SEARCH

SEARCH BY CITATION

Keywords:

  • Tissue-specific stem cells;
  • Conditional knockout;
  • Stem cell microenvironment interactions;
  • Differentiation;
  • Cre-loxP system;
  • Developmental biology

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Rodent incisors grow continuously throughout life, and epithelial progenitor cells are supplied from stem cells in the cervical loop. We report that epithelial Runx genes are involved in the maintenance of epithelial stem cells and their subsequent continuous differentiation and therefore growth of the incisors. Core binding factor β (Cbfb) acts as a binding partner for all Runx proteins, and targeted inactivation of this molecule abrogates the activity of all Runx complexes. Mice deficient in epithelial Cbfb produce short incisors and display marked underdevelopment of the cervical loop and suppressed epithelial Fgf9 expression and mesenchymal Fgf3 and Fgf10 expression in the cervical loop. In culture, FGF9 protein rescues these phenotypes. These findings indicate that epithelial Runx functions to maintain epithelial stem cells and that Fgf9 may be a target gene of Runx signaling. Cbfb mutants also lack enamel formation and display downregulated Shh mRNA expression in cells differentiating into ameloblasts. Furthermore, Fgf9 deficiency results in a proximal shift of the Shh expressing cell population and ectopic FGF9 protein suppresses Shh expression. These findings indicate that Shh as well as Fgf9 expression is maintained by Runx/Cbfb but that Fgf9 antagonizes Shh expression. The present results provide the first genetic evidence that Runx/Cbfb genes function in the maintenance of stem cells in developing incisors by activating Fgf signaling loops between the epithelium and mesenchyme. In addition, Runx genes also orchestrate continuous proliferation and differentiation by maintaining the expression of Fgf9 and Shh mRNA. STEM CELLS 2011;29:1792–1803


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Tooth development is characterized by sequential and reciprocal interactions between the dental epithelium and mesenchyme, as is the case for all epithelial appendages [1–3]. Rodent incisors represent a special tooth type as they grow continuously throughout the lifetime of the animal, and their regenerative capacity implies the presence of a stem cell pool. The cervical loop, which is located in the apical part of the incisor epithelium, has been shown to be the epithelial stem cell compartment in the mouse incisor [4]. The epithelial cells in the cervical loop proliferate and move toward the distal tip, and histological sections also show a characteristic gradient of ameloblast differentiation, enamel matrix formation, in the apical to incisal direction [5]. Many signaling molecules and growth factors have been implicated in the mediation of these interactions [6], and several molecules are specifically involved in the maintenance of the continuous regeneration of the incisors [7]. Among such molecules, Fgf10 is expressed in the mesenchyme that surrounds the cervical loop epithelium, as well as in the mesenchyme underlying the inner enamel epithelium, and Fgf10 deficiency leads to a reduction of the stem cell niche, which results in hypoplasia of the incisors. Phenotype analysis of Fgf10 null mice also indicated that Fgf10 is not directly involved in the early morphogenesis of the dental organ but is involved in the creation of the stem cell compartment in the cervical loop region [4, 8, 9]. Taken together, rodent incisors are a unique experimental model for examining the molecular mechanism involved in stem cell regulation and the subsequent stem cell differentiation.

Core binding factors (CBFs) are dimeric transcription complexes composed of an α unit (Cbfa) that binds to DNA and a stabilizing β subunit (Cbfb) [10, 11]. There is only one Cbfb subunit while the α subunit is encoded by three mammalian genes: Runx1, Runx2, and Runx3 [12] which all require Cbfb for their function. Targeted deletion of Runx genes has revealed distinct roles for these proteins in development, with Runx1 being required for hematopoiesis [13], Runx2 being required for osteogenesis [14, 15], and Runx3 being required for neurogenesis, thymopoiesis, and the control of gastric epithelial cell proliferation [16–18]. Some Runx family members show overlapping expression patterns, suggesting that they possess tissue-specific redundant functions, and Runx family genes also display cross-regulation because they contain Runx binding sites in their promoter regions [19]. Therefore, studies on the roles of Runx proteins are complicated.

During tooth development, all three Runx genes are expressed in the dental epithelium and/or mesenchyme and display distinct temporal-spatial patterns [20]. Runx2 is intensely expressed in the dental mesenchyme, and the development of Runx2 null mutant teeth arrests at the bud/cap stage [21]. In the dental mesenchyme, Runx3 expression overlapped with Runx2 expression, and Runx3 expression was upregulated in Runx2 null mutant upper teeth, indicating that Runx2 represses Runx3 expression in the upper molars and that Runx3 partly compensates for the function of Runx2 in teeth [22]. In the dental epithelium, Runx1 is also expressed in the epithelium in developing molars and incisors, as well as in hair follicles, another skin appendage organ. In a previous study, targeted deletion of Runx1 in the epithelial tissues under the K14 promoter affected hair shaft structure. Interestingly, epithelial deletion of Runx1 also affects the normal adult hair follicle cycle and maintains adult hair follicle stem cells (HFSCs) in a state of quiescence [23].

Here, we report that epithelial Runx genes are involved in the maintenance of the stem cell compartment and their subsequent continuous differentiation in rodent incisors. We analyzed the functional role of Cbfb in incisor development in vivo, using a conditional genetic approach in which the Cbfb gene was inactivated in mouse epithelial cells. As the β subunit of CBFs (Cbfb) acts as a binding partner for all Runx proteins, targeted inactivation of its expression abrogates the activity of all Runx complexes [24, 25]. Cbfb deficiency results in marked shortening of the incisors, and we have shown that these mutant incisors can be used as a model of how Runx genes function in the maintenance of the stem cell compartment. Cbfb, a cofactor of all Runx genes, serves to maintain the Fgf signaling loop between the epithelium and mesenchyme in the cervical loop region. In addition, we show that Runx genes also serve to regulate the continuous differentiation that occurs in growing incisors by maintaining both Fgf9 and Shh expressing cellular compartments. We also demonstrate that Fgf9 has an antagonizing effect on Shh expression.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

Mice

To generate K14-Cre/Cbfbfl/fl mice, we first mated heterozygous K14-Cre mice [26] and homozygous Cbfbfl/fl mice [27] to obtain F1 K14-Cre/Cbfbfl/+ mice. These progeny were subsequently bred with Cbfbfl/fl mice. Genomic DNA was isolated from each tail sample using a DNeasy kit according to the manufacturer's protocol (Qiagen, Valencia, CA). Genotyping was performed by the conventional polymerase chain reaction (PCR) method using each primer set to detect Cre (5′ CTCTGGTGTAGCTGATGATC 3′ and 5′ TAATCGCCATCTTCCAGCAG 3′ and the loxP site of Cbfb (5′ CCTCCTCATTCTAACAGGAATC 3′ and 5′ GGTTAGGAGTCATTGTGATCAC 3′). We used their littermates or knockout embryos that did not carry the K14-Cre/Cbfbfl/fl genotype as controls. Fgf9 null mutant mice were maintained on a C57/B6 background and genotyped as described previously [28].

X-Gal Staining

We mated K14-Cre and R26R conditional reporter allele R26R mice [29] to generate K14-Cre;R26R mice. Newborn mice were fixed in 0.2% glutaraldehyde in PBS containing 2 mM MgCl2, equilibrated in graded sucrose containing 2 mM MgCl2, and embedded in Tissue-Tek (OCT compound, Sakura, Tokyo, Japan). The samples were then sectioned into 10 μm sections and incubated in 5-bromo-4-chloro-3-indoxyl-β-D-galactopyranoside (X-Gal) solution [30] at 37°C for 12–16 hours.

5-Bromo-3-deoxy-uridine Labeling and Histology

5-Bromo-3-deoxy-uridine (BrdU) (Sigma-Aldrich, Poole, UK) was injected intraperitoneally at 250 μg/g body weight in saline buffer (PBS) on postnatal day (PD) 10, 12 hours before their sacrifice. After the mice had been sacrificed, we separated the maxilla and mandible using fine scissors, fixed them overnight at 4°C in 4% paraformaldehyde in PBS, and decalcified them with 0.5 M EDTA (pH 7.4) for 1 week. The samples were reacted overnight in 4°C with rat anti-BrdU (Oxford Biotechnology, Oxford, UK). Fluorescein-conjugated goat affinity-purified antibodies to rat IgG (Cappel, Cochranville, PA) or goat anti-rat IgG–horse radish peroxidase (HRP) (Invitrogen, Carlsbad, CA) were used as the secondary antibodies. Visualization of secondary antibody conjugated to HRP was performed using the DAB kit (Vector Laboratories, Burlingame, CA). Finally, the sections were counterstained with Hoechst 33258 (Sigma-Aldrich) or hematoxylin and mounted with fluorescent mounting medium (DakoCytomation, Carpinteria, CA).

In Situ Hybridization Analysis

In situ hybridization was performed using 10 μm frozen sections or paraffin sections. The digoxigenin-labeled RNA probes used in this study were prepared using a DIG RNA labeling kit according to the manufacturer's protocol (Roche, Mannheim, Germany) using each cDNA clone as the template. After hybridization, the expression patterns for each mRNA were detected and visualized according to their immunoreactivity with anti-digoxigenin alkaline phosphatase-conjugated Fab fragments (Roche), as previously reported [31, 32].

RNA Extraction and Reverse-Transcriptase PCR (RT-PCR) Analysis

Total RNA was extracted from the dental epithelium at E18.5 using Isogen (Nippon gene, Toyama, Japan), according to the manufacturer's protocol. Total RNA (500 ng) was reverse transcribed to cDNA using an oligo (dT) with avian myeloblastosis virus reverse transcriptase (Takara, Osaka, Japan). For realtime RT-PCR analysis, the cDNA was amplified with Blend-Taq Plus (Toyobo, Osaka, Japan) with a regular thermal cycler. The sets of synthetic primers used for the amplification are described below: mouse glyceraldehyde-3-phosphate dehydrogenase gene (gapdh), sense 5′-GTCCCGTAGACAAAATGGTG-3′ and antisense 5′-CAATGAAGGGGTCGTTGATG-3′; mouse ameloblastin, sense 5′-GGACCAATGGCACACAACAAAG-3′ and antisense 5′-AGTCTCCTAAGGGTTTGCCTG-3′; mouse amelogenin, antisense 5′-GCCTCCACTGTTCTCCATGC-3′, sense 5′-ATGTTAAGCGGATGCCTTGTC-3′. Each PCR was carried out and analyzed as described previously [33, 34].

Micro–Computed Tomography (Micro-CT) Analysis

For micro-CT analysis of the 4-week-old control and conditional knockout mice, a LaTheta LTC-200 was used to obtain X-ray images. Scans of 24-μm-thick layers were performed, and 3D reconstruction was performed using VGStudio MAX 1.2® software (Volume Graphics GmbH, Heidelberg, Germany).

Tissue Culture and Bead Treatment

The incisor tooth germ was dissected from the E16.0 embryo mandible and cultured on a Nuclepore track-etched membrane (Whatman, Middlesex, UK). Heparin-acrylic beads (Sigma-Aldrich) were incubated in FGF9 (50 ng/μl) and SHH (1 μg/μl; R&D Systems, Minneapolis, MN). Bovine serum albumin (BSA) (Sigma-Aldrich) was used instead of recombinant protein for the control beads. The beads were immersed in recombinant proteins or BSA at 37°C for 60 min and placed beneath the labial cervical loop of the explants using a pipette tube. After being cultured, the in vitro explants were fixed at each stage in 4% paraformaldehyde overnight and processed for whole-mount RNA in situ hybridization using appropriate RNA probes, as described previously [35].

Immunohistochemistry

The tissue samples were sectioned into 10 μm slice. After being dewaxed, the samples were treated with anti-rabbit polyclonal antibody to histone H3 (phosphor S10; Abcam, Cambridge, UK) overnight at 4°C and incubated with anti-rabbit A488 antibody (Invitrogen) as a secondary antibody for 1 hour at room temperature. The sections were then counterstained with Hoechst 33258 (Sigma-Aldrich) and mounted with fluorescent mounting medium (DakoCytomation).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

The Expression of Cbfb mRNA During Tooth Development

In this study, the expression of Cbfb mRNA was analyzed in detail in the developing teeth of E14.5, E16, and E18 wild-type mice using in situ hybridization (Fig. 1A–1E). In E14.5 mice, Cbfb mRNA was broadly expressed in the developing molar epithelium and the mesenchyme, including the dental papillae and the dental follicle (Fig. 1A). At E16, Cbfb transcripts were observed in a restricted region including the dental papillae, which underlie the secondary enamel knot (Fig. 1B, black arrowhead), but with less expression at the epithelium. In an E18 developing incisor, Cbfb mRNA was detected in the labial cervical loop epithelium (Fig. 1C, red arrowhead) and the incisor mesenchyme (Fig. 1C, black arrowhead). In a magnified view, the expression of Cbfb mRNA was evident, especially in the single cell layer of differentiation (Fig. 1D, red arrowhead) as well as the secretion stage of ameloblasts (Fig. 1E, red arrowhead), although the expression of lingual cervical loop was weak (Fig. 1F, red arrowhead). From previous studies, it is well known that Cbfb plays a critical role in bone development [36, 37]. Our study also confirmed the presence of Cbfb mRNA expression in osteoblasts (Fig. 1A–1D) and chondrocytes in the developing spinal column (Fig. 1G). According to these results, Cbfb mRNA is widely expressed in both the dental epithelium and the mesenchyme, and its temporospatial distribution suggests that Cbfb expression is associated with epithelial–mesenchymal interactions during molar and incisor development. Especially in developing incisors, the Cbfb expression pattern strongly suggests the critical role of Cbfb during ameloblast differentiation and enamel protein secretion.

thumbnail image

Figure 1. Cbfb mRNA was expressed in a temporospatial manner during tooth development. (A): In E14.5 molars, Cbfb mRNA was broadly expressed throughout the de, dp, df, and ab. (B): At E16, the expression pattern became restricted to the region in which odontoblasts were differentiating (black arrowhead). (C): In E18 incisors, Cbfb transcripts were detected in the epithelium of the laCL (red arrowhead) and the mesenchyme (black arrowhead) rather than the liCL (F). (D): In a magnified view of the laCL, Cbfb mRNA was evident in a single layer of the de (red arrowhead) and in the dm (black arrowhead) underlying the laCL. (E):Cbfb expression was also significant at the enamel protein secretion stage of ameloblasts. (G): Strong Cbfb expression was also detected in the E18 spinal column (sc). Scale bars = 200 μm in A, B, D, and E; 500 μm in C. Abbreviations: ab, alveolar bone; de, dental epithelium; df, dental follicle; dm, dental mesenchyme; dp, dental papillae; liCL, lingual cervical loop; laCL, labial cervical loop.

Download figure to PowerPoint

Disruption of Cbfb in the Epithelium Resulted in Significant Shortening of the Incisors and Enamel Defects

To analyze the functional role of Cbfb in the dental epithelium, we examined the tooth phenotypes of the conditional knockout mice, in which Cbfb expression was disrupted in the epithelium using Cre recombination. We confirmed the efficiency of the K14-Cre recombination using Rosa26R reporter mice and X-gal staining [29] in newborn mice. In both the incisors (Fig. 2T) and molars (Fig. 2U), X-gal staining was broadly detected in all dental epithelia. To generate mice in which Cbfb had been conditionally deleted in the epithelium (K14-Cre;Cbfbfl/fl), we mated K14-Cre and Cbfb loxP-containing (floxed) mice. In the PD 28 mice, Cbfb deficiency in the dental epithelium resulted in column-shaped incisors and loss of their transparency (Fig. 2A, 2B, and 2D) although the control mice developed fine edged shaped incisors (Fig. 2G, 2H, and 2J). Also, some Cbfb conditional mutants showed broken incisors that displayed increased fragility (data not shown). The Cbfb conditional mutant mice also developed molars with rough surfaces displaying mild enamel fractures (Fig. 2C) whereas the molar surfaces of the control mice were smooth (Fig. 2I). Micro-CT analysis revealed that the conditional Cbfb mutant mice developed extremely short lower incisors, the apex of which did not even reach the first molar (Fig. 2E, red arrowhead), whereas the control mice developed incisors that reached the third molar of the mandible (Fig. 2K, arrowhead). In the case of molars, there were no significant differences between the conditional Cbfb mutant mice and the control mice with regard to crown or root length or width (Fig. 2I and 2L).

thumbnail image

Figure 2. Epithelial Cbfb downregulation resulted in enamel hypoplasia and incisor shortening. Epithelial specific Cbfb null mice (K14-Cre;Cbfbfl/fl) showed severe enamel hypoplasia in their incisors and enamel chipping in their molars (A–C) compared with the control mice (G–I) although their head size was identical (A and G). (D–F, J–L): In micro-CT analysis, K14-Cre;Cbfbfl/fl mice showed abnormally shaped and extremely short incisors (D and E) compared with the control mice (J and K). The distal end of the lower incisor is indicated by a red arrowhead. On the other hand, there was no significant difference in molar root length or morphology (F and L). (M–S): Hematoxylin and eosin staining of mandibular incisor sagittal sections from PD28 K14-Cre;Cbfbfl/fl mice and control mice. K14-Cre;Cbfbfl/fl mice showed short incisors and defective ameloblast differentiation (M) whereas the control mice showed elongated incisors and normal ameloblast differentiation (Q). The boxes in (M) and (Q) show magnified views of the labial and lingual regions of each incisor. (N): Ameloblast cells in K14-Cre;Cbfbfl/fl mouse incisors show developmental defects from the beginning of differentiation (red arrowhead). (O): Secretion stage ameloblasts at labial surfaces of postnatal day 28 K14-Cre;Cbfbfl/fl mouse incisors showed flattened ameloblast cells (red arrowhead); however, the control mice developed well-polarized ameloblast cells (R, red arrowhead). On the other hand, the lingual sides of the incisors were identical at the histological level (P and S). X-gal staining was performed in the K14-Cre;R26R mice and confirmed that Cre recombination had been successfully performed in the newborn mouse incisor (T) and molar (U) epithelium. Scale bars = 500 μm in M and Q; 200 μm in N–P, R–U. Abbreviations: ab, alveolar bone; am, ameloblast; de, dentin; pdl, periodontal ligament.

Download figure to PowerPoint

We observed 43 PD28 conditional Cbfb mutant mice and found that 7 of them (16.2%) showed normal incisor development (data not shown). This was probably due to inefficient Cre-mediated gene recombination, and similar features were reported in a previous hair development study [23]. We did not use these inefficient Cre-mediated transgenic mice for any further analysis. Although phenotypes are evident in postnatal incisors, the morphology and size of the mutant incisor tooth germs had not been affected at E16 (Fig. 5A, 5D, and 5G).

Polarized Morphology of Incisor Ameloblasts Is Disrupted in Conditional Cbfb Mutant Mice

To investigate the causes of the short incisors and enamel defects seen in the conditional Cbfb mutant mice, we performed a histological analysis using mandible tissue sections. The hematoxylin and eosin stained mandible sections from the conditional Cbfb mutant mice revealed extremely short lower incisors, as observed by micro-CT (Fig. 2M and 2Q). As rodents continuously renew the dental epithelium-producing enamel matrix of their incisors throughout their life, incisor tooth germs exhibit a cell differentiation gradient from the apex to the incisal end. In particular, at the secretory stage, ameloblasts polarize and their nuclei align at the apical side of the cells (Fig. 2R, red arrowhead) [5]. However, no such polarization was evident in Cbfb mutant incisors, and the mutant epithelium was flattened from the primary differentiation stage through the secretion stage (Fig. 2N and 2O, red arrowhead). These findings clearly show that ameloblast differentiation is disrupted at an early stage because of Cbfb deficiency. In contrast, the histological appearance of the lingual side of these incisors was normal (Fig. 2O and 2S).

The Number of Proliferating Cells in the Incisor Cervical Loop Is Reduced in Conditional Cbfb Mutant Mice

To elucidate the mechanism by which short incisors are produced in conditional Cbfb mutant mice, we labeled proliferating cells using BrdU to assess whether the number of proliferative cells is affected by epithelial Cbfb deficiency. The conditional Cbfb mutant mice showed a significant reduction in the number of proliferating cells compared with the control mice in the developing cervical loop epithelium and the mesenchyme at PD10 (Fig. 3A and 3B). Furthermore, the labial cervical loops of the conditional Cbfb mutant mice were smaller (Fig. 3A and 3C, yellow dotted line) than those of the control mice (Fig. 3B and 3D, yellow dotted line). Quantification of the BrdU signals confirmed the reduction in the proliferation of the labial cervical loop epithelia of the conditional Cbfb mutant mice (Fig. 3E, p < .05).

thumbnail image

Figure 3. Epithelial Cbfb disruption caused a reduction in the number of proliferating cells in the laCL and underlying mesenchyme. Incorporated 5-bromo-3-deoxy-uridine (BrdU) is shown in green in (A) and (B) and in brown in (C) and (D). Nuclei are shown in blue in (A) and (B) and purple in (C) and (D). The yellow dotted line indicates the outline of the epithelium. (A): In the postnatal day 10 K14-Cre;Cbfbfl/fl mice, the number of cells incorporating BrdU was markedly reduced in the epithelium as well as the mesenchyme of the labial cervical loop (red arrowhead), compared with (B) the control mice (red arrowhead). The significance of the differences between the two groups (n = 3 individual/group) was determined by the Student's t test (*, p < .05) (E): In addition, the laCL was smaller in the K14-Cre;Cbfbfl/fl mice (C, D). Scale bar = 200 μm. Abbreviations: laCL, labial cervical loop; liCL, lingual cervical loop; TG, transgenic.

Download figure to PowerPoint

Epithelial Fgf9 as well as Mesenchymal Fgf3 and Fgf10 are Downregulated in the Developing incisors of Conditional Cbfb Mutant Mice

The reductions in the length and number of proliferating cells in the Cbfb mutant incisors suggest that the mechanism of incisor elongation is disrupted in these mice. As mentioned previously, it is well established that the epithelial–mesenchymal Fgf signaling loop in the developing incisor cervical loop plays a critical role in incisor elongation throughout the life of rodents [9]. Also it is well known that this mechanism is supported by Fgf9 expression in the epithelium, which upregulates Fgf3 and Fgf10 expression in the mesenchyme and vice versa [8, 38–40]. To investigate whether this Fgf signaling loop is related to the phenotype of conditional Cbfb mutant mice incisors, we analyzed the expression patterns of Fgf9, Fgf3, and Fgf10 in the two types of mice using in situ hybridization. In the normal E18 incisors, Fgf3 as well as Fgf10 were detected in the mesenchyme beneath the labial cervical loop (Fig. 4D and 4E, red arrowhead) and Fgf9 was observed in the epithelium (Fig. 4F, red arrowhead), as reported previously. In contrast, in the conditional Cbfb mutant mice, the expression of Fgf was downregulated at the same stage and in the same region (Fig. 4A, 4B, and 4C, red arrowhead). However, no such alteration of incisor Fgf expression was evident in the molars at E16 (data not shown).

thumbnail image

Figure 4. Fgf signaling loop as well as genes related to ameloblast differentiation (Shh, ameloblastin, amelogenin) are downregulated in developing K14-Cre;Cbfbfl/fl flox mice incisors. (A–F): The expression of Fgf family genes was analyzed by RNA in situ hybridization with the indicated probes in E18 mandibular incisor sagittal sections. The yellow dotted line indicates the outline of the epithelium. mRNA signals are shown in dark blue. Mesenchymal Fgf3 and 10 and epithelial Fgf9 were significantly downregulated in K14-Cre;Cbfbfl/fl mice (red arrowhead).Gene expression as analyzed by RNA in situ hybridization using the indicated probes in sagittal sections of the mandibular incisors of E16 (G and L) and E18 (H, J, M, and N) samples. The yellow dotted line indicates the outline of the epithelium. The expression levels of Shh as well as ameloblastin and amelogenin were significantly downregulated in the K14-Cre;Cbfbfl/fl dental epithelium (G, H, J, L–N). (I, K) Quantification of the relative mRNA amounts of ameloblastin and amelogenin in each developing incisor were determined by qRT-PCR. Both ameloblastin and amelogenin were significantly downregulated in the K14-Cre;Cbfbfl/fl developing incisor epithelium (p < .05 by Student's t test). Scale bars = 200 μm in A–F; 500 μm in G, H, J, L–N. Abbreviations: ab, alveolar bone; laCL, labial cervical loop; liCL, lingual cervical loop; TG, transgenic.

Download figure to PowerPoint

Epithelial Expression of Cbfb Is Necessary for Incisor Ameloblast Differentiation

To investigate the reason why the conditional Cbfb mutant mice showed defective enamel formation, we analyzed the expression levels of several genes that are important for ameloblast differentiation. It has been established that Shh signaling is crucial for ameloblast differentiation and enamel formation [26, 41–43]. As reported previously, Shh mRNA was found to be expressed throughout the preameloblast region of the control incisal epithelium at E16 using in situ hybridization (Fig. 4L). However, Shh expression was markedly downregulated in the epithelia of Cbfb mutants (Fig. 4G). These results strongly suggest that reduced Shh signaling in the Cbfb mutant incisor epithelium might be directly or indirectly involved in results in defective ameloblast differentiation.

Furthermore, we compared the expression pattern of ameloblastin and amelogenin, which are the main structural proteins of the enamel matrix [5, 44–47]. In situ hybridization and quantitative RT-PCR demonstrated a marked reduction in ameloblastin and amelogenin mRNA expression in conditional Cbfb mutant mice at E18 (Fig. 4H–4K, 4M, and 4N). In a previous study, mice in which Fgf signaling had been disrupted showed a similar ameloblast differentiation phenotype, including reduced Shh signaling [48]; however, still we saw a slight expression of these ameloblast differentiation-related genes (Shh, Ameloblastin, and Amelogenin) in mutants developing incisor epithelium (Fig. 4G–4K), presumably because the timing of Cre recombination by K14 transcription begins around E11 of development, which is too long before any epithelial cytodifferentiation [26].

Fgf9 Can Rescue the Incisor Developmental Defects of the Conditional CbfbMutant Mice

As epithelial Cbfb deficiency resulted in a significant downregulation of the expression levels of Fgf9, Fgf3, and Fgf10, which regulate cell proliferation in the incisor cervical loop and the adjacent mesenchyme [9, 38], we evaluated the possibility of rescuing the mutant phenotypes by applying FGF9 protein beads to the mutant explants. The E16 wild-type incisor explants elongated during 14 days of culture (5/5 cases; Fig. 5G, 5H), while the Cbfb mutant incisors had shrunk after 14 days culture (4/4 cases; Fig. 5D, 5E). Interestingly, beads soaked with FGF9 protein frequently (3/4 cases) rescued the developmental defects of the mutant incisors (Fig. 5A and 5B). SHH protein did not rescue the mutant phenotype (3/3 cases; data not shown).

thumbnail image

Figure 5. Exogenous FGF9 can rescue the incisor phenotype of K14-Cre;Cbfbfl/fl mice. (A–I): Organ culture was performed using E16 incisors. (C, F, and I): In immunohistochemistry of dissected tooth sections, the yellow dotted line indicates the outline of the epithelium, nuclei are shown in blue, and proliferating cell nuclei are shown in green. (G, H): Control mouse incisors incubated with BSA soaked beads were able to elongate for 14 days after the dissection. (I): Sections of cultured control incisors stained with phosphorylated histone H3 (a proliferation marker) showed proliferating cells in the epithelium (white arrowhead). However, in the K14-Cre;Cbfbfl/fl mice, the dissected incisor failed to survive and had shrunk at 14 days after the dissection (D and E). Also, there were no proliferating cells (F). Exogenous FGF9 protein rescued the phenotype of the K14-Cre;Cbfbfl/fl mice incisors at the morphological level (A and B) as well as cell proliferation in the epithelium (C, white arrowhead). Scale bar = 50 μm. Abbreviations: BSA, bovine serum albumin; laCL, labial cervical loop; liCL, lingual cervical loop.

Download figure to PowerPoint

To investigate whether cell proliferation was recovered by the implantation of beads soaked in FGF9 protein, we compared the expression pattern of phosphorylated histone H3 (mitosis marker). In mutant incisor explants, few mitosis marker signals were observed (Fig. 5F), while the wild-type explants showed that some signals were present in the cervical loop area (Fig. 5I, white arrowhead). The implantation of beads soaked in FGF9 protein into the Cbfb mutant incisor explants recovered cell proliferation in the labial cervical loop (Fig. 5C, white arrowhead). These results suggest that epithelial Cbfb induces cell proliferation in the cervical loop during incisor development via an Fgf signaling loop involving Fgf9.

Fgf9 Plays a Critical Role During Cervical Loop Development and Represses the Expression of Shh in the Dental Epithelium

From previous work, it seems that Fgf and Shh signaling interact during tooth development [8, 26]. In tooth, Shh expression was associated with Fgf in ameloblast progenitors [8, 29]. To investigate these interactions, we analyzed the incisor development of E18 Fgf9 null mutants. These mice die at birth and therefore postnatal phenotypes could not be analyzed [28]. In Fgf9 null mutants, the region of Shh expression during incisor development extended toward a more posterior position in the anterior–posterior (A–P) axis beyond the labial cervical loop (Fig. 6A, red arrowhead; n = 3 in each control and mutant group); that is, it was expanded compared with that of the control mice (Fig. 6B, red arrowhead). Also, we found that the labial cervical loop was smaller in the Fgf9 null mutants but the size reduction was milder than that observed in the Fgf10 null mice [9]. Identical sections revealed the relative distribution of Shh and Fgf9 mRNAs in the cervical loop regions. The Fgf9 expression domains overlapped with those of the Shh genes along but extended more posterior into the cervical loop regions (Fig. 6C and 6D). To further investigate the regulatory network involving these genes, dissected control E16 incisors were cultured with FGF9 or SHH recombinant protein, and whole mount RNA in situ hybridization was performed for the indicated genes (Fig. 6E–6H). We put beads soaked in recombinant FGF9 into the cervical loop region and examined the expression of Shh (n = 3, in each group). Interestingly, Shh expression was markedly downregulated by FGF9 in the dental epithelium (Fig. 6E and 6F), whereas SHH had no effect on Fgf9 expression (Fig. 6G and 6H). These results suggest that Fgf9 plays a critical role during cervical loop development by repressing the expression of Shh during incisor development.

thumbnail image

Figure 6. Fgf9 inhibits Shh signaling during incisor development. (A and B): The expression of Shh genes was analyzed by RNA in situ hybridization E16 Fgf9 null and control mice incisor sagittal sections. The expression of Shh was expanded in the E16 Fgf9 null mice incisor laCL (red arrowhead). (C and D): Identical sections were used for in situ hybridization of Shh (C) and Fgf [9]. (E–H): Dissected control E16 incisors were cultured with FGF9 or SHH recombinant protein, and whole mount RNA in situ hybridization was performed for the indicated genes. RNA transcripts are shown in dark blue. The white dotted line indicates the outline of the epithelium, and the yellow dotted line shows the location of the beads. (E and F) The expression of Shh was disrupted by exogenous FGF9. (G and H): However, exogenous SHH did not affect the expression of Fgf9. Abbreviations: BSA, bovine serum albumin; laCL, labial cervical loop; liCL, lingual cervical loop.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

We generated mice in which the Cbfb gene was deleted in the epithelium (K14-Cre;Cbfbfl/fl mice) thus leading to inactivation of the function of all Runx family target genes that were inactivated in the dental epithelium. The present results provide the first loss-of-function genetic evidence that Runx genes function in the maintenance and differentiation of stem cells in developing tooth. We show that Cbfb/Runx function is required for the Fgf signaling loop stem cell niche in the mouse incisor. In addition, Runx genes also contribute to the continuous growth of incisors by sustaining the expression of Fgf9 and Shh, whose complementary functions allow the retainment of a transit-amplifying domain and a differentiation domain.

Novel Roles of Runx Genes During Tooth Development

Rodent incisors are a unique experimental model for examining the mechanisms of stem cell maintenance. The location of the stem cell niche in rodent incisors is the cervical loop [4], and the progenitor pools that contribute to continuous incisor growth are expanded through continued proliferation along the A–P axis toward the tooth apex [5]. In this study, mice deficient in epithelial Cbfb failed to produce a fully developed cervical loop in their incisors which resulted in short incisors. In addition, the cellular proliferation in the cervical loop and the underlying mesenchyme was also markedly suppressed (Fig. 3). These findings indicate that Runx genes are involved in the maintenance of stem cells and the subsequent growth of the incisors (Fig. 7A). Epithelial Cbfb knockdown also results in absence of enamel in the incisors resulting from absence of ameloblasts (Fig. 4G–4N).

thumbnail image

Figure 7. Epithelial Runx family genes are necessary for ameloblast differentiation and maintenance of the stem cell compartment during incisor development. (A): Function of epithelial Runx/Cbfb genes during incisor development. The arrows indicate a stimulatory effect, and the ⟂ symbol indicates an inhibitory effect of one signaling molecule on the expression of another. In the wild-type mice, epithelial Runx/Cbfb genes upregulate the expression of Fgf9 and Shh in the dental epithelium. Epithelial Fgf9 stimulates mesenchymal Fgf3 and 10 expression and forms the Fgf signaling loop, which maintains cell proliferation in the laCL. At the same time, Fgf9 in the dental epithelium inhibits Shh signaling to form a compartment for ameloblast differentiation. Also, Shh itself is essential for ameloblast differentiation. (B): Proposed mechanism responsible for the K14-Cre;Cbfbfl/fl mice incisor phenotype. Epithelial disruption of Cbfb results in cell proliferation failure in the labial cervical loop and defective ameloblast differentiation. This phenotype is caused by the downregulation of epithelial Fgf9 and Shh, the expression of which is supported by epithelial Runx/Cbfb genes. Exogenous FGF9 protein rescued the cell proliferation in the laCL of K14-Cre;Cbfbfl/fl mice incisors by rescuing the Fgf signaling loop. Abbreviations: laCL, labila cervical loop; liCL, lingual cervical loop.

Download figure to PowerPoint

Our defective incisor phenotypes are analogous of the proposed role for Runx genes in a variety of stem cell systems. Runx1 deficient mice die between embryonic days 11.5 and 12.5 with no hematopoietic progenitors in the liver or yolk sac [49, 50]. HFSCs reside in the bulge and express Runx1 prior to their proliferation. Constitutive epithelial deletion of Runx1 prolongs the hair cycle in the quiescent phase and impairs stem cell colony formation [23]. The C. elegans seam cells display stem cell-like properties, and some of them give rise to male specific sense organs. Rnt-1 and Bro-1, which are homologues of Runx and Cbfb, respectively [51], promote self-renewal and proliferation as rate-limiting regulators [52, 53]. In addition to incisor phenotypes caused by insufficient stem cell activation, the Cbfb-deficient mice also showed defective enamel formation. Such phenotypes are consistent with the observed hair development phenotypes. For example, epithelial Runx1 deficiency resulted in mild hair shaft deformation [54].

Runx Maintains the Stem Cell Niche Through the Activation of Fgf Signaling

The incisor phenotypes that we observed in this study are similar to those of Fgf10-deficient mice. Fgf10 mRNA is continuously expressed in the mesenchyme surrounding the cervical loop [55], and mice deficient in Fgf10 fail to develop the incisor cervical loop, suggesting that Fgf10 plays a role in the maintenance of the incisor stem cell niche [9, 35, 40, 48]. Cbfb deficiency resulted in marked downregulation of Fgf10, as well as Fgf3, expression in the mesenchyme underlying the cervical loop (Fig. 4A, 4B, 4D, and 4E). These findings indicate that epithelial Runx genes are required for Fgf signaling and the maintenance of the stem cell niche (Fig. 7A and 7B).

In addition to mesenchymal Fgf10, Fgf9 mRNA, which is localized in a small epithelial domain just anterior to the cervical loop, was also downregulated in Cbfb mutant incisors (Fig. 4C and 4F). Fgf4, 8, and 9 have similar functions in the regulation of adjacent mesenchymal cell proliferation and/or the prevention of apoptosis during molar development [56]. However, in incisors, Fgf4 and Fgf8 are expressed during the early developmental stages, and apparently only Fgf9 expression is maintained into adulthood in the incisor epithelium. In the cervical loop regions, a positive-feedback signaling loop is established between epithelial Fgf9 and mesenchymal Fgf10 and Fgf3 in the cervical loop to ensure stem cell maintenance [8, 39]. A recent study using Fgfr2b mutant mice confirmed the importance of such Fgf signaling loops [48]. It has also been established that Fgf signaling by Fgf10 and Fgf9 is mediated by FgfR2b, and conditional FgfR2b null mice exhibit a reduced cervical loop with suppressed cellular proliferation [57]. In addition, the functional significance of epithelial Fgf9 in the survival of progenitors and the induction of epithelial proliferation has also been reported in testis development [58, 59]. Taken together, our findings provide novel molecular evidence that Runx genes are involved in the maintenance of stem cells through the activation of the Fgf signaling loop in growing incisors.

Our organ culture findings confirmed the significance of Runx genes in the maintenance of stem cells and the induction of the Fgf signaling loop. The wild-type incisor explants grew in length in organ culture (Fig. 5G and 5H), while the mutant incisor explants did not grow but rather shrank in the same experimental conditions (Fig. 5D and 5E). FGF9-soaked beads placed next to the cervical loops rescued the Cbfb mutant phenotype (Fig. 5A and 5B). A proliferation assay also confirmed that FGF9 rescued the suppressed proliferation in the mutant cervical loop epithelium caused by Cbfb deficiency (Fig. 5C, 5F, and 5I). These findings also suggest that Fgf9 lies downstream of Runx genes during the maintenance of epithelial stem cells (Fig. 7B). Fgf9-deficient mice showed a slightly reduced cervical loop (Fig. 6A and 6B) as compared to the Cbfb and Fgf10-deficient mice [9]. These findings suggest that other Fgfs which are regulated by Runx genes also play redundant roles in the epithelium during the formation of the Fgf signaling loop and maintenance of the cervical loop stem cells.

It is striking that defective phenotypes are evident in mutant incisors but less so in molars. While mutant incisors failed to grow and lacked enamel, only minor enamel chipping was noted in mutant molars (Fig. 2C) and the size and the morphology of the molars were minimally affected. These differences are apparently due to the fact that the molars do not grow continuously and they do not produce enamel continuously like incisors. Our findings indicate that Runx genes have a minor role in the early incisor and molar morphogenesis. The later incisor morphogenesis involves the activation and proliferation of stem cells in the cervical loop to obtain continuous growth and enamel formation. Hence, Cbfb deficiency produces marked phenotypes in the later stages of incisor development. The phenotype of Fgf10 mutant is similar in this respect. Such specific functions of Runx genes during stem cell activation, but not morphogenesis, are in line with the roles of Runx1 in hair development. In epithelial Runx1 knockout mice, Runx1 disruption resulted in a slightly deformed hair shaft but produced a normal hair follicle. However, Runx1 deficiency affects HFSC activation and hair cycling at the transition to adult skin homeostasis [23].

Epithelial Runx Genes Play Critical Roles During Ameloblast Differentiation Through Shh and Fgf Signaling

Another marked phenotype of the mutant incisors was enamel aplasia. Shh is crucial for ameloblast differentiation [26, 42, 60], and the expression of the enamel matrix proteins amelogenin and ameloblastin is induced by SHH in vitro [43]. In this study, we observed the marked suppression of Shh mRNA expression as well as ameloblastin and amelogenin mRNA in developing incisors (Fig. 4G–4N), indicating that Runx genes are involved in the continuous differentiation of epithelial progenitors to ameloblasts in growing incisors (Fig. 7A and 7B). It is also possible that the differentiation defect is a result of the progenitor defect [39]. On the other hand, Fgf signaling is also associated with enamel formation and/or ameloblast differentiation. Fgf3−/−, Fgf10+/−, and Fgf3−/− mice also showed enamel hypoplasia [40]. More severe phenotypes were observed by conditional Fgfr2b knockdown [48, 57], and these Fgfr2b mutant mice lacked ameloblasts and enamel with significant suppression of Shh expression in the inner enamel epithelium [48]. Hence, it is possible that the Fgf signaling loop in the cervical loop maintains Shh expression in differentiating ameloblasts. We cannot exclude the possibility of the direct induction of Shh by Runx family genes, but we can state that Runx genes directly or indirectly upregulated Shh in the inner epithelium and induced the subsequent induction of ameloblast differentiation. Another explanation may be that defective enamel phenotypes are associated with the Runx2-mediated signaling pathway. Runx2 is expressed in differentiated ameloblasts in both incisors and molars [61] and in vitro [62]. The importance of Runx2 in amelogenesis is evidenced by poorly differentiated ameloblasts in the incisor tooth germs of Runx2-deficient mice [63]. A recent study showed Runx2 control enamel-related genes MMP20 and odontogenic ameloblast-associated protein (ODAM) [62]. MMP20 null mutation caused impaired maturation of enamel that mimicked hypomineralized/hypomature amelogenesis imperfecta [64, 65] and ODAM was expressed in ameloblasts during the secretory and maturation stages [66]. An in vitro study showed that Runx2 regulates the expression of the ODAM protein level, which in turn regulates MMP20 promoter activity, thus suggesting that Runx2 serves an important regulatory function in the mineralization of enamel [62].

Runx Genes Retained Complementary Expression of Fgf9 and Shh, Whose Antagonistic Function Defines the Boundary Between the Amplifying and Differentiation Domains

In continuously growing incisors, the epithelial cell lineage is subdivided into several domains with increasing levels of differentiation along the A–P axis. There is a self-renewal compartment containing quiescent stem cells in the cervical loop, a proliferating compartment containing proliferating progenitors, and a series of differentiation compartments with specific markers. Our findings suggest that Fgf9 together with other epithelial Fgfs are responsible for the steady supply of the ameloblast lineage cells. In contrast, the Shh expressing domain represents the compartment of the differentiating progeny of the stem cells. Shh is required for ameloblast differentiation and it directly induces ameloblast differentiation markers [43]. Interestingly, we found that Fgf9 deficiency resulted in the expansion of the Shh-expressing domain toward a more posterior position in the A–P axis in Fgf9 null mutant incisors (Fig. 6A and 6B, red arrowhead), presumably due to the reduction in Shh-negative ameloblast progenitors. But, we also found that ectopic FGF9 beads significantly downregulated the expression of Shh mRNA in incisor explants (Fig. 6C and 6D) indicating that Fgf9 has an inhibitory effect on Shh expression. Also our RNA in situ hybridization analysis revealed that endogenous Fgf9 transcribe occurs at mutual region of Shh expression in growing incisors (Fig. 4F and 4L). From these findings, it is also possible that Fgf9 signaling prevents the progenitors from responding to Shh signals, thus holding them in an undifferentiated state in the transit-amplifying domain (Fig. 7A). This type of molecular arrangement might contribute to the distinctive cellular compartments; that is, a transit amplifying domain and a differentiation domain, in incisors. An undifferentiated and proliferating field of Fgf9-expressing epithelium produces the spatial pattern for continuous elongation, and these lineage cells subsequently enter into the Shh-expressing differentiating domain to undergo co-ordinated ameloblast differentiation. Recent findings showed that Shh in ameloblast maintained epithelial stem cell and established a positive-feedback loop [42]. Hence, it is also possible that the downregulation of Fgfs would prevent differentiation of the Shh-positive cells and prevent such positive effect on the progenitor maintenance. Taken together with the findings from the Cbfb null mutant mice, we suggest that Runx genes contribute to the continuous growth of incisors by sustaining the expression of Fgf9 and Shh, whose complementary functions allow the retainment of a transit-amplifying domain and a differentiation domain (Fig. 7A).

SUMMARY

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

The present results provide the first genetic evidence that Runx/Cbfb genes function in the maintenance of stem cells in developing incisors by activating Fgf signaling loops between the epithelium and mesenchyme. In addition, Runx genes also orchestrate continuous proliferation and differentiation by maintaining the expression of Fgf9 and Shh mRNA.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

We thank Irma Thesleff for critical reading of the manuscript, David M. Ornitz for providing the Fgf9 null mutant mice, and Shigeru Kuratani for access to the micro-CT equipment. This work was supported by grants-in-aid for scientific research program from the Japan Society for the Promotion of Science (to T.Y.).

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES

The authors indicate no potential conflict of interest.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. SUMMARY
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES