Gli2 and Gli3 synergistically mediate HH-TGF-β crosstalk in mesenchymal progenitor cells to orchestrate tooth root morphogenesis

Abstract

Hedgehog (HH) signaling is essential in directing the fate determination of postmigratory cranial neural crest cells (CNCCs) to ensure proper craniofacial development. Gli transcription factors (TFs) are established as primary effectors of HH signaling, yet their distinct roles and regulatory mechanisms in governing cell commitment and differentiation of postmigratory CNCCs remain poorly understood. Here, using tooth root as a model, we combined transgenic mouse models with bioinformatic analyses to interrogate the functions of Gli2 and Gli3 in CNCC-derived root progenitor cells of the mouse molar. We revealed that loss of Gli3 alone in dental mesenchymal root progenitor cells caused shortened roots and that concurrent loss of Gli2 and Gli3 exacerbated root malformations, concomitant with profound impairments in cell proliferation and multilineage differentiation, suggesting a synergistic interaction between Gli2 and Gli3 during tooth root development. Mechanistically, Gli2 and Gli3 cooperatively regulated the transcription of Acvr2b, thereby modulating the activity of TGF-β/SMAD signaling within the dental mesenchyme. This Gli2/Gli3-TGF-β signaling cascade was critical for the lineage specification of tooth root progenitor cells during molar morphogenesis. Collectively, this work uncovers synergistic interactions of Gli2 and Gli3 in orchestrating tooth root morphogenesis and provides a novel insight into HH-TGF-β crosstalk in cell fate decisions of postmigratory CNCCs.

Introduction

Cranial neural crest cells (CNCCs) are a migratory and multipotent population originating from the dorsal neural tube, giving rise to a broad spectrum of cells and tissues in the craniofacial complex including skeleton, teeth, cartilage, melanocytes, endocrine cells, and peripheral neurons and glia.¹ Defective development of CNCCs underlies a variety of congenital disorders, such as Treacher Collins syndrome, Cornelia de Lange syndrome, and Goldenhar syndrome, leading to micrognathia, cleft palate, tooth malformations, and other craniofacial anomalies.^2,3,4 Previous studies have demonstrated that the development of CNCCs is highly reliant upon the Hedgehog (HH) signaling gradient for proper migration, differentiation, and survival.⁵ Removal of HH-responsiveness in CNCCs results in the absence of many CNCC-derived skeletal and non-skeletal tissues, accompanied by defective cell proliferation and specification,^6,7 suggesting that HH signaling is essential for conveying intrinsic patterning cues during craniofacial development. However, despite the importance of HH-dependent CNCC patterning in orchestrating craniofacial structures, the transcriptional mechanisms underlying CNCC fate determination remain elusive.

Gli transcription factors (TFs), which encompass Gli1, Gli2 and Gli3, serve as critical mediators of HH signal transduction, regulating diverse biological processes ranging from organogenesis to disease. Genetic studies and clinical evidence have underscored the indispensable roles of Gli2 and Gli3 in embryonic development and tissue homeostasis. For instance, loss of Gli2 in CNCCs specifically disrupts secondary palate fusion, aligning with clinical observations of craniofacial anomalies correlated with GLI2 variants.^8,9 In addition, mutations in GLI3 are associated with Greig cephalopolysyndactyly syndrome (GCPS), which is characterized by polysyndactyly, macrocephaly, a prominent forehead, and a wide nasal base with mild hypertelorism.¹⁰ Intriguingly, although Gli2 and Gli3 are not functionally equivalent, concurrent loss of Gli2 and Gli3 leads to severe patterning defects such as micrognathia, absent condylar process, and limb malformations,^11,12 suggesting their functional redundancy in organogenesis. Currently, however, the exact functional interplay between Gli2 and Gli3 in CNCCs has not been fully elucidated.

Postmigratory CNCCs give rise to the dental mesenchyme after their arrival in the oral region of the first pharyngeal arch, thereby contributing to the postnatal establishment of the tooth root.¹³ Tooth root represents an ideal model for studying tissue morphogenesis, which involves reciprocal interactions between epithelium and CNCC-derived mesenchyme.¹⁴ In the developing root, dental epithelium transduces Sonic hedgehog (Shh) signal to the underlying mesenchymal cells expressing Patched1 (Ptch1) and Gli1, thereby activating target gene transcription to initiate root formation.¹⁵ Suitable levels of HH signaling are essential for tooth root development as both suppression or hyperactivation of HH signaling can lead to shortened roots.^16,17 However, the functional interaction and molecular basis of Gli2 and Gli3 in postnatal molar morphogenesis remain undefined.

In this work, we utilized the tooth root as a model to investigate the functions of Gli2 and Gli3 in CNCC-derived progenitor cells and elucidate the HH/Gli-driven transcriptional network underlying tissue-specific morphogenesis. We examined the spatiotemporal expression of Gli2 and Gli3 and found that they were predominantly co-expressed in the apical region at the onset of root formation. Importantly, deletion of Gli3 in tooth root mesenchymal progenitor cells led to shortened roots, and concurrent ablation of Gli2 and Gli3 exacerbated root dysplasia with severely compromised cell differentiation and proliferation, highlighting their functional synergy in orchestrating tooth root morphogenesis. At the molecular level, we demonstrated that concurrent loss of Gli2 and Gli3 downregulated TGF-β/SMAD signaling by regulating the transcription of Acvr2b. Moreover, restoration of TGF-β/SMAD signaling partially rescued shortened roots and abnormal differentiation of tooth root progenitor cells. Taken together, this study uncovers previously unknown synergistic interactions between Gli2 and Gli3 and reveals the molecular mechanisms governing CNCC-derived cell fate during tooth root morphogenesis.

Results

Gli2 and Gli3 are expressed in Gli1⁺ progenitor cells at the onset of tooth root morphogenesis

To examine the expression pattern of Gli2 and Gli3 during tooth root development, we first reanalyzed single-cell RNA sequencing (scRNA-seq) data of the tooth germ at P3.5,¹⁸ the onset of tooth root development in the mouse molar. Leveraging unsupervised marker analysis, dental mesenchymal and epithelial cell populations were clearly identified (Figs. 1a, S1a-b). We found that Gli TFs, including Gli1, Gli2 and Gli3, were ubiquitously expressed in both dental epithelium and mesenchyme at P3.5, with Gli3 displaying a higher expression level compared to Gli1 and Gli2 (Figs. 1b-c, S1c-f). To further validate the spatiotemporal colocalization of Gli2 and Gli3, we conducted RNAscope analysis across multiple stages of tooth root formation. Intriguingly, we observed that Gli2 and Gli3 were highly co-expressed in the apical area where tooth root progenitor cells reside (Fig. 1d–o), implying that they may exert important functional effects during tooth root morphogenesis.

Fig. 1

Spatiotemporal co-expression pattern of Gli2 and Gli3 in the developing root. a UMAP plot showing cell clusters of dental epithelium and mesenchyme in mouse molars at P3.5. b, c Feature plots showing the expression of Gli2 and Gli3 in the first molar at P3.5. d–g Colocalization of Gli2 and Gli3 by RNAscope in the first molar at P3.5. h–k Colocalization of Gli2 and Gli3 by RNAscope in the first molar at P7.5. l–o Colocalization of Gli2 and Gli3 by RNAscope in the first molar at P12.5. p, p’ Colocalization of Gli2 and tdTomato (tdT) at P7.5 after tamoxifen (TAM) induction at P3.5. q, q’ Colocalization of Gli3 and tdT at P7.5 after TAM induction at P3.5. White boxes represent the apical region. White dotted lines outline the Hertwig’s epithelial root sheath (HERS). White arrows denote positive signals of Gli2 or Gli3. White arrowheads indicate colocalized signals. Scale bars, 100 μm

Gli1⁺ cells are important root progenitors giving rise to the entire root complex including odontoblasts, pulp cells, and periodontal tissue.¹⁹ To assess whether Gli2 and Gli3 are expressed in Gli1⁺-derived cells, we performed colocalization analysis and found that both Gli2 and Gli3 were expressed in the progeny of Gli1⁺ cell populations in the apical area of the mouse molar at P7.5 (Fig. 1p-q, p’-q’).

Gli2 and Gli3 synergistically regulate tooth root morphogenesis in the dental mesenchyme

In order to understand the function of Gli2 and Gli3 during tooth root development, we generated Gli1-CreER;Gli2^fl/fl, Gli1-CreER;Gli3^fl/fl, and Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mouse models, in which Gli2 or Gli3 was deleted alone or simultaneously in Gli1⁺ root progenitor cells. We found that loss of Gli2 had no obvious impact on the tooth root development in mouse molars (Fig. 2a-j, u). In contrast, deletion of Gli3 resulted in shortened roots, decreased alveolar bone volume as well as delayed tooth eruption (Fig. 2k-o, 2v), suggesting Gli3 plays a more prominent role in root development. Notably, concurrent loss of Gli2 and Gli3 led to marked exacerbation of tooth root dysplasia, with an approximately 50% reduction in root length (0.39 mm ± 0.022 mm) compared to control mice (0.735 mm ± 0.037 mm) (Fig. 2p-t, w). We also found that tooth eruption was delayed in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice, along with reduction in alveolar bone mass at the furcation region (Fig. 2s-t). These findings revealed a synergistic interaction of Gli2 and Gli3 in mediating tooth root morphogenesis.

Fig. 2

Concurrent deletion of Gli2 and Gli3 in Gli1⁺ progenitor cells exacerbates root dysplasia. a–e Micro-CT sagittal sections, 3D reconstruction images, and HE staining of molars in control mice at P14.5 and P21.5. f–j Micro-CT sagittal sections, 3D reconstruction images, and HE staining of molars in Gli1-CreER;Gli2^fl/fl mice at P14.5 and P21.5. k–o Micro-CT sections, 3D reconstruction images, and HE staining of molars in Gli1-CreER;Gli3^fl/fl mice at P14.5 and P21.5. p–t Micro-CT sections, 3D reconstruction images, and HE staining of molars in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice at P14.5 and P21.5. u–w Quantitative analysis of root length in control and mutant mice at P21.5 (n = 4). White arrows denote the measurement of root length. Black arrows indicate normal alveolar bone. Red arrows indicate compromised alveolar bone. Data are presented as mean ± SD. **P < 0.01, ***P < 0.001, ****P < 0.000 1, ns, no significance. Scale bars, (b, e, g, j, l, o, q, t), 200 μm; others, 500 μm

As Gli2 and Gli3 were also expressed in dental epithelium, to rule out their potential roles in epithelium during tooth root development, we generated K14-CreER;Gli2^fl/fl, K14-CreER;Gli3^fl/fl, and K14-CreER;Gli2^fl/fl;Gli3^fl/fl mouse models, enabling epithelial-specific deletion of Gli2, Gli3, or both. Importantly, loss of Gli2, Gli3 or both in dental epithelium had no apparent impact on tooth root length, alveolar bone mass, dentin thickness, or periodontal ligament (PDL) alignment (Fig. 3a-s), indicating that Gli2 and Gli3 are dispensable for regulating tooth root morphogenesis in dental epithelium.

Fig. 3

Deletion of Gli2, Gli3 or both in dental epithelium has no effect on tooth root development. a–d Micro-CT images and HE staining of molars in control mice at P21.5. e–h Micro-CT images and HE staining of molars in K14-CreER;Gli2^fl/fl mice at P21.5. i-l Micro-CT images and HE staining of molars in K14-CreER;Gli3^fl/fl mice at P21.5. m–p Micro-CT images and HE staining of molars in K14-CreER;Gli2^fl/fl;Gli3^fl/fl mice at P21.5. q–s Quantitative analysis of root length in control and mutant mice (n = 4). White arrows point to root length measurement. Black boxes represent the furcation region of the first molar. Black dotted lines indicate the area of periodontal ligament. Data are presented as mean ± SD. ns, no significance. Scale bars, (c, g, k, o), 200 μm; (d, h, l, p), 100 μm; others, 500 μm

Gli2 and Gli3 collaboratively control cell proliferation and differentiation of root progenitor cells

To interrogate the cellular changes underlying the tooth root defects, we first examined the differentiation of root progenitor cell-derived lineages. We observed that formation of odontoblasts (Dspp), the PDL (Periostin), and osteoblasts (Sp7) remained largely unaffected in Gli2 mutant mice (Fig. 4a-f), consistent with the histological analysis shown above. In contrast, Gli3 deficiency resulted in compromised PDL and odontoblast lineage differentiation, particularly in the root furcation region (Fig. 4g-h). These defects were further exacerbated upon concurrent deletion of Gli2 and Gli3, as the expression of Dspp and Periostin was nearly absent within the tooth root furcation region in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice (Fig. 4j-k). Moreover, the number of Sp7⁺ cells in alveolar bone decreased following the loss of Gli3 alone or concurrent loss of Gli2 and Gli3 (Fig. 4i, l), suggesting osteogenic differentiation is also affected in these mutant models. To assess the fate of Gli1⁺ root progenitor cells more definitively, we performed lineage tracing using tdT reporter mice. Interestingly, we observed an increased proportion of tdT⁺ cells in the coronal area accompanied by a reduced contribution to the periodontal region in Gli3 and double mutant molars at P14.5, suggesting disrupted fate determination of Gli1⁺-derived cells following loss of Gli3 or in combination with Gli2 (Fig. 4m-p). Given the essential role of Hertwig’s epithelial root sheath (HERS) as a signaling center in guiding tooth root formation, the morphology of HERS was examined by Krt14 immunofluorescence. Our analysis revealed that the bilayer structure and subsequent dissociation of HERS were not affected in the mutant molars (Fig. S2), suggesting that loss of Gli2 and Gli3 has little impact on HERS during tooth root formation.

Fig. 4

Concurrent loss of Gli2 and Gli3 profoundly disrupts cell differentiation and proliferation. a–c RNAscope staining of Dspp, and immunostaining of Periostin and Sp7 of molars in control mice at P21.5. d–f RNAscope staining of Dspp, and immunostaining of Periostin and Sp7 of molars in Gli1-CreER;Gli2^fl/fl mice at P21.5. g–i RNAscope staining of Dspp, and immunostaining of Periostin and Sp7 of molars in Gli1-CreER;Gli3^fl/fl mice at P21.5. j–l RNAscope staining of Dspp, immunostaining of Periostin and Sp7 of molars in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice at P21.5. m–p Visualization of tdT⁺ cells in molars from control, Gli1-CreER;Gli2^fl/fl, Gli1-CreER;Gli3^fl/fl and Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice at P14.5. q–t, q’–t’ Immunostaining of Ki67⁺ in the first molar in control and mutant mice at P5.5. u–v Quantification of Ki67⁺ cells in epithelium and mesenchyme in control and mutant mice (n = 4). White boxes represent the apical region. White dotted lines in (c, f, i, l) indicate the area of alveolar bone, and in (q’–t’) outline HERS. White arrows indicate positive signals, while yellow arrows denote impaired signals. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ns, no significance. Scale bars, (a–t), 100 μm; (q’–t’), 50 μm

In addition, we evaluated cell proliferation and apoptosis after loss of Gli2 or Gli3 in root progenitor cells. We found a reduction in Ki67⁺ cells in apical mesenchyme while epithelial proliferation was unaffected in Gli2 or Gli3 mutants. In contrast, Ki67⁺ cells were significantly decreased in both epithelium and mesenchyme of double mutant mice (Fig. 4q-v, q’-t’). TUNEL analysis indicated that there was no significant variation of apoptotic cells between the control and mutant groups (Fig. S3). Collectively, these data suggested that Gli2 and Gli3 work together to control the cell differentiation and proliferation of tooth root progenitor cells.

Concurrent loss of Gli2 and Gli3 leads to downregulated TGF-β/SMAD signaling

To assess HH signaling activity after deletion of Gli2, Gli3, or both, we examined the expression of Gli1, a well-established transcriptional readout of HH signaling activation. The expression of Gli1 was reduced in all three mutant models, suggesting that individual and concurrent loss of Gli2 and Gli3 in root progenitor cells leads to attenuated HH signaling activity (Fig. S4).

To further identify the downstream molecular mechanisms synergistically regulated by Gli2 and Gli3, we performed bulk RNA-seq to profile transcriptomic alterations. Heatmap and Volcano plots showed distinct differentially expressed genes (DEGs) in three mutant models (Fig. 5a-d). Consistent with the severe histological and cellular abnormalities, the double mutant models exhibited pronounced transcriptional alterations, with 518 genes upregulated and 1077 genes downregulated (|log2(fold change)|>1.5, P < 0.05) (Fig. 5d). Given that Gli2 and Gli3 act as transcription activators to influence HH signaling in tooth root development, we focused on downstream genes that are directly regulated by them. KEGG pathway analysis was performed to identify signaling pathways downregulated in Gli3 and double mutant mice, both of which displayed apparent tooth root defects. This analysis revealed a consistent downregulation of mTOR signaling, HH signaling, TGF-β signaling, and FoxO signaling in both mutant models (Fig. 5e-f). TGF-β signaling is a critical pathway that not only facilitates dentinogenesis and osteogenesis in the dental mesenchyme to ensure proper tooth root development,^20,21 but also regulates the proliferation and differentiation of CNCC-derived osteogenic progenitor cells.²² Therefore, we next performed molecular experiments to explore the relationship between HH/Gli signaling and TGF-β signaling. Western blot and immunostaining showed reduced levels of phosphorylated SMAD2 and SMAD3 in apical mesenchyme in double mutant models (Fig. 5g-o), confirming that Gli2 and Gli3 cooperate to mediate TGF-β signaling activity in mesenchymal root progenitor cells.

Fig. 5

Simultaneous deletion of Gli2 and Gli3 downregulates TGF-β signaling activity. a Clustered heatmap of bulk RNA-seq data from control and mutant groups at P5.5. b–d Volcano plots illustrating differentially expressed genes (DEGs) between control and mutant groups. e, f KEGG pathway analysis of downregulated genes showing top enriched signaling pathways in Gli3 mutant and double mutant groups. g Western blot of p-SMAD2, p-SMAD3, and total SMAD2/3 protein levels of tooth germ in control and mutant groups at P5.5. h–o Immunofluorescence staining of p-SMAD3 and p-SMAD2 in the first molar in four groups at P5.5. White boxes represent the apical region of the molar. White dotted lines outline HERS. White arrows indicate positive signals, while yellow arrows denote impaired signals. Scale bars, (h–o), 100 μm

To identify the potential mediator linking HH and TGF-β signaling, we analyzed DEGs commonly downregulated and associated with TGF-β signaling using Venn analysis (Fig. 6a). Among these candidates, Acvr2b, the primary receptor of Activin signaling, emerged as a compelling mediator as it was widely expressed in mouse molars (Fig. 6b) and has been previously implicated in embryonic mandible tooth formation.²³ Notably, its expression was significantly downregulated in Gli3 and double mutant models (Fig. 6c). We then examined Acvr2b expression using RNAscope and observed a specific reduction in the apical mesenchyme in double mutant mice (Fig. 6d-g). To investigate whether Gli2 and Gli3 directly regulate the transcription of Acvr2b, we performed CUT&Tag assay to compare binding profiles between the IgG control and Gli2/Gli3 group. Analysis of differential peaks demonstrated a markedly increased signal at Chr9:119401452-119402005 in the Gli2 group and Chr9: 119401686-119402067 in the Gli3 group, both of which correspond to the promoter region of Acvr2b (Fig. 6h-j). To further validate the direct binding of Gli2 and Gli3 to this region, we designed primers spanning the identified promoter locus and performed qPCR. The results showed significantly higher binding of Gli2 or Gli3 to the Acvr2b promoter compared to the IgG control (Fig. 6k). Together, these findings indicated that Acvr2b serves as a direct target of Gli2 and Gli3 to mediate HH-TGF-β crosstalk in the dental mesenchyme during tooth root development.

Fig. 6

Gli2 and Gli3 bind to the promoter of Acvr2b to activate its expression. a Venn diagram showing commonly downregulated genes associated with TGF-β signaling among three comparison groups. b Feature plot of Acvr2b in the mouse molar at P3.5. c RT-qPCR showing the relative mRNA expression of Acvr2b in control and mutant molar mesenchyme (n = 4). d–g RNAscope staining of Acvr2b expression in control and mutant mice. h Heatmap of CUT&Tag showing the binding profiles of Gli2 and Gli3 around ±3kb of the transcription start site (TSS). i Motif enrichment analysis for Gli2 and Gli3 in the molar mesenchyme. j Peak calling analysis showing enhanced binding of Gli2 and Gli3 to the promoter region of Acvr2b. The annotated peak is enlarged in the black box below. k CUT&Tag-qPCR showing enrichment at the Acvr2b promoter relative to the IgG control (n = 3). White dotted lines outline HERS. *P < 0.05, **P < 0.01, ns, no significance. Scale bars: (d–g), 50 μm

Upregulation of TGF-β/SMAD signaling partially rescues root dysplasia in Gli1-CreER;Gli2 ^fl/fl ;Gli3 ^fl/fl model

To verify whether the compromised TGF-β pathway underlies the root defects in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice, we pharmacologically upregulated TGF-β/SMAD signaling through intraperitoneal injection of the agonist SRI-011381.²⁴ Immunostaining demonstrated increased phosphorylation of SMAD3 in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice following SRI-011381 administration (Fig. S5), indicating partial restoration of TGF-β/SMAD signaling in the mouse molar. Importantly, we found that pharmacological activation of TGF-β signaling partially rescued defective roots in Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice, including restored root length, increased alveolar bone mass and resumption of tooth eruption at P21.5 (Fig. 7a-j). Moreover, the impaired differentiation of odontoblasts and PDL lineages was also partially restored, especially in the furcation area (Fig. 7k-r). These findings highlighted that Gli2/Gli3-mediated activation of TGF-β/SMAD signaling in root progenitor cells plays a critical role in postnatal tooth morphogenesis.

Fig. 7

Upregulation of TGF-β signaling activity partially rescues tooth root defects in double mutant mice. a–i Micro-CT images and HE staining of molars from control mice treated with vehicle, Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice treated with vehicle, and Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice treated with SRI-011381. Red arrows indicate the sites of tooth eruption. White arrows indicate the measurement of the tooth root length. j Quantitative analysis of root length among three treatment groups (n = 4). k–n Immunofluorescence staining and quantitative analysis of Periostin in three treatment groups (n = 3). o–r RNAscope staining and quantitative analysis of Dspp in three treatment groups (n = 3). White arrows indicate positive signals, yellow arrows indicate impaired signals, and white arrowheads indicate restored signals. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.000 1. Scale bars, (a, b, d, e, g, h), 500 μm; (c, f, i), 200 μm; (k–q), 100 μm

Discussion

In this study, we revealed synergistic interactions between Gli2 and Gli3 in regulating the proliferation and differentiation of CNCC-derived root progenitor cells in the mouse molar. We demonstrated that Gli2 and Gli3 cooperatively promote the transcription of Acvr2b, thereby modulating TGF-β/SMAD signaling activity in Gli1⁺ mesenchymal progenitor cells, highlighting the significance of HH-TGF-β signaling crosstalk in postnatal tooth development (Fig. 8). Our findings uncover a previously unidentified transcriptional mechanism synergistically mediated by Gli2 and Gli3 during tooth root morphogenesis, providing novel insights into HH/Gli-mediated regulation of CNCC-derived progenitor cells during craniofacial development.

Fig. 8

Schematic diagram illustrating HH-TGF-β signaling crosstalk in orchestrating tooth root development

Postmigratory CNCCs arriving at the pharyngeal arch continue to undergo dynamic repositioning and gradually establish distinct mesenchymal populations, a process which is controlled by complex transcriptional networks.¹⁸ HH signaling is important for the proliferation and differentiation of CNCCs to direct the patterning of CNCC derivatives.^25,26 Importantly, Gli2 and Gli3 maintain the appropriate level of HH signaling activity required for the osteogenic specification of CNCC-derived mesenchyme.⁷ In addition, HH/Gli2 signaling also participates in chondrocyte-to-osteoblast fate conversion and is responsible for cartilage heterotopic ossification in CNCC-derived mesenchyme.²⁷ In this study, we revealed that loss of Gli2, Gli3, or both attenuated HH signaling activity, thereby impairing proliferation, osteo/odontogenic, and PDL lineage differentiation of CNCC-derived root progenitor cells. It is well recognized that proper SHH-Gli1 signaling activity, modulated by signals from dental epithelium and mesenchyme, regulates epithelial stem cell maintenance and mesenchymal cell differentiation in the mouse molar.^28,29,30 Our findings further reinforce the central role of HH/Gli-driven transcriptional programs in dental mesenchymal progenitor cells in orchestrating tooth root morphogenesis. Interestingly, Nagata et al. recently reported that transient activation of HH signaling governs the fate of PTHrP⁺ dental follicle progenitor cells at the onset of tooth root formation,³¹ underscoring a critical temporal window of HH signaling in mediating tooth root development. Building upon previous studies, our research extends the understanding of precise spatiotemporal dynamics of HH signaling mediated by Gli2 and Gli3, which are indispensable for initial postnatal molar morphogenesis.

Notably, Gli2 and Gli3 exhibit highly context-dependent functional interactions that profoundly influence tissue morphogenesis. In some cases, they exert distinct or even opposite roles. The differential requirement for Gli2 and Gli3 is classically exemplified during neural tube patterning, where Gli2 participates in ventral-most patterning while Gli3 mediates intermediate-level patterning.³² A recent study revealed their counter-regulatory roles in osteogenesis, in which Gli2 promotes osteoblast differentiation, whereas Gli3 inhibits this process.³³ Despite the repressive function, Gli3-mediated transcriptional mechanism is essential in establishing anteroposterior limb axis, determining digit identities, and maintaining cranial suture patency.^34,35 Conversely, other studies have revealed that Gli2 and Gli3 function synergistically in organogenesis, constituting a conserved mechanism that drives multiple embryonic developmental processes. For instance, activator forms of Gli2 and Gli3 cooperatively mediate ventral neuronal patterning by regulating the segregation of neural progenitor cells in spinal cord development.³⁶ Similarly, foregut development is also highly dependent on this coordinated interplay to form the esophagus, trachea, and lungs.³⁷ More recently, this functional synergy of Gli2 and Gli3 in CNCCs has been implicated in midfacial development, with embryos exhibiting midfacial widening, bilateral cleft palate and a bifurcated nasal septum upon loss of both Gli2 and Gli3.⁹ Consistent with this finding, our study further elucidates the interaction of Gli2 and Gli3 in postmigratory CNCCs, demonstrating their synergistic roles in regulating postnatal molar morphogenesis. We found that concurrent loss of Gli2 and Gli3 in root progenitor cells results in a more severe phenotype than loss of either gene alone, characterized by significantly shortened tooth roots, reduced alveolar bone, and delayed tooth eruption. These defects arise from impaired proliferation and multi-lineage differentiation of root progenitor cells. Both odontogenic and periodontal differentiation of progenitor cells are severely inhibited following loss of Gli2 and Gli3. Importantly, lineage tracing analyses further revealed that Gli2 and Gli3 differentially regulate periodontal and odontogenic lineage migration of Gli1⁺ progenitor cells during tooth root development. In parallel, we also observed a reduction in cell proliferation. As Ki67 labels only a subset of Gli1⁺ progenitor cells, reduced Ki67⁺ cells may not be able to reverse the overall differentiation capacity of Gli1⁺ cell population, ultimately leading to defective tooth root morphogenesis.

It is well established that Gli2 and Gli3 are bifunctional transcription factors that contain a C-terminal activation domain and an N-terminal repression domain, with Gli2 predominantly acting as an activator (GliA), whereas Gli3 functions primarily as a repressor (GliR) of the HH pathway.^38,39 Interpretation of HH signals by Gli involves a finely tuned balance between GliA and GliR, both within individual proteins and among different Gli family members, which ultimately determines developmental process.⁴⁰ In our study, although Gli2 acts as an activator of HH signaling, the absence of an overt phenotype in the Gli2 mutant model suggests functional redundancy within the Gli TFs, with Gli3 potentially compensating for Gli2-dependent mechanisms. Interestingly, in hair follicle morphogenesis, keratinocyte-specific deletion of Gli2 disrupts hair follicle formation with defective proliferation of sebaceous gland progenitor cells, while ablation of Gli3 shows no detectable phenotype.⁴¹ These observations highlight the differential contributions of Gli2 and Gli3 across different organs. Notably, despite being a repressor, our findings suggest that Gli3 functions as an activator of HH signaling to promote mesenchymal progenitor cell differentiation and proliferation, thereby orchestrating tooth root morphogenesis. A similar activator role of Gli3 has also been reported in ventral spinal cord patterning.⁴²

TGF-β superfamily signaling directs a wide spectrum of cellular processes throughout the lifespan from embryonic patterning to maintenance of tissue homeostasis.^43,44 TGF-β signaling influences the specification of CNCCs and directly modulates gene circuits essential for craniofacial skeletal formation.⁴⁵ Importantly, crosstalk between HH and TGF-β signaling regulates the proliferation and differentiation of mesenchymal progenitor cells in organogenesis. In particular, HH/Smo signaling has been implicated in regulation of Tgfbr2 expression during nephrogenesis, as well as modulation of mesenchymal niche function in hair follicle.^46,47 Moreover, Gli proteins are indispensable for the timely induction of Tgfb2, thereby dictating the temporal patterning of neural stem cells.⁴⁸ In the prostate, stromal HH/Gli signaling regulates epithelial stem and progenitor cell activity by suppressing TGF-β signaling, highlighting a HH-TGF-β axis as a key modulator of epithelial renewal and differentiation.⁴⁹ Our findings demonstrate that Gli2 and Gli3 synergistically regulate phosphorylated SMAD2/3 levels in the dental mesenchyme, thereby establishing an HH-TGF-β axis that governs cell proliferation and lineage commitment of CNCC-derived root progenitor cells. However, Gli3 mutant models display shortened roots without a detectable change in TGF-β signaling activity, implying that alternative regulatory mechanisms contribute to Gli3-mediated tooth root morphogenesis.

Interestingly, we found that Activin signaling receptor Acvr2b serves as a potential mediator for the crosstalk between HH and TGF-β signaling. Our data demonstrate that Gli2 and Gli3 directly bind to the promoter region of Acvr2b, indicating that Acvr2b is a direct transcriptional target of HH signaling. Furthermore, simultaneous ablation of Gli2 and Gli3 caused a more pronounced reduction in Acvr2b expression in the dental mesenchyme than deletion of Gli2 or Gli3 alone, suggesting a cooperative regulatory mechanism between Gli2 and Gli3. Consistently, the phenotypes observed in three mutant groups also indicate the synergistic function of Gli2 and Gli3 during tooth root development. However, we cannot exclude the involvement of other molecules in this signaling interaction, which deserves further investigation.

In summary, our study elucidates a Gli2/Gli3-dependent synergistic transcriptional mechanism in orchestrating tooth root morphogenesis. We demonstrate that Gli2 and Gli3 synergistically regulate the proliferation and lineage specification of CNCC-derived dental mesenchymal progenitor cells by maintaining appropriate TGF-β signaling output. These findings extend our knowledge of HH/Gli-mediated tissue morphogenesis and uncover an intrinsic Gli2/Gli3-TGF-β cascade that governs progenitor cell fate decisions. Our study also sheds light on dental defects and has important implications for developing approaches for tooth root regeneration.

Materials and methods

Animals

Gli2^fl/fl and Gli3^fl/fl mouse strains were purchased from Gem Pharmatech Co., Ltd (China). Gli1-CreER and B6.Cg-Gt(ROSA)26Sortm14^{(CAG-tdTomato)Hze/}J were kindly given by Dr. Hu Zhao, Chinese Institute for Brain Research, Beijing. K14-CreER were purchased from Cygen Biomodels, LLC (China). Tamoxifen (T5648, Sigma, USA) was dissolved in corn oil (20 mg/mL, A561403, Ambeed, USA) and administered via intraperitoneal injection (i.p.) for CreER activation at P3.5 (1.5 mg per 10 g). All neonatal pups including both control and mutant mice received the same dose of tamoxifen. SRI-011381 (HY-100347, MedChemExpress, USA) was dissolved in DMSO at 20 mg/mL and stored at −80 °C for long-term use. Control mice were injected with vehicle solution containing 40% PEG300 and 5%Tween-80, Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice were injected with SRI-011381 via i.p. at a final dosage of 10 mg/kg every two days from P5.5 to P21.5. For clarity, Gli2^fl/fl;Gli3^fl/fl mice were designated as the control group in this study.

All animals were maintained under specific-pathogen-free (SPF) conditions at 20-26 °C; 10%-70% relative humidity, and a 12 h light/dark cycle, with free access to standard water and feeds. Animal experiments received approval from the Research Ethics Committees of the West China School of Stomatology, Sichuan University (WCHSIRB-AT-2025-410) and complied with ARRIVE2.0 guidelines.

Micro-CT

All mandibles were scanned using a SCANCO μCT45 (Switzerland) device with the following parameters: voxel size: 10 μm, energy: 55 kVp, intensity: 145 μA. Root length measurements and three-dimensional (3D) reconstructions of molars were performed using Mimics software. Root length was defined as the vertical distance between the furcation region and the apical region of the root.

Histological and immunofluorescence staining

Mouse mandibles were collected and fixed in 4% paraformaldehyde (PFA) at 4 °C for 24-48 h,and then decalcified in 10% EDTA (Biosharp, China) for 1-3 weeks. For histological staining, the dental tissues were dehydrated through a graded ethanol series, embedded in paraffin, and sectioned at 5 μm using a microtome. The tissues were stained with hematoxylin and eosin following the manufacturer’s protocol (G1120, Solarbio, China). For immunofluorescence staining, the mandibles were dehydrated in 15% sucrose for 2 h and 30% sucrose overnight, then embedded in OCT (6502, Epredia, USA) and prepared into 10 μm sections. The cryosections underwent antigen retrieval at 95 °C for 10 min if necessary, followed by incubation in blocking solution (FP1012, PerkinElmer) for 1 h at room temperature. Subsequently, they were incubated with primary antibodies at 4 °C overnight and cross-reacted with secondary antibodies for 1 h. Images were acquired using a STELLARIX5 confocal microscope (Leica, Germany). The primary antibodies used were as follows: tdTomato (AB8181, SICGEN, Portugal), Periostin (ab14041, Abcam, UK), Sp7 (ab209484, Abcam, UK), p-SMAD2 (44-24 G, Invitrogen, USA), p-SMAD3 (ET1609-41, Huabio, China), and Ki67 (ab15580, Abcam, UK). The secondary antibodies used were as follows: Alexa FluoR^TM 488 goat anti-rabbit (1:400, A-11034, Invitrogen, USA), Alexa FluoR^TM 555 goat anti-rabbit (1:400, RGAR003, Proteintech, China).

RNAscope in situ hybridization

Cryosections were prepared at a thickness of 10 μm, and RNAscope staining was performed using RNAscope Multiplex Fluorescent Kit v2.0 (323100, Advanced Cell Diagnostics, USA) or PinpoRNATM Multiplex Fluorescent RNA kit (PIF3000, Pinpoease, China). According to the manufacturer’s protocol, sections were pretreated with H₂O₂ for 15 min, subjected to antigen retrieval at 95 °C for 10 min, and treated with protease for 20 min at 40 °C. Then probes were hybridized to sections for 2 h at 40 °C, and paired with TSA Vivid Fluorophore 520 (PG-323271, PerkinElmer), 570 (PG-327232, PerkinElmer), or 650 (PG-323273, PerkinElmer). Images were acquired using a STELLARIX5 confocal microscope (Leica). Probes were synthesized as follows: Probe-Mm-Gli1 (311001), Probe-Mm-Gli2 (405771), Probe-Mm-Gli3 (445511), Probe-Mm-Dspp (448301), Probe-Mm-Acvr2b (114811).

TUNEL assay

A TUNEL BrightRed Apoptosis Detection Kit (A113, Vazyme, China) was used to detect apoptosis according to the instructions.

Bulk RNA sequencing

The mesenchyme of the first molar was separated carefully from control, Gli1-CreER;Gli2^fl/fl, Gli1-CreER;Gli3^fl/fl, Gli1-CreER;Gli2^fl/fl;Gli3^fl/fl mice at P5.5 with microforceps under a stereo microscope (Leica, Germany). RNA was extracted using the FastPure cell/tissue total RNA isolation kit (RC101-01, Vazyme, China) and stored at -80 °C. DNA libraries were prepared with AMPure XP system (Beckman Coulter, Beverly) and qualified using an Agilent 2100 bioanalyzer, subsequently they were submitted for sequencing on the Illumina NovaSeq 6000. Raw sequencing data quality was assessed using FastQC software. Gene-level quantification was performed using Feature Counts (v1.5.0-p3), and FPKM values were calculated to quantify gene expression levels. Differential expression analysis was performed with significance thresholds set at P < 0.05 and |log₂(fold change)| >1.5.

Western blot

Total protein from mesenchymal tissue of the first molar at P5.5 was extracted by ultrasonic lysis using the Total Protein Extraction Kit (KGB5303, KeyGEN, China) for 30 min at 4 °C. Subsequently, the supernatant was collected and denatured by heating at 99 °C for 10 min. The extracted proteins were loaded onto 7.5%-10% polyacrylamide gels (PG212, Epizyme, China) and transferred to PVDF membranes (ISEQ00005, Millipore, Germany). Membranes were blocked in 5% non-fat dry milk for 1 h and incubated with primary antibodies on a shaker overnight at 4 °C. After washing in TBST for three times, the membranes were incubated with secondary antibodies for 1 h and the bands were detected using a ChemiDoc Imaging System (Bio-Rad, USA). The following primary antibodies were used: p-SMAD2 (18338 T, Cell Signaling Technology, USA), p-SMAD3 (ET1609-41, Huabio, China), SMAD2/3 (14L86C22, Enzyme, China), β-actin (EM21002, Huabio, China). The secondary antibodies included HRP-conjugated goat anti-rabbit IgG polyclonal antibody (HA1001, Huabio, China), HRP-conjugated goat anti-mouse IgG polyclonal antibody (HA1006, Huabio, China).

Quantitative reverse transcription PCR (RT-qPCR)

Total RNA from mesenchymal tissue of the first molar at P5.5 was extracted using the FastPure cell/Tissue Total RNA isolation kit (RC101-01, Vazyme, China) and reverse transcribed to cDNA using HiScript®III RT Super Mix (R323-01, Vazyme, China). qRT-PCR was conducted using Taq pro Universal SYBR qPCR Master Mix (Q712-02/03, Vazyme, China) on a LightCycler480 system (Roche, Sweden). The relative expression levels of genes were calculated with the 2^-ΔΔCt method. Primers sequences used for RT-qPCR were listed below: Gapdh: Forward: AAGAAGGTGGTGAAGCAGGCATC; Reverse: CGGCATCGAAGGTGGAAGAGTG. Acvr2b: Forward: ACAAGAAGATGAGGCCCACG; Reverse: GCTCTTCTACACAGCCTGCA.

CUT&Tag

The mesenchyme of the first mandibular molars from wildtype mice at P3.5 was isolated using microforceps under a stereomicroscope (Leica, Germany). Nuclei were then isolated and purified using Bioyou Nuclei Isolation Kit (52009-10, WEMatrix Biotechnology, China) and numbered with Trypan Blue (T10282, Invitrogen, USA). For CUT&Tag, libraries were prepared according to the manufacturer’s instructions of Hyperactive Universal CUT&Tag Assay Kit for Illumina Pro (TD904, Vazyme, China) and sequenced on the Illumina NovaSeq 6000 platform with paired-end reads of 150 bp (performed by Hangzhou Kaitai Biotechnology Co., Ltd.). Raw reads were filtered using fastp (v0.23.4) and subsequently aligned to the mouse reference genome (GRCm38/mm10) using Bowtie2 (v2.5.3). Peak calling was performed using MACS3 (v3.0.1) with P < 0.05. Differential peaks were annotated using ChIPseeker, and motif analysis was conducted using MEME with E-value ≤ 0.05. For qPCR validation, stop buffer was added to release DNA fragments by incubation at 95 °C for 5 min, and the eluted DNA was subsequently used for qRT-PCR analysis with locus-specific primers. Relative enrichment over the IgG control was calculated using the 2^-ΔΔCt method. The antibodies used for CUT&Tag were: Gli2 (1:50, NB-600-874, Novus Biologicals, USA), Gli3 (5 μg/10⁶ cells, AF3690, R&D systems, USA), rabbit control IgG (1:50, AC005, Abclone, China), goat anti-rabbit IgG H&L (Ab207, Vazyme, China), and rabbit anti-goat IgG (H + L) (1:100, AS081, Abclone, China). The sequence of qPCR primers designed to detect the Acvr2b binding site were: Forward: AGTGACTGGCTTTTCCAGTGTA; Reverse: CAGATCCCTCCATTTCAGATGTTAC.

Statistical analysis

All data are presented as mean ± standard deviation (SD) and analyzed using GraphPad Prism10.0. Statistical significance was assessed using two-tailed Student’s t-test for two-group comparisons or one-way ANOVA for multiple comparisons. Immunofluorescence staining was quantified using ImageJ software. At least three independent experimental replicates were conducted, and representative images were selected. P < 0.05 was considered statistically significant.