Many reptiles are able to continuously replace their teeth through life, an ability attributed to the existence of epithelial stem cells. Tooth replacement occurs in a spatially and temporally regulated manner, suggesting the involvement of diffusible factors, potentially over long distances. Here, we locally disrupted tooth replacement in the leopard gecko (Eublepharis macularius) and followed the recovery of the dentition. We looked at the effects on local patterning and functionally tested whether putative epithelial stem cells can give rise to multiple cell types in the enamel organs of new teeth. Second generation teeth with enamel and dentine were removed from adult geckos. The dental lamina was either left intact or disrupted in order to interfere with local patterning cues. The dentition began to reform by 1 month and was nearly recovered by 2–3 months as shown in μCT scans and eruption of teeth labeled with fluorescent markers. Microscopic analysis showed that the dental lamina was fully healed by 1 month. The deepest parts of the dental lamina retained odontogenic identity as shown by PITX2 staining. A pulse-chase was carried out to label cells that were stimulated to enter the cell cycle and then would carry BrdU forward into subsequent tooth generations. Initially we labeled 70–78% of PCNA cells with BrdU. After a 1-month chase, the percentage of BrdU + PCNA labeled cells in the dental lamina had dropped to 10%, consistent with the dilution of the label. There was also a population of single, BrdU-labeled cells present up to 2 months post surgery. These BrdU-labeled cells were almost entirely located in the dental lamina and were the likely progenitor/stem cells because they had not entered the cell cycle. In contrast fragmented BrdU was seen in the PCNA-positive, proliferating enamel organs. Homeostasis and recovery of the gecko dentition was therefore mediated by a stable population of epithelial stem cells in the dental lamina.

Investigating the molecular basis for tooth shape variation provides an important glimpse into the evolution of tooth function. We recently showed that loss of mesenchymal BMP7 is sufficient to alter morphology and function of the toothrow. Here we report on the underlying mechanism. Expression of mesenchymal Bmp7 is observed at sites where mineralization is initiated, in tooth cusps of developing molars. Neural crest-specific deletion of Bmp7 (Bmp7^ncko) resulted in a complete lack of dentin/enamel formation at birth, the time when mineralization is normally initiated in the upper molars, similar to what was observed in Bmp2^ncko mice. Unlike loss of Bmp2, loss of Bmp7 did not affect odontoblast polarization and did not significantly alter the levels of pSmad1/5/8, but almost completely abolished canonical Wnt signaling in (pre)-ameloblasts. Tooth mineralization resumed with a 48-h delay allowing for additional mesenchymal proliferation. Enamel volume was still reduced at P4 and P8, but was comparable in erupted teeth, which were broader and had altered cusp shapes. Tooth eruption was also delayed. Overall, enamel appeared inconspicuous, although some structural changes along with reduced mineral density could be observed. Loss of Bmp7 led to an increase in mesenchymal Bmp6 suggesting an interplay between Bmp6 and Bmp7 in the regulation of mineralization initiation. Our findings show that regulation of the onset of tooth mineralization is a hitherto unsuspected mechanism controlling tooth shape variation. Initiation of tooth mineralization is regulated by a complex epithelial-mesenchymal Bmp/Wnt-signaling network to which Bmp7 contributes. This network is separate and independent of the Bmp2-signaling network regulating odontoblast cell polarization. From an evolutionary perspective, addition of Bmp7 as initiator of tooth mineralization might be akin to an upgrade of an existing computer operating system. While not essential, it provides obviously sufficient advantage warranting its evolutionary incorporation.