SEARCH

SEARCH BY CITATION

Keywords:

  • Gsk3β;
  • cleft palate;
  • palatal elevation;
  • palate development

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

In Wnt/β-catenin signaling pathway, Gsk3β functions to facilitate β-catenin degradation. Inactivation of Gsk3β in mice causes a cleft palate formation, suggesting an involvement of Wnt/β-catenin signaling during palatogenesis. In this study, we have investigated the expression pattern, tissue-specific requirement and function of Gsk3β during mouse palatogenesis. We showed that Gsk3β is primarily expressed in the palatal epithelium, particularly in the medial edge epithelium overlapping with β-catenin. Tissue-specific gene inactivation studies demonstrated an essential role for Gsk3β in the epithelium for palate elevation, and disruption of which contributes to cleft palate phenotype in Gsk3β mutant. We observed that expression of Aixn2, a direct target gene of Wnt/β-catenin signaling, is ectopically activated in the mutant tongue, but not in the palate. Our results indicate that Gsk3β is an intrinsic regulator required in the epithelium for palate elevation, and could act through a pathway independent of Wnt/β-catenin signaling to regulate palate development. Developmental Dynamics 239:3235–3246, 2010. © 2010 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Cleft palate is a common birth defect in humans, with a prevalence ranging between 1/1,000 and 1/700 (Koillinen et al., 2005). The development of mammalian secondary palate is a multiple-step process and disruption at any step, environmentally or genetically, could result in a cleft palate formation. In mice, the palatal shelves protrude from bilateral maxillary processes at embryonic day (E) 11.5, and grow vertically along the developing tongue from E12.5 through E14.0. Subsequently, they elevate, meet, and fuse at the midline, ultimately form an intact palate shelf at E14.5 (Ferguson,1988). Palatal shelf elevation is a rapid process, accompanied by extracellular matrix accumulation within the palatal shelf and movement of other craniofacial structures. Impaired palatal shelf elevation, along with defective palate growth and fusion, comprises major causes for cleft palate formation (Ferguson,1988; Gritli-Linde,2007). It has been proposed that palate elevation is triggered by the intrinsic erectile forces, from both palatal mesenchyme and epithelial covering, and the extrinsic influence, such as coordinated movement of adjacent craniofacial and oral structures including descending of the tongue (Ferguson,1988). It is thought that mesenchymal cells provide erectile elevation force and the epithelial covering constrains and guides the force to appropriate direction to manage the elevation (Bulleit and Zimmerman,1985; Ferguson,1988). Studies using gene-targeting mice have demonstrated that failed palatal shelf elevation could be attributed either to intrinsic elevation defects or to other oral structure deformation (reviewed in Gritli-Linde,2007). In addition, anomalous fusion of the palatal shelves with the tongue or the mandible also leads to elevation defect (Rice et al.,2004; Alappat et al.,2005; Casey et al.,2006; Xiong et al.,2009).

The β-catenin–mediated Wnt canonical (Wnt/β-catenin) signaling pathway plays pivotal role in regulating multiple events of embryonic development. This signaling cascade is activated by Wnt ligand binding to the Frizzled (Fz) transmembrane receptors and LRP5/6 co-receptors. The binding disassembles in the cytoplasm the β-catenin destruction complex, which is composed of glycogen synthase kinase-3β (Gsk3β), Axin, and adenomatous polyposis coli (APC). Consequently, β-catenin accumulates and translocates into the nucleus, where it binds to LEF/TCF co-transcription factors and activates target genes expression (van Amerongen and Nusse,2009). In the absence of Wnt ligands binding, β-catenin is phosphorylated by Gsk3β and is further processed to degradation. Gsk3β acts as a negative modulator of Wnt/β-catenin signaling.

Two highly related Gsk3 protein homologues have been identified in mammals: Gsk3α and Gsk3β, which are encoded by two genes (Frame and Cohen,2001; Doble and Woodgett,2003). First recognized by its function in regulating glycogen synthase activity, Gsk3β has been shown to play a role in several cellular processes, including modulating the Wnt/β-catenin signaling (Embi et al.,1980; Frame and Cohen,2001; Grimes and Jope,2001; Doble and Woodgett,2003). During mouse embryonic development, Gsk3β is essential for liver cell survival as well as palate formation (Hoeflich et al.,2000; Liu et al.,2007). Although the specific function for Gsk3α remains unknown, it has been shown that Gsk3α and Gsk3β function redundantly in regulating Wnt/β-catenin signaling in embryonic stem cell lines (Doble et al.,2007).

Wnt/β-catenin signaling has been implicated in the development of various tissues and organs, including the craniofacial structures (Grigoryan et al.,2008). Although a direct role for β-catenin in the regulation of palate development awaits evidence, β-catenin expression was indeed observed in the palatal epithelium and possibly in the mesenchyme as well (Martinez-Alvarez et al.,2000; Tudela et al.,2002; Nawshad and Hay,2003; He et al.,2008). In fact, many lines of recent evidence suggest an involvement of the β-catenin–mediated Wnt canonical signaling in palatogenesis. For example, mice deficient for either Wnt9b, or Lrp6, or Wnt5a exhibit a cleft palate phenotype (Yang et al.,2003; Carroll et al.,2005; He et al.,2008; Song et al.,2009). Wnt11 was found to be expressed in the palatal epithelium and its function is indispensible for palate fusion (Lee et al.,2008). In addition, mutations in WNT5A or WNT3 have been correlated to a cleft palate defect in humans (Blanton et al.,2004; Niemann et al.,2004). All these lines of evidence suggest a role for Wnt canonical signaling in palatogenesis.

We report here that Gsk3β is primarily expressed in the palatal epithelium, particularly in the medial edge epithelium (MEE) region overlapping with β-catenin. To determine the tissue-specific requirement of Gsk3β in palatogenesis, we have generated Gsk3β conditional knockout mice and demonstrated that Gsk3β is specifically required in the epithelium for normal palatogenesis. Mice deficient for Gsk3β exhibit an impaired elevation of the palatal shelves, leading to a complete cleft of secondary palate. We also detected Axin2 expression in the developing palate, and showed that its expression is not altered in the Gsk3β mutant palate. Our results support the idea that Gsk3β is an intrinsic regulator of palate elevation, and its absence does not alter the Wnt/β-catenin signaling pathway in the developing palatal shelves.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Expression of Gsk3β Is Enriched in the Epithelium of Developing Palate

Although it was proposed that Gsk3β is ubiquitously expressed in all cells, tissue-specific expression of this gene has been reported (Ciaraldi et al.,2006). To investigate the role of Gsk3β during palatogenesis, we first examined its expression pattern in the developing palate by in situ hybridization. Our results showed that, at E12.5, Gsk3β is expressed both in the anterior and posterior palate, at a higher level in the palatal epithelium, and an above background level of expression in the palatal mesenchyme (Fig. 1A,C). At E13.5, Gsk3β is expressed in a pattern similar to that in E12.5 embryo, but its expression is mainly restricted to the MEE (Fig. 1B,D). This expression pattern overlaps with that of Catnb, which encodes β-catenin, the central mediator of the Wnt canonical signaling (He et al.,2008). In addition, Gsk3β expression was also found in the epithelium of the developing tongue (Fig. 1). These results suggest that Gsk3β may modulate Wnt/β-catenin signaling function during palate development.

thumbnail image

Figure 1. Gsk3β expression pattern in the developing palate. A,C: At embryonic day (E) 12.5, Gsk3β expression is primarily detected in the epithelium of both the anterior (A) and posterior (C) palatal shelf. B,D: In the E13.5 palatal shelf, Gsk3β expression remains in the palatal epithelium, particularly in the medial edge epithelium (MEE), of the anterior (B) and posterior (D) regions. In addition, Gsk3β is also expressed in the epithelium of the developing tongue. PS, palatal shelf; T, tongue.

Download figure to PowerPoint

Gsk3β Is Specifically Required in the Epithelium to Regulate Palate Development

It was reported previously that Gsk3β-deficient mice exhibit a cleft palate defect, but the mechanisms underlying this phenotype still remained unknown (Liu et al.,2007). To extend the finding and examine the morphological and molecular defects that lead to the complete cleft palate formation in Gsk3β mutant mice, we created by gene targeting in embryonic stem (ES) cells a Gsk3β conditional allele (Gsk3βF/+) in which exon 2 was flanked with the LoxP sequences (Fig. 2). The targeted locus in ES cells and germline transmission were confirmed by Southern blotting (Fig. 2B), mice heterozygous or homozygous for the conditional allele were genotyped by a polymerase chain reaction (PCR) -based method (Fig. 2). To create a null Gsk3β allele (Gsk3β+/−), we crossed Sox2Cre mice to Gsk3βF/+ mice to delete exon 2, which encodes the Gsk3β protein domain responsible for the kinase activity. While mice heterozygous for Gsk3β mutation appear normal and fertile, Gsk3β−/− mice, obtained by intercrosses between Gsk3β+/− mice, showed a complete cleft palate defect and died within 24 hours after birth (Fig. 3B), a phenotype identical to what was reported previously (Liu et al.,2007).

thumbnail image

Figure 2. Creation of Gsk3β conditional mice. A: Targeting scheme. Exon 2 of Gsk3β was flanked by loxP sites (closed arrowheads). PGK-Neo cassette used for isolation of clones was flanked by FRT sites (open arrowheads), and removed by in utero exposure to Flp recombinase. Homology arms used for recombineering are shaded boxes. Primers used for genotyping are denoted as F1, F2, F3, R1. B: Southern blot of correctly targeted embryonic stem (ES) cell clones. Genomic DNA was digested with SpeI. WT locus, 12 kb; targeted locus, 6 kb. Targeted clones are denoted by asterisks (*).C: Polymerase chain reaction (PCR) genotyping of Gsk3β floxed mice after breeding to Sox2Cre mice. Product for WT allele (using F1/R1 primer pair) is present in all animals. Animals transgenic for Cre recombinase results in loss of F2/R1 fragment, while generating F3/R1 fragment, demonstrating loxP sites are functional in vivo. MW, molecular weight ladder; +/−, Gsk3β heterozygote; +/loxP, WT mouse containing one floxed allele.

Download figure to PowerPoint

thumbnail image

Figure 3. Epithelial inactivation of Gsk3β leads to cleft palate formation. A: A wild-type mouse embryo shows an intact palate at embryonic day (E) 18.5. B,C: Global inactivation of Gsk3β (B) or epithelial-specific ablation of Gsk3β (C) generates identical cleft palate phenotype in mice at E18.5. D,G: Coronal sections of an E13.5 wild-type embryo show the secondary palatal shelves growing vertically along the tongue at the anterior (D) and posterior (G) level. E,F,H,I: Gsk3β−/− and Pitx2Cre;Gsk3βF/F palatal shelves exhibit morphology comparable to the wild-type controls at E13.5 in both the anterior and posterior domains. J: At E14.5, the wild-type palatal shelves have elevated and fused to form an intact palatal shelf. K,L: At the same stage, however, the palatal shelves in Gsk3β−/− (K) and Pitx2Cre;Gsk3βF/F (L) remain vertically oriented and fail to elevate. M: Coronal section of an E18.5 wild-type palatal shelf at the posterior level. N.O: At E18.5, the posterior palatal shelves of both Gsk3β−/− and Pitx2Cre;Gsk3βF/F still do not elevate. NS, nasal septum; PS, palatal shelf; T, tongue; asterisks highlight clefting.

Download figure to PowerPoint

Histological analyses of the developing palate of Gsk3β−/− mice at multiple stages revealed a normal development of the palatal shelves before E13.5, as compared to the wild-type controls (Fig. 3D,E,G,H). At E14.5, the palatal shelves have elevated to a horizontal position and fused in the middle line in the wild-type controls (Fig. 3J). In Gsk3β−/− embryo, however, the palatal shelves remained in the vertical position and failed to elevate to the position above the tongue (Fig. 3K). At E18.5, the mutant anterior palatal shelves eventually elevated to a horizontal position but failed to make contact (data not shown), and the posterior palatal shelves remained in a vertical orientation and never elevated (Fig. 3M,N). The delayed/failed elevation of palatal shelves is, therefore, the cause of cleft palate phenotype in Gsk3β−/− mice.

Because Gsk3β expression is detected primarily in the palatal epithelium, and in the mesenchyme as well, we next sought to address the tissue-specific requirement of Gsk3β function during palate development. We chose to use two established transgenic lines Pitx2Cre and Osr2Cre to abort Gsk3β function in tissue-specific manners in the developing palate. Pitx2Cre and Osr2Cre mice have been used in previous studies to delete LoxP-flanked genes in the palatal epithelium and mesenchyme, respectively (Lan et al.,2007; Xiong et al.,2009). We found that Osr2Cre;Gsk3βF/F mice survived postnatally and showed normal palate formation (5/5). In contrast, among 14 Pitx2Cre;Gsk3βF/F mice examined, 11 exhibited a complete cleft of the secondary palate, phenocopying the defect in the Gsk3β−/− mice (Fig. 3C; and data not shown). The incomplete penetrance of cleft palate phenotype in Pitx2Cre;Gsk3βF/F mice could be attributed to variation in Cre activity or expression timing by the Pitx2Cre transgenic allele in different embryos. Nevertheless, histological examination of Pitx2Cre;Gsk3βF/F embryos revealed the development of the palatal shelves do not exhibit any malformation until E13.5 (Fig. 3F,I). However, at E14.5, similar to defects found in Gsk3β−/− mice, the Pitx2Cre;Gsk3βF/F palatal shelves failed to elevate (Fig. 3L). While the anterior palatal shelves did elevate at E18.5, the posterior palatal shelves remained at the vertical position (Fig. 3O). These results indicate that Gsk3β function is required in the palatal epithelium during palate development, and Gsk3β acts in the palatal epithelium to control palatal shelf elevation.

To rule out the possibility that the failure of palate shelf elevation in Gsk3β mutants is due to extrinsic defects such as malformation of the tongue and mandible, we performed in vitro organ culture on a rotary apparatus. Individual embryonic head with removal of mandible of E13.5 embryos harvested from mating of Gsk3β+/− mice was cultured for 24 hr, and then processed for histological examination after genotyping. As shown in Figure 4, after 24 hr in culture, the wild-type controls exhibited elevated palatal shelves (9/9); however, the mutant palatal shelves remained at the vertical position (10/10). These results indicate that intrinsic defects in the absence of Gsk3β in fact contribute to the failure of palate elevation.

thumbnail image

Figure 4. Gsk3β is an intrinsic regulator of palatal shelf elevation. A,B: Coronal sections of embryonic day (E) 13.5 wild-type (A) and mutant (B) heads before culture show palatal shelves at the vertical position. C: Coronal section of an E13.5 wild-type heads after 24 hr in organ culture shows elevated palatal shelves. D: Coronal section of an E13.5 mutant head after 24 hr in organ culture shows failure of palatal shelf elevation. Scale bar = 200 μm.

Download figure to PowerPoint

Gsk3β Is Required for Normal Cell Proliferation and Palatal Epithelial Cell Survival

Defects in cell proliferation and/or survival have been shown to impair normal development of the palatal shelves, leading to the cleft palate formation (Gritli-Linde,2007). To determine if these cellular processes were affected in Gsk3β−/− palate, we performed bromodeoxyuridine (BrdU) labeling and TUNEL assays on E12.5 and E13.5 palatal shelves of wild-type controls and mutants. Our found comparable cell proliferation rates (P > 0.05) in the mesenchyme between the controls and mutant at E12.5 and E13.5 (Fig. 5A,B; and data not shown). In contrast, cell proliferation rates in the epithelium showed significant difference statistically, with a enhanced rate at E12.5 and a decreased level at E13.5 (Fig. 5G). This altered cell proliferation rate in the mutant epithelium is consistent with Gsk3β expression domain in the developing palatal shelf. In additional, we also observed significantly increased apoptotic cells in the palatal epithelium, including the MEE of the Gsk3β−/− palate at E13.5 (Fig. 5D,F). These results indicate that Gsk3β is implicated in the regulation of cell proliferation and survival during palatogenesis.

thumbnail image

Figure 5. Gsk3β is required for normal cell proliferation and epithelial cell survival during palate development. A,B: Bromodeoxyuridine (BrdU) labeling shows cell proliferation levels in the wild-type controls (A) and Gsk3β−/− palatal shelves (B) Black line demarcate the palatal shelf region for counting BrdU-positive cells and apoptotic cells. C–F: Increased apoptotic cells are detected in the Gsk3β−/− palatal region, including the medial edge epithelium (MEE), at embryonic day (E) 12.5 (D) and E13.5 (F) as compared to the wild-type controls (C,E). G: Comparison of the ratio of BrdU-labeled cells in the epithelium of developing palatal shelves in E12.5 and E13.5 wild-type controls and Gsk3β mutants. *P < 0.01; **P < 0.005. H: Comparison of numbers of apoptotic cells found in the palatal epithelium in wild-type controls and mutants at E12.5 and E13.5.

Download figure to PowerPoint

Gsk3β Does Not Regulate Palate Elevation Through Shh, Osr2, and PDGF Signaling

To determine the molecular defects in the Gsk3β−/− palate, we examined the expression of several genes that either exhibit an overlapped expression pattern with Gsk3β in the developing palate or have been shown to regulate palate elevation. Among them, Shh is expressed in the MEE and oral palatal epithelium (Fig. 6A,E; Zhang et al.,2002; Rice et al.,2004). Epithelial-specific inactivation of Shh results in abnormal cell proliferation and apoptosis, as well as failed palatal shelf elevation (Rice et al.,2004; Gritli-Linde,2007). Our in situ hybridization results showed that the expression of Shh, as well as that of its direct target Ptch1, was not altered in the Gsk3β−/− palate (Fig. 6B,F; and data not shown), suggesting that the Shh signaling functions upstream of or in parallel to Gsk3β during palate development. Osr2 is expressed in the oral side palatal mesenchyme and inactivation of Osr2 leads to a defect in palatal shelf elevation in mice (Fig. 6C,G; Lan et al.,2004). We found that Osr2 expression was not changed in the Gsk3β−/− palate (Fig. 5D,H). The platelet-derived growth factor (PDGF) signaling has been shown to regulate palate development (Soriano,1997; Tallquist and Soriano,2003; Ding et al.,2004). Pdgfa and Pdgfc, encoding two of the PDGF ligands, were specifically expressed in the palatal epithelium overlapping with Gsk3β (Fig. 6I,K; Ding et al.,2004). Inactivation of Pdgfc in mice causes delayed palatal shelf elevation, and Pdgfa;Pdgfc double knockout mice exhibit a more severe cleft face phenotype, indicating an essential role for these two genes in craniofacial development (Ding et al.,2004). We, therefore, examined the expression of Pdgfa and Pdgfc in the Gsk3β−/− palate. As a result, we did not observe an altered expression of either gene (Fig. 6J,L,N,P). We concluded that Gsk3β did not regulate expression of these genes during the palatal shelf elevation.

thumbnail image

Figure 6. Gsk3β does not regulate Shh, Osr2, Pdgfa, and Pdgfc expression during palate development. A,E: At E13.5, Shh is expressed in the MEE and oral epithelium in the wild-type palate. B,F: In the Gsk3β−/− palate, Shh is expressed comparably to that in the wild-type controls. C,D,G,H: Osr2 is expressed in the oral side palatal mesenchyme of the wild-type palatal shelf along the anterior–posterior axis (C,G, and its expression is not unaltered in the Gsk3β−/− palate (D,H). I–P: Pdgfa and Pdgfc are expressed throughout palatal and oral epithelium in the wild-type controls (I,M,K,O), the expression of these genes remains unaltered in the Gsk3β−/− palate (J,N,L,P). PS, palatal shelf; T, tongue.

Download figure to PowerPoint

Inactivation of Gsk3β Does Not Alter Activity of Wnt/β-Catenin Signaling in Developing Palate

It was reported previously that Catnb expression is detected in the palatal epithelium, overlapping with that of Gsk3β (Martinez-Alvarez et al.,2000; Tudela et al.,2002; Nawshad and Hay,2003; He et al.,2008). It is thus intriguing to determine if Gsk3β modulates the activity of Wnt/β-catenin signaling, and if Wnt/β-catenin signaling is involved in palate development. To this end, we have taken advantage of TOPGAL transgenic reporter line, which has been widely used to detect the activity of Wnt/β-catenin signaling (DasGupta and Fuchs,1999). To detect potentially enhanced and/or ectpoic Wnt/β-catenin signaling activity in the Gsk3β−/− palate, we compounded the TOPGAL transgenic allele onto the Gsk3β−/− background, and compared the β-galactosidase reporter activity in the mutant mice with that in the TOPGAL controls. In the TOPGAL transgenic control at E13.5, while intensive LacZ expression was detected in the lip and incisor regions, we did not observe any β-galactosidase activity in the developing palate (Fig. 7A), consistent with the previous report (He et al.,2008). If Gsk3β regulates palate development by modulating Wnt/β-catenin signaling embryos, one would expect to see an ectopic activity of the Wnt canonical signaling in the palatal epithelium. Surprisingly, we observed neither enhanced β-gal staining in the lip and incisor regions nor ectopic LacZ expression in the developing palatal shelves (Fig. 7B). These results suggest that either Wnt/β-catenin signaling is not operating in the developing palate or the TOPGAL transgenic reporter has a limited sensitivity. To address these concerns, we chose to examine the expression of Axin2, a direct target gene of Wnt/β-catenin signaling and a reliable indicator of active Wnt/β-catenin signaling (Jho et al.,2002; Ontiveros et al.,2008; Wang et al.,2008). In the wild-type controls, we observed a specific Axin2 expression in the MEE, with some expression in the immediately adjacent mesenchyme (Fig. 7C). The expression of Axin2 in the MEE overlaps with that of Catnb and Gsk3β, indicating the presence of Wnt/β-catenin signaling activity in the MEE region. At the level of Wnt/β-catenin signaling activity in the developing palatal shelves, the TOGAL transgenic allele appears to have limited sensitivity. Interestingly, we did not observe an elevated or ectopic Axin2 expression in the Gsk3β−/− palate (Fig. 7D). However, ectopic Axin2 expression was detected in the mesenchyme adjacent to the epithelium in the developing tongue of Gsk3β−/− mice (Fig. 7D). These results thus suggest that the function of Gsk3β in modulating Wnt/β-catenin signaling is compensated in the palate but not in the tongue. Since Gsk3α has been shown to have functional redundancy with Gsk3β in modulating the Wnt/β-catenin signaling in stem cell lines (Doble et al.,2007), we reasoned that Gsk3α could have an overlapped expression pattern with Gsk3β in the developing palate where they function redundantly. In situ hybridization assays showed that Gsk3α was expressed in the palatal epithelium, primarily in the MEE and the oral side epithelium (Fig. 7E). This expression pattern in the developing palate overlapped almost perfectly with that of Gsk3β (Fig. 1). In the developing tongue, Gsk3α expression was detected only in the mesenchyme and the epithelium in the ventral–lateral region, complementary to the ectopic Axin2 expression domain in the Gsk3β−/− tongue (Fig. 7D,E). In the developing palate and tongue of the Gsk3β mutant, we did not observe an altered expression of Gsk3α (Fig. 7F). These data suggest that Gsk3α may compensate for the loss of Gsk3β function in modulating Wnt/β-catenin signaling in the developing palate. We thus concluded that a Wnt/β-catenin signaling independent pathway regulated by Gsk3β is essential for palate elevation during normal palate development.

thumbnail image

Figure 7. The Wnt/β-catenin signaling activity is not affected in the Gsk3β−/− palate. A: At embryonic day (E) 13.5, the TOPGAL reporter activity is detected in the lip and the upper incisors (blank arrows), but not in the palatal shelves (arrow). B: The I activity is also not detected in the Gsk3β−/−;TOPGAL palate (arrows). C: In the E13.5 wild-type embryo, Axin2 is expressed in the developing palatal shelves, particularly in the MEE and underlying mesenchyme, as well as the mesenchyme immediately adjacent to the epithelium in the ventral-lateral region of the developing tongue (arrow). D: In the E13.5 Gsk3β−/− embryo, Axin2 expression is detected in the developing palatal shelves, with expression pattern and level comparable to that in the wild-type control. However, ectopic Axin2 expression is found in the mesenchyme adjacent to the epithelium in the developing tongue (red arrows). E: Gsk3α is expressed in the palatal epithelium, including the medial edge epithelium (MEE) but not in the mesenchyme in the dorsal–lateral region of the tongue (arrow) in the wild-type control. F: Gsk3α expression is unaltered in the Gsk3β−/− palate. PS, palatal shelf; T, tongue; U, upper incisor.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The β-catenin–mediated Wnt canonical signaling has been implicated in many developmental events. However, direct evidence for an involvement of this signaling pathway in palate development is still lacking. In this study, we reported that Gsk3β is primarily expressed in the palatal epithelium, and its major function in palate development is to regulate palatal shelf elevation. Our Gsk3β conditional knockout mice enable us to determine that Gsk3β is specifically required in the palatal epithelium for palatal shelf elevation. We further demonstrate the presence of Wnt/β-catenin signaling activity in the normal developing palatal shelves, and the activity remains unaltered in the Gsk3β mutant palate. Our studies thus support the idea that Gsk3β is an intrinsic epithelium-expressed regulator of palatal shelf elevation and functions independently of Wnt/β-catenin signaling.

Gsk3β Is an Epithelially Expressed Intrinsic Regulator of Palatal Shelf Elevation

The phenomenon of palatal shelf elevation has been recognized for a long time, but the mechanism of this process remains far from being well understood. It is now widely accepted that palatal shelf elevation is a rapid process, undergoing in an anterior to posterior sequence and being controlled by both intrinsic erectile force and extrinsic influences from other craniofacial structures (Lazzaro,1940; Ferguson,1988). Several mechanisms have been proposed for this process. Considering the intrinsic factors, it has been suggested that hyaluronan differential accumulation, type I collagen distribution, mesenchymal cell alignment, and epithelial covering all provide elevation force during palatal shelf elevation (reviewed in Ferguson,1988). Although the mesenchymal tissue was thought to provide major driving force for the elevation, removal of oral epithelium was shown to inhibit the reorientation of palatal shelves in vitro. The palatal epithelium is thus considered to constrain and direct the elevation force generated by the mesenchymal core during palatal shelf elevation (Bulleit and Zimmerman,1985). Temporally and spatially, palatal shelf elevation occurs coincidentally with the descending of the tongue, and the growth of the mandible and basicranium. Malformation of these structures, therefore, could also prevent palatal shelf elevation as extrinsic factors (Ferguson,1988).

Corresponding to the above-proposed mechanisms for palatal shelf elevation, defects in palatal shelf elevation in gene-targeting mice could be categorized into three major classes. First, the palatal shelf elevation can be impaired by intrinsic defects, as seen in Osr2 or Pdgfc mutant mice (Ding et al.,2004; Lan et al.,2004). Second, palatal shelf elevation could be hindered by abnormal fusion between palatal shelf and adjacent oral structures. For example, in mice deficient for Jagged2, or Fgf10, or Hand2, the palatal shelves abnormally fuse with the developing tongue or the mandible, preventing palatal shelf elevation and causing cleft palate formation (Alappat et al.,2005; Casey et al.,2006; Xiong et al.,2009). Third, malformation of the tongue or mandible could physically block palatal shelf elevation, as it is seen in Hox2a and Gli3 mutant mice (Barrow et al.,1999; Huang et al.,2008). In this study, we show that Gsk3β is expressed and functions in the palatal epithelium. Inactivation of Gsk3β impairs palatal shelf elevation without obvious extrinsic defects (for example, tight contact with the tongue or malformation of the mandible) or abnormal fusion with adjacent tissues. Moreover, the observation that the failure of palatal shelf elevation in the mutant head without mandible in organ culture rules out the possibility of extrinsic effects. These results support the conclusion that Gsk3β is an intrinsic regulator of palatal shelf elevation.

Gsk3β has been implicated in cell proliferation regulation (Force and Woodgett,2008; Kerkela et al.,2008). The absence of Gsk3β results in cardiomyocyte hyperproliferation without influence on β-catenin signaling (Kerkela et al.,2008). Consistent with this line of discoveries, we found an altered cell proliferation rate in the mutant palatal epithelium where Wnt/β-catenin signaling activity remained unaltered. It appears that the role of Gsk3β on cell proliferation regulation in the palatal shelves depends largely on developmental stages. We certainly cannot rule out the possibility that this altered cell proliferation rate, together with enhanced cell apoptosis in the palatal epithelium, may also contribute to the cleft palate formation in Gsk3β mutants.

Gsk3β has been shown to modulate activities of several signaling pathways (Frame and Cohen, 2001). Among them is Shh signaling, which is critical for palate elevation (Gritli-Linde,2007; Lan and Jiang,2009). Our findings show that Gsk3β is expressed overlapping with Shh in MEE (Fig. 1A,C) and the mutants also exhibit palate elevation defect (Fig. 3K,N). Shh signaling is thus potential related to the defective palate development of the Gsk3β mutant. However, our in situ hybridization studies showed an unaltered expression of Shh and its transcriptional target Patch1 in the mutant palate, indicating that Gsk3β does not regulate palatal shelf elevation by controlling Shh signaling activity. The expression of Osr2, Pdgfa, and Pgdfc is neither affected in the Gsk3β mutant palatal shelves. Future studies comparing gene expression profiles in the wild-type and Gsk3β mutant palatal shelves would help to identify downstream targets of Gsk3β in the developing palate.

The role of Gsk3β in palate development seems to vary slightly in different genetic background. In the present study, we have shown that Gsk3β controls palate elevation in an inbred C57BL/6 background. In an outbred CD-1 background; however, Gsk3β deficiency does not affect palate elevation, although it also causes cleft palate defect in the null mutant (Liu et al.,2007). Of interest, in another study by Hoeflich and colleagues (2000), Gsk3β homozygous null mice were reported to die during mid-gestation, a phenotype we never observed in our studies. Thus genetic background appears to exert influence on phenotype severity in the absence of Gsk3β. Such correlations between genetic background and phenotype have been reported for other genes (Dansky et al.,1999; Yang et al.,2005).

Implication of the Wnt/β-Catenin Signaling in Palate Development

The Wnt/β-catenin signaling has been implicated in the formation of many organs (Grigoryan et al.,2008). Several Wnt ligands, including Wnt9b, Wnt5a, and Wnt11, have been shown to play a role in palate development (Yang et al.,2003; Carroll et al.,2005; He et al.,2008; Lee et al.,2008). However, these lines of evidence do not support a direct involvement of the Wnt/β-catenin signaling. This is because, in Wnt9b mutant mice, the cleft palate phenotype is a secondary defect (Lan et al.,2006). Wnt5a was shown to act through a β-catenin–independent signaling pathway to regulate palate development (He et al.,2008). Stronger evidence, though still not direct, came from a recent publication in which cleft lip and cleft palate defects are observed in Lrp6 mutant mice (Song et al.,2009). While the TOPGAL reporter activity is not detectable in the developing palate (He et al.,2008; this study), here we show that Axin2, a direct target gene of Wnt/β-catenin signaling and a sensitive indicator of active Wnt/β-catenin signaling (Jho et al.,2002; Ontiveros et al.,2008; Wang et al.,2008), is expressed in the palatal epithelium and restricted to the MEE region. This expression pattern is overlapped with that of Gsk3β and β-catenin (Fig. 1; He et al.,2008). These results demonstrate the presence of active Wnt/β-catenin signaling in the developing palate. Together with the cleft palate phenotype found in Lrp6 knockout mice, these observations support a role for the Wnt/β-catenin signaling in palate development.

Functional Redundancy of Gsk3β and Gsk3α in Palate Development

The mammalian Gsk3 genes have two homologues, Gsk3α and Gsk3β. Although Gsk3β alone is able to modulate the Wnt/β-catenin signaling, in some cases, inactivation of Gsk3β does not necessarily lead to Wnt/β-catenin signaling up-regulation. For instance, in ES cell lines, Gsk3β and Gsk3α have been shown to function redundantly in modulating Wnt/β-catenin signaling (Doble et al.,2007). Wnt/β-catenin signaling activity, which is assessed by the expression of Axin2 transcripts and TCF reporter, is not significantly elevated until three of the four Gsk3 alleles are depleted (Doble et al.,2007). In mouse embryonic fibroblasts, on the other hand, loss of Gsk3β function itself enhances Wnt/β-catenin signaling activity (Kapoor et al., 2008). The differential outcome of Gsk3β ablation in the embryonic stem cells and fibroblasts may reflect the differential expression of Gsk3α in these two types of cells. Of interest, we found that Gsk3α is expressed in a pattern overlapped with that of Gsk3β in the developing palate, but not in the tongue. Consequently, an ectopic Wnt/β-catenin signaling activity is found in the Gsk3β−/− tongue, but not in the Gsk3β−/− palate. These observations suggest a redundant role for Gsk3β and Gsk3α during palate development. The absence of Gsk3β does not alter the Wnt/β-catenin signaling, suggesting that Gsk3β could also mediate a signaling pathway independent of Wnt/β-catenin signaling during palate development.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Generation and Genotyping of Gsk3β Conditional Knockout Mice

Targeting construct for Gsk3β was created using a recombineering strategy (Liu et al.,2003). Briefly, a bacterial artificial chromosome (BAC) clone containing the murine Gsk3β locus, from Children's Hospital of Oakland Research Institute (CHORI), was transformed into EL350 cells. Polymerase chain reaction (PCR) was used to retrieve approximately 400-bp homology arms flanking exon 2 of Gsk3β by 5-kb on each side. Deletion of this exon has been shown to result in a nonfunctional kinase (Hoeflich et al.,2000). After cloning homology arms into a retrieval vector (pBluescript modified to contain a thymidine kinase [TK] cassette), this construct was transformed into the EL350 BAC-containing cells, and induced homologous recombination by means of gap repair, resulting in a 10-kb fragment of Gsk3β being retrieved into the pBluescript derivative. Next, mini-targeting vectors were created by introducing identical 34-bp loxP sites on both sides of a neomycin resistance cassette (neomycin phosphotransferase gene; NeoR). The loxP/NeoR cassette was then introduced into gap repaired retrieval plasmids using a similar approach as just described above for retrieving the Gsk3β fragment. Transformation of these constructs into bacteria with arabinose-inducible Cre allowed recombination, removing the NeoR cassette and one loxP site. An additional round of this cloning approach was used to introduce a second loxP/NeoR cassette, now flanked by FRT sites, on the 3′ side of exon 2. Plasmids were linearized with NotI, gel purified, and electroporated into mouse ES cells by the ES Cell and Transgenic Core Facility at The Research Institute at Nationwide Children's Hospital. Targeted ES cells were grown in the presence of G418 (350 μg/ml) and FIAU (5-iodo-2′-fluoro-2′-deoxy-arabinouridine; 200 nM) for the selection of cells which contained correct gene targeting.

Southern blotting of G418/FIAU-resistant clones revealed 11 of 288 colonies (3.8%) contained a correctly targeted Gsk3β floxed allele. Two clones were selected for injection into blastocysts, from which we obtained several male chimeric mice. These mice were mated to either ROSA26-FLPe mice, a well-characterized transgenic mouse that ubiquitously expresses Flp at high levels (Farley et al.,2000) resulting in the highly efficient removal of the NeoR cassette, or wild-type C57BL/6 females. Primer information for PCR genotyping is available upon request.

Other Animals

The conventional Gsk3β mutant (Gsk3β+/−) mice were generated by intercrossing Gsk3β conditional (Gsk3βF/F) mice with the Sox2Cre transgenic mice. Sox2Cre and TOPGAL mice were obtained from the Jackson Laboratories (Bar Harbor, ME) and were genotyped as described previously (DasGupta and Fuchs,1999; Hayashi et al.,2002). The generation and genotyping of Pitx2Cre and Osr2Cre transgenic mice have been reported previously (Lan et al.,2007; Xiong et al.,2009). All the genetically engineered mice used in this study were maintained on the B6/C57 background. The age of embryos was defined as embryonic day 0.5 (E0.5) in the morning of the day when a vaginal plug was discovered. All animal studies were approved by the Tulane University Institutional Animal Care and Use Committee.

In Vitro Organ Culture

E13.5 embryos were harvested from Gsk3β+/− crosses and decapitated. After removal of mandible, each embryonic head was placed in 2 ml DMEM culture medium supplemented with 20% fetal calf serum in a 20-ml glass bottle. Bottles were fixed in vertical position on a rotary apparatus rotating at a speed of 4 rpm in an incubator at 37°C and 5% CO2. Samples were harvested after 24 hr in culture, and processed for histological examination.

Histology and In Situ Hybridization

Staged embryos were harvested in ice-cold phosphate buffered saline (PBS), and embryonic heads were removed and fixed in 4% paraformaldehyde (PFA)/PBS overnight at 4°C before dehydration through a graded ethanol series and embedded in paraffin. Samples were processed into 10-μm coronal sections for histological analysis by standard hematoxylin/eosin staining or for nonradioactive in situ hybridization, as described previously (St. Amand et al.,2000). For the expression of each gene, three independent experiments of in situ hybridization were performed.

X-gal Staining

Embryos were collected, and embryonic heads were dissected in ice-cold PBS. The mandible of each head was carefully removed with a pair of fine forceps and discarded. Samples were washed in PBS, and fixed in 1% glutaraldehyde/PBS at 4°C for 1 hr. After three washes in PBS, samples were stained in X-gal solution (400 μg/ml, with potassium ferricyanide and potassium ferrocyanide) overnight at room temperature for β-galactosidase activity.

Cell Proliferation and TUNEL Assays

Cell proliferation rate was measured using the Bromodeoxyuridine (BrdU) Labeling and Detection Kit (Roche Diagnostics Corporation, Indianapolis). Briefly, timed pregnant female mice were injected intraperitoneally with BrdU solution (1.5 ml/100 g body weight) 1 hr before embryos were harvested. Embryonic heads were fixed in Carnoy's fixative, dehydrated, embedded in paraffin, and sectioned at 5 μm for immunostaining according to manufacturer's instruction (He et al.,2008). BrdU-labeled cells were counted within arbitrarily defined areas and presented as percentage of total nuclei of the same area. With data collected from nine sections of three individual samples of wild-type controls and mutants, respectively, were subjected to Student's t-test to determine the significance of differences. For TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling) assays, samples were fixed in 4% PFA/PBS, dehydrated, and embedded in paraffin. Coronal sections were made at 5 μm thickness. Apoptotic cells were detected as described previously (Alappat et al., 2005).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Dr. Hao Ding of University of Manitoba, Canada and Dr. Andras Nagy of Mount Sinai Hospital, Canada, for providing Pdgfa and Pdgfc probes. We thank Dr. Rulang Jiang of the University of Rochester for providing Osr2-Cre mice. C.J.P. and Y.C. were funded by the NIH.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES