Stem Cell

MAST4 regulates stem cell maintenance with DLX3 for epithelial development and amelogenesis – Experimental & Molecular Medicine

Time Stamp: June 30, 2024 8:00 PM
Source Node: 626166

Incisor morphology, strength, and enamel composition in Mast4 KO mice

Mast4 KO mice with targeted deletion of 71 base pairs in exon 1, which resulted in a premature stop codon resulting from a frameshift mutation, were generated using the CRISPR/Cas9 system (Supplementary Fig. 1). Mast4 KO mice were grossly similar to their WT littermates after birth in terms of size and shape, particularly those of the cranial skeleton (Supplementary Fig. 2a). However, the maxillary and mandibular incisors of Mast4 KO mice showed asymmetrical attrition at postnatal week 3 (Supplementary Fig. 2b, arrow). At 6 weeks, the asymmetrical attrition in Mast4 KO mice became more severe, and bending began to occur (Fig. 1a, b). At 18 weeks, the Mast4 KO mice were all viable, but they possessed opaque mandibular incisors with chalky surfaces, while the mandibular incisors of WT mice were transparent and glossy (Fig. 1c, d).

**Fig. 1: Deletion of *Mast4* leads to an amelogenesis imperfecta phenotype.**

We also observed twisting to one side between the maxillary and mandibular incisors in 80 to 90% of the Mast4 KO mice; the incisors were overgrown, and the enamel of the maxillary incisors was peeled off (Fig. 1b, arrow). Microcomputed tomography (μCT) analysis of the head more clearly revealed the incisor malocclusion in Mast4 KO mice (Fig. 1e, f). To investigate the developmental patterns of molar dentition, histological analysis was performed at the bell stage (embryonic day 18.5; Supplementary Fig. 3a, b, d, e). Relative to that in the molar tooth germs of WT mice, the ameloblast layer arrangement in the molar tooth germs of Mast4 KO mice exhibited an irregular shape (Supplementary Fig. 3e, arrowheads). However, there was no difference in morphogenesis, including in cuspal patterning. In addition, there were no differences in molar dentition between WT and Mast4 KO mice until postnatal week 10 (Supplementary Fig. 3c, f).

Scanning electron microscopy (SEM) analysis was performed to compare the developing enamel between WT and Mast4 KO mice (Fig. 1g, h and Supplementary Fig. 4). This revealed that the enamel of developing Mast4 KO mouse incisors contained a collapsed enamel rod arrangement in the inner enamel (Fig. 1h, arrows), specifically in the area adjacent to the outer enamel (Supplementary Fig. 4, arrowheads). Furthermore, electron probe microanalysis (EPMA) was performed to investigate the mineral content of the incisors (Supplementary Fig. 5a, b). We divided the calcified portion of incisors from 6-week-old mice into two parts (the secretory and maturation regions) and proceeded with EPMA. Interestingly, in the maturation region of Mast4 KO incisors, calcium (Ca), magnesium (Mg), and phosphorous (P) were reduced, whereas there were no significant differences in the secretory region. We then performed a Vickers test to determine the hardness of the incisors of 6-week-old WT and Mast4 KO mice. The Vickers test revealed that the loss of MAST4 expression significantly reduced enamel hardness (Supplementary Fig. 6). This altered intensity cannot balance the incisor bite force and may lead to a twisted phenotype. Based on these results, the expression pattern of MAST4 was confirmed in various developmental stages of molars and incisors (Supplementary Fig. 7) as well as in three different regions of juvenile (3-week-old) and adult (6-week-old) incisors (Fig. 1i–n). At the E12.5, E14.5, and E16.5 stages, MAST4 was not detected in molars (Supplementary Fig. 7a–c). At postnatal 1 day, while MAST4 was expressed and dispersed in the stellate reticulum (SR), its expression was not detected in ameloblasts (Supplementary Fig. 7d). In the incisors, there was no MAST4 expression at any developmental stage (Supplementary Fig. 7e, f). At 3 weeks, MAST4 was generally weakly expressed, but strong MAST4 expression was observed in a single cell in the stem cell niche of the apical bud (Fig. 1i). However, MAST4 was expressed at proximal and distal terminal web complexes in the secretory and maturation stages (Fig. 1j, k, PTW, DTW). At 6 weeks, MAST4 expression was observed on the apical side (Fig. 1l, asterisks). In particular, MAST4 was detected at Tomes’ processes (Fig. 1m, TP) as well as

terminal web complexes (Fig. 1m, PTW, DTW) in the secretory stage and at the ruffled border of ameloblasts (Fig. 1n, RB) and terminal web complexes (Fig. 1n, PTW, DTW) in the maturation stage. These results suggest that the weakness of the incisors of Mast4 KO mice compared to those of WT mice may be caused by dysregulation during amelogenesis, not during development.

Amelogenesis dysregulation in the incisors of Mast4 KO mice

The apical buds of mouse incisors contain an epithelial stem cell niche that provides the inner dental epithelium (IDE) cells that differentiate into ameloblasts. To investigate stem cell differentiation, apical bud regions were dissected from the mandibles of 6-week-old WT and Mast4 KO mice (Supplementary Fig. 8a, b). The initiation of enamel matrix secretion was shifted to the apical side of the Mast4 KO incisors (Supplementary Fig. 8b). Interestingly, the apical bud of the Mast4 KO incisor appeared to be reduced (Supplementary Fig. 8b, dotted line). In particular, the TA zone seemed to have disappeared. We investigated the histology of the incisor teeth in decalcified sagittal sections from WT and Mast4 KO mice at 3 and 6 weeks of age (Fig. 2a–d). In WT mice, secretory stage ameloblasts were tall and columnar, and enamel matrix proteins were present in the forming enamel (Fig. 2a). During the maturation stage, WT mice exhibited characteristic shortened ameloblasts. In contrast, epithelial cells in the TA zone of the apical bud region appeared to have been transformed into secretory ameloblasts in the KO incisors (Fig. 2b). The initiation of enamel matrix secretion was shifted to the apical bud region in the Mast4 KO incisors, while the initiation of secretion started in secretory stage ameloblasts in the WT incisors (Fig. 2a, b, arrowheads). In particular, ectopic atypical enamel matrix deposition without underlying dentin formation was detected in 40 to 50% of Mast4 KO mice (Fig. 2b, circle). In the maturation stage, severe hypomineralization was detected in the Mast4 KO incisor enamel (Fig. 2b Maturation). The initiation position of dentin matrix secretion did not differ between the WT and KO incisors. At 6 weeks, the initiation of enamel deposition in the KO incisors was shifted to the apical as observed at 3 weeks (Fig. 2c, d, arrowheads). Additionally, compared with the normal Tomes’ processes in WT mice (Fig. 2c, TP), the Tomes’ processes in Mast4 KO mice were stretched through the apical bud region to the secretory region at 6 weeks (Fig. 2d, asterisks). The degradation and absorption of enamel matrix proteins, important functions of maturation stage ameloblasts, were performed properly, and the enamel space appeared empty in the WT incisors (Fig. 2c, arrow). However, the eosinophilic staining of the enamel space in Mast4 KO incisors indicated that these functions were not performed properly in KO incisor ameloblasts and that significant amounts of enamel proteins remained (Fig. 2d, arrow). Enamel matrix proteins are substituted during enamel maturation, and the proportion of these proteins in mature enamel is known to be less than 2% m/m⁴². This feature of excessive remaining enamel proteins was consistent with the low mineral composition (high protein composition) and peeled enamel observed in Mast4 KO mice (Fig. 1b, arrow and Supplementary Fig. 4). Ectopic enamel secretion was verified by evaluating ameloblastin expression (Fig. 2e, f), and ameloblastin was found to be expressed and released to the enamel layer on the apical side in Mast4 KO incisors (Fig. 2f, arrowhead), suggesting that enamel matrix secretion is accelerated temporarily which resulted in advanced spatially in Mast4 KO incisors. Proliferating Cell Nuclear Antigen (PCNA), a TA cell marker in incisors, was expressed in the TA region of the epithelium and adjacent mesenchyme of the WT incisors, whereas PCNA was scattered throughout the epithelium and sparsely detected in the mesenchyme of the KO incisors (Fig. 2g, h). SOX2, which is a marker of dental epithelial stem cells (DESCs), was expressed throughout the WT apical bud, including in TA cells (Fig. 2i). However, SOX2 was detected in a limited area known as the stem cell niche in the apical bud of the KO incisors (Fig. 2i, j). Schematic diagrams showing these features of WT and Mast4 KO incisors and a summary of the histological observations are presented (Fig. 2k).

**Fig. 2: Reduced apical bud region size and accelerated ameloblast differentiation in *Mast4* KO incisors.**

To determine and identify additional molecules involved in the disrupted amelogenesis in Mast4 KO mice, we screened MMP20 and FAM83H by immunofluorescence staining in 6-week incisors from WT and Mast4 KO mice (Fig. 3a–h). Interestingly, the expression of MMP20, which aids in enamel protein alignment in the enamel matrix and ameloblasts, did not differ between WT and Mast4 KO mice (Fig. 3a, b). However, FAM83H secretion into the enamel matrix was significantly increased in the Mast4 KO secretory region (Fig. 3c, d). In the maturation region, the quantity of both proteins in the enamel matrix was lower in Mast4 KO mice than in WT mice (Fig. 3e-h). In particular, the expression of FAM83H, which plays a key role in enamel maturation, shifted to an early stage in the secretory region (Fig. 3d, g, arrows and arrowheads). These results are consistent with the shifted initiation of enamel secretion and reduced size of the apical bud region in the Mast4 KO incisors.

**Fig. 3: Enamel maturation dysregulation in *Mast4* KO incisors.**

We further performed RNA sequencing analysis of two regions apical buds and secretory ameloblasts, to compare 6-week-old WT incisors and Mast4 KO incisors (Fig. 3i). The major challenge can be considered to be the definition of the Mast4 KO apical bud region (properties of secretory ameloblasts and apical buds) and the secretory ameloblast region (properties of maturative ameloblasts) due to the reduction in the apical bud size in Mast4 KO mice. Consequently, transcriptome analysis of secretory ameloblasts (Fig. 3i) represented an area of an exact match, as shown in Fig. 3a–d. In the apical bud region, the expression of enamel matrix protein-encoding genes, including Klk4, Enam, and Mmp20, was increased in Mast4 KO mice. In the secretory ameloblast region, the expression of enamel matrix maturation-related genes, including Fam83h, Dlx3, and Wdr72, was increased in Mast4 KO mice. These results suggest that enamel secretion and maturation end at an early stage of amelogenesis in the Mast4 KO incisor compared to the WT incisor.

Effects of Mast4 deficiency on the Wnt signaling pathway

Via RNA sequencing analysis of secretory ameloblasts and apical buds isolated from 6-week-old WT and Mast4 KO incisors, changes in the Wnt signaling pathway profile and stem cell maintenance were detected. Changes in gene expression identified through analysis of differentially expressed genes (DEGs) were visualized on heatmaps, and large decreases in the expression of canonical Wnt signaling genes and stem cell maintenance-related genes were observed in both the apical bud and secretory ameloblast regions in Mast4 KO incisors (Fig. 4a, c). Gene Ontology (GO) analysis was performed on the RNA sequencing data (Fig. 4b, d). Canonical Wnt signaling was significantly downregulated in both the apical buds and ameloblasts of Mast4 KO mice. Interestingly, ossification was downregulated in Mast4 KO ameloblasts. Some tooth mineralization markers were upregulated in apical buds (Fig. 4b). We confirmed these changes in the Wnt signaling pathway in Mast4-null mHat9d cells, a dental epithelial stem cell line derived from the apical bud epithelium of a mouse incisor, finding that β-catenin expression was decreased in both the nucleus and cytoplasm (Fig. 4e). Mast4 ablation markedly decreased the transcription of Wnt signaling molecules, including Wnt-3a, β-catenin, and Lrp5 (Fig. 4f). This result suggests that ameloblast differentiation is accelerated due to altered Wnt signaling and stem cell maintenance in the apical bud region.

**Fig. 4: Disruption of stem cell maintenance and the *Wnt* signaling pathway in *Mast4* KO mice.**

MAST4 promoted the nuclear translocation of DLX3 by mediating serine/threonine phosphorylation of DLX3 within the nuclear localization signal (NLS)

Considering that the phenotypes of enamel disruption and the significant reduction in the mineral content found in Mast4 KO mice were similar to the phenotypes observed in conditional Dlx3 KO mice³⁰, we focused on examining the relationship between MAST4 and the DLX3 transcription factor, which plays an essential role in amelogenesis. First, we checked the distribution pattern of DLX3 in WT and Mast4 KO mice. The expression of DLX3 showed distinct localization patterns depending on the differentiation stage in the incisors of 6-week-old WT mice. DLX3 was predominantly located in the nuclei of ameloblasts at the secretory stage in the incisors of WT mice (Fig. 5a, arrowheads). At the maturation stage, DLX3 expression remained high in the nuclei of ameloblasts (Fig. 5b, arrowheads). However, in the incisors of Mast4 KO mice, DLX3 localization to the nucleus was decreased at both the secretory and maturation stages (Fig. 5c, d, blank arrowheads). These results suggest that the nuclear translocation of DLX3 might be impaired in Mast4 KO mice.

**Fig. 5: MAST4 regulates the translocation of DLX3 by mediating its phosphorylation near the NLS region.**

Based on this observation, we examined whether MAST4 regulates the nuclear translocation of DLX3. Because of the high molecular weight (>285 kDa) of full-length MAST4 and its relatively low expression in mHat9d cells, we used a truncated Mast4 construct (MAST4 PDZ) containing the DUF, kinase, and PDZ domains that was shown to function normally in two previous studies^37,38. To explore the relationship between MAST4 and DLX3, immunocytochemistry was performed for MAST4 and DLX3 in control, MAST4 PDZ-overexpressing, and MAST4-depleted mHAT9d cell lines. In the control (empty LPCX vector-transfected) cells, several MAST4 and DLX3 proteins directly bound to each other in the cytoplasm (Fig. 5e, yellow arrows and Supplementary Fig. 9a, d). Some DLX3 was detected in the nucleus, while most MAST4 was located in the cytoplasm. In MAST4 PDZ-overexpressing mHAT9d cells, DLX3 was translocated into the nucleus, and the abundance of MAST4 was increased in the cytoplasm (Fig. 5f and Supplementary Fig. 9b, e). A direct interaction between MAST4 and DLX3 was observed in these cells, but the number of complexes was decreased compared to that in the control cells. In MAST4-depleted cells, the abundances of MAST4 and DLX3 in the cytoplasm were not decreased (Fig. 5g and Supplementary Fig. 9c, f). The cross-sectional z-stack images acquired from these three cell types (LPCX, PDZ over, and KO), along with the related quantitative analysis, demonstrated a positive correlation between the MAST4 level and the degree of DLX3 nuclear translocation (Supplementary Fig. 9a–c, g). Additionally, an immunoprecipitation assay was performed in HEK293T cells, and interaction between MAST4 (both MAST4 PDZ and MAST4-Full) and DLX3 was detected (Fig. 5h). We next confirmed that DLX3 nuclear translocation increased when MAST4 was stably overexpressed in mHat9d cells (Supplementary Fig. 10a) or transiently overexpressed in HEK293T cells (Fig. 5i and Supplementary Fig. 10b). Considering that MAST4 functions as a serine/threonine kinase and that DLX3 is targeted by MAST4 for phosphorylation to regulate its DNA binding activity³², we examined whether MAST4 induces DLX3 phosphorylation. Interestingly, immunoprecipitation assays revealed that MAST4 overexpression significantly increased both the serine and threonine phosphorylation of DLX3 in both mHat9d (Supplementary Fig. 10e) and HEK293T cells (Fig. 5j and Supplementary Fig. 10c, d, f).

Next, to determine the sites of MAST4-mediated phosphorylation, multiple Dlx3 deletion mutants were generated, and we found that MAST4 bound to the C-terminus of the ND of DLX3, which is adjacent to the NLS region (aa 124–150) (Fig. 5k, l). In a previous report, it was revealed that the NLS region of DLX3 is a critical region for its nuclear translocation⁴³. In addition, the mechanism by which phosphorylation adjacent to the NLS region regulates protein translocation has been elucidated by performing serine-to-alanine mutations of the target residues^44,45. Therefore, to investigate whether MAST4-mediated phosphorylation adjacent to the NLS of DLX3 is necessary for its nuclear translocation, we generated a nonphosphorylatable mutant with alanine substitutions of the serine and threonine residues (T134, S137, and S138) within the NLS region (DLX3^AAA, Supplementary Fig. 16a). Interestingly, while MAST4 overexpression significantly increased serine and threonine phosphorylation of DLX3^WT, the phosphorylation of the DLX3^AAA mutant was not affected by MAST4 overexpression in HEK293T cells (Fig. 5m and Supplementary Fig. 11). More DLX3^AAA than DLX3^WT was localized to the cytoplasm, but the nuclear translocation of DLX3^AAA was not promoted by MAST4 (Fig. 5n and Supplementary Fig. 12). Furthermore, we confirmed that the expression of carbonic anhydrase genes, such as Car6 and Car12, and ion transporter genes, such as Slc26a1 and Slc34a2, in Mast4 KO mHat9d cells was rescued through the overexpression of DLX3 (Supplementary Fig. 13). These results suggest not only that phosphorylation of the DLX3 NLS by MAST4 is important for the nuclear translocation of DLX3 but also that the abnormal enamel secretion and maturation phenotypes observed in Mast4 KO mice are mediated through the regulation of DLX3 by MAST4.

Considering the previously reported correlation between DLX3 and Wnt signaling, we investigated whether the alterations in Wnt signaling observed in Mast4 KO mice are associated with the abnormal translocation of DLX3. Interestingly, in mHat9d cells, transient overexpression of DLX3 did not differentially regulate Wnt signaling, and DLX3 did not exhibit any specific effect on the inhibition of Wnt signaling caused by Mast4 KO (Supplementary Fig. 14). Therefore, further research is needed to elucidate the relationship between DLX3 and Wnt signaling. Additionally, we confirmed that the changes in Wnt signaling and DLX3 activity observed in Mast4 KO mice were independent of each other.

Phosphorylation of the DLX3 NLS by MAST4 regulated the activation of DLX3 target genes involved in pH regulation

To understand the functional implications of phosphorylation of the NLS of the DLX3 transcription factor, a luciferase reporter assay using pGL3-3xDRE, which contains three copies of the DLX3-responsive elements, was performed to assess the transcriptional activity of both DLX3^WT and DLX3 mutants in HEK293T cells⁴⁶. As expected, the basal transcriptional activity of DLX3^WT was greater than that of the DLX3^AAA mutant (Fig. 6a and Supplementary Fig. 15). In particular, the transcriptional activity of DLX3^WT was further increased when MAST4 was overexpressed, whereas the effect of MAST4 overexpression on the transcriptional activity of the DLX3^AAA mutant was relatively not significant. In addition, the phosphomimetic DLX3^EEE mutant with glutamic acid substitutions further increased the basal transcriptional activity of DLX3 without being affected by MAST4 overexpression, indicating that the phosphorylation status of the NLS of DLX3 is critical for activation of its target genes (Supplementary Fig. 16b). Next, we investigated whether the occupancy of each target gene promoter by DLX3 is regulated by MAST4-mediated NLS phosphorylation. In a previous report, direct target genes of DLX3, such as carbonic anhydrase and ion transporter genes involved in pH regulation, were identified, and DLX3 binding sites in the promoter of each target gene were also identified³⁰. With reference to a previous report, chromatin immunoprecipitation (ChIP) assays were conducted to examine the binding of DLX3 to its target gene promoters in both HEK293T and mHAT9d cells. Interestingly, in the case of carbonic anhydrase genes (CA6 and CA12) and ion transporter genes (CFTR, SLC24A1, and SLC26A1), DLX3^WT exhibited increased occupancy compared with that of DLX3^AAA (Fig. 6b–f and Supplementary Fig. 17). In particular, the occupancy of DLX3^WT on the target gene promoters was further increased by MAST4 PDZ overexpression, while that of DLX3^AAA was not affected. Next, to confirm whether the mRNA levels of the target genes are regulated by altered translocation of DLX3 via MAST4, RT‒qPCR was performed in mHat9d cells. The activation of the carbonic anhydrase target genes Car6 and Car12 and the ion transporter target genes Slc26a1 and Slc34a2 was significantly downregulated by the expression of DLX3^AAA, while expression of the DLX3^EEE mutant further increased target gene activation (Fig. 6g and Supplementary Fig. 16c). Consistently, a MAST4 PDZ-mediated increase in target gene activation was observed when DLX3^WT was transiently co-overexpressed, while neither of the two DLX3 mutants was further differentially regulated, confirming the NLS phosphorylation site-specific role of Mast4. These results indicate that phosphorylation of the DLX3 NLS by MAST4 plays an important role in promoting the nuclear translocation of DLX3 and subsequent activation of the target genes.