Sox2 Controls Periderm and Rugae Development to Inhibit Oral Adhesions

Abstract

In humans, ankyloglossia and cleft palate are common congenital craniofacial anomalies, and these are regulated by a complex gene regulatory network. Understanding the genetic underpinnings of ankyloglossia and cleft palate will be an important step toward rational treatment of these complex anomalies. We inactivated the Sry (sex-determining region Y)–box 2 (Sox2) gene in the developing oral epithelium, including the periderm, a transient structure that prevents abnormal oral adhesions during development. This resulted in ankyloglossia and cleft palate with 100% penetrance in embryos examined after embryonic day 14.5. In Sox2 conditional knockout embryos, the oral epithelium failed to differentiate, as demonstrated by the lack of keratin 6, a marker of the periderm. Further examination revealed that the adhesion of the tongue and mandible expressed the epithelial markers E-Cad and P63. The expanded epithelia are Sox9-, Pitx2-, and Tbx1-positive cells, which are markers of the dental epithelium; thus, the dental epithelium contributes to the development of oral adhesions. Furthermore, we found that Sox2 is required for palatal shelf extension, as well as for the formation of palatal rugae, which are signaling centers that regulate palatogenesis. In conclusion, the deletion of Sox2 in oral epithelium disrupts palatal shelf extension, palatal rugae formation, tooth development, and periderm formation. The periderm is required to inhibit oral adhesions and ankyloglossia, which is regulated by Sox2. In addition, oral adhesions occur through an expanded dental epithelial layer that inhibits epithelial invagination and incisor development. This process may contribute to dental anomalies due to ankyloglossia.

Introduction

Ankyloglossia is a common condition affecting up to 10% of newborns (Chandrashekar et al. 2014; Veyssiere et al. 2015; Yoon et al. 2017). The condition presents with an inappropriate fusion of the tongue to the alveolar ridge or floor of the mouth and can have several debilitating consequences, including difficulty with breastfeeding and problems with speech, swallowing, and gaining weight (Chandrashekar et al. 2014; Ferrés-Amat et al. 2016). Although the condition is readily treatable in human patients, the genetic underpinnings of ankyloglossia remain largely uninvestigated.

Clefting is one of the most common congenital defects observed in humans and has, in some cases, been shown to arise from oral adhesion (Richardson et al. 2009; Paul et al. 2017). Another defect observed in humans associated with ankyloglossia is tooth abnormalities, as reported in several families (Chandrashekar et al. 2014; Lenormand et al. 2018).

During craniofacial development, the oral cavity is usually protected from inappropriate oral adhesions by the formation of a transient structure called the periderm. The periderm is derived from the basal layer of the oral mucosa and lines the oral cavity. These flat Keratin 6 (K6)–, Keratin 17–positive cells have been shown to provide a barrier function in the developing oral cavity and skin. Several genes have been implicated in the development of the periderm, including Irf6, Grhl3, Ikka, and Sfn, and the genetic ablation of any of these genes in the oral ectoderm results in oral adhesions leading to cleft palate (Richardson et al. 2014; Kousa et al. 2017; Hammond et al. 2019). These previous studies indicate that ankyloglossia may be related with the abnormal formation of periderm during development.

We recently reported that the inactivation of Sox2 in the oral epithelium with the Pitx2^Cre resulted in tooth agenesis and defective invagination of developing tooth buds, which remained at the surface of the oral epithelial layer (Sun et al. 2016). Further examination of these embryos revealed that they also had a cleft palate and ankyloglossia. We found that the periderm was not well formed and poorly organized in this mouse model, resulting in ankyloglossia. Additionally, we show that the oral adhesion is composed of the inappropriate adhesion of the tongue and dental epithelia, linking ankyloglossia with defects in dental development.

Materials and Methods

Mouse Lines and Embryonic Staging

Mouse experiments were performed in accordance with rules provided by the Office of Animal Resources at the University of Iowa. The Sox2^Flox/Flox, Pitx2^Cre, and Shh^Cre were described in a previous publication (Liu et al. 2003; Taranova et al. 2006; Sun et al. 2016). The Pitx2^Cre/Sox2^F/F mice are termed Sox2^CKO.

Genotype Primers

Sox2 WT:

• Forward: 5′ GCTCTGTTATTGGAATCAGGCTGC 3′

• Reverse: 5′ CTGCTCAGGGAAGGAGGGG 3′

Sox2 CKO:

• Forward: 5′ CAGCAGCCTCTGTTCCACATACAC 3′

• Reverse: 5′ CAACGCATTTCAGTTCCCCG 3′

Tissue Fixation and Slide Preparation

Embryos were fixed in 4% PFA (ChemCruz) and embedded in paraffin; samples were sectioned at 7 µm with Thermo (HM325) microtome per previous reports (Sun et al. 2016).

Hematoxylin and Eosin Staining

Sections were stained by hematoxylin and eosin as previously described (Sun et al. 2016).

Immunofluorescence Staining and Confocal Imaging

Slides were subjected to citric acid antigen retrieval, washed 2 times with 1X PBS for 5 min, and blocked with 20% donkey serum in PBS for 30 min at room temperature. Slides were then incubated with primary antibody (Pitx2 antibody, 1:50, AF7388 [R&D Systems]; Sox2 antibody, 1:50, AF2018 [R&D Systems]; P63 antibody, 1:50 [Biocare Medical]; Loricrin, 1:500, 905101 [BioLegend]; Keratin 6A, 1:500, 905702 [BioLegend]; Desmoplakin, 1:50 [Bio-Rad]) overnight at 4 °C. Slides were washed with 1X PBS, and secondary antibody (Life Technologies) was applied for 30 min, incubated with DAPI solution (Thermo Scientific), and mounted with cover slips. Confocal pictures were taken with a ZEISS 700 confocal microscope and Zen imaging software.

IdU/CIdU Labeling Assay

CIdU was injected 24 h prior to harvesting embryonic day 15.5 (E15.5) embryos (i.e., E14.5), and IdU was injected 1 h prior to harvesting embryos. Both analogues were injected at 100-µg/g body weight of the pregnant female mouse. Embryos were then harvested and embedded into paraffin with the approach described previously. Staining for IdU/CIdU was performed in accordance with a previous study (Tuttle et al. 2010; Sun et al. 2016).

Quantitative Reverse Transcription Polymerase Chain Reaction

Total RNAs were extracted with the RNeasy Mini Kit from Qiagen and reverse transcribed into cDNA with PrimeScript RT Reagent Kit (TaKaRa).

Polymerase Chain Reaction Primers

SOX2:

• Forward: 5′ GCCGAGTGGAAACTTTTGTCG 3′

• Reverse: 5′ GGCAGCGTGTACTTATCCTTCT 3′

KRT6A:

• Forward: 5′ TGCTGCCTACATGAACAAGG 3′

• Reverse: 5′ TGTCTGAGATGTGGGTCTGC 3′

ChIP

The chromatin immunoprecipitation (ChIP) assay was performed on GMSMK cells with a ChIP assay kit (17-295; Millipore) and Sox2 Ab (R&D systems). ChIP primers were selected around the predicted Sox2 binding site and a control region, and regions were amplified by polymerase chain reaction.

ChIP Primers

Negative control:

• Forward: 5′ CTTCTTCCAAATATGCCCGTCAGTG 3′

• Reverse: 5′ CACATTGAGTTTGACGCATGTTC 3′

SOX2 BS:

• Forward: 5′ CCAAATGTTGGAGAAATGGGACTG 3′

• Reverse: 5′ GTTTACAGAGGAATGAGCTTCACTTCTCC 3′

Statistical Analysis for Experiments

Each experiment was repeated 3 times with at least 3 mutant and control embryos for mouse studies. The results are shown as the mean ± standard error of the mean (SEM), and the analyses were performed with an independent 2-tailed t test.

Results

Sox2^CKO Embryos Develop Ankyloglossia due to Defects in Periderm Formation

We examined the E13.5, E14.5, E15.5, and E18.5 developmental stages of murine embryos with sagittal sections of control (Pitx2^Cre) and Pitx2^Cre/Sox2^F/F (Sox2^CKO) embryos. Interestingly, a delay was noted in the invagination of the dental epithelium in the Sox2^CKO embryos (Fig. 1A, B). By E14.5, the delay in dental epithelial invagination may have caused the fusion of the tip of the tongue to the mandible (Fig. 1D, arrow). The apparent fusion of the dental and tongue epithelium results in a further disruption of incisor dental epithelial cell invagination. This ankyloglossia phenotype was present at E15.5 and E18.5 as compared with the control embryos (Fig. 1F, F′, H, H′). The ankyloglossia phenotype was also observed in the Shh^Cre-Sox2^F/F embryos, demonstrating that 2 independent murine Cre lines created similar Sox2^CKO phenotypes (Fig. 4E′).

Figure 1.

Sox2^CKO embryos exhibit ankyloglossia beginning at embryonic day 14.5 (E14.5). (A, A′) Sagittal sections and hematoxylin and eosin staining of E13.5 embryos reveal normal development of the tongue in control (Pitx2^Cre) embryos. The tongue is connected to the mandible in the posterior region of the mandible. (B, B′) In Sox2^CKO embryos, the tongue shows no signs of fusion in the anterior region at E13.5. (C, C′) Hematoxylin and eosin staining reveals that in control embryos, the tongue and the mandible are separated in the anterior region at E14.5. (D, D′) In Sox2^CKO embryos, the anterior portion of the tongue is fused to the mandible with an oral adhesion (black arrow). (E, E′) At E15.5, the tongue is elongated and clearly separated from the mandible in control embryos. (F, F′) Fusion of the tongue to the mandible persists in Sox2^CKO embryos. The fusion is located at the anterior region of the tongue above the tooth bud region. (G, G′) In control E18.5 embryos, tongue development proceeds, and the anterior tongue is separated from the mandible. (H, H′) In Sox2^CKO embryos, ankyloglossia persists at E18.5 (black arrow). Dotted lines in A′-D′ represent the developing dental epithelium, which is smaller and fails to invaginate in Sox2^CKO embryos. F′ Arrows denote ankyloglossia. Scale bar = 100 μm.

To determine the nature of the tissue connecting the tongue and the mandible, we performed immunostaining for markers of different cell types found in the developing oral mucosa, and we determined that the adhesion tissue connecting the mandible epithelium to the tongue epithelium in Sox2^CKOembryos expressed E-cadherin (Fig. 2A, B). We next stained for K6, a protein expressed by the transient periderm layer. While the E15.5 control embryos had a well-formed layer of K6⁺ periderm cells on the surface of the developing mandible (Fig. 2C, C′), this layer was poorly formed in Sox2^CKOembryos (Fig. 2D, D′), and only a few K6⁺ cells were identified in the anterior mandibular oral epithelium. Interestingly, the K6⁺ periderm layer identified in the controls at E15.5 also expressed Sox2, suggesting a potential role for Sox2 in the formation of the periderm. To verify the results of the K6 staining, we next stained for Grhl3, a reported marker of the periderm (Fig. 2E, F). In Sox2^CKO embryos, Grhl3 protein levels were significantly reduced in regions where Sox2 is expressed as compared with the control.

Figure 2.

Loss of Sox2 results in oral adhesion and impaired formation of the periderm. (A) In control embryos (embryonic day 15.5 [E15.5]), staining for E-cadherin labels the tongue, oral, and dental epithelial tissues. (B) When Sox2 is ablated, E-cadherin labels the same structures, including the lower incisor, which fails to invaginate in these embryos. The oral adhesion, which is indicated with a black arrow, is also clearly expressing E-cadherin, demonstrating that the adhesion is epithelial tissue. (C, C′) In control embryos, Sox2 protein is labeled in the tongue and oral epithelial tissues, including the vestibule lamina. Staining for Keratin 6 also labels the periderm, as indicated by the green arrow. (D, D′) When Sox2 is ablated with Pitx2^Cre, Sox2 protein is no longer labeled in the tongue and oral epithelial tissues. Additionally, the Keratin 6–positive periderm structure is almost completely ablated. (E) Grhl3 protein is expressed in multiple oral epithelia layers and highly expressed in the periderm of control embryos. (F) In Sox2^CKO embryos, Grhl3 expression was diminished. (G) In control embryos, Desmoplakin protein is not expressed on the apical side of the most superficial oral epithelia. (H) In Sox2^CKO embryos, Desmoplakin is ectopically expressed on the apical layer of the tongue and oral epithelia and within the oral adhesion connecting the tongue and mandible. (I, I′) P63 and Sox2 double staining reveals a well-ordered P63⁺ basal layer (arrow). Superior to the basal layer are Sox2 and P63 double-positive cells. Finally, the periderm, the most superficial layer, contains only Sox2 protein. (J, J′) When Sox2 is ablated, the well-ordered basal layer is disrupted. The oral adhesion is P63 positive. (K) Pregnant mice timed to E14.5 were injected with CIdU, were injected with IdU 23 h later, and sacrificed 1 h later. (L) In Sox2^CKO littermates, the oral adhesion is CIdU positive, indicating that the oral epithelial layer was actively proliferating at E14.5. (M) The number of CIdU+ cells is significantly higher in Sox2^CKO embryos as compared with the control. **P < 0.01. Values are presented as mean ± SEM. (N) Loricrin is expressed in the space separating the tongue and the mandible at E15.5 (yellow arrow). (O) In Sox2^CKO embryos, Loricrin staining is lost the oral adhesion (yellow arrow). CIdU positive cells were quantified in the oval region. LI- Lower incisor, Md- Mandible, T- Tongue. Scale bar = 100 μm.

Previous work has demonstrated that cell adhesion complexes, including desmosomes, are localized between basal cells but are not present on the apical surface of the epidermis, thus preventing the abnormal adhesion of opposing epithelial layers in the developing skin (Richardson et al. 2014). To determine if ectopic apical organization of desmosomes contributed to the ankyloglossia phenotype that we observed in the Sox2^CKO embryos, we stained for Desmoplakin, a component of desmosomes. In control embryos, staining for Desmoplakin revealed that it was not localized apically on either the mandibular or tongue epithelia (Fig. 2G). Interestingly, in Sox2^CKO embryos, Desmoplakin expression was no longer restricted, and the tongue and mandibular epithelial layers were connected by cells with unpolarized Desmoplakin localization (Fig. 2H). These data demonstrate that the malformation of the periderm resulting from the inactivation of Sox2 allows for the mislocalization of desmosomes, resulting in epithelial adhesion.

Costaining of P63 and Sox2 in control E15.5 embryos (Fig. 2I, I′) revealed a well-organized single layer of P63⁺, Sox2^- basal cells and a suprabasal Sox2⁺, P63⁺ layer. The periderm layer separates the suprabasal layer from the oral cavity. Interestingly, in Sox2^CKO embryos, the P63⁺ basal layer was poorly organized. The disorganized P63⁺ cells did not give rise to the K6⁺ periderm layer observed in control embryos but instead expanded and became attached to the tongue epithelium (Fig. 2J, J′).

Deletion of Sox2 Results in Increased Proliferation of the Dental Epithelium within the Oral Adhesion

We examined E15.5 control and Sox2^CKO embryos for differences in proliferation using injections of the 2 thymidine triphosphate analogues, CIdU (24 h prior to sacrifice) and IdU (1 h prior to sacrifice), coupled with immunofluorescence staining to label proliferating cells (Fig. 2K, L). Neither the control nor the Sox2^CKO oral epithelial region had large numbers of proliferating cells 1 h prior to sacrifice (green), although the Sox2^CKO embryo had a much greater number of CIdU⁺ cells (red), indicating that the oral epithelium was proliferating 24 h prior to harvesting the embryos (quantification, Fig. 2M).

Loricrin is a differentiation marker expressed in mammalian stratified epithelia that contributes to barrier function (Nithya et al. 2015). Interestingly, while the control embryos had a coherent layer of Loricrin protein marking the superior layer of the oral and dental epithelium, Sox2^CKO embryos failed to develop the Loricrin⁺ cell layer in the region of the mandible fused to the tongue (Fig. 2N, O), the same adhesion region previously shown to have greater numbers of CIdU⁺ cells. Thus, we concluded that the oral adhesion disrupts the formation of the Loricrin+ layer specifically at the adhesion region of the oral mucosa.

Oral Adhesion Is Composed of Dental Epithelial Cells

A major unanswered question concerning oral adhesions is the tissue type that directly interacts with the tongue epithelium. We stained for known dental epithelial marker expression in the oral adhesion because the oral adhesion connecting the tongue and the mandible occurs at the lower incisor region comprising the oral and dental epithelium. Other groups have demonstrated that pathogenic oral adhesions resulting in cleft can occur at the molar tooth buds (Richardson et al. 2006; Richardson et al. 2014; Paul et al. 2017). To determine if the dental epithelial tissue might participate in the formation of adhesions, we stained for Sox9, which is specifically expressed in the enamel organ of the developing lower incisor in control embryos (Fig. 3A, A′). Indeed, we demonstrate that many of the oral adhesion cells express the Sox9 protein (Fig. 3B, B′).

Figure 3.

Oral adhesions are composed of dental epithelial cells. (A, A′) Staining for Sox9 marks dental epithelial cells in control lower incisors. The white boxed region is shown in higher magnification in A′. (B) In Sox2^CKO embryos, the cells composing the oral adhesion are also Sox9+, indicating that they are derived from dental epithelial tissue. The white boxed region is magnified in B′. (C, C′) In control embryos, Pitx2 is strongly expressed by the oral and dental epithelial tissue, including the developing lower incisor. (D, D′) In the Sox2^CKO embryos, oral adhesion cells are also Pitx2+. (E) The tongue epithelium expresses Tbx1, which is also expressed by dental epithelial cells. (E′) In the Sox2^CKO embryo, the oral adhesion is composed by Tbx1+ cells.

We stained for Pitx2 and Tbx1, which are both markers of the dental and oral epithelial tissue in control embryos (Fig. 3C, C′, E). Interestingly, the cells composing the oral adhesion expressed these markers associated with the dental epithelium (Fig. 3D, D′, E′). Thus, we conclude that the oral adhesion is composed of expanded dental epithelial cells.

Sox2 Regulates the Palate Rugae Signaling Center and Palatogenesis

We sectioned the heads of E15.5 control and Sox2^CKO embryos in the coronal orientation and found that whereas control E15.5 embryos had a well-formed, completely fused palate (Fig. 4A–C), palate closure had not occurred in Sox2^CKO embryos. Shelf extension, the process by which the 2 palate shelves meet at the midline at E15.5, had not occurred, nor did it occur at later stages (Fig. 4A′–C′). Furthermore, the Shh^Cre-GFP/Sox2^fl/fl E18.5 embryos had a cleft palate (Fig. 4D, D′) and ankyloglossia (Fig. 4E, E′). In addition, these embryos lacked palate rugae. Thus, using the Pitx2^Cre and Shh^Cre lines resulted in similar Sox2^CKO phenotypes.

Figure 4.

Ablating Sox2 in the oral epithelium results in cleft palate. (A–C) Coronal sectioning of a control embryonic day 15.5 (E15.5) embryo in the anterior, middle, and posterior of the palate reveals complete fusion of the shelves at this stage. (A′–C′) Coronal sectioning of the Sox2^CKO E15.5 embryos reveals cleft palate. The asterisks in panel B′ designate molar tooth buds, which fail to develop in Sox2^CKO embryos. (D, D′, E, E′) Shh^CreGFP/Sox2^fl/fl embryos have a cleft phenotype and ankyloglossia (black arrow denotes the oral adhesion), respectively. (F, G) Sagittal sectioning of an E15.5 control embryo and an enlargement of the palate. Palatal rugae are signified by black arrows. (F′, G′) A similar plane of section is shown for Sox2^CKO E15.5 embryos, which demonstrates clefting of the palate. An enlargement of the palate reveals that no rugae are formed at this stage. (H) Staining for Lef-1 and Sox2 in E14.5 control embryos. Sox2 is a marker for the oral epithelial tissue, and Lef-1 marks the developing rugae. (H′) Immunofluorescence staining for Sox2 and Lef-1 in the E14.5 Sox2^CKO embryos reveals the loss of Sox2 expression in the oral epithelial tissue as well as the loss of Lef-1 expression in the palate rugae. (I) P63/Sox2 double staining of the control embryos shows that P63 is expressing on palatal epithelium at E15.5. The most outer layer of the epithelium is expressing Sox2 but not P63. (J) Staining of the control for Sox2 and the periderm marker K6 reveals that the outer-most epithelial layer is periderm and that the periderm expresses Sox2. (K) Double staining for P63 and K6 demonstrates that the periderm lacks P63 expression. (I′) Double staining for P63 and Sox2 in Sox2^CKO embryos reveals that Sox2 has been successfully ablated from the palatal epithelium and that ablating Sox2 negatively effects the organization of the P63⁺ cells. (J′) Double staining for Sox2 and K6 demonstrates that periderm formation is greatly decreased in Sox2^CKO embryos. (K′) Double staining for P63 and K6 in Sox2^CKO embryos demonstrates defects in periderm formation in the Sox2^CKO embryos. (L, L′) Grhl3 protein is highly expressed in the periderm of control embryos and greatly reduced in Sox2^CKO embryos. PS, Palate Shelf. Scale bar = 400 μm.

Further investigation with sagittal sections taken from the midline of control (Fig. 4F, G) and Sox2^CKO embryos (Fig. 4F′, G′) revealed that the mutant embryos lacked well-defined palate rugae (denoted with arrows in control embryos), which are sources of signaling molecules during palatogenesis (Welsh and O’Brien 2009; Lin et al. 2011). Lef-1, which is specifically expressed in the palate rugae of E14.5 control embryos (Fig. 4H), distinguishes the rugae from the surrounding Sox2⁺ oral epithelium. In Sox2^CKO embryos, Sox2 and Lef-1 proteins were undetectable in the oral epithelium, and the rugae failed to form (Fig. 4H′).

Sox2 Is Required for Palate Periderm Formation

To test if Sox2 regulated periderm differentiation of the palate as well, we examined the oral mucosa by staining for P63, Sox2, and K6. In control embryos, we found that the oral mucosa was composed of a layer of P63⁺ basal cells and that a subset of these cells expressed Sox2. On the apical side of the basal cell layer, we identified a population of P63^-, Sox2⁺ cells (Fig. 4I). This outermost layer also expressed K6, demonstrating that the periderm of the palate expresses Sox2 (Fig. 4J, K). In Sox2^CKO embryos, we found that the P63⁺ basal layer was severely disrupted (Fig. 4I′) and that K6 expression was almost completely undetectable (Fig. 4J′, K′). This demonstrates that Sox2 is required for the differentiation of periderm in the palate and has a role in basal cell maintenance. We confirmed that periderm formation was decreased in Sox2^CKO palates by staining for Grhl3, which is strongly expressed in the periderm of control embryos (Fig. 4L, L′).

Sox2 Regulates K6 Expression

Because K6 was reduced in Sox2^CKO embryos (Figs. 2D and 4I′, J′), we reasoned that the role for Sox2 in periderm differentiation might include promoting the expression of periderm-specific genes, so we tested the ability of Sox2 to regulate K6 expression. In GMSM-K cells overexpressing Sox2, we observed a ~10-fold increase in K6 transcripts, indicating that Sox2 does have a role in regulating K6 expression (Fig. 5A, B). We further confirmed that K6 was upregulated at the protein level by Western blotting (Fig. 5C).

Figure 5.

SOX2 directly regulates K6 expression in oral epithelial cells through promoter binding. (A) The GMSM-K cell line was transfected with empty vector and Sox2 overexpression plasmid, and Sox2 mRNA levels were detected through quantitative reverse transcription polymerase chain reaction (PCR). (B) From the same reverse transcription sample, levels of KRT6 mRNA were detected with quantitative PCR. Cells overexpressing Sox2 had a significant increase in KRT6 mRNA. (C) Western blotting was used to determine the protein levels of empty vector and Sox2-transfected GMSMK cells. Cells overexpressing Sox2 had higher K6 protein levels when compared with control. Bands were quantitated, and relative values are shown. (D) To determine if SOX2 directly interacts with the KRT6A promoter, a ChIP assay was performed with nonspecific IgG and a SOX2 antibody. Input samples and precipitated chromatin were then used in a PCR assay with primer flanking the SOX2 binding site (left) or a negative control region (right). Primers flanking the SOX2 binding site amplified a PCR product in input, and SOX2 antibody precipitated samples but not when chromatin was precipitated with IgG. Primers flanking the negative control region amplified the input sample but failed to amplify when the specific or nonspecific antibody was used. (E) The KRT6A promoter (3,000 bp prior to transcription start site) was cloned into TK luciferase reporter. Either the control reporter or KRT6A promoter luciferase reporter was cotransfected with empty vector or Sox2 overexpression plasmid into LS-8 cells. Overexpressing Sox2 increased the normalized KRT6A promoter luciferase signal as compared with the control. Values are presented as mean ± SEM. *P < 0.05. **P < 0.01.

To determine if SOX2 directly regulates K6 expression, we examined the K6 promoter for the Sox2 binding sequence previously identified by our group (Sun et al. 2016) and noted a match 1.2 kb upstream of the transcription start site (Fig. 5D). To test if endogenous SOX2 protein interacts with the binding site, we performed a ChIP assay using a Sox2 antibody and tested for enrichment with primers flanking the SOX2 binding site, as well as a pair of negative control primers that did not flank a SOX2 site. The specific and negative control primers amplified the correct product from an input sample, and neither amplified a product when the pull-down was performed with a nonspecific IgG. Only the primer pair specific for the SOX2 binding site was found to amplify a product when the specific Sox2 antibody was used, demonstrating that SOX2 protein indeed binds the K6 promoter. To confirm the direct regulation of K6 by SOX2, we cloned the K6 promoter sequence upstream of a luciferase reporter vector, as well as a promoter reporter with a mutated SOX2 binding motif. The overexpression of SOX2 significantly increased the luciferase signal of the WT K6 promoter reporter as compared with the mutated K6 reporter, demonstrating that SOX2 activates the K6 promoter (Fig. 5E).

Discussion

Ankyloglossia, the fusion of the mandible to the tongue, occurred after E13.5 and persisted until birth in the Sox^CKO embryos. Our report describes how Sox2 regulates periderm formation and oral adhesions due to expanded surface dental epithelium and delayed incisor epithelial cell invagination and associated tooth agenesis. Furthermore, Sox2 regulates development of the palate rugae, and a loss of this palate signaling center may contribute to clefting.

Sox2 Is Required for the Formation of the Periderm Layer by Regulating K6 Expression

It is well known that Sox2 is a dental epithelial stem cell marker (Juuri et al. 2012; Li et al. 2015; Sun et al. 2016), and in this report, we show that it functions to promote the differentiation of the periderm structure. Sox2 is also expressed in the tongue epithelium (Arnold et al. 2011). Sox2 was expressed in progenitor cells superior to the basal cell layer in the mandible of WT embryos superior to the tooth bud region, as well as in the periderm. In Sox2^CKO embryos, which lacked a periderm layer, the mandibular basal cell layer was disorganized, and P63-positive cells are expanded (Fig. 2) due to Sox2-regulating K6 expression and periderm formation (Fig. 5).

Oral Adhesions Associated with Ankyloglossia Form between an Expanded Dental Epithelium and Tongue Epithelium

Sox2 deletion in the developing mouse caused invagination defects in the developing lower incisor and, combined with a lack of the periderm structure, resulted in oral adhesions. In several other mouse models, oral adhesions occur near tooth buds of the molars (Richardson et al. 2006; Richardson et al. 2014; Paul et al. 2017). Taken together, these data suggest that the dental epithelium plays a role in the formation of oral adhesions. We determined that the oral adhesion tissue was dental epithelium and that the oral adhesion is formed by proliferating surface dental epithelium, due to a delayed invagination of the dental epithelium, which results in dental anomalies and tooth agenesis. Because ankyloglossia can affect up to 10% of newborns (Chandrashekar et al. 2014; Veyssiere et al. 2015; Yoon et al. 2017), these oral adhesions can also affect tooth development and may cause tooth anomalies in children.

Sox2 Controls Formation of the Palate Rugae Signaling Center and Palatogenesis

Sox2 played a related but slightly different role in the palate. In the basal cell layer, many of the palatal P63⁺ cells coexpressed Sox2 in control embryos, and ablating Sox2 resulted in the disorganization of the basal cell layer. These data suggest that Sox2 functions to promote periderm differentiation in the mandible but has a dual function of regulating basal cell maintenance in the palate. The periderm covering the palate was also poorly formed, demonstrating the conserved role of Sox2 in periderm development. Interestingly, although the periderm covering the palate was defective, oral adhesions between the tongue and the palatal shelves did not occur due to the restricted ability of the tongue to contact the palate. Abnormal fusion between the palatal shelf and mandible has been reported in multiple mouse models (Xiong et al. 2009; Richardson et al. 2014; Kousa et al. 2017; Hammond et al. 2019).

While clefting in other mouse models with periderm defects usually resulted from oral adhesions between the palate and the tongue that prevent shelf elongation, this was not the case in the Sox2^CKO embryo. Instead, we found that the ablation of Sox2 resulted in a loss of palatal rugae (Fig. 4). Others have shown that the rugae serve as Shh signaling centers during palate development (Welsh and O’Brien 2009; Lin et al. 2011). In a ß-catenin conditional knockout mouse, the palate rugae did not form, and a lack of Lef-1, Pitx2, and Tcf-1 was shown (Lin et al. 2011). Shh expression, which is localized to the rugae, was also reduced, linking Shh and Wnt signaling in the formation of the rugae (Lin et al. 2011). The rugae provide morphogens such as Shh to ensure that the palate extends to the midline (Welsh and O’Brien 2009; Lin et al. 2011). The loss of Sox2 prevented these structures from forming and deprived the developing palatal shelves from the signals driving cell proliferation and shelf elongation. While we show a lack of rugae formation and associated cleft palate in the Sox2 mutant embryos, it is not clear how periderm formation affects palate formation. We speculate that the rugae may be required for periderm formation and proliferation of the palate epithelium. Conversely, the periderm layer may interact with the palate rugae during palatogenesis. Interestingly, we have shown that ablation of Sox2 also results in a lack of Lef-1 expression in the palate rugae, suggesting that Sox2 is required to form and maintain the palate rugae.

This study is the first to report that the dental epithelial tissue participates in the formation of inappropriate oral adhesions and provides an explanation for defects in the permanent dentition observed in several families with severe hereditary ankyloglossia (Chandrashekar et al. 2014; Lenormand et al. 2018). Clinically, these results suggest that the presence of ankyloglossia can be used to diagnose potential defects in dental development.

Author Contributions

Y.Y. Sweat, M. Sweat, W. Yu, S. Eliason, B.A. Amendt, contributed to conception, design, and data analysis, drafted and critically revised the manuscript; M. Sanz-Navarro, L. Zhang, Z. Sun, O.D. Klein, F. Michon, Z. Chen, contributed to conception, design, and data analysis, critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.