Mesenchymal Wnt/β-catenin signaling induces Wnt and BMP antagonists in dental epithelium

ABSTRACT

Previous studies indicated that the elevated mesenchymal Wnt/β-catenin signaling deprived dental mesenchyme of odontogenic fate. By utilizing ex vivo or pharmacological approaches, Wnt/β-catenin signaling in the developing dental mesenchyme was suggested to suppress the odontogenic fate by disrupting the balance between Axin2 and Runx2. In our study, the Osr2-cre^KI; Ctnnb1^ex3f mouse was used to explore how mesenchymal Wnt/β-catenin signaling suppressed the odontogenic fate in vivo. We found that all of the incisor and half of the molar germs of Osr2-cre^KI; Ctnnb1^ex3fmice started to regress at E14.5 and almost disappeared at birth. The expression of Fgf3 and Msx1 was dramatically down-regulated in the E14.5 Osr2-cre^KI; Ctnnb1^ex3f incisor and molar mesenchyme, while Runx2transcription was only diminished in incisor mesenchyme. Intriguingly, in the E14.5 Osr2-cre^KI; Ctnnb1^ex3f incisor epithelium, the expression of Noggin was activated, while Shh was abrogated. Similarly, the Wnt and BMP antagonists, Ectodin and Noggin were also ectopically activated in the E14.5 Osr2-cre^KI; Ctnnb1^ex3f molar epithelium. Recombination of E13.5 Osr2-cre^KI; Ctnnb1^ex3f molar mesenchyme with E10.5 and E13.5 WT dental epithelia failed to develop tooth. Taken together, the mesenchymal Wnt/β-catenin signaling resulted in the loss of odontogenic fate in vivo not only by directly suppressing odontogenic genes expression but also by inducing Wnt and BMP antagonists in dental epithelium.

INTRODUCTION

Wnt/β-catenin signaling is involved in multiple organogenesis, such as hair follicle, glands, bone, lung, etc.^1–5 During tooth initiation and morphogenesis, Wnt/β-catenin signaling is only detected in dental epithelium but excluded from dental mesenchyme.^6,7 However, when Wnt/β-catenin signaling is inactivated in dental mesenchyme by deleting β-catenin, tooth development was arrested at the bud stage,⁸ implicating that Wnt/β-catenin signaling is indispensable for tooth morphogenesis. Since β-catenin is not only a transducer mediating Wnt/β-catenin signaling, but also a key component of cytoskeleton,⁹ the deletion of β-catenin may disable tooth morphogenesis by impairing cell mobility. On the other hand, constitutive activation of Wnt/β-catenin signaling in mouse dental mesenchyme results in the bone-like pulp and compromised enamel,⁸ indicating a negative regulation on the differentiation of ameloblasts and odontoblasts. Our previous study demonstrated that the Wnt/β-catenin signaling in dental mesenchyme was suppressed by FGF3 to maintain the odontogenic fate of dental mesenchymal cells.⁷ The latest study also verified that the activation of Wnt/β-catenin signaling in mouse embryonic mesenchyme limited the tooth number by suppressing the formation of the 2^nd and 3^rd molars.¹⁰As an inhibitor of Wnt/β-catenin signaling, Axin2 deficiency in human was associated with oligodontia.¹¹Taken together, Wnt/β-catenin signaling in dental mesenchyme is currently suggested to exert inhibitory effects on odontogenic capability and require a fine-tuning during tooth development.

Through the ex vivo and pharmacological approaches, the Wnt/β-catenin signaling in dental mesenchyme was suggested to be finely tuned by FGF/AKT signaling or the antagonism between Runx2 and Axin2.^7,10However, the supplement of pharmacological chemicals ex vivo activates Wnt/β-catenin signaling in both the epithelium and mesenchyme, instead of solely in dental mesenchyme. Since tooth development was accomplished through the reciprocal epithelial-mesenchyme interactions,¹²the elevated Wnt/β-catenin signaling in dental epithelium would trigger the extra responses in dental mesenchyme, which confuses the real influences of the elevated mesenchymal Wnt/β-catenin signaling on odontogenic capability. Therefore, the Osr2-cre;Ctnnb1^ex3f mouse was employed in this study to address the in vivo effects of the mesenchymal Wnt/β-catenin signaling on dental epithelium and mesenchyme, respectively.

RESULTS

The impaired morphogenesis in Osr2-cre^ki;Ctnnb1^ex3f tooth germs

The Osr2-cre^KI mice were mated with Ctnnb1^ex3f mice to specifically express a constitutively stabilized form of β-catenin in the developing dental mesenchyme.^6,8 At E14.5, when WT incisors and molars developed the typical cage-like enamel organs (Fig. 1A–C), the tooth germs of the Osr2-cre^KI;Ctnnb1^ex3f mouse seemed in the comparable stage (Fig. 1A’, 1B’, 1C’), except the mandibular incisor germs retarded in a bud shape (Fig. 1B’). Intriguingly, when WT incisor germs got into the bell stage and secreted dentin at the E16.5 (Fig. 1D, 1E), the Osr2-cre^KI;Ctnnb1^ex3f maxillary incisors were undergoing severe regression and losing the typical bell- or cap-like shape (Fig. 1D’), and the Osr2-cre^KI;Ctnnb1^ex3f mandibular incisor germs diminished completely (Fig. 1E’). Compared with the WT molar germs (Fig. 1F, 1G), about 50% (15/31) of E16.5 Osr2-cre; Ctnnb1^ex3f mice lost their maxillary and mandibular molars completely. Even in the Osr2-cre;Ctnnb1^ex3f molar germs developing into bell stage, their sizes are smaller than those in WT control (Fig. 1F’, 1G’). These findings indicated that the mesenchymal Wnt/β-catenin signaling impaired the morphogenesis of Osr2-cre; Ctnnb1^ex3f tooth germs, especially in the incisor germs.

FIGURE 1.

Histological analysis on tooth germs of Osr2-cre^KI;Ctnnb1^ex3f mouse. Azon staining shows that the enamel organs of both the maxillary (A) and mandibular incisor germs (B) of the E14.5 WT mouse were typical cage-like. While in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f mouse, the maxillary incisor germs (A’) were relatively normal, and the mandibular incisor germs (B’) retained in bud stage. Compared with the E14.5 WT molar germs (c), the E14.5 Osr2-cre^KI;Ctnnb1^ex3f molar germs showed a smaller size (C’). At E16.5, both the WT maxillary (D) and mandibular incisor germs (e) developed into bell stage, while the Osr2-cre; Ctnnb1^ex3f maxillary incisor (D’) only degenerate to epithelial residues and mandibular incisor (E’) disappeared completely. Compared with the E16.5 WT maxillary (F) and mandibular molars (G), the E16.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (F’) and mandibular molar germs (G’) developed into bell-stage, but in a smaller size. (Dashed lines delineated the boundary between epithelium and mesenchyme; scale bar: 200μm).

The cell proliferation and survival in Osr2-cre^KI;Ctnnb1^ex3f tooth germs

To ensure the elevated Wnt/β-catenin signaling in dental mesenchyme, the distributions of the Wnt/β-catenin signaling effector, Axin2, and the mediator, Lef1 were examined in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f tooth germs. As shown by immunohistochemistry, compared with the WT control (Fig. 2A,B,C), the Axin2 distribution was throughout the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary and mandibular incisor and molar mesenchyme, even extended into the palatal and mandibular mesenchyme (Fig 2A’,B’,C’). Similarly, the Lef1 domain in the Osr2-cre^KI;Ctnnb1^ex3f mice also extended from the mesenchyme surrounding tooth epithelial buds in the WT control (Fig. 2D,E,F) to the palatal and mandibular mesenchyme (Fig 2D’,E’,F’). The expanded Axin2 and Lef1 domain indicated that the Wnt/β-catenin signaling was indeed constitutively activated in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f tooth mesenchyme.

FIGURE 2.

The Axin2 and Lef1 distribution in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f tooth germs. Immunohistochemistry with the antibody against Axin2 showed that Aixn2 distributed in the both the WT maxillary (A) and mandibular incisor (B) epithelium and mesenchyme surrounding the epithelial buds, but not in the medial mesenchyme between incisor germs (red arrows in A and B). In contrast, the Axin was not only detected in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (A’) and mandibular incisor (B’) epithelium and the surrounding mesenchyme, but also in the mesenchyme between incisor germs (red arrows in A’ and B’). In the E14.5 WT maxillary and mandibular molar germs (C), Aixn2 was detected in the epithelium and the mesenchyme adjacent to epithelium (red arrows in C), but absent from the palatal and mandibular mesenchyme (red asterisks in C). The Axin2 staining was detected in E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary and mandibular epithelium (C’) and all the condensed molar mesenchyme (red arrows in C’), even extended into the palatal and mandibular mesenchyme (red asterisks in C’). Similarly, the immunohistochemistry with the antibody against Lef1 exhibited the similar pattern in the maxillary and mandibular incisor germs of E14.5 WT (D,E) and Osr2-cre^KI;Ctnnb1^ex3f mice (D’,E’). The Lef1 was detected in the inner enamel epithelium of WT (F) and Osr2-cre^KI;Ctnnb1^ex3f molar germs (F’). In the WT(F) and Osr2-cre^KI;Ctnnb1^ex3f mesenchyme (F’), Lef1 distribution was almost same as that of Axin2. (red arrows in F and F’ indicated the molar mesenchyme; red asterisks in F and F’ indicated the palatal and mandibular mesenchyme; scale bar: 200μm).

BrdU labeling test revealed that in the E13.5 Osr2-cre^KI;Ctnnb1^ex3f tooth germs, the BrdU-positive cells were reduced remarkably in the incisor epithelium and mesenchyme (Fig 3A’,B’), especially in the mandibular incisor epithelium (Fig 3B’). The decreases of cell proliferation in Osr2-cre^KI;Ctnnb1^ex3f maxillary and mandibular molar epithelium were also evident as that in incisors (Fig. 3C,C’). However, the mesenchymal proliferation of Osr2-cre^KI;Ctnnb1^ex3f maxillary molar germs differed little from the control groups, while the Osr2-cre^KI;Ctnnb1^ex3f mandibular molar mesenchyme exhibited a decreased cell proliferation (Fig. 3C,C’,G). TUNEL positive cells in the E13.5 Osr2-cre^KI;Ctnnb1^ex3f incisor mesenchyme were increased, especially in the maxillary incisor (Fig 3D’,E’). By contrast, the TUNEL positive cells in the Osr2-cre^KI;Ctnnb1^ex3f maxillary and mandibular molar germs were comparable to those in WT counterparts (Fig. 3F,F’). These results suggested that the impaired morphogenesis of Osr2-cre^KI;Ctnnb1^ex3f tooth germs was mainly attributed to the reduced cell proliferation in both dental epithelium and mesenchyme.

FIGURE 3.

The cell proliferation and death in the E13.5 Osr2-cre^KI;Ctnnb1^ex3f tooth germs. BrdU labeling showed the BrdU positive cells in both the WT maxillary incisor epithelium and mesenchyme (A) were more than those in Osr2-cre^KI;Ctnnb1^ex3f maxillary incisor germs (A’). Similarly, the numbers of BrdU positive cells in the WT mandibular incisor epithelium and mesenchyme (B) were also larger than those in Osr2-cre^KI;Ctnnb1^ex3f mandibular incisor germs (B’). Although the BrdU positive cells in the WT maxillary and mandibular molar epithelium (C) was obviously more than those in Osr2-cre^KI;Ctnnb1^ex3f molar epithelium (C’), there was no difference in the BrdU cell numbers between WT (C) and Osr2-cre^KI;Ctnnb1^ex3f maxillary molar mesenchyme (C’). In contrast, the BrdU-positive cells in WT mandibular molar mesenchyme were more than that of Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme (C’). TUNEL assay indicated the cell death in the WT maxillary (D) and mandibular (E) incisor mesenchyme was more than those in Osr2-cre^KI;Ctnnb1^ex3f maxillary (D’) and mandibular incisor mesenchyme (E’). However, the cell death in the WT maxillary and mandibular molar germs (F) was comparable to those in Osr2-cre^KI;Ctnnb1^ex3f maxillary and mandibular molar mesenchyme (F’). Statistical analyses on the percentage of BrdU positive cells in the WT and Osr2-cre^KI;Ctnnb1^ex3f dental mesenchyme were shown in G. (Blue and white dot lines indicated the boundary between the epithelium and mesenchyme; * meant p< .05; ** meant p< .001; scale bar: 200 μm).

Persistent Wnt/β-catenin signaling in dental mesenchyme suppressed the expression of odontogenic markers

To explore how the mesenchymal Wnt/β-catenin signaling suppressed the odontogenic fate, the expression pattern of odontogenic markers was examined in Osr2-cre^KI;Ctnnb1^ex3f tooth germs. At E14.5, Fgf3 was intensively expressed in the maxillary and mandibular incisor mesenchyme (Fig. 4A,B), however, neither the maxillary nor mandibular incisor mesenchyme expressed Fgf3 in Osr2-cre^KI;Ctnnb1^ex3f mouse (Fig. 4A’,B’). Even in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f molar germs (Fig. 4C’₁,C’₂), the expression of Fgf3 was significantly reduced compared with the WT control (Fig. 4C). Another mesenchymal odontogenic marker, Runx2 was expressed robustly in maxillary bone and relatively mildly in maxillary incisor mesenchyme of the E14.5 WT mouse (Fig. 4D). In the WT mandible, Runx2 expression was silenced in incisor mesenchyme but activated in the surrounding bone (Fig. 4E). In contrast, the Runx2 expression was excluded from the regressing E14.5 Osr2-cre^KI;Ctnnb1^ex3f incisor germs though still detected in the surrounding bone (Fig. 4D’,E’). In the E14.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme, the expression of Runx2 was a little stronger than that in the surrounding bone (Fig. 4F’₁,F’₂), which was opposed to the weaker Runx2 intensity in WT molar mesenchyme compared with the bone (Fig. 4F). Moreover, compared with the E14.5 WT maxillary and mandibular incisors (Fig. 4G), the Msx1-expressing domain in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary incisor mesenchyme was reduced and even disappeared in the mandible (Fig. 4G’). Similarly, the intensity of Msx1 transcription in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme (Fig. 4H’) was also remarkably decreased than that in WT molar mesenchyme (Fig. 4H). The quantitative PCR further confirms that in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme, the transcription of Fgf3, and Msx1 were decreased, while the Runx2 transcription was increased in contrast to the corresponding transcripts in WT molar mesenchyme (Fig. 4I). These results suggested that the mesenchymal Wnt/β-catenin signaling suppressed the odontogenic fate in vivo.

FIGURE 4.

The expression of mesenchymal odontogenic markers in E14.5 Osr2-cre^KI;Ctnnb1^ex3f tooth germs. In situ hybridization indicated that compared with the robust expression in the E14.5 WT maxillary (A) and mandibular incisors (B), the expression of Fgf3 was not detected in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (A’) and mandibular incisors (B’); though the Fgf3 expression could be detected in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (C’₁) and mandibular molar germs (C’₂), it was significantly weaker than those in WT molar germs (C). In the E14.5 WT mouse, Runx2 was expressed in maxillary incisor mesenchyme (D), and silenced in mandibular incisor mesenchyme (E), while both the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (D’) and mandibular incisor mesenchyme (E’) were devoid of Runx2 expression; the expression of Runx2 in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (F’₁) and mandibular molar mesenchyme (F’₂) was mildly stronger than that in the WT molar germs (F). Msx1 expression was evident in the E14.5 WT maxillary and mandibular incisor mesenchyme (G). In contrast, Msx1 expression was reduced in the maxillary incisor mesenchyme, and diminished in the mandibular incisor of the E14.5 Osr2-cre^KI;Ctnnb1^ex3f mouse (G’). Similarly, Msx1 transcription in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme (H’) was also weaker than that in WT molar mesenchyme (H). The Q-PCR showed the significant alterations in Fgf3, Msx1 and Runx2 expression in the E14.5 WT and Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme (I). (Dashed lines delineated the boundary between epithelium and mesenchyme; * meant p< .05; ** meant p< .01; scale bar: 200μm).

Induced BMP and Wnt antagonists in Osr2-cre^KI;Ctnnb1^ex3fdental epithelium

To clarify the epithelial responses to the persistent Wnt/β-catenin signaling in the Osr2-cre^KI;Ctnnb1^ex3f dental mesenchyme, gene expression was checked in the E14.5 tooth germs. In the E14.5 Osr2-cre^KI;Ctnnb1^ex3f incisor epithelium, the Pitx2 expression was still active throughout the regressing enamel organs (Fig. 5A’,A’’), displaying no difference from the WT control (Fig. 5A). However, Noggin, a BMP antagonist which was silenced in WT oral epithelium (Fig. 5B), was ectopically activated in the Osr2-cre^KI;Ctnnb1^ex3f oral epithelium, especially the regressing incisor epithelium (Fig. 5B’). In contrast, although actively transcribed in the inner enamel epithelium of WT incisors (Fig. 5C,D), the expression of Shh was abrogated in the Osr2-cre^KI;Ctnnb1^ex3f incisor epithelium (Fig. 5C’,D’).

FIGURE 5.

The expression of epithelial odontogenic markers in the E14.5 Osr2-cre^KI;Ctnnb1^ex3ftooth germs. As assessed by in situ hybridization, Pitx2 expression was detected in the maxillary and mandibular incisor epithelium of the E14.5 WT mouse (A), as well as in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (A’) and mandibular (A’’) incisor epithelium. Noggin expression was inactivated in the E14.5 WT incisor germs (B), but ectopically activated in the Osr2-cre^KI;Ctnnb1^ex3f oral epithelium (B’). Although robust in the E14.5 WT maxillary (C) and mandibular incisor epithelium (D), Shh expression was diminished in the E14.5 Osr2-cre^KI;Ctnnb1^ex3f maxillary (C’) and mandibular incisor epithelium (D’). However, Shh expression in the hair follicle of Osr2-cre^KI;Ctnnb1^ex3f mouse (arrows in D’) remained robust. In Osr2-cre^KI;Ctnnb1^ex3f mouse, Ectodin expression was diminished in the molar mesenchyme, but ectopically activated in the entire enamel organ (F, F’), compared with the WT molar germs (E, E’). Noggin was inactivated in the WT molar germs (G, G’), but ectopically activated in the Osr2-cre^KI;Ctnnb1^ex3f oral and palatal epithelium (H, arrow in H’). Compared with the WT control (J, J’), Sfrp2 expression was excluded from palatal mesenchyme (I) and the superficial oral mesenchyme (I’) in the Osr2-cre^KI;Ctnnb1^ex3f mouse. The Q-PCR showed the significant alterations in Ectodin, Noggin and Shh transcription in the E14.5 WT and Osr2-cre^KI;Ctnnb1^ex3f molar epithelium were statistically analyzed (K). (Dashed lines delineated the boundary between epithelium and mesenchyme; E’, F’, G’, H’, I’ and J’ were the enlarged box areas in E, F, G, H, I and J, respectively; ** meant p< .01; *** meant p< .001; scale bar: 200μm).

Similar to the incisor germs, the constitutive Wnt/β-catenin signaling in molar mesenchyme also activated the ectopical expression of Wnt and BMP antagonists. Ectodin, the dual antagonist to Wnt and BMP signaling, was expressed surrounding the E14.5 WT molar germs (Fig. 5E,E’). However, Ectodin was diminished in the Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme, but distributed intensively throughout the oral epithelium, even including the epithelia of tongue and palatal shelves (Fig. 4F, 4F’). Although hard to be detected in E14.5 WT molar germs (Fig. 5G,G’), Noggin was also specifically activated in the epithelium throughout molar, tongue and palatal shelves in Osr2-cre^KI;Ctnnb1^ex3f mouse (Fig. 5H,H’). In contrast, another Wnt inhibitor, Sfrp2, which was not expressed in the E14.5 WT molar mesenchyme (Fig. 5I,I’), was not activated in the Osr2-cre^KI;Ctnnb1^ex3f oral epithelium (Fig. 5J,J’). Quantitative PCR was performed and confirmed the significant increases of Noggin and Ectodin in the Osr2-cre^KI;Ctnnb1^ex3f molar epithelium (Fig. 5K). The Shh transcription in Osr2-cre^KI;Ctnnb1^ex3f molar epithelium was down-regulated, which was similar to that in the incisor epithelium (Fig. 5K). Therefore, the persistent mesenchymal Wnt/β-catenin signaling was suggested to induce ectopic expression of Noggin and Ectodin in the dental epithelium.

The loss of odontogenic capability in Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme

Tissue recombination was operated to further confirm that the decreased odontogenic capability of molar mesenchyme resulted from the persistent mesenchymal Wnt/β-catenin signaling. When the E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar germs were dissociated into single cell suspension and re-aggregated into a mass, no tooth structure was found after 2 weeks of sub-renal culture (Fig. 6A’). When combined with the E10.5 dental epithelium, the E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme still failed to form tooth (Fig. 6B’). Even when recombined with the intact E13.5 WT molar epithelium, the E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme gave rise to only bone and keratinzed tissues after sub-renal culture (Fig. 6C’). These results indicated that the constitutive activation of Wnt/β-catenin signaling in dental mesenchyme disabled the odontogenic capability of dental mesenchyme.

FIGURE 6.

Tissue recombination by utilizing E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme. After 2 weeks of sub-renal culture, the re-aggregates from the single cell suspension of the E13.5 WT molar germs produced tooth structure (A), while the re-aggregates from Osr2-cre^KI;Ctnnb1^ex3f molar germs only gave rise to keratinized beads and bone (A’). When combined with the E10.5 dental epithelium, the E13.5 WT molar mesenchyme formed tooth structure (B), but the E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme still failed to form tooth (B’). Even when recombined with the intact E13.5 WT molar epithelium, the E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme gave rise to only bone and keratinized tissues after sub-renal culture (C). In contrast, the typical tooth structure was detected in the recombinants of E13.5 WT molar mesenchyme and epithelium (C). (Dashed lines delineated the boundary between epithelium and mesenchyme; epi – epithelium; mes – mesenchyme; scale bar: 200μm).

DISCUSSION

Previous report showed the bone-like dental pulp and the compromised differentiation of the odontoblasts and ameloblasts in the Osr2-cre^KI; Ctnnb1^ex3f molars.⁸ By utilizing the organ culture or ex vivo approaches, several studies showed that by elevating Wnt/β-catenin signaling during differentiation stage, the differentiation of odontoblasts was promoted.^13,14 In contrast, during tooth morphogenesis, the enhanced Wnt/β-catenin signaling disturbed the normal morphology and cusps patterning of molar germs, as well as the odontoblast differentiation.¹⁵ In our study, all of the incisor germs and 50% of the molar germs in Osr2-cre^KI; Ctnnb1^ex3f mice were arrested at E13.5 or E14.5, and regressed eventually. All these results indicated an inhibitory effect of mesenchymal Wnt/β-catenin signaling on tooth morphogenesis. Such conclusion was verified by the latest study using Dermo1-cre; Ctnnb1^ex3f mouse, whose possessed the normal first molar, but lacked the second and third molars.¹⁰ Since the Osr2-cre^KI; Ctnnb1^ex3f mouse died of the cleft palate immediately after birth, to further confirm if the second and third molars were absent, the sub-renal culture of the intact molar germs will be operated in the future study.

How the increased Wnt/β-catenin signaling in the dental mesenchyme exerted the inhibitory effect during tooth morphogenesis attracts the concerns. Our previous study demonstrated that the elevated Wnt/β-catenin signaling in dental mesenchymal deprived the odontogenic fate and transformed the cells into osteogenic program.⁷ This ex vivo conclusion was consistent with the in vivo decreased Fgf3 expression in the Osr2-cre^KI; Ctnnb1^ex3f molar mesenchyme. As a key molecule repressing the mesenchymal Wnt/β-catenin signaling, FGF3 maintained the odontogenic fate of dental mesenchyme.⁷ Combined with the decreased Msx1 transcription, the odontogenic capability of the Osr2-cre^KI; Ctnnb1^ex3f molar mesenchyme was supposed to be attenuated by the active Wnt/β-catenin signaling.¹⁶ Correspondingly, the elevated Runx2 expression might represent the enhanced osteogenic tendency in Osr2-cre^KI; Ctnnb1^ex3f molar mesenchyme.

Tooth development is accomplished by the reciprocal signal exchanges between dental epithelium and mesenchyme. However, the effects of the constitutive activated mesenchymal Wnt/β-catenin signaling on the odontogenic capability of dental epithelium have never been addressed. The regressed incisor enamel organs in our study and the disabled in vivo ameloblast differentiation reported previously suggested that the mesenchymal Wnt/β-catenin signaling impeded not only the morphogenesis of enamel organs but also ameloblast differentiation.^8,17 Because previous studies demonstrated that over-expression of the inhibitors to Wnt or BMP signaling, such as Dkk1 or Noggin, in the dental epithelium disrupted tooth morphogenesis,^6,18 the ectopic activation of Noggin and Ectodin in Osr2-cre^KI; Ctnnb1^ex3f dental epithelium was believed to suppress the morphogenesis and ameloblast differentiation. Since Wnt/β-catenin signaling could activate Bmps or Wnts transcription directly,^6,19 the ectopic Noggin and Ectodin activation in epithelium was most likely a feedback to the overdosed mesenchymal Wnt/β-catenin signaling. Although the inhibition on Shh indeed suppressed tooth development,^20,21 if the eliminated Shh expression in the incisor epithelium resulted from the ectopic Ectodin or Noggin expression remains to be investigated.

An interesting finding in this study was the differential responses to the mesenchymal Wnt/β-catenin signaling between the incisor and molar germs. Our study found that the incisor mesenchyme was more susceptible to Wnt/β-catenin signaling.⁷ This phenomenon might result from the differential expression of Syndecan-1 between the incisor and molar mesenchyme. The eliminated and attenuated Fgf3 expression in the Osr2-cre^KI;Ctnnb1^ex3f incisor and molar mesenchyme was coincided with the explanation. Of course, if there are other factors leading to the difference still required exploration.

Neither the WT E10.5 nor the E13.5 molar epithelium was able to form tooth with the E13.5 Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme. Since the E10.5 molar epithelium usually was regarded to possess the odontogenic potential which could induce non-odontogenic mesenchyme into odontogenic program,²² the active Wnt/β-catenin signaling in the Osr2-cre^KI;Ctnnb1^ex3f molar mesenchyme was supposed to not only disrupt the odontogenic capability in mesenchyme but also resist the induction from the odontogenic epithelium.

Our study explored the tooth development in Osr2-cre^KI;Ctnnb1^ex3f mouse, which characterized the influence on tooth morphogenesis and gene expression. Taking the advantage of the in vivo model, we addressed the real outcomes of the elevated mesenchymal Wnt/β-catenin signaling to the odontogenic capability in both the epithelium and mesenchyme. Compared with the ex vivo or in vitro approaches, our study provided the convincing results that the mesenchymal Wnt/β-catenin signaling not only directly suppressed the ondontogenic gene expression in dental mesenchyme but also disrupted tooth morphogenesis by inducing ectopic expression of Wnt and BMP antagonists.

METHODS

Animals

The Osr2-cre knock-in mouse (Osr2-cre^KI) was generated in Dr. Jiang’s lab by replacing the coding sequence of Osr2 with Cre sequence.⁸ Both Osr2 -cre^KI and the Ctnnb1^ex3f line were obtained from Dr. Yiping Chen’s lab at Tulane University. The genotyping of Osr2-cre^KI andCtnnb1^ex3f lines follows the protocols as described previously.^8,23 All these mice were fed and mated in the Specific Pathogenic Free System of the Institute of Genome Engineered Animal Models for Human Diseases at Dalian Medical University. To generate Osr2-cre^KI;Ctnnb1^ex3f embryos, the Osr2-cre^KI line was crossed with Ctnnb1^ex3f mice and when vaginal plug was detected, the day was recorded as Embryonic Day 0.5 (E0.5). To collect the embryos, the timed pregnant female mice were euthanized by cervical dislocation after carbon dioxide inhalation. All animal work has been conducted according to the ethics guidelines of Dalian Medical University, and the protocol was approved by the Ethics Committee of Dalian Medical University (No. AEE17038).

Histological section and staining

The mouse embryos were dissected in the ice-cold phosphate buffer solution to collect the heads. After fixed in 4% paraformaldehyde overnight, the heads were dehydrated with gradient ethanol and then, embedded with paraffin for 10 μm section. For morphological analyses, the sections were stained with the Azon dichromic staining as previously described.⁷

In situ hybridization

The heads of mouse embryos were harvested in the ice-cold phosphate buffer solution treated with diethyl pyrocarbonate. The 4% paraformaldehyde used for fixation and the gradient alcohol for dehydration were treated with diethyl pyrocarbonate to eliminate RNase. The embryonic heads were embedded in paraffin and sectioned at 10 μm. The antisense RNA probes applied in the hybridization were synthesized as mentioned previously. The hybridization with each probe was repeated for at least three times. After color development, the section was counter-stained with Eosin.

Quantitative polymerase chain reaction

he dental epithelium and mesenchyme were separated from each other as described previously.⁷ The RNA extraction from the separated mandibular incisor and molar epithelium or mesenchyme with RNAiso Plus reagent (TaKaRa, Otsu, Japan), the synthesis of complementary DNA with PrimerScript™ RT reagent (TaKaRa, Otsu, Japan), and the quantitative polymerase chain reaction (PCR) using SYBR® Premix ExTaq™ II (DRR081A, TaKaRa, Otsu, Japan) on a Bio-Rad iQTM5 system (BioRad, Hercules, CA, USA) followed the previous protocols.²⁴

Tissue recombination and subrenal culture

The epithelium from the E10.5 and E13.5 molar germs was separated from the mesenchyme by 30 min digestion in the dispase (2mg/ml) at 37°C. The recombination of different epithelium and mesenchyme was performed with fine forceps. The recombined tooth germs were first cultured in Trowell dishes overnight and then, transplanted to subrenal capsule for 2 weeks culture as previously described.⁷

Funding Statement

This work is supported by National Natural Science Foundation of China (grant no. 81771055 to CL and grant no. 81570962 to JX).

Abbreviations

Bmp	Bone mophogenic protein
Fgf	Fibroblast growth factor
Shh	Sonic hedgehog
Osr2	Odd-skipped related transcription factor 2
Axin 2	Axis inhibition protein 2
Runx2	Runt-related transcription factor 2
Msx1	Msh homeobox transcription factor 1
Pitx2	Paired like homeodomain transcription factor 2
Sfrp2	Secreted frizzled related protein 2
E13.5	Embryonic Day 13.5
DE	Dental epithelium

Disclosure of potential conflicts of interests

No potential conflicts of interests were disclosed.

Acknowledgments

We are grateful to Dr. Yiping Chen (Tulane University, New Orleans, USA) for providing Osr2-cre^KI and Ctnnb1^ex3f mice.

Author contributions

XYC and JL designed and performed experiments. NL, YW and LZ performed experiments. NZ, YDS and YZW interpreted data and commented on the manuscript. JX and CL conceived the study, interpreted data and wrote the manuscript. commented on the manuscript. All authors have read and approved the manuscript.