Wnt signaling from Gli1-expressing apical stem/progenitor cells is essential for the coordination of tooth root development

Summary Stem cell regulation plays a crucial role during development and homeostasis. Here, an essential source of Wnts from Gli1+ stem/progenitor cells was identified in the murine molar. Loss of Wnt production in Gli1+ apical stem/progenitor cells led to loss of Axin2 at the root apex, mis-regulation of SOX9, loss of BMP and Hh signaling, and truncation of root development. In the absence of Wnt signals, the root epithelium lost its integrity and epithelial identity. This phenotype could be partially mimicked by loss of Sox9 in the Gli1 population. Stabilization of Wnt signaling in the apical papilla led to rapid unordered differentiation of hard tissues and fragmentation of the epithelial root sheath. Wnt signaling from Gli1+ stem/progenitor cells, therefore, orchestrates root development, coordinating mesenchymal and epithelial interactions via SOX9 to regulate stem/progenitor cell expansion and differentiation. Our results demonstrate that disparate stem/progenitor cell populations are unified in their fundamental signaling interactions.


INTRODUCTION
Wnt signaling is essential for stem cell self-renewal, proliferation, and differentiation (Clevers et al., 2014). This has been demonstrated in diverse tissues and organs such as bone (Liu et al., 2009), kidney (Barker et al., 2012), gut (Farin et al., 2016), and mammary gland (Plaks et al., 2013). In the gut, Wnts are key to maintaining the stem/ progenitor cells at the base of the crypts. Wnt responding cells form all epithelial cell lineages found in the gut (Valenta et al., 2016). More recently, Gli1-expressing mesenchymal stem cells have been shown to secrete Wnt ligands crucial for the maintenance of the epithelial stem cell niche in the colon (Degirmenci et al., 2018). Gli1 is a stem cell marker in many tissues including bone (Shi et al., 2017), kidney (Kramann et al., 2016), and gut (Valenta et al., 2016). Interestingly, Wnts and Gli1 play important roles in tooth development. Wnts play diverse roles in specification and differentiation of odontogenic cells in early stages of dental organogenesis (Sarkar and Sharpe, 1999;Jussila and Thesleff, 2012). At later stages, Wnts are essential for root development. Both canonical (Lohi et al., 2009;Kim et al., 2013;Zhang et al., 2013;Bae et al., 2015) and non-canonical pathways (Ma et al., 2020) have been shown to be active during root development. In particular, Wnts are essential for the differentiation of odontoblasts forming root dentin . Gli1 is an essential stem cell marker in the developing molar and can give rise to all lineages forming the postnatal molar root .
Like many organ systems, teeth develop through a series of carefully orchestrated epithelial-mesenchymal interactions. Mice present two distinct dental developmental models: continuously growing non-rooted incisors and rooted molars. The mouse proximal incisors house a persistent stem cell niche that supports continuous odontogenesis, thus providing a good model system to study developmental mechanisms during homeostasis. The mouse molars, on the other hand, develop roots during early postnatal life with the organization of a transient stem/progenitor cell cluster at the base of the crown known as SCAPs (stem cells from the apical papilla). SCAPs are lost at the end of the first postnatal month, by which time the tooth has erupted into the oral cavity and root development is complete ( Figure 1A). The apical papilla is enclosed on both sides by epithelial cells that create a structure called Hertwig's epithelial root sheath (HERS). HERS is bilayered and comprised of closely packed epithelial cells that are formed by apical proliferation in the cervical loop of the developing tooth germ (Zeichner-David et al., 2003;Huang et al., 2009a). HERS plays a crucial role in guiding root development and forms a physical boundary between the root-forming cells and the periodontium (Huang et al., 2009a;Li et al., 2017;Fons Romero et al., 2017). As roots near completion, most epithelial cells undergo apoptosis or epithelial-to-mesenchymal transition to form cementoblasts (Huang et al., 2009a). A small subset of these epithelial cells have been shown to consolidate and persist as epithelial cell rests of Malassez (ERMs), which may function as local stem cell niches within the periodontal ligament (Oka et al., 2012).
Given the relationship between Wnts and Gli1 in the gut, the relationship between the Gli1-expressing cells and Wnt signaling was investigated during root development to explore their role in maintaining the molar root stem cell niche. As Wnts are important in gut and tooth development, the possible conserved relationships between Wnt action in the gut and tooth root was explored.
Here, we highlight the key role of the Wnt signaling pathway in the maintenance of the dental apical stem/progenitor population and in control of the integrity of the surrounding epithelial sheath. The findings demonstrate the conserved use of Gli1, Sox9, and Wnts to regulate stem/progenitor cell niches in both the tooth and gut, two distinct and highly specialized organ systems.

RESULTS
Wnt signaling is active in apical root stem/progenitor cells To chart canonical Wnt activity during root development, Axin2LacZ reporter mice and inducible Axin2creERT2; R26mTmG mice were investigated. Axin2 is a direct intracellular target of Wnt; therefore, Axin2 reporter mice have been shown to reliably label Wnt responding cells (Lustig et al., 2002). At root initiation, Axin2 + cells were located at the base of the crown ( Figure 1B). During root elongation, these cells were localized to the apex of the molar root, at the advancing developmental front (Figures 1C and 1F). Wnt responding cells were found in both the apical HERS and adjacent mesenchymal cells, confirmed by the presence of activated b-CATENIN ( Figure 1D).
The apex of the developing tooth root has previously been shown to house stem/progenitor cells that form the tooth root and its surrounding supporting structures (Sonoyama et al., 2008;Liu et al., 2015). Given the expression of Axin2 in this region, the fate of the apical Wnt-responsive cells was followed using Axin2-creERT2;R26mTmG (Axin2Lin) mice during the stages of root elongation, when radicular hard and soft tissues are rapidly developing. Tamoxifen was administered to Axin2-Lin mice at P12 (postnatal day 12) and was traced for 10 days (until P22) ( Figure 1E). The majority of root-forming cells giving rise to the dental pulp at the core, as well as dentinogenic odontoblasts at the periphery, were found to be derivatives of the initial apical Wnt-responsive stem/ progenitor cell population, along with a subset of cells in the surrounding periodontium ( Figures 1F and 1G).
The contribution of the Axin2 lineage was compared in the incisors of P12 tamoxifen-injected Axin2Lin pups. The mouse incisor lacks a well-defined root and has a crown  Figure S1. and root analog-the labial and lingual cervical loops, respectively, which, along with the proximal mesenchyme, house the stem cells that support its continuous growth (Renvoisé and Michon, 2014). At the beginning of the chase, Axin2-expressing Wnt responding cells were sparsely interspersed in mesenchymal tissue of the proximal pulp with a greater density adjacent to the lingual cervical loop. After 10 days of tamoxifen injection, Axin2 lineage cells formed the bulk of the proximal pulp, including dentinogenic odontoblasts, and surrounding periodontal tissues ( Figure S1). In contrast to molars, Axin2-expressing cells and their progeny had no significant contribution to the epithelial cervical loops ( Figure S1C).
Wnts produced by Gli1 + stem/progenitor cells are necessary for root development Having charted Wnt activity, the source of the Wnt ligands in developing roots was traced. In the developing root, Gli1 broadly labels a proliferative stem/progenitor population that gives rise to diverse cell lineages forming roots and periodontium ( Figure S2). The Axin2-expressing stem/progenitor cells formed a subset of the broader Gli1 population at the advancing front of developing roots (compare Figures 1C and S2A 0 ). Given this close association, it was hypothesized that the Gli1 population could be a crucial source of Wnt ligands. Analysis of a publicly available single-cell RNA sequencing (scRNA-seq) dataset generated from the murine incisor mesenchyme highlighted several Wnt ligands expressed by the Gli1 dental population (Figures 2A and 2B) (Krivanek et al., 2020). Of particular note, Wnt10a and Wnt9a (both reported to act as canonical Wnts) were expressed in the apical pulp, along with Wnt5b (a non-canonical Wnt), while Wnt6 was more closely associated with the odontoblast lineage, and Wnt4 was expressed in the Gli1 À distal pulp ( Figures 2C-2G). To test the effect of loss of Wnts during root development, a loss-of-function transgenic murine model was utilized to specifically target Wntless production in Gli1-expressing cells (Gli1 creERT2/+ ;Wls tm1.1Lan , hereafter referred to as Gli1creERT2;Wls fl/fl ). Abrogation of Wntless affects both canonical and non-canonical pathways (Bänziger et al., 2006) and would be predicted to lead to loss of WNT10a, -9a, -6, and -5b. Confirming the loss of canonical Wnt activity in the root, Axin2 expression was specifically lost in the apical region ( Figures S3A and S3B). Interestingly, Axin2 was still robustly expressed in the periodontium, suggesting that there is another source of Wnt ligands in this region of the tooth ( Figure S3C). In keeping with loss of Wnts from Gli1 + cells, the mutant mice also showed gut defects, mimicking previously reported findings (Degirmenci et al., 2018) (Figures 2J and 2K). This deletion arrested root formation from the point of tamoxifen injection ( Figures 2H and 2I). Loss of Wnt signaling at the initiation of root formation resulted in rootless teeth (Figures 2LÀ2O). In keeping with the presence of Axin2, the surrounding periodontal ligament (PDL) still formed, with fibers surrounding the unerupted, rootless molar ( Figure 2O). Deletion at the stage of root elongation caused a significant reduction in tooth root length (Figure 2Vi) with wide, immature apices and defective cementum deposition (Figures 2PÀ2S). These short-rooted teeth erupted into the oral cavity and reached occlusion, in contrast to rootless third molars, which failed to erupt (Figures S3F and S3G). In situ staining for Dspp, a marker of odontoblasts, showed well-differentiated dentin-producing cells ( Figures 2T and 2U). Our mutation, therefore, appears to specifically target radicular progenitors causing truncated root formation from the time the deletion is induced.
The impact of the deletion was also evident in the calvarium in regions adjacent to the cranial sutures ( Figures S3D and S3E), indicating that Gli1 progeny originating from cranial sutures  also produce essential Wnts. By P30, the alveolar bone surrounding the tooth was brittle and irregularly deposited compared with littermate controls ( Figures 2R and 2S). Inactivation of Wntless in developing bone has previously been shown to increase osteoclast activity (Zhong et al., 2012). Bone defects might, therefore, impact root formation. To investigate this, roots were analyzed at P18 after induction at P13, before any overt bone defects. In these mice, the roots were still shorter, and no difference in osteoclast activity was evident (Figures S4AÀS4H). The short root phenotype therefore predates any later bone defects.
In keeping with the formation of truncated roots, loss of apical Wnts produced by Gli1 + cells resulted in an almost complete loss of proliferation of apical cells in the root mesenchyme and epithelium ( Figures S5AÀS5D 0 ), while the surrounding forming periodontium appeared to be largely unaffected (Figures S5B 0 and S5D 0 ).
Agreeing with the role of Wnts in the molar root, mutants exhibited overall shorter incisors (Figures 2Vii-2Viii, 2W, and 2X; Figures S3H and S3I), with a relatively greater disruption to the lingual cervical loop, which is known as the root analog ( Figure 2Viv).

Loss of Wnt signaling disrupts signaling networks in the apical papilla
The BMP signaling cascade is responsible for regulation of cell proliferation and differentiation in many tissues, including skeletal hard tissues (Beederman et al., 2013). In the tooth root, activation of BMP signaling is required for the differentiation of dentin-forming odontoblasts (Feng et al., 2017). During normal root development, pSMAD1,5,9, a readout of BMP signaling, was observed in the apical HERS and was widespread in radicular mesenchymal tissue, barring the apical papilla ( Figure 3A). Loss of apical Wnts produced by Gli1 + cells resulted in a reduction in mesenchymal BMP activity and a complete loss within HERS ( Figure 3B). Furthermore, members of the Shh signaling pathway were shown to be downregulated in the developing roots of mutant Gli1creERT2;Wls fl/fl molars ( Figures 3C-3J), and Gli1 expression was reduced in the mutant teeth ( Figures 3C-3F). Reduced apical Gli1 expression suggests potential autoregulation of these cells by the Wnt ligands they produce. Another target of the Shh pathway important for molar root development, Ptch1 (Nakatomi et al., 2006), was also downregulated in mutant teeth ( Figures 3G and 3H). Shh expression, which is normally found in the apical HERS (Nakatomi et al., 2006) and is crucial for maintaining epithelial-mesenchymal interaction at the root apex (Huang et al., 2009b), was also lost in mutant teeth ( Figures 3I and 3J). These findings suggest that BMP and Shh signaling pathways are downstream targets under the control of Wnt signaling in apical stem/progenitor cells, although whether they are direct or indirect targets is unclear, as they may be consequences of the loss of specific populations.
Loss of apical Wnt ligands results in disruption to HERS HERS shapes and guides root development (Fons Romero et al., 2017), and manipulating Wnt activity in HERS has an impact on the integrity of this important epithelial layer (Yang et al., 2021). The structure and gene expression of HERS was, therefore, investigated in the Gli1-creERT2;Wls fl/fl mice. Confirming the key role of Wnts in HERS, abrogation of Wntless resulted in a breakdown of epithelial integrity, with the formation of aberrant cellular projections evident a week after tamoxifen induction (Figures 4AÀ4B 0 ). These projections were linked to a defect in the basement membrane surrounding HERS, demonstrated by loss of LAMININ staining (Figures S6AÀS6D 0 ). By P30, control root formation was complete, and most epithelial cells were lost, barring remnants that form the ERMs (Figures 4C and 4D). In Gli1creERT2;Wls fl/fl mutants, however, numerous persistent apical epithelial cells were found within the periodontal stroma ( Figures 4E and 4F). The change in HERS was accompanied by loss of E-CADHERIN expression, while the expression of pan-CYOKERATIN was maintained (Figures 4GÀ4H 00 ). This suggests a convergent role of Wnts, potentially acting through b-catenin, a known component of the cadherin complex (Nelson and Nusse, 2004), thereby orchestrating cellular signaling while maintaining epithelial integrity.

Loss of apical Wnt signaling leads to misexpression of Sox9
Sox9 is a known marker of progenitor cells in numerous tissues, including hard tissues (Blache et al., 2004). In the gut, Sox9-expressing progenitor cells are essential for development and homeostasis and are regulated by Wnt ligands of epithelial and mesenchymal origin (Blache et al., 2004;Valenta et al., 2016). In teeth, Sox9 labels pro-genitor cells in the enamel organ of the embryonic tooth germ (Kawasaki et al., 2015). In the continuously developing incisor, SOX9 expression is observed in the proximal epithelial and mesenchymal cells ( Figures 5A and  5D), confirming previously reported homeostatic dental expression (Krivanek et al., 2020). Interestingly, a zone of mesenchyme devoid of SOX9 is located adjacent to the lingual cervical loop (root analog), while the juxtaepithelial mesenchymal cells of the labial cervical loop (crown analog) express SOX9 (Figures 5A and 5D). Much like the lingual cervical loop of the incisor, the molar root also exhibits a narrow zone of SOX9-devoid mesenchymal cells adjacent to HERS, sandwiched between the SOX9 + epithelium and a broader SOX9 + mesenchymal domain (Figures 5B and 5E). Interestingly, this SOX9 À mesenchymal domain expresses Axin2, suggesting Figure 5. Loss of apical Wnts disregulates SOX9 expression in the developing molar (A) In the mouse incisor, the juxta-epithelial mesenchymal cells of the labial cervical loop (crown analog) express SOX9 (pink), while (D) the lingual cervical loop (root analog) and molar (B and E, magnified) exhibit a narrow zone (yellow asterisk) of SOX9-devoid mesenchyme adjacent to the SOX9 + apical HERS.
(C-C 00 ) Axin2creERT2:R26mTmG reporter mice 1 day after Tmx injection (see schedule, Figure 1E) showing Axin2 expression in green (GFP) (P13) (C and C 0 ). Expression of Axin2 is found in the apical region sandwiched between the SOX9-expressing mesenchyme and epithelium in pink (C and C%). (F and H) Wls deletion in Gli1 + cells resulted in an overall downregulation of the broad mesenchymal SOX9-expressing domain (F and H, magnified) and upregulation of SOX9 in the normally SOX9-devoid apical zone (blue asterisk). (G and I) Downregulation of SOX9 was noted within the fragmenting apical epithelium (yellow arrowhead) with concomitant loss of cell adhesion labeled with low P-CADHERIN (green) expression (G and I, magnified). n R 3 tissue sections were examined from n R 3 mice from mutant and control samples. Scale bars in (A)-(I): 100 mm. Same scale in (C) as (C 0 ) and (C 00 ). See also Figures S6EÀS6H.

compartmentalization of the apical mesenchyme into
Axin2 and SOX9 domains (Figures 5CÀ5C 00 ). Loss of Wnt ligand production in Gli1-expressing cells, and loss of the Axin2 domain, resulted in expansion of SOX9 expression in the mesenchyme up to HERS (Figures 5F and 5H), suggesting that Axin2 normally restricts the expression of SOX9 in the molar. In contrast to the expansion of SOX9 in the apical mesenchyme, SOX9 expression was lost within the apical epithelium of HERS ( Figures 5G  and 5I). Loss of Wnt signaling thus leads to loss of E-CADHERIN and SOX9 in the root epithelium (Figures S6EÀS6H).
Sox9 expression in Gli1 + stem/progenitor cells is essential for molar root development Having observed a downregulation of SOX9 expression in the epithelium of Gli1creERT2;Wls fl/fl mutant roots, we then asked if apical Sox9 expression is necessary for root formation. To investigate this, Gli1 creERT2/+ ; Sox9 tm2Crm (hereafter referred to as Gli1creERT2;Sox9 fl/fl ) transgenic mice were generated. Loss of SOX9 in the forming roots was confirmed by immunofluorescence ( Figures S7B and  S7C), with the mutant mice displaying severe hair defects, as would be expected given the essential role of Sox9 in the hair stem cell compartment ( Figure S7A) (Vidal et al., 2005). To capture early effects on root formation while maintaining vitality of the experimental model, the mutation was induced during early postnatal life at P7 ( Figure 6A). The molar roots of the Gli1creERT2;Sox9 fl/fl mice were shorter than littermate controls ( Figures 6B-6F (Figures 6GÀ6H 00 ). Overall, this indicates that Sox9 is an essential transcription factor for molar root formation, playing an important role in HERS stability, although its abrogation produced a milder phenotype compared with the dramatic rootless teeth produced by loss of Wnt ligand production in Gli1 + stem/progenitor cells.
Overexpression of Wnts in the developing root apex results in rapid differentiation of odontogenic precursor cells Given the importance of Wnt for root development, the effect of Wnt overactivation during root formation was explored. In the gut, ectopic Wnt secretion disrupts crypt epithelial stem cells and results in increased instances of polyposis and neoplasia (Harada et al., 1999). To test the effect of Wnt overactivity in the developing root, a transgenic mouse line that conditionally expressed an activated form of Ctnnb1 in Axin2-expressing cells, Axin2 creERT2/+ ;R26 mTmG/+ ; Ctnnb1 lox(ex3)/+ , was generated. To target stages of active root formation, a Cre-driven gain-offunction mutation was conditionally induced by administering tamoxifen at P15. Tissues were harvested at P28 for phenotypic assessment and at P20 to investigate the cellular and molecular changes leading to the phenotypic changes ( Figures 7A and 7G). In the developing root, stimulation of canonical Wnt signaling resulted in rapid differentiation of dental hard tissue-forming cells, both odontoblasts and cementoblasts ( Figures 7B-7E and 7HÀ7I 0 ). The resultant root surface was irregular with thickened apical margins ( Figures 7C and 7E). The odontoblasts (especially in the apex) appeared to have lost their polarization, were more numerous, and laid down irregular masses of dentin, indicating rapid disorderly deposition ( Figure 7I 0 ). Isolated masses of dentinoid were observed within the pulp, suggestive of premature odontogenic differentiation of precursor/ progenitors found in the apical dental pulp (Figure 7I 0 ). The surrounding cementum was irregular and very cellular, reflective of the rapid rate of hard tissue deposition with increased cellular entrapment. Similar effects have been reported in murine gain-of-function models with targeted constitutive activation of b-CATENIN in odontoblasts (Wang et al., 2021). Interestingly, the apical HERS lost its  (legend continued on next page) bilayer architecture (Figures 7J and 7K). This phenotype was a direct consequence of the genetic overactivation of the Wnt pathway, evidenced by the distribution of Axin2 lineage cells specifically in the newly formed root following induction of the mutation (Figure 7L 0 ).

DISCUSSION
Development and homeostasis utilize a conserved set of biochemical signaling pathways to activate intracellular transcriptional networks that guide cell proliferation, specification, and quiescence. Here, we examined the role of the Wnt signaling pathway in postnatal tooth development and found similarities in the way these developmental pathways and their intracellular transcriptional targets are used across two morphologically, functionally, and developmentally distinct organ systems-teeth and gut.
Stem/progenitor cells at the base of the developing molar root were shown to be Wnt-responsive Axin2-expressing epithelial and mesenchymal cells that gave rise to all dental hard and soft tissue-forming cells. Axin2, therefore, could be a marker for the root apical stem/progenitor population. This parallels the gut, where Axin2-labeled stem cells at the base of the epithelial crypts form all cell lineages of the intestinal crypts and villi (Flanagan et al., 2018;Harnack et al., 2019). Unlike the intestine, postnatal molar roots are primarily mesenchymal derivatives. Here, the epithelium guides root form (Fons Romero et al., 2017) and maintains intercellular signaling crosstalk with the underlying mesenchymal cells during development but then largely disappears (Huang et al., 2009a;Zeichner-David et al., 2003;Li et al., 2017). Similar to molars, Axin2 + cells were found in the mesenchyme of the proximal incisors, which form the putative stem/progenitor cell compartment of this homeostatic model; however, unlike molars, the proximal epithelium was largely devoid of Axin2-expressing cells and their progeny.
Wnt ligands are short-range glycoproteins (Farin et al., 2016); hence, the sources of these ligands are usually positioned close to responding cells. This has been demonstrated in the murine colon (Valenta et al., 2016;Degirmenci et al., 2018). By abrogating Wnt ligand production specifically in the apical Gli1 + stem/progenitor cells, we demonstrated that Wnt signaling is essential for postnatal dental development. The effect on the developing molar roots was stark, with cessation of root development upon abrogation of Wntless production. This is likely to involve both canonical and non-canonical Wnt activity, with both canonical and non-canonical Wnt ligands shown to be produced by the Gli1 population. In this article, we have focused on Axin2 as a canonical readout, but it would be interesting to explore the impact on non-canonical targets, which may be linked to the changes in the integrity of HERS. Of the Wnt ligands identified in the Gli1 population, loss of Wnt10a is known to lead to root defects, with the formation of non-bifurcating taurodont teeth in mice and patients . The root cessation phenotype described here is more severe, representing the accumulated of loss of all Wnt ligands in the Gli1 population. The relative roles of the other Wnts, such as Wnt9a, will be an interesting area for exploration in the future.
Since Gli1 is a broad stem/progenitor cell marker in the developing root, abrogation of Wnt ligand production was achieved in both epithelial and mesenchymal apical cells. Therefore, it is unclear whether Wnts produced by epithelial or mesenchymal cells are independently sufficient for root development. The gut exhibits regional variability in its dependence on epithelial vs. mesenchymal Wnt ligands. Mesenchymal Wnts are indispensable in the colon, where their loss results in complete disintegration of the epithelium, while the duodenum is spared unless both sources are lost (Degirmenci et al., 2018;Valenta et al., 2016). Future investigations to establish the effect of epithelial and mesenchymal Wnt ligands on postnatal dental development will be useful. In the gut, Sox9-expressing progenitor cells are essential for development and homeostasis and are regulated by Wnt ligands (Blache et al., 2004;Valenta et al., 2016). Here, we show a similar essential role for Sox9 in the developing tooth and a reliance of Wnt signaling for its epithelial expression. In the apical mesenchyme of the molar root, SOX9 was excluded from the Axin2 expression domain. A similar absence of SOX9 in the Axin2 domain was observed in the lingual loop of the incisor. In keeping with this, loss of Axin2 in Gli1creERT2;Wls fl/fl mutants led to an expansion of SOX9 into this juxta-epithelial (H-I 0 ) Hematoxylin and eosin-stained histology sections show exuberant and irregular dentin (d) and cellular cementum (c) deposition in the mutant root (I and I 0 ) compared with control roots (H and H 0 ). Isolated dentinoid islands (black arrowhead) interspersed within the pulp of mutant root (I 0 ). (J and K) Cytokeratin (pink) expression in the mutant root shows prolific apical epithelium with loss of bilayered architecture (K), otherwise seen in the control root (J). (L and L 0 ) GFP (green) immuno-expression labels Axin2 lineage cells seen in abundance at the apex of the irregularly developing mutant root. n = 3 mice with mutant (Axin2 creERT2/+ ;R26 mTmG/+ ; Ctnnb1 lox(ex3)/+ ) and control (non-Cre littermates) genotypes. (F) Unpaired t-test. P < 0.0007. Scale bars in (H)-(L): 100 mm. mesenchymal domain. Regulation of SOX9 in the molar epithelium and mesenchyme is, therefore, distinct with potential opposing roles of Wnt signaling.
In developing molar roots, the Gli1creERT2;Wls fl/fl loss-of-function mutation selectively hampered radicular progenitors, while the periodontal progenitors were spared. This is particularly interesting since both these progenitor cell populations arise from the same apical Gli1-expressing stem cell domain (Men et al., 2020). Our findings indicate a reliance of the periodontium on Wnt ligands from alternate sources, highlighted by the persistence of Axin2 expression and proliferation in this region. Wnt ligands may therefore originate from part of the larger Thy1-expressing progenitor cell population (Zhao et al., 2021).
Differential regulation of dental progenitor cells was also observed between developing molars and homeostatic incisors and between the labial and lingual cervical loops within the proximal incisor when Wnt ligand production by Gli1-expressing cells was lost. The different outcomes observed on the labial and lingual side of the mutant incisors were interesting since Gli1 labels proximal epithelial and mesenchymal stem cells across both cervical loops of the murine incisor. These findings highlight the molecular similarities between the molar root and the incisor lingual cervical loop, supporting the classification of the lingual loop as a root analog.
Similar to the gut and ectodermal organs such hair and salivary glands, tooth development is guided by interactions between the key developmental signaling pathways. Both gut and teeth have Wnts at the helm, maintaining the stem/progenitor cell pool and interacting with SHH, transforming growth factor b (TGF-b), and BMP signaling pathways, guiding proliferation and differentiation. Here, we demonstrated how derangement in Wnt signaling led to lowered SHH and BMP signaling, although it is unclear whether this is through a direct or indirect mechanism. Co-localization of Axin2, Gli1, Shh, and activated pSMAD1,5,8 in the apical HERS suggests that this bilayered epithelial sheath is the core signaling center where the Wnt, SHH, and BMP signaling pathways act in parallel and perhaps converge to regulate root formation.
This study highlights the conserved nature of molecular interactions across two distinct tissue types-tooth and gut. Our results demonstrate how lessons in development and homeostasis can provide potential clues about cellular interactions occurring during development and homeostasis in health and disease across distinct tissues. Furthermore, this new understanding lends further support to the vital role of Wnts in the maintenance of the tooth stem/progenitor cell niche and radicular hard tissue formation and provides new insight into the diverse mechanisms of their action. A clearer understanding of the molecular mechanisms of stem/progenitor cell behavior and regulation during development will help us decipher the etiopathogenesis of developmental anomalies and disease states, which will form the basis for novel precision diagnostic and therapeutic modalities.

Resource availability Corresponding author
Further requests for resources and reagents should be directed to and will be fulfilled by the corresponding author, Abigail Tucker (Abigail.tucker@kcl.ac.uk).

Materials availability
The study did not generate new unique reagents.

Data and code availability
Data will be shared with the research community upon request to the authors. No code or standardized datasets were generated.

Murine models
Transgenic (maintained on a mixed background) and wild-type mice used in this investigation were housed in controlled conditions at the King's College London Biological Services Unit. Wild-type CD1 mice were obtained from CRL (Charles River Laboratory, Edinburgh, UK) and transgenic mouse lines Axin2creERT2 (Lustig et al., 2002); Axin2LacZ (Lustig et al., 2002); Gli1creERT2 (Ahn and Joyner, 2004); R26R-mTmG (Muzumdar et al., 2007); R26R-tdTomato (Madisen et al., 2010); Wls fl/fl (Carpenter et al., 2010); Sox9 fl/fl (Kist et al., 2002); and Ctnnb1 lox(ex3)/+ (Harada et al., 1999) have been previously described. Postnatal mice were culled following schedule-one methods as approved by the UK Home Office and were performed by trained individuals. Use of genetically modified mice was approved by the local GMO committee at King's, under personal and project licenses in accordance with the Animal (Scientific Procedures) Act of 1986, UK. This project conforms with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

Immunofluorescence
Sections were deparaffinized and rehydrated. Heat-induced antigen retrieval was carried out in 0.1 M sodium citrate (pH 6) buffer. Tissue sections were then blocked in 1% BSA, 0.1% Triton-X, and 1% serum. Sections were then treated with primary antibodies overnight at 4 C. For BrdU immunofluorescence, sections were additionally treated for 30 min with 2 M HCl at 40 C before incubation with primary antibody. After secondary antibodies, slides were mounted with Fluoroshield containing DAPI (Abcam). Sections were imaged with a Leica TCS SP5 confocal microscope. To ensure true and specific protein expression, positive and negative (without primary antibodies) procedural controls were maintained. Images were processed using Fiji (Schindelin et al., 2012). See supplemental experimental methods for antibody details. Tartrate-resistant acid phosphatase (TRAP) staining was carried out as described in Grigoriadis et al. (1994).

In situ hybridization and RNAscope
Dig-labeled antisense RNA probes were synthesized against mouse Gli1 and Ptch1 mRNA, and in situ hybridization was carried out to detect the expression in sagittal sections of wild-type mice, as previously described (Fons Romero et al., 2017). For the analysis of Axin2, RNAscope multiplex fluorescence assay (400,331-C3) (Advance Cell Diagnostics [ACD], a BioTechne brand) was performed following the manufacturer's instructions.
MicroCT and 3D reconstruction PFA fixed mouse heads (n R 5 samples each of mutant and control genotype) were dissected and scanned with a GE Locus SP microcomputed tomography (microCT) scanner and a Scanco microCT 50 scanner. 3D reconstructions and hard tissue image analysis of molars, incisors, and whole head were generated using Micro-View and Amira. Statistics involved non-paired t tests to compare mutant vs. wild-type samples.

AUTHOR CONTRIBUTIONS
R.L. and A.S.T. conceptualized and designed the study. R.L. and J.K. collected data. R.L., N.A., and A.S.T. analyzed and interpreted data. R.L., N.A., and A.T. drafted and revised the manuscript. All authors critically revised the manuscript and approved the final version.