revealed eleven Wnt response elements (Lef/ Tcf binding sites) within the murine Fuz 2.4 kb promoter (Fig. 10A). A chromatin immunoprecipitation (ChIP) assay demonstrates  endogenous b-catenin associating with the Fuz promoter. Nontransfected LS-8 cells were used as these cells endogenously express Lef-1, b-catenin and Fuz. The Fuz promoter chromatin was amplified by PCR using primers specific for the Fuz promoter flanking the Lef/Tcf binding site (Fig. 10A,B). The primers amplified a 201 bp product from the chromatin input and antibody IP (Fig. 10B, lanes 3 and 5, respectively). The primers did not produce a PCR product from primers only or normal rabbit IgG IP control (Fig. 10B, lanes 2 and 4, respectively). To test whether this association could lead to functional activity, the Fuz 2.4 kb promoter was cloned into a luciferase vector and transfected into LS-8 and CHO cells. Addition of LiCl (10 mM) to the cell culture medium stimulates b-catenin nuclear localization and caused significant increase of Fuz promoter activity in both cell lines (Fig. 10C). These results demonstrate that b-catenin directly targets the Fuz promoter and activates its transcription. Combined with the Fuz repression of Wnt/b-catenin signaling, we conclude that Fuz constitutes a negative feedback loop that controls Wnt/b-catenin signal activity.
Sox9 expression is increased in Fuz 2/2 Meckel's cartilage A previous study reported that Wnt/b-catenin signaling repressed Sox9 expression both in vitro and in vivo [46]. Given the increased Wnt/b-catenin signaling in the Fuz null mice, we expected reduced expression of Sox9 in Fuz null Meckel's cartilage. However, immunofluorescence with Sox9 antibody in E14.5 Meckel's cartilage revealed that Sox9 expression was increased in the null embryos (Fig. 11A). This increase was validated by Realtime PCR using RNA from E14.5 Meckel's cartilage (Fig. 11B). Type II Collagen (Col2a1) is a downstream target gene of Sox9 [47]. Col2a1 expression was also increased in Fuz null Meckel's cartilage measured by Real-time PCR (Fig. 11B). Over-expression of Fuz in SW1353 cells confirmed the repression of Sox9 expression. The human chondrocyte cell line SW1353 was transfected with the Fuz Immunofluorescence of Gli2 in sagittal sections of E14.5 Fuz +/2 heterozygous embryos. Sections in the bottom panels are higher magnification of those in the top panels. Gli2 is highly expressed in the oral and dental epithelium and mesenchyme, but weakly expressed in the Meckel's cartilage. The expression pattern of Gli2 is similar to Ptch1, but Gli2 has increased expression in the oral mesenchyme compared to Ptch1. D) In the Fuz 2/2 embryos, Gli2 expression was decreased overall as well as in Meckel's cartilage. E) The wild type and Fuz 2/2 Meckel's cartilage and surrounding mesenchyme were dissected from E14.5 embryos and the mRNA was extracted, followed by reverse transcription. The real-time PCR was performed with indicated probes. Results are shown as normalized relative expression. bactin served as the reference gene. Experiments were repeated three to five times each from multiple samples. Error bars indicate S.E. *: p-values,0.05; **: p-values,0.01. doi:10.1371/journal.pone.0024608.g008 expression vector and total RNA was harvested two days after transfection. Real-time PCR revealed that endogenous SOX9 and COL2A1 expression was significantly repressed in the Fuz transfected cells compared to those with control vector. These results indicate that Fuz represses Sox9 and cartilage expansion during craniofacial development. Loss of Fuz leads to increased Sox9 expression in the null mice, which maintains cartilage expansion in the presence of increased Wnt/b-catenin signaling.

Discussion
In this study we have analyzed the role of the Fuz gene during craniofacial development. Fuz loss of function analyses revealed a critical function of this gene in the development of multiple craniofacial tissues and structures. Fuz is required for the formation of eyes, bone, tongue and incisors. These affected organs and craniofacial structures show essential roles of Fuz in tissue patterning and cell proliferation.

Fuz regulates Hh signaling
In the Fuz null mice, down regulation of Hh signaling has been shown in the neural tube and limb buds [7,8]. In this study we have shown a decrease in Hh downstream gene expression in early and late craniofacial development. Hh signaling is essential for dorsal-ventral patterning and incisor development, but apparently not for molar development, as the molar tooth buds form normally in the Fuz 2/2 mice. A loss of Hh signaling in the oral epithelium results in a lack of epithelial cell proliferation and tooth bud formation [48]. This would explain the absence of incisor development, but not molar development. We suspect that this is a timing issue as incisors develop earlier than molars, however more experiments are required to understand the defect. Hh signaling is required for mouse brain and craniofacial morphogenesis and loss or gain of Hh function is associated with midline facial anomalies [49,50]. However, the overall craniofacial defects observed in the Fuz null mice do not resemble the other cilia related gene defects or Shh mutant mice, with the exception of brain and cleft palate defects. A recent report revealed a role for Fuz in the development of hair follicles [51]. Hair follicle development was impaired due to inhibition of cilia formation and Hh signaling, revealing a direct role for Fuz in regulating Hh signaling through defective cilia formation. Interestingly, Fuz regulated Hh signaling does not appear to regulate Meckel's cartilage growth but may affect patterning and placement of Meckel's cartilage in the mandible. Thus, Fuz could be regulating Hh signaling which in turn restricts the normal growth and patterning of Meckel's cartilage by influencing the surrounding tissue formation (such as incisor development) and other craniofacial structures.

Fuz regulates Wnt signaling
In the Fuz null mice, the Wnt/b-catenin pathway was up regulated whereas the non-canonical Wnt5a/Ror2 pathway was down regulated. These data suggest that Fuz is a potent repressor of Wnt/b-catenin signaling. The data also indicate that Fuz is critical to maintain the non-canonical Wnt signaling pathway, which is overlapped with PCP signaling. It provides a possible explanation of the neural tube defect (a typical phenotype in noncanonical Wnt deficient mice) previously reported in Fuz null mice [7,52]. The increased Meckel's cartilage growth in the Fuz 2/2 mice can be attributed to increased Wnt/b-catenin signaling. Thus, PCP, Wnt/b-catenin and Hh signaling pathways may converge to provide cues for and instruct specific cell differentiation, tissue patterning and morphogenesis.
Fuz regulation of Wnt/b-catenin signaling may be mediated by primary cilia and interaction of multiple signaling pathways. Previous studies have reported that the basal component of primary cilium could repress Wnt/b-catenin signaling through interaction of Dishevelled protein [29,30]. The loss of Fuz could cause the secondary up-regulation of Wnt/b-catenin signaling. Previous studies have shown that Hh signaling antagonizes Wnt/ b-catenin during development of the tongue and cartilage [37,38,53]. The increase in Wnt/b-catenin signaling could be secondary to decreased Hh signaling in the Fuz null mice. Noncanonical Wnt signaling is known to antagonize the canonical Wnt/b-catenin activity [17,45]. In particular, Wnt5a was shown to repress Wnt/b-catenin signaling in the limb buds [54]. An in vitro study revealed that the Ror2 receptor was necessary for Wnt5a repression of Wnt/b-catenin activity [55]. In the Fuz null mice, down-regulation of Wnt5a and Ror2 could attribute to activation of Wnt/b-catenin signaling as well. Whether it is one of those mechanisms or a combination requires further investigation.

Primary cilia defects associated with increased Hh signaling
In contrast to loss of Fuz, mice mutants for cilia intraflagellar transport (IFT) proteins; IFT88/polaris and Kif3a have increased Hh activity [50,56]. Mice with mutations in IFT88 lack cilia on all cells and present with severe neural tube defects, polydactyly, asymmetry defects and ectopic tooth formation [56,57,58]. Kif3a mice mutants display a range of similar developmental defects and a conditional knockout of Kif3a in the neural crest results in an increase in Hh activity associated with truncated cilia [50]. However, other groups have reported that mutations in IFT proteins including Kif3a demonstrate decreased Hh signaling [59,60,61]. These differences could be attributed to the differential tissue expression of the Gli proteins during development [50,56,60,62,63]. Our data reveal Fuz expression is predominantly expressed in the oral and dental epithelium at early stages and in epithelial cell lines. The Fuz 2/2 mice do not present with a wide facial prominence, which is indicative of increased Hh signaling [50]. Furthermore, as a PCP effector it may play an extended role in regulating signaling pathways and gene expression independent of cilia formation. However, unlike the Wnt and Hh signaling mechanisms, which act at the level of transcription, the PCP pathway controls cell morphology [17,45]. It is not inconceivable that changes in cell morphology induce changes in signaling mechanisms. The Fuz 2/2 mice have reduced and truncated cilia, but not a complete loss of primary cilia. The presence of truncated cilia in the Fuz null mice may play a limited role in the signaling activities. Fuz interacts with a Rab-similar GTPase (RSG1) to affect trafficking from the cytoplasm to basal bodies and to cilia tips [7]. Fuz may have a bigger role in exocytosis than ciliogenesis and could promote a cell type-specific regulation of signaling mechanisms and gene expression [7,17,64]. Transport and exocytosis facilitated by Fuz maybe be more important and explain the mouse craniofacial phenotypes of the Fuz null mice more than defective cilia.
Craniofacial Bone and Meckel's Cartilage Defects. Runx2 plays a major role in bone development and specifically intramembranous bone formation during craniofacial development. Runx2 null mice have a complete loss of bone tissue and regulation of Runx2 dosage during development can result in the cleidocranial dysplasia phenotype [65]. These hypomorphic Runx2 mutant mice have reduced ossification of the calvarial bones, reduced basisphenoid bone and non-osseous tissue between the parietal bones, similar to the Fuz null mice [65]. However, Runx2 expression is increased in the Fuz null mice and Runx2 over-expression newborn mice do not reveal defects in calvarial bone development [66]. It is not known if defects occurred at earlier stages of development in these mice. The Fuz null mice demonstrated a delayed calvarial bone development, which may appear normal at later stages of newborn mice (Fuz null mice die at birth, we are unable to determine if bone development is normal at later stages). We speculate that increased Runx2 expression may compensate for reduced Hh expression, which is also required for bone development.
Meckel's cartilage is a transient structure responsible for mandible formation and undergoes endochondral-like ossification to produce several bones during craniofacial development [67,68,69,70,71]. Meckel's cartilage is derived from ectomesenchymal cells from cranial neural crest (CNC) cells and non-CNCderived cells [72,73,74]. Sox9 is a critical transcription factor required for cartilage formation and mutations in Sox9 are associated with campomelic dysplasia, which is characterized by skeletal defects and cranial dismorphology [75,76,77]. Sox9 is expressed early in during development in the cranial neural crest cells and regulates Col2a1 required for chondrocyte differentiation and Sox9 appears to be a master regulator of cartilage development [78,79]. Sox9 null mice die midway through gestation, however the Sox9 conditional knockout with Wnt1-Cre reveal craniofacial defects including a short mandible and a large cleft palate [80,81]. Sox9 is required for Meckel's cartilage growth and inactivation of Sox9 leads to reduced Meckel's cartilage growth. Sox9 and Col2a1 expression are increased in the Fuz null mice, which correlates with increased Meckel's cartilage growth. The expanded Meckel's cartilage would cause increased endochondral ossification as well giving rise to the boney mandible structure observed in the Fuz null mouse mandible. Sox9 expression also occurs in the mouse head skeleton and some of the bone defects could be due to the activity of Sox9 in these tissues [77]. TGF-b signaling regulates chondrogenesis during mandible development and may be regulated by Fuz [82,83]. We are currently exploring these mechanisms.
Human Ciliopathies. There are multiple ciliary genes and proteins associated with ciliopathies, (for reviews, [22,84,85]). Human FUZ mutations have not been reported or associated with syndromes with known ciliary pathology. However, Oral-Facial-Digital syndrome, type1 (OFDI; OFD 1; OMIM311200) appears to represent similar phenotypic features to the craniofacial defects associated with the Fuz null mouse [86]. Although the tongue is present but is malformed in these patients. This syndrome is caused by mutations in the Cxorf5 transcript termed OFD1 [87]. It belongs to a group of developmental disorders known as oralfacial-digital syndromes and is characterized by malformations of the oral cavity, face and digits [88]. OFD1 is also mutated in Xlinked Joubert syndrome [89]. OFD1 is a centrosomal protein localized at the basal bodies of the primary cilia. The mouse OFD1 mutants reproduce the main features of the human syndromes and have some features not observed in the Fuz null mouse [90,91,92]. OFD1 is nuclear localized and part of the human TIP60 chromatin-remodeling complex [93]. Recent data reveals that Fuz is not in the nucleus but distributed throughout the cytoplasm (unpublished data, Amendt laboratory). Fuz does not appear to be associated with the plasma or nuclear membranes or specific organelles. The role of FUZ in human disorders and the molecular mechanisms of FUZ are currently being investigated.
In summary, Fuz is critical for Hh, PCP, and Wnt signaling regulation and may serve to connect these distinct signaling pathways. A tentative model linking Fuz to Hh, Wnt and PCP signal pathways based on our data and others are shown in figure 12. Fuz acts downstream of Frizzled and Dishevelled to regulate PCP signaling. Non-canonical Wnt signaling represses Wnt/b-catenin signaling and Fuz is involved in a negative feedback loop of Wnt/b-catenin signal regulation. Loss of Fuz leads to impaired PCP signaling and up-regulation of Wnt/bcatenin signaling. In addition, it results in impaired actin cap and cilia, which inhibits the activation of Gli transcription factors and weakens Hh signaling. Loss of Fuz also causes up regulation of Sox9. These factors together contribute to the complex defects in craniofacial structures, delayed bone development and the hyperplastic and malformed Meckel's cartilage.

Ethics Statement
All animals were housed at the Institute of Biosciences and Technology under the care of the Program of Animal Resources, and were handled in accordance with the principles and procedure of the Guide for the Care and Use of Laboratory Animals. All experimental procedures were approved by the Texas A&M Health Science Center, Institutional Animal Care and Use Committee. Protocol number 09001, mouse models for tooth development.

Animals
The Fuz gene trap mice were generated by the Texas A&M Institute for Genomic Medicine and were described previously [7]. These mice were maintained in the C57BL/6 background. Fuz LacZ targeted ES cells corresponding to clone PG00134_Z_E03_2 were purchased from the Knock Out Mouse Project (KOMP) Repository and injected into blastocysts by the Texas A&M Institute for Genomic Medicine. Four chimeras were obtained and mated to wild-type C57BL/6 mice. Two chimeras yielded germ-line transmission. Offspring from both chimeras were used for X-gal staining and exhibited the same expression pattern. Embryos were collected at various time points, considering the day of observation of a vaginal plug to be embryonic day (E) 0.5. Genotyping PCR primers for Fuz +/2 and Fuz 2/2 mice and embryos were described previously [7]. Genotyping PCR primers for Fuz LacZ are as below: forward, 59-CTCTCCTGGCAGCATGTCCT -39; reverse, 59-TCCTCCTACATAGTTGGCAGTG -39. The resulting PCR product represents the LacZ knockin allele, whereas the wild-type allele does not generate any products. All PCR products were sequenced to confirm their identity.

Real-time PCR analyses
Meckel's cartilage and surrounding mesenchyme were dissected from E14.5 mouse embryos. The Real-time PCR was performed with different probes listed in the table 1. Experiments were repeated three times each from multiple samples and p-values are shown. Total RNA was extracted using the RNeasy mini kit from Qiagen. Total RNA was reverse transcribed into cDNA by iScript Select cDNA Synthesis kit (BioRad). Real-time PCR was carried out in a total reaction of 25 ml containing 12.5 m iQ SYBR Green Supermix, 0.1 mM forward primer, 0.1 mM reverse primer, 0.25 ml cDNA template in the MyiQ Singlecolor Real-Time Detection System and analyzed by the MyiQ Optical System Software 2.0 (BioRad). The Real-time PCR was performed with gene specific probes. b-actin served as a reference gene. The thermal cycling profile consisted of 95uC for 4 min followed by 40 cycles of denaturation at 95uC for 30 sec, annealing at 60uC for 30 sec and elongation at 72uC for 18 sec. Samples were run in triplicate. Experiments were repeated three times each from multiple samples and p-values were calculated. No-template control was run in each experiment. Melting curve analyses were performed to confirm amplification specificity of the PCR products. All PCR products were sequenced to confirm their identity.

Expression and reporter constructs
The expression plasmid containing the cytomegalovirus (CMV) promoter linked to the mouse Fuz full length cDNA reversetranscribed from total RNA of NIH-3T3 cells was constructed into pcDNA3.1 vector (Invitrogen). The 7xTopFlash reporter plasmid was constructed into luciferase vector by inserting seven Lef/Tcf binding sites upstream of the minimal TK promoter [94]. The mouse Fuz 2.4 kb promoter was constructed into the TKluciferase vector by replacing the minimal TK promoter [95].
Cell culture, transient transfections, luciferase and bgalactosidase assays CHO, HEK 293 FT, LS-8 and SW1393 cells (all cells were purchased from the ATCC, except LS-8 cells [96]) were cultured in DMEM supplemented with 10% FBS and penicillin/streptomycin and transfected by electroporation. The method of transient transfections, luciferase and b-galactosidase assays were described previously [95]. LiCl was added to the specified cells at a final concentration of 10 mM, 23 h before harvest. The pcDNA3.1 empty vector was added to equalize the total amount of cotransfected expression vectors. All the plasmids were doublebanded CsCl purified.

Chromatin Immunoprecipitation (ChIP) analyses
The ChIP analyses were performed as described [97] using the ChIP Assay Kit (Upstate) with the following modifications. LS-8 cells were fed for 24 h, harvested and plated in 60 mm dishes. Cells were cross-linked with 1% formaldehyde for 10 m at 37uC the next day. Samples were incubated with rabbit anti b-catenin polyclonal antibody (Upstate 06-734) overnight at 4uC. An aliquot of the immunoprecipitated DNA (3 ml) from non-transfected cells were used for PCR (32 cycles). All reactions were done under an annealing temperature of 61uC. Two primers for amplifying the Lef/Tcf binding site in the Fuz promoter are as follows: sense-59-GCAACACCTTAGCACCATCA -39 and antisense, 59-GGCT-AAATTCCTGCCTTCATC -39. All the PCR products were evaluated on a 1% agarose gel in 16 TBE for appropriate size (201 bp) and confirmed by sequencing. As controls the primers were used without chromatin, normal mouse IgG was used replacing the b-catenin antibody to reveal non-specific immunoprecipitation of the chromatin.

Statistics
Statistics were performed by two-sample t-test. P-values less than 0.05 were considered to be significant.