Persistent Wnt/β-catenin signaling disables soft palatogenesis and palatal osteogenesis by inducing mesenchymal condensation

Abstract

Introduction

Mammalian palates are composed of the anterior hard palate and the posterior soft palate. However, the correlation of the genesis, pattern formation, and morphogenesis between the hard and soft palates remains elusive.

Methods

In this study, we explicated the complicated palatal defects in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice, in which canonical Wnt activity was persistent due to constitutively active β-catenin in the palatal mesenchyme. Osr2-cre ^KI ;Ctnnb1 ^ex3f palates displayed an ectopic mesenchymal condensation extending from the proximal–posterior area to the distal–anterior area, along with impaired osteogenesis and agenesis of soft palate.

Results

Immunohistochemistry showed the overlapping active canonical Wnt domain with the ectopic mesenchymal condensation, indicating that the condensation was induced by persistent canonical Wnt signaling. Wnt5a, a chemokine that induces posterior–anterior migration of palatal mesenchymal cells, was activated in the anterior and middle palatal mesenchyme of Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. Exogenous supplementation of Wnt5a into wild-type (WT) palates recapitulated the mesenchymal condensation. These findings indicate that the persistent canonical Wnt signaling in the palatal mesenchyme extended Wnt5a expression, which enforced posterior mesenchymal migration toward the anterior to form the convoluted condensation, thereby impairing the genesis of the soft palate in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. Moreover, the medially osteogenic markers Sox9, Runx2, and Osx; the laterally Shh, Foxf1, and Fgf10; and another Wnt inhibitor, Sfrp2, were significantly reduced or even diminished in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. In contrast, the condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme displayed the medial markers Dlx5 and p-Smad1/5/8, along with the fibrosis/dermal markers ɑ-SMA and Tbx15. The Wnt and TGF-β/BMP inhibitors Ectodin and Noggin were also ectopically activated in the palatal epithelium overlying the condensed mesenchyme Osr2-cre ^KI ;Ctnnb1 ^ex3f mice.

Discussion

These findings indicate a transition of palatal mesenchymal cells from an osteogenic fate into fibrosis commitment, along with disrupted mediolateral patterning of the palatal shelves due to persistent canonical Wnt activity. Our study provides molecular clues that fine-tuning the mesenchymal canonical Wnt activity and Wnt5a-directed cell migration correlates with the morphogenesis of hard palates and the genesis of soft palates.

Introduction

The mammalian palate is composed of the anterior two-thirds hard palate, which is contributed by the primary palate and most of the secondary palate, and the posterior one-third soft palate, which originates from the posterior part of the secondary palate (Bush and Jiang, 2016). The hard palate contains the bony maxillary and palatine processes that physically separate the nasal and oral cavities, while the soft palate includes aponeurosis and paired muscles to not only create the nasal–oral separation but also allow speaking and swallowing (Li et al., 2019). The development of the hard palate includes the outgrowth of secondary palatal shelves from the lateral walls of the maxillary arches, the downward growth beside the tongue, horizontal re-orientation, contact and fusion with the contralateral palatal shelves, and osteogenic differentiation of the maxillary and palatine processes (Bush and Jiang, 2016; Li et al., 2017a). Unlike the dramatic morphogenesis in the secondary palatal shelves of hard palates, the development of soft palates is initiated from the most posterior secondary palatal shelves when the shelves are elevated horizontally. With the posterior extension, the shelves of the soft palates protrude horizontally from the lateral pharyngeal wall and grow to fuse with each other (Janečková et al., 2019; Li et al., 2019). Although the clear difference in histology makes the hard and soft palates appear as independent organs, the anatomy and temporal order of their development strongly indicate correlations in the genesis, pattern formation, and morphogenesis between the hard and soft palates. However, there are few studies on such a correlation during palatogenesis.

As one of the fundamental growth factors, Wnt ligands and canonical Wnt signaling were demonstrated to play essential roles in the development of secondary palates (Chen et al., 2009; He et al., 2011; Liu et al., 2015; Li et al., 2017b; Reynolds et al., 2019; Zhang et al., 2022). Previous studies exploiting Wnt reporter mouse lines indicated that the activity of canonical Wnt signaling was restricted to the palatal epithelium and excluded from the palatal mesenchyme (He et al., 2011). Both inactivation and constitutive activation of canonical Wnt signaling in the palatal epithelium resulted in cleft palates (He et al., 2011), indicating that fine-tuning of canonical Wnt signaling was critical for palatogenesis. However, the inactivation of canonical Wnt signaling in palatal mesenchyme by deleting β-catenin, the pivotal factor forming the trans-activator of canonical Wnt signaling with Tcf/Lef, also led to cleft palates (Chen et al., 2009; Janečková et al., 2023). Consistently, genotype and haplotype analyses in humans also indicated that mutations in Wnt genes were causative of cleft palates (Chiquet et al., 2008). Although cleft palates observed in loss-of-function mouse models indicated an indispensable role of the mesenchymal canonical Wnt signaling in palatogenesis, this conclusion was challenged by the fact that, in addition to mediating canonical Wnt signaling, β-catenin also contributes to cell behavior as a component of cytoskeleton (Salinas, 2007; Perez-Moreno and Fuchs, 2006). Thus, deleting β-catenin in the palatal mesenchyme was implied to cause cleft palates by disabling the cell cytoskeleton and migration. Therefore, the role of mesenchymal canonical Wnt signaling in palatogenesis still requires further exploration. On the other hand, although cleft palates caused by the activation of canonical Wnt signaling in palatal mesenchyme have been reported (Chen et al., 2009), the mechanism of how the elevated canonical Wnt activity disrupts palatal development has never been clarified. A recent study reported that the constitutive activation of canonical Wnt signaling in the palatal mesenchyme altered the expression pattern of genes involved in the cytoskeleton along the lateral–medial orientation (Wang et al., 2022), which implied that elevated canonical Wnt signaling caused cleft palates by disrupting cell migration. Furthermore, active Wnt signaling was also detected in the mesenchyme of soft palates, indicating a role of Wnt activity in soft palatogenesis (Janečková et al., 2019). A recent study on Osr2-Cre;Ctnnb1 ^f/f mice indicated that canonical Wnt signaling regulated mesenchymal cell proliferation and subsequently myogenesis by mediating ciliogenesis during soft palatogenesis (Janečková et al., 2023). Thus, in our study, by crossing the Osr2-cre knock-in allele (Osr2-cre ^KI ) with the Ctnnb1 ^ex3f knock-in allele, we activated canonical Wnt signaling constitutively in mouse palatal mesenchyme to examine how the elevated canonical Wnt activity impacts the hard and soft palatogenesis.

Materials and methods

Mouse lines

The Osr2-cre knock-in mouse (Osr2-cre ^KI ), Shh-cre, Rosa26R-mT/mG, Ctnnb1 ^ex3f , and pMes-Noggin mouse lines, which were obtained and described in previous studies (Wu et al., 2015; Li et al., 2021; Liu et al., 2022), were bred in the Specific Pathogen-Free System of the Institute of Genome-Engineered Animal Models for Human Diseases at Dalian Medical University. To constitutively activate Wnt signaling in the palatal mesenchyme, Osr2-cre ^KI mice were crossed with Ctnnb1 ^ex3f mice to obtain Osr2-cre ^KI ;Ctnnb1 ^ex3f mouse embryos. The morning of the day of vaginal plug identification was regarded as embryonic day 0.5 (E 0.5). After inhaling carbon dioxide, timed-pregnant mice were euthanized through cervical dislocation. This research was approved by the Ethics Committee of Dalian Medical University (Protocol No. AEE18011), and all the animal experimental procedures strictly complied with the ethical guidelines of Dalian Medical University.

Bulk RNA-seq analysis

Palates from E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f mice were dissected for RNA sequencing. The RNA-seq libraries were constructed according to our previously published protocol (Chen et al., 2025), with three biological replicates prepared for each sample. Differential expression analysis to identify differentially expressed genes (DEGs) between E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates was performed using the DESeq2 package, followed by Gene Ontology (GO) enrichment analysis of these DEGs using the clusterProfiler R package.

Paraffin section and Masson’s trichrome staining

Embryonic mouse heads were collected from timed-pregnant mice in ice-cold phosphate-buffered solution (PBS) and fixed in 4% paraformaldehyde (PFA) overnight at 4 °C. After being dehydrated with gradient ethanol, staged heads were embedded in paraffin and consecutively sectioned at 10 μm for Masson’s trichrome staining, in situ hybridization, or immunohistochemistry, as described previously (Liu et al., 2022).

Cell proliferation and apoptosis assays

For BrdU labeling, the timed-pregnant mice were intraperitoneally injected with BrdU labeling solution (10 mg/mL) at a dosage of 1 mL/100 g body weight. After an hour of observation, embryonic mouse heads were harvested and fixed in Carnoy’s fixative for 2 h. Following dehydration through an ethanol series, the heads were embedded in paraffin. Coronal sections were made at 10 µm and stained using the Detection Kit II (Roche Applied Science, Roche, Switzerland), following the manufacturer’s instructions. For the Ki67 immunofluorescence analysis, the antibody against Ki67 (1:2,000; Abcam, ab15580) was applied with the secondary antibody in the MaxVision TM HRP Polymer anti-Mouse/Rabbit IHC Kit (No. KIT5020, Maixin Ltd., Fuzhou, China). The TUNEL labeling assay was performed to detect cell apoptosis using the In Situ Cell Death Detection Kit, POD (No. 11684817910, Roche).

Cryostat section and phalloidin staining

After being fixed in a mixture containing 4% PFA and 15% sucrose overnight at 4 °C, embryonic mouse heads of the selected stages were immersed in 30% sucrose and embedded in O.C.T. compound (Tissue-Tek, Sakura®Finetek, VWR, Torrance, CA, United States) for coronal sections at 10 µm. For phalloidin (ab176753, Abcam, Cambridge, MA, United States) staining, cryo-sections were permeabilized for 1 h in 0.3% Triton-X 100 at room temperature prior to incubation with the phalloidin conjugate reagent for 1 h. Sections were counterstained with DAPI and imaged under an Olympus DP72 microscope.

In situ hybridization

Embryos were dissected in diethyl pyrocarbonate (DEPC)-treated PBS and fixed with 4% PFA in 0.1% DEPC-treated PBS overnight at 4 °C. The fixed heads were dehydrated through graded ethanol, embedded in paraffin, and sectioned at 10 µm for in situ hybridization (ISH). For whole-mount ISH, fixed heads were also dehydrated through a methanol series. All sense and anti-sense RNA probes were transcribed from the plasmids containing the mouse-linearized cDNA templates using an RNA Labeling Kit (Roche, Indianapolis, IN, United States), as described previously (Liu et al., 2022). The alkaline phosphatase-conjugated anti-digoxigenin (DIG) antibody was used to incubate whole mounts or sections. The hybridization signal was detected by BM Purple (Roche, Indianapolis, IN, United States). Nuclear fast red was used for counter-staining in all sections.

Immunohistochemistry and immunofluorescence

Immunohistochemistry and immunofluorescence were carried out on the 10-µm-thick paraffin sections. The primary antibodies against Lef1 (ab137827, Abcam, Cambridge, MA, United States), integrin ɑv (bs-1310R, Bioss Antibodies Inc., Beijing, China), ColII (K0921, Santa Cruz), Foxf1 (ab308633, Abcam), Dlx5 (10592-1-AP, Proteintech Group Inc., Wuhan, China), Runx2 (sc-101145, Santa Cruz Biotechnology Co., Ltd., Shanghai, China), Sox9 (ab185966, Abcam), Osterix (ab209484, Abcam), phospho-Smad1/5/8 (13820S, Cell Signaling Technology, Danvers, MA, United States), ɑ-SMA (ab7817, Abcam), ColI (P28372-BIF, Abmart, Shanghai, China), p-ERK1/2 (4370T, Cell Signaling Technology), Etv4 (10684-1-AP, Cloud-Clone Corp., Wuhan, China), Tbx15 (YN1399, Immunoway, Thermo Fisher Scientific, Shanghai, China), and Scx (ab307722, Abcam) were applied for immunofluorescence. The TSA Fluorescence Triple Staining Kit (G1236-100T, ServiceBio Co. Ltd., Wuhan, China) was applied for fluorescence development using the secondary antibody of IF Tyramide (1:1,000; ServiceBio, G1236-50T), showing green fluorescence at a wavelength of 488 nm and red fluorescence at 647 nm. The sections were counterstained with DAPI.

Agarose bead implantation and organ culture

The recombinant mouse Wnt5a protein (URPP549Mu01, Cloud-Clone Corp., Wuhan, China) was dissolved in phosphate-buffered saline at 0.5 mg/mL. The agarose beads (1537302, Bio-Rad) were incubated with the Wnt5a working solution for 1 hour prior to implantation into the E13.5 mouse palatal shelves. After 12 h of organ culture in Trowell dishes, the palatal shelves were fixed and sectioned for immunofluorescence staining with an antibody against Tbx15. The implantation was performed three times with different litters.

Statistical assay

For cell density, cell proliferation, and apoptosis, the Ki67-positive, BrdU-positive, TUNEL-positive, and total nuclei in bilateral palatal shelves were counted using ImageJ (version 1.54 g). For measuring the color intensity in ISH and the percentages of active domains to the entire palatal shelves in immunofluorescence, the images were converted to 8-bit grayscale with ImageJ, in which the color/fluorescence from the adjacent unstained or fluorescence-free areas was set as the background. Three independent replicates were performed in the paired WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f littermates. A two-tailed Student’s t-test was conducted using GraphPad Prism 9 (GraphPad Software Inc., United States) and presented as the mean with standard deviation (SD), with statistical significance set at p < 0.05.

Results

Osr2-cre ^KI ;Ctnnb1 ^ex3f mice displayed ectopic condensed mesenchyme in the palatal shelves

Cryostat sections of Osr2-cre ^KI ;Rosa26R-mT/mG mice showed that the Cre activity was only detected in the mesenchyme of hard and soft palates but was absent from the palatal epithelium (Supplementary Figure S1). To explore how constitutively activated canonical Wnt signaling in the palatal mesenchyme caused cleft palates, we first explicated the histological characteristics of Osr2-cre ^KI ;Ctnnb1 ^ex3f palates. Masson’s trichrome staining showed that compared to the anterior, middle, and posterior areas of the E13.5 palatal shelves (Figures 1A–C), the Osr2-cre ^KI ;Ctnb1 ^ex3f palatal shelves were enlarged (Figures 1D–F), with visible mesenchymal condensation (Figures 1B′,E′,F′). At E14.5, when the WT palate shelves were reoriented horizontally and even contacted at the middle and posterior levels (Figures 1G–I), the Osr2-cre ^KI ;Ctnnb1 ^ex3f palate shelves were still vertically beside the tongue (Figures 1J–L), with the more evident mesenchymal condensation through the palatal shelves (Figures 1H′,K′,L′). In contrast to the fused E16.5 WT palates with noticeably ossifying centers in the middle and posterior areas (Figures 1M,N,Q), the enlarged Osr2-cre ^KI ;Ctnnb1 ^ex3f palate shelves became more severe (Figures 1P,Q,R), with the mesenchymal condensation almost throughout the distal anterior, middle, and posterior shelves (Figures 1N′,Q′,R′). These findings indicated an association of the constitutively active canonical Wnt signaling in palatal mesenchyme with ectopic mesenchymal condensation.

FIGURE 1.

Masson staining of the hard palates in Osr2-cre ^KI ; Ctnnb1 ^ex3f mice. Masson staining showed the histology of E13.5 WT hard palatal shelves (A–C) and E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f hard palatal shelves (D–F); (B′,E′) the corresponding boxed areas in (B,E); (F′) statistical assay showed the increased cell density in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves compared to that in WT counterparts (WT: 118.7 ± 6.81/2500 π μm² vs. Osr2-cre ^KI ; Ctnnb1 ^ex3f : 287.3 ± 14.05/2500 π μm², p < 0.001). (G–I) Histology of the E14.5 WT hard palate; (J–L) histology of E14.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f hard palatal shelves; (H′,K′) the corresponding boxed areas in (H,K); (L′) statistical assay showed the increased cell density in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves compared to that in WT counterparts (WT: 138.0 ± 6.25/2500 π μm² vs. Osr2-cre ^KI ; Ctnnb1 ^ex3f : 335.0 ± 5.08/2500 π μm², p < 0.001). (M–O) Histology of the E16.5 WT hard palate; (P,O,R) histology of E16.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f hard palatal shelves; (N′,Q′) the corresponding boxed areas in (N,Q); (R′) statistical assay showed the increased cell density in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves compared to that in WT counterparts (WT: 225.3 ± 5.03/2500 π μm² vs. Osr2-cre ^KI ; Ctnnb1 ^ex3f : 522.3 ± 20.06/2500 π μm², p < 0.001). (The dashed blue circles in (B′,E′,H′,K′,N′,Q′) show the fixed areas for cell counting; asterisks and arrowheads indicate the condensed mesenchyme in Osr2-cre ^KI ;Ctnnb1 ^ex3f hard palatal shelves; scale bars: 500 μm; ****, p < 0.0001).

The enhanced cell adhesion convoluted cell arrangement but reduced cell proliferation in the condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme

To reveal the correlation between the activated canonical Wnt signaling and mesenchymal condensation, immunohistochemical staining of Lef1, a nuclear mediator of canonical Wnt signaling, was performed. In E13.5 WT palates, Lef1 staining was only detected in a few mesenchymal and epithelial cells of the anterior (Figure 2A) and middle shelves (Figure 2B), but it was absent from the posterior palatal shelves (Figure 2C). In contrast, the Lef1-positive cells in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves were not only absent from palatal epithelium but also coincided with the location of the condensed mesenchyme in the anterior (Figure 2D), middle (Figure 2E), and posterior (Figure 2F) shelves, which strongly indicated the induction of mesenchymal condensation by the active canonical Wnt signaling. Ki67 staining indicated the remarkably reduced proliferation in both the condensed mesenchymal and epithelial cells of Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Figures 2D–G) compared to that in the WT counterparts (Figures 2A–C,G). Moreover, BrdU labeling assays indicated that the tip of the E13.5 anterior Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves, which was not reached by mesenchymal condensation, exhibited more BrdU-positive nuclei than the WT control (Supplementary Figures S2A,B). In contrast, there were fewer BrdU-positive nuclei in the condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f anterior–mid palatal shelves than in the WT counterparts (Supplementary Figures S2C,D). These phenomena indicated that the proliferating cells were pushed into the most anterior tip of the palates by the condensed cells, which increased the proliferation ratio in the anterior tip. On the other hand, the TUNEL assay showed no difference in cell apoptosis between WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Supplementary Figures S2E–J). Therefore, the active canonical Wnt signaling most likely enhanced the condensation but reduced cell proliferation in the palatal mesenchyme.

FIGURE 2.

Canonical Wnt signaling distribution, cell proliferation, adhesion, and arrangement in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. (A–F) Immunofluorescence staining of Lef1 and Ki67 in E13.5 WT palatal shelves (A–C) and E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice (D–F). (G) Ki67 positive percentages reduced from 88.64% ± 4.04% and 69.38% ± 4.51% in WT to 78.67 ± 3.59 (p < 0.0001) and 59.26 ± 10.66 (p < 0.05) in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium and mesenchyme, respectively. (H–M) Immunofluorescence staining of integrin ɑv and ColII in E13.5 WT palatal shelves (H–J) and E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice (K–M). (N) The percentage of integrin ɑv-positive areas increased from 8.30% ± 6.48% in WT to 40.64% ± 22.50% (p < 0.05) in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. (O–T) Immunofluorescence of phalloidin showed the cytoskeletal orientation in the E13.5 WT palate shelves (O–Q) and the E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palate shelves (R–T). (Asterisks in (R–T) represent the centers of the convoluted mesenchymal loops; red short lines in (O–T) delineated the longitudinal axes of mesenchymal cells; *, p < 0.05; ****, p < 0.0001; scale bars: 100 μm).

Furthermore, integrin ɑv staining in the mesenchymal condensation of E13.5 Osr2-cre ^KI ; Ctnnb1 ^ex3f palatal shelves was more extensive and robust (Figures 2K–N) than that in WT controls (Figures 2H–J), indicating enhanced cell adhesion in the condensed mesenchymal cells. Both the cytoskeletal F-actin marked by phalloidin and nuclei orientation by DAPI staining indicated a convoluted loop with a center of condensed mesenchymal cells in the anterior (Figure 2R), middle (Figure 2S), and posterior (Figure 2T) of the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves compared to the distal–proximal and lateral–medial arrangement in the WT counterparts (Figures 2O,P,Q). Taken together, this indicates that persistently active canonical Wnt signaling in the palatal mesenchyme most likely induced mesenchymal adhesion that condensed cells into a convoluted loop but reduced proliferation.

Ectopic Wnt5a activation and diminished Sfrp2 expression in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves

To explicate the alteration in the gene expression profile of palatal mesenchymal cells, the E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves were collected for bulk RNA-seq. The DEGs were read out using the DESeq2 package (Figure 3A; Supplementary Material 1). Furthermore, GO analysis enriched the DEGs in regulation in the Wnt signaling pathway/planar cellular polarity pathway, Wnt-protein binding, cell adhesion-mediator activity, cell fate specification, and intermediate filament cytoskeleton organization (Figure 3B; Supplementary Material 2). Bulk RNA-seq showed significant downregulation of Sfrp2 and upregulation of Sfrp5 expression, but no difference was observed in Wnt5a transcription between Osr2-cre ^KI ;Ctnnb1 ^ex3f and WT control palates (Figure 3C). However, in situ hybridization revealed that Wnt5a expression, which was detected in the anterior and middle lateral mesenchyme of E13.5 WT palatal shelves (Figures 3D,F,H), was significantly enhanced in the anterior and even extended throughout the middle mesenchyme of Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Figures 3E,G,I,J). The in situ hybridization in E14.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates confirmed the ectopic activation of Wnt5a (Figures 3K–M), along with the diminished Sfrp2 expression in Osr2-cre ^KI ;Ctnnb1 ^ex3f middle palatal shelves (Figures 3N–P).

FIGURE 3.

Suppressed Sfrp2 and enhanced Wnt5a expression in Osr2-cre ^KI ; Ctnnb1 ^ex3f palatal shelves. (A) Volcano plot visualization of DEGs between E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates. (B) GO enrichment analysis of DEGs between E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates, as presented in (A). The length of each horizontal bar corresponds to the enrichment fold. The dot paired with each bar represents −log10 (p.adjust) (negative logarithm of the adjusted p-value). The dashed vertical line denotes the significance threshold (−log10 (0.05)), and all terms presented in this plot satisfy the criterion of p. adjust <0.05. (C) Boxplots comparing DESeq2-normalized expression of Wnt5a, Sfrp2, and Sfrp5 between E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates, with statistical significance evaluated using the Wald test in the DESeq2 package. (D–I) In situ hybridization of Wnt5a in E13.5 WT anterior (D), middle (F), and posterior (H) palatal shelves and in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f (E) middle (G) and posterior (I) palatal shelves. (J) Statistical assay showed the normalized signal intensity of Wnt5a transcription (D–I). (K,L) In situ hybridization of Wnt5a in E14.5 WT (K) and Osr2-cre ^KI ;Ctnnb1 ^ex3f middle palates (L). (M) Statistical assay of Wnt5a transcription normalized from (K,L). (N,O) In situ hybridization of Sfrp2 in E14.5 WT (N) and Osr2-cre ^KI ;Ctnnb1 ^ex3f middle palates (O). (P) Statistical assay of Sfrp2 transcription normalized from (N,O). (Black arrowheads in (D–G) indicate the signals; scale bars: 100 μm; *, p < 0.05; ***, p < 0.001; ****, p < 0.0001).

The Wnt5a-induced cell migration and condensation led to the agenesis of the soft palate in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice

To examine whether the ectopic Wnt5a activation played a role in palatal mesenchymal condensation, Wnt5a-soaked agarose beads were grafted into the E13.5 WT middle–posterior palatal shelves for organ culture. The sagittal sections of the cultured palatal shelves implicated a remarkable cell condensation around the Wnt5a-soaked beads (Figures 4B,B′) instead of around the BSA-soaked beads (Figures 4A,A′). Further immunofluorescence staining displayed an anterior extension of Tbx15 expression, a dermal marker in the soft palate, to Wnt5a-soaked beads from the most posterior palatal shelves (Figures 4B,B′) compared to BSA beads (Figures 4A,A′), indicating that the supplemented Wnt5a chemo-attracted the mesenchymal cells to migrate anteriorly and form the condensation. Consistently, the cross-sections in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves exhibited the ectopic Tbx15 cells in the mesenchymal condensation (Figures 4C,C′,C″,D,D′,D″,E), indicating that ectopic Wnt5a was likely important for the migration of posterior mesenchymal cells toward the anterior.

FIGURE 4.

Wnt5a-induced cell migration and condensation impaired the genesis of Osr2-cre ^KI ;Ctnnb1 ^ex3f soft palates. (A,B,A′,B′) Sagittal views showing Tbx15 immunofluorescence staining in E13.5 WT palatal shelves supplemented with BSA (A) or exogenous Wnt5a (B). (A′,B′) Corresponding amplified boxed areas in (A,B). (C,D,C′,D′,C′′,D′′) Cross-sections displaying the immunofluorescence of Tbx15 and Scx in E13.5 WT soft palates (C,C′,C′′) and E13.5 Osr2-cre ^KI ; Ctnnb1 ^ex3f soft palatal shelves (D,D′,D′′). (E) Percentage of the Tbx15-positive areas to entire palatal shelves increased from 0.18% ± 0.2% in WT to 8.66% ± 4.99% (p < 0.05) in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (F–Q,F′–Q′,F′′–Q′′) Cross and transverse histology features of Osr2-cre ^KI ;Ctnnb1 ^ex3f soft palates. Cross-sections of E13.5 WT soft palates (F,F′,F′′) and E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f soft palatal shelves (G,G′,G′′). Cross-sections of E14.5 WT soft palates (H,H′,H′′) and E14.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f soft palatal shelves (I,I′,I′′). Cross-sections of E16.5 WT soft palates (J,J′,J′′) and E16.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f soft palatal shelves (K,K′,K′′). Transverse sections of E14.5 WT (L–N) and Osr2-cre ^KI ;Ctnnb1 ^ex3f hard and soft palatal shelves (L′–N′). Transverse sections of E16.5 WT (O–Q) and Osr2-cre ^KI ;Ctnnb1 ^ex3f hard and soft palatal shelves (O′,P′,Q′). (White arrowheads in (A,B) indicate the anterior margin of Tbx15 domains; white arrowheads in (D,D′,D′′) indicate the Tbx15 positive cells; TVP, tensor veli palatini; LVP, levator veli palatini; PLP, palatopharyngeus; PP, pterygoid process; SPC, superior pharyngeal constrictor; Epi, epiglottis; Apo, aponeurosis. The dashed lines in (H,I) delineate PP; black arrows indicate TVP, LVP, and PP; black arrowheads point to the condensed mesenchyme in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves; asterisks indicate soft palate; *, p < 0.05; scale bars in (A–D): 100 μm; scale bars in (F–Q): 500 μm).

We further assessed the impact of the ectopic Wnt5a activation and mesenchymal condensation on palatogenesis and found the agenesis of soft palates in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. The cross-sections showed that the soft palatal shelves were downward at the tensor veli palatini (TVP) level, horizontally protruded at the levator veli palatini (LVP) level, and absent at the palatopharyngeus (PLP) level in E13.5 WT mice (Figures 4F,F′,F″). In contrast, the soft palatal shelves were completely absent in the TVP, LVP, and PLP levels of E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice (Figures 4G,G′,G″). When the soft palatal shelves contacted and even fused at the TVP and LVP levels and protruded at the PLP level in E14.5 WT mice (Figures 4H,H′,H″), there were no soft palatal shelves protruding from the maxillary and pharyngeal walls in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice (Figures 4I,I′,I″). At E16.5, the TVP, LVP, and PLP were notable in the WT soft palates (Figures 4J,J′,J″), while Osr2-cre ^KI ;Ctnnb1 ^ex3f mice exhibited not only the absence of TVP, LVP, and PLP but also penetrated oral and nasal cavities (Figures 4K,K′,K″). These consequences indicated the agenesis of soft palates in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. Furthermore, transverse sections were performed to further reveal the mesenchymal condensation of Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. At the top level of E14.5 maxillary molars, the fusing hard palatal shelves of WT mice showed the condensed maxillary and palatine mesenchyme (Figure 4L), while the Osr2-cre ^KI ;Ctnnb1 ^ex3f mice displayed sporadic condensation in the posterior palatal shelves (Figure 4L′). At the medial level of E14.5 maxillary molars, the mesenchymal condensation was excluded in WT hard palates but was detected around the pterygoid processes (PP) in soft palates (Figure 4M). In contrast, both the posterior hard palatal shelves and the mesenchyme surrounding PP in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice were occupied by the condensation (Figure 4M′). At the bottom level of E14.5 maxillary molars, few mesenchymal condensation were found in WT soft palates (Figure 4N), while the enlarged distal palatal shelves in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice displayed condensations located in the anterior and middle shelves but absent in the posterior palatal shelves (Figure 4N′). Similarly, at the top level of E16.5 maxillary molars, both the WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates phenocopied their E14.5 counterparts (Figures 4O,O′). At the median and bottom levels of E16.5 maxillary molars, the PP, aponeurosis, and PLP were notable in WT soft palates without the condensed mesenchyme (Figures 4P,Q). In contrast, the condensation was extended throughout the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Figures 4P′,Q′). Thus, the mesenchymal condensation was extended from top (proximal)–posterior to bottom (distal)–anterior in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves.

The condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme was deprived of osteogenic capacity

The von Kossa staining showed that the ossified bones were extending to the secondary ossification centers in the E16.5 distal WT palate (Figures 5A,A′,A″) but were absent in the mesenchymal condensation of the Osr2-cre ^KI ;Ctnnb1 ^ex3f palate shelves (Figures 5B,B′,B″), indicating an impaired palatal osteogenesis correlated to the constitutively active canonical Wnt signaling in the palatal mesenchyme. To further address the molecular characteristics of the condensed mesenchymal cells in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves, the expression pattern of several osteogenic markers was first checked by immunofluorescence staining. The Sox9 staining was detected in the distal tip of the anterior and the medial side of the middle and posterior shelves in E13.5 WT palatal mesenchyme (Figures 5C,C′,C″) but was almost completely absent in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme (Figures 5D,D′,D″,E). Similarly, the osteogenic marker Runx2, which was located in the medial mesenchyme of the middle and posterior E13.5 WT palatal shelves (Figures 5C,C′), was only sporadically detected in the condensed mesenchyme of Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves without extending into the distal side of the medial mesenchyme (Figures 5D,D′,D″,E). Another osteogenic marker, Osx, which displayed a similar but reduced domain as Runx2 did in E13.5 WT palatal shelves (Figures 5F,F′,F″), was also diminished in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme (Figures 5G,G′,G″,H). Surprisingly, the mesenchymal condensation in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves showed robust p-Smad1/5/8 staining (Figures 5G,G′,G″), which was only detected in the distal mesenchyme of the middle and the medial mesenchyme of the posterior WT palatal shelves (Figures 5F,F′,F″,H). Thus, the ectopic condensed mesenchyme in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves was indicated to lose osteogenic capacity. Further immunofluorescence staining indicated that the ectopic condensed mesenchyme was devoid of Col1 staining but activated the fibrosis markers ɑ-SMA in the lateral condensation (Figures 5J,J′,J″). In contrast, E13.5 WT palatal shelves were devoid of ɑ-SMA staining, with widespread Col1 staining in the mesenchyme (Figures 5I,I′,I″,K). The FGF signaling indicators Etv4 and p-Erk1/2 showed little difference between E13.5 WT (Figures 5L,L′,L″) and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Figures 5M,M′,M″,N). Therefore, these results strongly indicated that the condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme was deprived of osteogenic capacity and was most likely transformed into fibrogenic fate.

FIGURE 5.

Impaired osteogenesis in the condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme. (A,A′,A′′,B,B′,B′′) von Kossa staining displayed the ossified maxillary and palatine bones in the E16.5 WT hard palate (A,A′,A′) and Osr2-cre ^KI ;Ctnnb1 ^ex3f hard palatal shelves (B,B′,B′). (C,C′,C′′,D,D′,D′′) Immunofluorescence of Runx2 and Sox9 in E13.5 WT anterior (C), middle (C′), and posterior (C′′) palatal shelves and in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f (D) middle (D′) and posterior (D′′) palatal shelves. (E) The percentages of Runx2- and Sox9-positive areas to the entire palatal shelves decreased from 14.88% ± 5.80% and 26.15% ± 2.08% in WT to 5.89% ± 3.75% (p < 0.05) and 4.35% ± 2.35% (p < 0.0001) in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice, respectively. (F,F′,F′′,G,G′,G′′) Immunofluorescence of Osx and p-Smad1/5/8 in the anterior (F), middle (F′), and posterior (F′′) palatal shelves of E13.5 WT mice and the anterior (G), middle (G′), and posterior (G′′) palatal shelves of E13.5 Osr2-cre ^KI ; Ctnnb1 ^ex3f mice. (H) The percentage of the Osx-positive area to the entire palatal shelves decreased from 3.37% ± 2.45% in WT to 5.89% ± 3.75% (p < 0.05) and 0.31% ± 0.42% (p < 0.05) in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice, while the percentage of the p-Smad1/5/8-positive area to the entire palatal shelves increased from 4.35% ± 2.35% in WT to 26.15% ± 2.08% (p < 0.0001) in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (I,I′,I′′,J,J′,J′′) Immunofluorescence of Col1 and ɑ-SMA in the anterior (I), middle (I′), and posterior (I′′) palatal shelves of E13.5 WT mice and the anterior (J), middle (J′), and posterior (J′′) palatal shelves of E13.5 Osr2-cre ^KI ; Ctnnb1 ^ex3f mice. (K) The percentage of the ColI-positive area to the entire palatal shelves decreased from 15.67% ± 5.90% in WT to 7.51% ± 2.62% (p < 0.01) in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice, while the percentage of the a-SMA-positive area to the entire palatal shelves increased from 5.64% ± 0.92% in WT to 45.43% ± 20.31% (p < 0.05) in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (L,L′,L′′,M,M′,M′′) Immunofluorescence of p-Erk1/2 and Etv4 in the anterior (L), middle (L′), and posterior (L′′) palatal shelves of E13.5 WT mice and the anterior (M), middle (M′), and posterior (M′′) palatal shelves of E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (N) The percentages of the p-Erk1/2- and Etv4-positive areas to the entire palatal shelves were slightly different between WT (3.13% ± 1.35% and 6.65% ± 4.25%) and Osr2-cre ^KI ;Ctnnb1 ^ex3f mice (7.48% ± 2.38%, p > 0.05 and 3.15% ± 1.10%, p > 0.05), respectively. (White arrowheads point to the immunofluorescence signal in palatal shelves; *, p < 0.05; **, p < 0.01; ****, p < 0.0001; scale bars in (A,B) 500 μm; scale bars in (C–M) 100 μm).

The ectopic activation of BMP-inhibitors suppressed palatal osteogenesis

To explore how persistent canonical Wnt signaling disrupted the osteogenic fate of the palatal mesenchyme, in situ hybridization was performed on E14.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palates. The Wnt/BMP inhibitor Ectodin and the TGF-β/BMP inhibitor Noggin, both of which were excluded from the E14.5 WT palatal mesenchyme (Figures 6A,D), were ectopically activated in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium (Figures 6B,C) and in both the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium and condensed mesenchyme (Figures 6E,F), respectively. Runx2 transcription indicated the secondary ossification centers in the E14.5 WT palates (Figure 6G), while it displayed no transcripts in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Figures 6H,I). Msx1 transcription was detected in neither the E14.5 WT (Figure 6J) nor the Osr2-cre ^KI ;Ctnnb1 ^ex3f palates (Figures 6K,L). These results indicated impaired osteogenesis in Osr2-cre ^KI ;Ctnnb1 ^ex3f palates correlated with the ectopic activation of Ectodin and Noggin. To verify the repression of epithelial-derived Noggin on the osteogenic specification of the palatal mesenchyme, we studied Runx2 and Sox9 expression in the palatal shelves of Shh-cre;pMes-Noggin mice. Immunofluorescence staining showed significantly reduced Runx2 distribution, especially in the posterior shelf, and the almost absent Sox9 domain in the Shh-cre;pMes-Noggin palatal shelves (Figures 7M–O) compared to that in the E12.5 WT controls (Figures 7M–O), which indicated the suppression of osteogenic specification of palatal mesenchyme by epithelial-derived Noggin.

FIGURE 6.

Ectopically activated BMP inhibitors in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium suppressed palatal osteogenesis. (A–L) In situ hybridization showed the Ectodin transcription in E14.5 WT (A) and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (B). (C) Corrected in situ hybridization signal intensity of Ectodin showed the notable increase from 79.15 ± 6.03 in WT to 133.30 ± 3.73 (p < 0.001) in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium; Noggin transcription in E14.5 WT (D) and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (E). (F) Corrected in situ hybridization signal intensity of Noggin showed the significant increase from 48.94 ± 5.38 and 17.88 ± 2.92 in WT to 121.70 ± 4.42 (p < 0.0001) and 40.11 ± 8.91 (p < 0.05) in the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium and mesenchyme, respectively; Runx2 transcription in E14.5 WT (G) and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (H). (I) Corrected in situ hybridization signal intensity of Runx2 markedly decreased from 41.59 ± 8.39 in WT to 17.63 ± 5.80 (p < 0.05) in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme; Msx1 transcription in E14.5 WT (J) and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (K). (L) Corrected in situ hybridization signal intensity of Msx1 showed slight difference between WT (18.27 ± 4.31) and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme (14.86 ± 3.68, p > 0.05). (M–R) Immunofluorescence of Runx2 and Sox9 in E12.5 WT (M–O) and Shh-cre;pMes-Noggin palatal shelves (P–R). Arrowheads indicate the positive signals of in situ hybridization; *, p < 0.05; ***, p < 0.001; ****, p < 0.0001; scale bars: 100 μm.

FIGURE 7.

Suppressed Shh expression and related gene expression in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. (A,B) Whole-mount in situ hybridization indicated the Shh transcription in E13.5 WT rugae of the hard palate and knots of the soft palate (black arrowheads in (A)), while Osr2-cre ^KI ; Ctnnb1 ^ex3f palatal shelves were devoid of Shh expression (B). (C–H) Immunofluorescence of Foxf1 and Dlx5 in E13.5 WT anterior (C), middle (E), and posterior (G) palatal shelves and in E13.5 Ctnnb1 ^ex3f (D), middle (F), and posterior (H) palatal shelves. White arrowheads in (F) and (H) indicate the ectopic Dlx5 staining. (I–N) In situ hybridization of Fgf10 in E13.5 WT anterior (I), middle (K), and posterior (M) palatal shelves and in E13.5 Ctnnb1 ^ex3f (J) middle (L) and posterior (N) palatal shelves. (O) DESeq2-normalized Shh, Foxf1, Dlx5, and Fgf10 expression from the bulk RNA-seq data showed no significant difference between WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. (P) Corrected signal intensity from the whole-mount in situ hybridization of Shh in E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. (Q) Percentages of the Foxf1- and Dlx5-positive areas to the entire palatal shelves were markedly reduced and increased, respectively, in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves compared to that in WT controls. (R) Corrected signal intensity from the in situ hybridization of Fgf10 in E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. (White arrowheads in (F,H,K) indicate the signals; *, p < 0.05; **, p < 0.01; scale bars: 100 μm).

Diminished Shh expression and disrupted mediolateral patterning in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves

To further understand the impacts of persistent canonical Wnt signaling on palatogenesis, Shh expression and SHH signaling were examined. Whole-mount in situ hybridization revealed the abrogation of Shh transcription in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium (Figure 7B), which is in contrast to the specific restriction in WT rugae (Figure 7A). Consistently, the immunofluorescence of Foxf1, a downstream target of mesenchymal SHH signaling, which was distributed in the lateral side of the E13.5 WT palatal mesenchyme (Figures 7C,E,G), was reduced to the proximal side in the middle and posterior Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme, though it had little impact in the anterior side (Figures 7D,F,H). In contrast, the immunofluorescence of Dlx5, which marked the medial side of the palatal mesenchyme (Figures 7C,E,G), extended to the lateral mesenchyme in the middle and posterior Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme (Figures 7D,F,H). In situ hybridization indicated that another downstream target of mesenchymal SHH signaling, Fgf10, which was activated in the lateral mesenchyme of middle palatal shelves (Figures 7I,K,M), had noticeably faded in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme (Figures 7J,L,N). Although the DESeq2-normalized expression of Shh, Foxf1, Dlx5, and Fgf10 showed no difference when comparing bulk RNA-seq data of the E13.5 WT and Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves (Figure 7O), the quantified signal intensity by ImageJ verified the reduced transcription of the Shh, Fgf10, and Foxf1 domains (Figures 7P–R), along with the increased Dlx5 domain in E13.5 Osr2-cre ^KI ; Ctnnb1 ^ex3f palatal shelves (Figure 7R), which indicated that the persistent canonical Wnt activity was most likely important for the suppressed Shh transcription and SHH signaling and the disrupted mediolateral patterning in the palatal shelf.

Discussion

The cleft palates in Osr2-cre ^KI ;Ctnnb1 ^f/f and Osr2-cre ^KI ;Ctnnb1 ^ex3f mice have been reported previously (Chen et al., 2009; Janečková et al., 2023); however, how persistent canonical Wnt activity in the palatal mesenchyme causes cleft palates remains elusive. In this study, we explicated a series of defects in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatogenesis, including the failed elevation of palatal shelves, ectopic condensed palatal mesenchyme, impaired palatal osteogenesis, and agenesis of the soft palate. Unlike the cleft soft palates in Wnt1-cre;Fam20b ^f/f or Osr2-cre ^KI ;pMes-Noggin mice (Deng et al., 2021; Li et al., 2021; Chen et al., 2023), the soft palatogenesis in Osr2-cre ^KI ;Ctnnb1 ^ex3f mice was never initiated, which provided an ideal tool to study the genesis of soft palates. We integrated the posteriorly extended Wnt5a-expressing domain, the ectopic condensed mesenchyme, and the agenesis of soft palates into a comprehensive interpretation. The constitutively activated canonical Wnt signaling in palatal mesenchymal cells was indicated to extend the anterior Wnt5a expression into the middle of the palatal mesenchyme. Since Wnt5a acted as a chemokine and induced palatal mesenchymal migration along the posterior-to-anterior orientation (He et al., 2008), the increased and extended Wnt5a transcription most likely enhanced the posterior–anterior migration in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. The increased cell density and anteriorly extended Tbx15 expression from WT soft palates with Wnt5a supplement supported the enhanced posterior-to-anterior migration. With more posterior mesenchymal cells, including the mesenchymal cells for the presumptive soft palates, migrating toward and condensing in the anterior palates, the genesis of soft palates was indicated to be disabled. Furthermore, a recent study utilizing Sox9-cre ^ERT ;Ctnnb1 ^ex3f mice also showed cleft palates with the condensed mesenchyme, along with increased ɑ-actin-1/4 and F-actin, indicating a model in which upregulated cell adhesion was caused due to canonical Wnt signaling (Wang et al., 2022), which was consistent with the highly increased integrin ɑv in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. These consequences indicated that the ectopic mesenchymal condensation in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves most likely originated from the anteriorly migrated palatal mesenchymal cells condensed by the enhanced cell adhesion.

As a non-canonical Wnt ligand, Wnt5a has been proven to promote cell migration and fibrosis by enhancing Sfrp5 transcription (Chatani et al., 2015; Zou et al., 2021; Trinh-Minh et al., 2024), which coincided with the increased ɑ-SMA-expressing domain and Sfrp5 expression in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves. Further analysis revealed that although canonical Wnt signaling would be antagonized by Wnt5a in tumor cells (Yuzugullu et al., 2009), Wnt5a activates or suppresses canonical Wnt activity depending on the receptor context (Mikels and Nusse, 2006). However, the effects of canonical Wnt activity on the non-canonical Wnt signaling or Wnt5a expression are still unknown. Our study indicated the activation of Wnt5a due to persistent canonical Wnt activity in the palatal mesenchyme, which requires further exploration of the molecular regulation. Notably, the enhanced Wnt5a transcription in the Osr2-cre ^KI ;Ctnnb1 ^ex3f middle palatal shelves was excluded from the DEGs of bulk RNA-seq, most likely due to the local altered gene expression being shared by the entire palatal shelves.

The impaired palatal osteogenesis in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves could also be attributed to persistent canonical Wnt activity. Although Wnt5a was reported to enhance canonical Wnt signaling during osteoblastogenesis (Okamoto et al., 2014), the secreted Wnt inhibitor, Sfrp2, could promote osteogenic differentiation by antagonizing canonical Wnt signaling (Jin et al., 2017; Wen et al., 2020; Yang et al., 2020; Wang et al., 2021; Akova Ölken et al., 2022). Thus, the diminished Sfrp2 expression in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme strongly indicated the association of persistent canonical Wnt activity with impaired palatal osteogenesis by suppressing Sfrp2 transcription. Meanwhile, the ectopic activated Ectodin and Noggin in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal epithelium was also implicated in depriving the underlying mesenchymal cells of the presumptive palatal bones of osteogenic capacity. Our previous study had reported that the overexpression of Noggin in the palatal mesenchyme strongly suppressed palatal osteogenesis (Li et al., 2021). Furthermore, in this study, overexpressed Noggin in the palatal epithelium was also associated with disrupted osteogenic commitment of the palatal mesenchyme, which supported the impaired palatal osteogenesis caused by the epithelium-derived Noggin or Ectodin. Such ectopic activation of Noggin by the canonical Wnt activity in adjacent tissues has been reported in our previous study on tooth development (Chen et al., 2019) and the recent study on muscle stromal progenitors (Kajabadi et al., 2023), though the reciprocal interactions between the canonical Wnt activity-processing tissue and the Noggin-expressing tissue still require exploration. Moreover, in Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal condensation, a series of osteogenic markers, namely, Runx2, Osx, and ColI, were significantly reduced, but the fibrosis/dermal markers Tbx15 and ɑ-SMA were activated, indicating a transition of osteogenic fate of palatal mesenchyme into fibrosis/dermal specification.

Eventually, the disrupted mediolateral patterning of palatal shelves was also associated with the persistent canonical Wnt activity. Shh and Fgf10, activated in the lateral epithelium and mesenchyme, respectively, of the middle palatal shelves, and Dlx5, p-Smad1/5/8, and Sox9 in the medial mesenchyme were regarded as hallmarks of the lateral–medial patterning of palatal shelves (Lan et al., 2004; Han et al., 2009; He et al., 2010). In Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves, the diminished Shh, Foxf1, and Fgf10 expression indicated a loss of lateral identity, while the medial-to-lateral extension of Dlx5, Runx2, and p-Smad1/5/8 domain in the condensed Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal mesenchyme indicated an expansion of medial identity. Although the Sox9 domain was also diminished throughout the Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves, it could be interpreted as the loss of osteogenic fate of the medial mesenchyme instead of the loss of medial identity because Sox9 marked the chondro-osteogenic progenitors of craniofacial mesenchyme (Akiyama et al., 2002; Mori-Akiyama et al., 2003; Dash and Trainor, 2020). Thus, the lateral–medial patterning of Osr2-cre ^KI ;Ctnnb1 ^ex3f palatal shelves was indicated to be interrupted by the persistent canonical Wnt activity. Combined with the lateral–medial patterning of many other ECM in palatal shelves (Chiquet et al., 2016; Wang et al., 2020), it indicated that the lateral–medial patterning of palatal shelves contributed to palatal elevation through the ECM pattern. However, whether and how the lateral–medial patterning of palatal shelves contributes to palatal elevation is still unknown. Our previous study reported cleft palates in Osr2-cre ^KI ;Rosa26R-Fgf8 mice associated with the disrupted lateral–medial axis of the palatal shelves despite increased cell proliferation in the palatal mesenchyme (Wu et al., 2015). Although the previous study indeed demonstrated a posterior-to-anterior migration of palatal mesenchymal cells along the Wnt5a gradient (He et al., 2008), to date, there is no convincing evidence supporting the contribution of cell migration to palatal elevation.

In summary, our present study reported a disabled genesis of the soft palate with failed osteogenesis and mediolateral patterning in the palatal shelves when canonical Wnt signaling was constitutively activated in the palatal mesenchyme. An ectopic Wnt5a extension by the persistent canonical Wnt signaling was associated with the extra-anterior migration of posterior mesenchymal cells into the condensed mesenchyme, which most likely impaired the genesis of soft palates. Moreover, the persistent canonical Wnt activity appears to deprive the palatal mesenchyme of osteogenic capacity that was correlated to the suppressed Sfrp2 expression, ectopically activated Noggin and Ectodin in the palatal epithelium, and even the transformation of the osteogenic fate into fibrosis/dermal specification. Meanwhile, the mediolateral pattering of palatal shelves was also indicated to be disrupted by the persistent canonical Wnt activity in the palatal mesenchyme.

Funding Statement

The author(s) declared that financial support was received for this work and/or its publication. This research was funded by the National Natural Science Foundation of China (82270949).

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding authors.

Ethics statement

The animal study was approved by the Ethics Committee of Dalian Medical University. The study was conducted in accordance with the local legislation and institutional requirements.

Author contributions

BW: Investigation, Writing – original draft, Resources, Validation, Formal analysis, Methodology. JuX: Investigation, Writing – original draft, Formal analysis, Visualization, Data curation, Validation. YB: Writing – original draft, Formal analysis, Validation, Investigation. JD: Writing – original draft, Resources. BL: Resources, Writing – original draft. NL: Visualization, Writing – original draft, Investigation. LZ: Writing – review and editing, Visualization. JiX: Conceptualization, Writing – review and editing, Funding acquisition. CL: Project administration, Conceptualization, Funding acquisition, Writing – review and editing. HL: Supervision, Project administration, Conceptualization, Writing – review and editing, Data curation.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcell.2026.1740081/full#supplementary-material

SUPPLEMENTARY FIGURE S1

Cryostat section of the palates in Osr2-cre ^KI ; Rosa26R-mT/mG mice. (A) Middle palatal shelves of E13.5 Osr2-cre ^KI ; Rosa26R-mT/mG mice. (B) Middle hard palate of E16.5 Osr2-cre ^KI ; Rosa26R-mT/mG mice. (C) Posterior hard palate of E16.5 Osr2-cre ^KI ; Rosa26R-mT/mG mice. (D) Soft palate of E16.5 Osr2-cre ^KI ;Rosa26R-mT/mG mice.

SUPPLEMENTARY FIGURE S2

BrdU labeling and TUNEL assays in the palates of Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (A) Most anterior palate in front of mesenchymal condensation in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (B) Anterior palate with condensed mesenchyme in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. (C-H) TUNEL assay of the palates in E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f mice. Scale bars: 100 μm.

SUPPLEMENTARY DATA SHEET 1

DEGs found by comparing the gene expression profile of E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palates with that of WT palates.

SUPPLEMENTARY DATA SHEET 2

Enriched functions and signaling pathways via GO analysis of the DEGs by comparing the gene expression profile of E13.5 Osr2-cre ^KI ;Ctnnb1 ^ex3f palates with that of WT palates.