Continuous tooth generation in mouse is induced by activated epithelial Wnt/β-catenin signaling

Abstract

The single replacement from milk teeth to permanent teeth makes mammalian teeth different from teeth of most nonmammalian vertebrates and other epithelial organs such as hair and feathers, whose continuous replacement has been linked to Wnt signaling. Here we show that mouse tooth buds expressing stabilized β-catenin in epithelium give rise to dozens of teeth. The molar crowns, however, are typically simplified unicusped cones. We demonstrate that the supernumerary teeth develop by a renewal process where new signaling centers, the enamel knots, bud off from the existing dental epithelium. The basic aspects of the unlocked tooth renewal can be reproduced with a computer model on tooth development by increasing the intrinsic level of activator production, supporting the role of β-catenin pathway as an upstream activator of enamel knot formation. These results may implicate Wnt signaling in tooth renewal, a capacity that was all but lost when mammals evolved progressively more complicated tooth shapes.

Results and Discussion

Mice with a conditional β-catenin allele (22) were crossed with K14Cre mice (16) and the resulting K14cre/+;β-catenin^Δex3fl/+ mice, hereafter called β-cat^Δex3K14/+, thus expressed one WT allele of β-catenin and one mutated allele in oral and dental epithelium. Activity of the Cre recombinase was confirmed by crossing the K14cre mice with LacZ reporter mice. Cre recombinase was weakly active in the mandibular ectoderm already before dental placode formation in embryonic day (E) 11-E12 embryos and at E12–15 Cre activity was very strong [supporting information (SI) Fig. 6].

The mutant mice died perinatally, and we therefore focused first on the embryonic stages of tooth development. The first signs of abnormal tooth morphogenesis were detected in histological sections of E13 molar tooth buds. The β-cat^Δex3K14/+ buds were normal in size but had irregular contours (Fig. 1a). At E14 regular morphogenesis into cap stage was not detected. Instead the bud had become more irregular and small epithelial buds had formed at the surface (Fig. 1b). The extra buds grew larger and at E17 when the WT teeth had reached the bell stage of crown development, numerous long epithelial structures protruded into the mesenchyme in the region of molar development (Fig. 1 c and d). The mandibular and maxillary teeth showed similar phenotypes (Fig. 1, SI Fig. 7, and data not shown).

As the first step to resolve the mechanism whereby activated Wnt signaling caused abnormal morphogenesis we determined the localization of Wnt signal activity by crossing the mutant mice with BAT-gal reporter mice, expressing the LacZ gene under the control of β-catenin/Tcf responsive elements (23). In E14 WT mandibles only weak activity was observed in the molar bud epithelium although intense Wnt reporter activity was present, e.g., in tongue papillae. In the β-cat^Δex3K14/+ mandibles the reporter activity was more intense and was localized to discrete spots in the dental and oral epithelium (Fig. 1 e and f). Because high Wnt signaling activity in the dental epithelium has been localized to enamel knots (24, 25), important signaling centers regulating tooth morphogenesis (10), we next analyzed the expression of some known enamel knot marker genes. In WT embryos the first, primary enamel knot appears at the tip of the epithelial bud (Fig. 2a) and mediates the transition to cap stage, and at the bell stage secondary enamel knots determine the positions of the future cusps (26). In the mandibular and maxillary tooth buds of E13 and E14 β-cat^Δex3K14/+ embryos the enamel knot marker genes including Shh, Fgf4, Edar, Wnt10b (11), and Epiprofin (27) all were expressed in a similar pattern in dental epithelium as BAT-gal activity (Fig. 2 b and c, SI Fig. 7, and data not shown), indicating that the spots were supernumerary enamel knots. The number of supernumerary enamel knots increased markedly with advancing development in the mutant tooth germs; however, the intensity of the expression of most enamel knot genes appeared quite normal. Ectodin, a bone morphogenetic protein antagonist with Wnt modulating activity (5, 28, 29), showed a similar inverse expression pattern with the enamel knot markers as in the WT teeth (Fig. 2d). In addition, Fgf3, which is a specific marker for dental papilla identity (30), was strongly expressed below the supernumerary enamel knots in the mesenchyme (Fig. 2e) but it was not detected around the numerous buds outside the tooth region nor in the buds of the tongue. This finding implies that the odontogenic region had retained the competence to form teeth.

We analyzed the patterning of the supernumerary enamel knots in E17 molars by a 3D reconstruction of Shh expression domains, which revealed an increased number of knots throughout the dental epithelium (Fig. 2f). Furthermore, whereas WT tooth shows secondary enamel knots forming within the first molar epithelium surrounding a large dental papilla, the mutant showed multiple small knots, some of which appeared to colocalize with a small mesenchymal papilla, possibly indicating the formation of an individual tooth (Fig. 2f). The perinatal death of the mutant mice, however, left open the fate of the supernumerary knots.

To test whether teeth actually would form in the mutant mice, we transplanted tooth germs dissected from E14 WT and mutant embryos under the kidney capsules of immunodeficient nude mice. After 3 weeks of kidney capsule culture, one WT molar tooth germ gave rise to three molars with normal shape, size, and mineralized tissues similarly as shown earlier in intraocular grafts (31). In contrast, the transplanted molar of a β-cat^Δex3K14/+ embryo had formed a large cyst with a uniform ring pattern on the surface (Fig. 3a and SI Fig. 8). When the cyst was opened, these rings on the surface were observed to be ends of the growing roots. Inside, numerous tooth crowns were present, mostly pointing internally and making the structure resemble a geode (Fig. 3 b–e). We counted 42 teeth with well mineralized crowns (Fig. 3f). The largest individual tooth crowns were close to the same height as the first molars of adult WT mice (0.70/0.85 mm) and roughly equal in height to the second WT molars. However, whereas some teeth were large and multicusped, the majority (86%) were unicusped and conical in shape. We confirmed this by using histological sections, which revealed that the more advanced teeth, both unicusped and multicusped, also had normal looking roots indicative of molar tooth identity (Fig. 3 c–e). The paucity of multicusped mutant molars may be explained by the small size of the supernumerary enamel knots (Fig. 2h), as in mouse molar the cusp number correlates with the size of the primary enamel knot (4, 32). When two molar buds were transplanted together under the kidney capsule, >80 teeth were collected after 3 weeks. When the kidney capsule culture was repeated with incisor tooth germs of E14 mutant embryos, multiple incisor-type teeth characterized by sharp and elongated crowns formed. Especially in incisor experiments, often one large incisor and multiple smaller-sized incisors were detected (SI Fig. 9).

It was not possible to conclude from the morphology and histology of the teeth formed in the kidney transplants whether they had formed successionally, or whether the teeth had been initiated by subdivision of the dental epithelium at an early stage. To analyze the dynamics of supernumerary tooth formation we followed the development of dissected E14 and E15 molar tooth buds of β-cat^Δex3K14/+ mutant embryos and their WT littermates in organ cultures. Daily observation under the dissecting microscope revealed that the mutant tooth bud gave rise to several conical-looking teeth with enamel and dentin matrices, which formed successionally during 11 days of culture. The first forming tooth was usually clearly larger than the successively forming teeth. The WT tooth bud developed to a typical multicusped first molar followed by a smaller and multicusped second molar (Fig. 4a–e). The developmental capacity of the in vitro-formed successional tooth buds of β-cat^Δex3K14/+ embryos was tested by subculturing two tooth buds dissected from a mutant explant after 6 days of culture. During 19 days of continued culture, five additional teeth developed successionally with a concomitant increase in the tissue mass (Fig. 4g). The histological sections of the explants revealed small tooth buds interspersed between forming teeth (Fig. 4f), indicating, together with the in vitro results, that the supernumerary teeth developed by a continuing renewal process where new teeth bud off from the existing dental epithelium.

The restricted and high Wnt reporter activity detected in the supernumerary enamel knots in the mutants indicates a direct role for Wnt signaling in the formation of supernumerary enamel knots. However, despite the β-catenin stabilization throughout the mutant epithelium, only enamel knots and the ectodermal buds showed Wnt reporter activity. This finding suggests that besides Wnt signals additional components are required for the activation of the Wnt pathway and for the formation of enamel knots. Gat et al. (15) showed that stabilized β-catenin could activate Wnt signaling in hairs only after induction of localized Lef1 expression. Interestingly, these mice survived to adulthood and no tooth phenotype was reported. Compared with our mutants, the unaffected tooth development may be caused by the different strategy of introducing the transgene where strongly expressing founder mouse lines may have been lost by prenatal death. Until now, mice with fully formed supernumerary teeth have been limited to cases with one extra molariform or incisiform tooth adjacent to molars or incisors, respectively. Developmentally, these individual extra teeth result from enlarged enamel knots (5, 4) and, at the level of signaling, at least from changes in the induction and inhibition of enamel knots by the bone morphogenetic protein and FGF pathways (3, 5).

The dynamics of tooth development, including basic enamel knot induction and inhibition, has been reproduced by computer model simulations (33). Based on the dynamics of activator–inhibitor systems, an increase in the production of activator and a consequent induction of enamel knots might, in principle, be expected to produce supernumerary teeth. The extensive modification of tooth development in β-cat^Δex3K14/+ mice, however, raises questions about whether a single developmental factor, in this case β-catenin stabilization, could explain directly all of the observed changes in patterning or whether multiple, indirect effects are likely to be involved. To test the underlying principle of β-catenin function using computer simulations, we used the published computer model (33) to simulate WT and β-cat^Δex3K14/+ molar development. The mathematical model is a morphodynamic derivative of activator–inhibitor models where, in addition to affecting the production of each other, the activator and inhibitor molecules also affect cell division and differentiation (34). Activator–inhibitor models have been previously used successfully to simulated patterns of multiple organs, including teeth and feathers (34), but they have not yet been used in connection to extensively altered mutant morphologies. Because all of the epithelial cells of the β-cat^Δex3K14/+ mice have stabilized β-catenin we increased in the model the intrinsic production rate of activator (parameter k₃), most closely matching the hypothetical effect of β-catenin stabilization in vivo.

Whereas continuing increase in activator levels will ultimately result in uniform differentiation, the simulations show that at intermediate k₃ levels activator-induced inhibitor is able to prevent knot formation laterally and an iterative knot formation dynamics is obtained (Fig. 5). Furthermore, the results show that initially large enamel knots are followed by numerous small knots (Fig. 5), which suggests that the observed β-cat^Δex3K14/+ tooth formation patterns, where initially few larger teeth are followed by multiple smaller teeth, could be a direct result of elevated levels of intrinsic enamel knot activator production.

Additionally, agreeing with the computer simulations, we did not detect clear directionality in the formation of the successional teeth. In mammals, the dentition is replaced by budding of new teeth from the lingual aspect of the previous, primary, or milk teeth. On the other hand, molars develop successionally to the back of the developing tooth row and are initiated from the posterior end of the last molar (8). Both of these modes are more open-ended in most other vertebrates where dentition is replaced continuously and successional teeth are added posteriorly as the jaws grow larger (8). Our experimental and simulation results indicate that shared molecular mechanisms, in particular Wnt signaling, could, in principle, be involved in these different modes of tooth renewal. Furthermore, the possibility that the two modes of successional tooth formation, i.e., tooth replacement and addition along the tooth row, involve a similar molecular mechanism is supported by the phenotype of the human syndrome cleidocranial dysplasia (CCD) characterized by multiple supernumerary teeth, which form both as a partial third dentition and as additional posterior molars (35). CCD is caused by mutations in the Runx2 gene (35, 36), and it remains to be tested whether activated Wnt signaling may underlie tooth renewal in this human syndrome.

The composition of the geode-like outgrowths that developed from the transplanted mutant tooth buds was reminiscent of that in human compound odontomas, where multiple small teeth, sometimes >100 teeth, are found within the jaws (37). Because in some cases odontomas are caused by activated Wnt signaling due to loss-of-function mutations in the Wnt signal modulator adenomatous polyposis coli (18, 19) it is possible that supernumerary teeth are formed by the same mechanism in odontomas and our mutants. Furthermore, the odontomas are considered to be congenital anomalies rather than tumors, and it is possible that the teeth in odontomas also form successionally as we showed in our mutants, thus the mechanism of their pathogenesis involves unlocking the potential of tooth renewal.

In conclusion, we have shown that the stabilization of β-catenin in the dental epithelium leads to continuous tooth generation and that the mechanism involves the iterative formation of ectopic enamel knot signaling centers where enamel knot activation and lateral inhibition is the underlying key mechanism. Intriguingly, a trend in mammalian evolution has been a reduction of number and renewal of teeth, concomitant with the evolution of progressively more complex, multicusped teeth (8, 9). It remains to be tested whether Wnt signaling may have played a role in the loss of tooth renewal and increase in crown complexity during evolution. Taken together, our results may have implications for organ regeneration and bioengineering of teeth and the understanding of the genetic basis of the evolution of teeth.

Materials and Methods

Supplementary Material

Acknowledgments

We thank Stefano Piccolo for the BAT-gal mice; Natalia Soshnikova (Max-Delbrueck-Center for Molecular Medicine, Berlin, Germany) for the first samples of β-cat^Δex3K14/+ mice; and Riikka Santalahti, Merja Mäkinen, and Heidi Kettunen for excellent technical help. This work was supported by the Academy of Finland (I.T. and J.J.) and the Sigrid Juselius Foundation (I.T.).

Abbreviation

En

embryonic day n.

Footnotes

References

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.