A Biphasic Feature of Gli1⁺-Mesenchymal Progenitors during Cementogenesis That Is Positively Controlled by Wnt/β-Catenin Signaling

Abstract

Cementum, a specialized bony layer covering an entire molar root surface, anchors teeth into alveolar bone. Gli1, a key transcriptional activator in Hedgehog signaling, has been identified as a mesenchymal progenitor cell marker in various tissues, including the periodontal ligament (PDL). To address the mechanisms by which Gli1⁺ progenitor cells contribute to cementogenesis, we used the Gli1^lacZ/+ knock-in line to mark Gli1⁺ progenitors and the Gli1^CreERT2/+; R26R^tdTomato/+ line (named Gli1^Lin) to trace Gli1 progeny cells during cementogenesis. Our data unexpectedly displayed a biphasic feature of Gli1⁺ PDL progenitor cells and cementum growth: a negative relationship between Gli1⁺ progenitor cell number and cementogenesis but a positive correlation between Gli1-derived acellular and cellular cementoblast cell number and cementum growth. DTA-ablation of Gli1^Lin cells led to a cementum hypoplasia, including a significant reduction of both acellular and cellular cementoblast cells. Gain-of-function studies (by constitutive stabilization of β-catenin in Gli1^Lin cells) revealed a cementum hyperplasia. A loss of function (by conditional deletion of β-catenin in Gli1⁺ cells) resulted in a reduction of postnatal cementum growth. Together, our studies support a vital role of Gli1⁺ progenitor cells in contribution to both types of cementum, in which canonical Wnt/β-catenin signaling positively regulates the differentiation of Gli1⁺ progenitors to cementoblasts during cementogenesis.

Introduction

Cementum (a key component of periodontium) is critical for tooth attachment by anchoring teeth to surrounding alveolar bone via the unmineralized periodontal ligament (PDL) and provides a mechanical barrier to root resorption (Beertsen et al. 1999; Bosshardt 2005). To date, the cellular origin and regulatory mechanisms of cementogenesis remain elusive, hampering efforts to achieve cementum regeneration.

The development of cell lineage tracing technology has enabled the dynamic tracking of specific types of cells in vivo, providing an approach to determine the cell fate of stem/progenitor cells during tissue growth and repair (Braun et al. 2003; Tumbar et al. 2004; Blanpain and Simons 2013). Gli1, a transcriptional activator in Hedgehog signaling, has been identified as a marker for mesenchymal stem/progenitor cells in various tissues (Jing et al. 2021). Gli1⁺-mesenchymal progenitors are responsible for both bone formation and repair (Zhao et al. 2015; Shi et al. 2017). Of note, Gli1⁺ cells have also been identified as the multipotential periodontal ligament stem cells (PDLSCs) in vivo (Men et al. 2020). Although having distinct features in structure and function, bone and cementum share many similarities, particularly in their biochemical compositions and biomechanical properties (Fong et al. 2005; Foster, Ao, et al. 2015). However, it is largely unknown how Gli1⁺-mesenchymal PDL progenitors contribute to cementum growth.

The canonical Wnt/β-catenin pathway is highly conserved in evolution, and is involved in many biological processes such as growth, development, metabolism, and stem cell maintenance (Yan et al. 2013). Previous studies showed that Wnt/β-catenin signaling plays a vital role in cementogenesis by promoting cementoblast proliferation and inhibiting cementoblast differentiation (Nemoto et al. 2009; Cao et al. 2015). Interestingly, Bae et al. (2017) showed that the type of cementum (acellular vs. cellular) is not determined by its specific location but by levels of Wnt activity during cementum formation. To be specific, acellular cementum can be switched to cellular cementum when a high level of Wnt is induced in Osteocalcin-expressing cells, indicating a negative relationship between Wnt signaling and acellular cementum formation in cementoblasts. Nevertheless, whether and how Wnt signaling acts as a niche factor to control PDL stem/progenitor cell function remains unclear.

In the present study, we attempted to investigate 1) the roles and contribution of Gli1⁺-mesenchymal progenitors to postnatal cementum growth, 2) whether and how the canonical Wnt/β-catenin signaling defines the cell fate of Gli1⁺ progenitors during cementogenesis using both loss- and gain-of-function approaches in the Gli1^Lin tracing background, and 3) whether systemic administration of antisclerostin antibody (Scl-Ab; sclerostin is a potent antagonist of Wnt signaling) reactivates Gli1⁺ cells and accelerates cementogenesis. Our studies demonstrated a high level of Gli1⁺-mesenchymal progenitors in PDL at postnatal day (P) 21 when cementogenesis starts, which gradually reduces and maintains at a low level at P56 when cementogenesis is largely completed. The lineage-tracing data revealed a positive correlation between Gli1-derived progeny cell number and cementogenesis. The genetic studies showed an essential role of the Wnt/β-catenin signaling pathway in regulation of the Gli1⁺ cell-derived cementogenesis.

Materials and Methods

Transgenic Mice

All experimental protocols followed ARRIVE (Animal Research Reporting of In Vivo Experiments) guidelines and were approved by the Animal Care and Use Committees at Texas A&M University College of Dentistry and Sichuan University West China School of Stomatology.

Gli1^lacZ/+ (JAX#008211), Gli1^CreERT2/+ (JAX#007913), R26R^tdTomato/+ (JAX#007905), R26R^DTA/+ (JAX#006331), β-catenin^flox/flox (JAX#004152), and β-catenin^flox(Ex3)/+ (Harada et al. 1999) mice were housed in a temperature-controlled room with 12:12-h light/dark cycles.

To examine the expression profile of Gli1 during cementogenesis, Gli1^lacZ/+ reporter mice were sacrificed at P21, P42, and P56, separately. To trace the fate of the Gli1⁺ PDL cells, we performed cell lineage tracing using Gli1^CreERT2/+; R26R^tdTomato/+ mice. A single dose of tamoxifen (75 mg/kg body weight; T5648, Sigma-Aldrich) was administrated at P21. The mice were harvested at P23, P28, P42, and P56, separately, corresponding to 2, 7, 21, and 35 d after tamoxifen injection.

Gli1^CreERT2/+; R26R^DTA/+; R26R^tdTomato/+ mice were used to study the impact of ablation of Gli1⁺ cells on cementum development. Tamoxifen was injected once daily for 2 consecutive days starting from P21, and the mice were sacrificed at P42.

To investigate the regulatory roles of Wnt/β-catenin signaling in Gli1⁺-mesenchymal progenitors, the following mouse models were established: 1) downregulation of Wnt/β-catenin signaling in Gli1⁺ cells (loss of function) and 2) upregulation of Wnt/β-catenin signaling by constitutive stabilization of β-catenin or by using sclerostin-neutralizing antibody (Scl-Ab; sclerostin is a potent inhibitor of Wnt signaling) (gain of function).

X-gal Staining

The mandibles of Gli1^lacZ/+ reporter mice were fixed in 0.5% glutaraldehyde in phosphate-buffered saline (PBS) with 2 mM MgCl₂ at 4°C overnight before being decalcified with 10% EDTA, followed by being infused with 30% sucrose for 24 h at 4°C. Samples were then embedded in OCT compound and cryosectioned. X-gal staining was performed following standard protocol, as previously described (Brugmann et al. 2007).

Tissue Preparing, Histology, and Immunostaining

The specimens were fixed in 4% paraformaldehyde (PFA) and decalcified in 10% EDTA at 4°C. Samples for cell lineage tracing were dehydrated with 30% sucrose overnight and embedded in OCT embedding medium (Sakura Tissue-Tek) followed by Cryojane frozen sections as previously described (Jiang et al. 2005; Xie et al. 2019). Samples for β-catenin immunostaining were embedded in paraffin and cut into 5-µm-thick serial sections.

Immunostainings were carried out with the following antibodies: anti-BSP rabbit polyclonal antibody (generously provided by Dr. Renny Franceschi, University of Michigan; 1:200), anti-DMP1 rabbit polyclonal antibody (generously provided by Dr. Chunlin Qin, Texas A&M University College of Dentistry; 1:400), antiosterix rabbit polyclonal antibody (ab22552, Abcam; 1:400), or anti–β-catenin mouse monoclonal antibody (PY489, DSHB; 2 µg/mL).

Histomorphometric Analyses

Histomorphometric measurements were performed using ImageJ software (National Institutes of Health). The number of Gli1⁺ cells in the PDL adjacent to cementum was counted based on X-gal staining images of Gli1^lacZ/+ mice. All fluorescent images were captured using a SP5 Leica confocal microscope at wavelengths ranging from 488 (green) to 561 (red) nm. Red tdTomato fluorescent signal was observed in Gli1⁺ cells and their descendants; green color represented the corresponding immunofluorescent staining; blue color was 4′,6-diamidino-2-phenylindole (DAPI) staining for nuclei. The number of tdTomato⁺ and DAPI⁺ cells in a specified region of interest (ROI, the cementum of the first mandibular molar distal roots) was counted. Cellular cementum area and acellular cementum thickness were measured as previously described (Foster, Sheen, et al. 2015; Xie et al. 2019). For quantitative analyses, 4 animals were used in each group (n = 4), with at least 5 comparable sections from each mouse.

Micro–Computed Tomography Analysis

Micro–computed tomography (µCT) scans of mandibles were performed on 42-d-old Gli1^CreERT2/+; β-catenin^flox(Ex3)/+; R26R^tdTomato/+ mice and their littermates (Gli1^CreERT2/+; R26R^tdTomato/+) using a µCT 35 imaging system (Scanco Medical) as previously reported (Wang et al. 2017).

Statistical Analysis

The results were expressed as the mean ± standard deviation. All statistical analyses were performed with SPSS 17.0 software (SPSS, Inc.). The level of significance was determined by an independent-sample t test or one-way analysis of variance (ANOVA) in combination with Dunnett’s test. A value of P < 0.05 was considered statistically significant.

Results

Roles and Contribution of Gli1⁺ Cells to Cementogenesis

Gli1^lacZ/+ knock-in reporter mice have been widely used to visualize cells expressing Gli1 gene (Gli1⁺ cells) in vivo (Bai et al. 2002; Zhao et al. 2015). Here we showed that Gli1⁺ cells (by X-gal staining) were distributed throughout the entire PDL with few on the cementum surface at P21 when cementogenesis took place. As cementum rapidly expanded from P21 to P56, Gli1⁺ cells were sharply reduced and mainly restricted to the apical region (Fig. 1A, B). Quantitative data displayed a negative correlation between the number of Gli1⁺ PDL cells along the cellular cementum area and schematic cementum volume during growth (Fig. 1C, upper panel). Similarly, the number of Gli1⁺ PDL cells along the acellular cementum area was also significantly reduced during cementogenesis (Fig. 1C, lower panel). The above data indicate a negative correlation between Gli1⁺ PDL progenitor cells and cementum growth, which agrees with the recent finding that shows the negative regulation of Hedgehog signaling on cementogenesis (Choi et al. 2020).

Figure 1.

A progressive decrease of Gli1⁺ PDL cells during postnatal cementogenesis. The Gli1^lacZ/+ knock-in mice were harvested at postnatal day (P) 21, P42, and P56, respectively. X-gal stain visualized the distribution of Gli1⁺ PDL cells (yellow arrows) along the cellular (A) and acellular cementum (B) surface. (C) Quantification of Gli1⁺ PDL cells along cellular cementum (n = 4; P < 0.01, ^a′P < 0.001 vs. P21 group; ^bP < 0.01 vs. P42 group; upper panel, red line) and quantification of Gli1⁺ PDL cells along acellular cementum (n = 4; P < 0.01, ^a′P < 0.001 vs. P21 group; ^bP < 0.001 vs. P42 group; lower panel, red line). The dot line reflects cellular cementum volume (C, upper panel). ALB, alveolar bone; CC, cellular cementum; D, dentin; DP, dental pulp; PDL, periodontal ligament. Scale bars indicate 50 µm.

To investigate the roles of Gli1⁺ PDL progenitor cells during postnatal cementum growth, we studied the fate of Gli1⁺ progenitor cells using a compound tracing line (combined Gli1^CreERT2/+ and R26R^tdTomato/+). First, we showed that in the absence of tamoxifen, there was no tdTomato signal in the periodontium of Gli1^Lin at P23, excluding the “leaky” expression of the tdTomato reporter in the periodontium (Fig. 2A, left panel). Next, we performed one-time tamoxifen induction at P21 (the initiation stage of cementogenesis) followed by mandible collections at stages of P23, P28, P42, and P56, respectively (Fig. 2A, Appendix Fig. 1). At P23 (+2 d), there were few Gli1^Lin cells in the apical and middle third of the PDL with no Gli1^Lin-cementoblasts/cementocytes. By P28 (+7 d), there were more Gli1^Lin-PDL cells and Gli1^Lin-cementoblasts with a few Gli1^Lin-cementocytes. By P42 (+21 d), there were much more Gli1^Lin-PDL cells, Gli1^Lin-cementoblasts, and Gli1^Lin-cementocytes. By P56 (+35 d), there were abundant Gli1^Lin-cementoblasts and Gli1^Lin-cementocytes. Quantitative data showed that both the number of tdTomato⁺ cells (including Gli1^Lin-PDL cells and Gli1^Lin-cementoblasts) along the cellular cementum of the first mandibular molar distal roots (Fig. 2B, left panel) and the ratio of tdTomato⁺/DAPI⁺ cells (Fig. 2B, right panel) in the same area were gradually increased with a statistically significant difference at the ages of P42 and P56. Similarly, the significant increase of the number of tdTomato⁺ cells (including Gli1^Lin-PDL cells and Gli1^Lin-cementoblasts) along the acellular cementum surface took place (Fig. 2C). Importantly, after 35 d of chasing, over 60% of cementocytes were Gli1^Lin cells. The immunostaining of osterix (OSX), an essential transcription factor in cementum formation (Cao et al. 2012, 2015), showed that Gli1^Lin cells were initially OSX negative at P23, but after a 7-d chase, many of Gli1^Lin cells were OSX⁺ and expanded to both cementoblasts and cementocytes (Fig. 2D). Together, the chasing data indicate that the Gli1⁺ progenitor cells are responsible for expansion of PDL cells and cementoblasts along both cellular and acellular cementum areas, with some of them differentiating into cementocytes.

Figure 2.

Lineage tracing of Gli1⁺ progenitor cells and their progeny in cementum growth. Gli1^CreERT2/+; R26R^tdTomato/+ (CTR) mice were induced with tamoxifen at postnatal day (P) 21 and harvested at P23, P28, P42, and P56, separately. (A) The confocal images of BSP (upper panel) and DMP1 (middle-lower panel) immunostainings showed a gradual increase of tdTomato⁺ cells along the cementum surface and tdTomato⁺ cells entrapped in cellular cementum matrix over time. (B) The quantification of the number of tdTomato⁺ cells/mm² in the cellular cementum of the first mandibular molar distal roots (n = 4; ^aP < 0.05, ^a′P < 0.01 vs. P23 group; ^bP < 0.05 vs. P28 group; ^cP < 0.05 vs. P42 group) is shown in the left panel. The ratio of tdTomato⁺/DAPI⁺ cells in the cellular cementum of the first mandibular molar distal roots (n = 4; ^aP < 0.01, ^a′P < 0.001 vs. P23 group; ^bP < 0.05, ^b′P < 0.01 vs. P28 group) is shown in the right panel. (C) The quantification of the number of tdTomato⁺ cells/mm along the surface of acellular cementum of first mandibular molar distal roots (n = 4; ^aP < 0.05, ^a′P < 0.01, ^a′′P < 0.01 vs. P23 group; ^bP < 0.05, ^b′P < 0.01 vs. P28 group). (D) Expression profile of osterix (OSX) in Gli1 lineage cells (white arrows). AC, acellular cementum; ALB, alveolar bone; CC, cellular cementum; D, dentin; DP, dental pulp; PDL, periodontal ligament. Scale bars indicate 50 µm.

To further support the above hypothesis, we partially ablated Gli1⁺ cells by inducing expression of cytotoxic diphtheria toxin A (DTA). Specifically, both Gli1^CreERT2/+; R26R^DTA/+; R26R^tdTomato/+ (DTA) mice and Gli1^CreERT2/+; R26R^tdTomato/+ (CTR) mice were treated with 75 mg/kg tamoxifen for 2 consecutive days (P21 and P22) through intraperitoneal (i.p.) injection. Mice were harvested at P42 (Fig. 3A, Appendix Fig. 2). DMP1 and BSP immunostaining images revealed a striking cementum hypoplasia phenotype in DTA mice. Three weeks after tamoxifen induction, the extracellular matrix (ECM) masses in both cellular cementum and acellular cementum of DTA mice were sharply reduced, in which there were a lack of tdTomato⁺ cementocytes with an extremely low level of DMP1 and BSP compared to CTR mice (Fig. 3B, C). Quantitative analyses showed that the cellular cementum area and acellular cementum thickness of DTA mice were significantly reduced by ~71% (Fig. 3D, left) and ~56% (Fig. 3D, right), respectively. OSX immunostaining showed a lack of tdTomato⁺/OSX⁺ cementoblasts in both DTA cellular and acellular cementum (Fig. 3E), suggesting that OSX is a key downstream molecule of the Gli1 lineage.

Figure 3.

Genetic ablation of Gli1⁺ cell leads to cementum hypoplasia. (A) Schematic diagram of Gli1^CreERT2/+; R26R^DTA/+; R26R^tdTomato/+ mice with tamoxifen administration from postnatal day (P) 21 (once daily for 2 consecutive days) and harvest at P42. (B, C) DMP1 and BSP immunostainings showed fewer Gli1^Lin cells and a reduction of both cellular (B) and acellular (C) cementum in the diphtheria toxin A (DTA) mice (right panels). (D) The quantification data displayed a significant reduction of the cellular cementum area (n = 4; **P < 0.01) and the acellular cementum thickness (n = 4; *P < 0.05) in the DTA mice. (E) Osterix (OSX) immunostaining showed a lack of tdTomato⁺/OSX⁺ cementoblasts (white arrows) in the DTA cellular (left panel) and acellular cementum (right panel). AC, acellular cementum; ALB, alveolar bone; CC, cellular cementum; D, dentin; PDL, periodontal ligament. Scale bars indicate 50 µm.

Vital Roles of Wnt/β-Catenin Signaling in Gli1⁺ Progenitor Cells during Cementogenesis

To address the impact of Wnt/β-catenin signaling on the cell fate of Gli1⁺ progenitor cells, we first showed a gradual reduction of β-catenin level in PDL from P21 to P56 (Appendix Fig. 3) as previously reported (Xie et al. 2019). This expression pattern is similar to a progressive reduction of Gli1-expressing cells (Fig. 1), indicating a possible correlation between Wnt/β-catenin signaling and Gli1⁺ cells during cementogenesis.

To test this hypothesis, we first generated a gain-of-function model using Gli1^CreERT2/+; β-catenin^flox(Ex3)/+; R26R^tdTomato/+ (CA-β-cat mice) (Fig. 4A). Three-dimensional (3D) µCT reconstruction images revealed that constitutive activation of β-catenin in Gli1^Lin cells led to a cementum hyperplasia phenotype, featured by excessive deposition of cementum along the roots (Fig. 4A, left). Immunostaining displayed sharply increased ECM masses in CA-β-cat cellular and acellular cementum, with a dramatic increase of tdTomato⁺ cementocytes embedded in cementum matrix (Fig. 4A, middle; Appendix Fig. 4), and a great increase of the number of tdTomato⁺/OSX⁺ cementoblasts lining the root surface (Appendix Fig. 5), indicating a positive correlation between Wnt activity and the differentiation of Gli1⁺ cells to cementoblasts. The quantitative data showed that the cellular cementum area was significantly enlarged by ~3-fold in the CA-β-cat mice (P < 0.001; n = 4) with a similar increase trend in the acellular cementum thickness (P = 0.07) (Fig. 4A, right).

Figure 4.

Activation or inhibition of Wnt/β-catenin signaling greatly changes cementogenesis. (A) The impact of the constitutive activation of β-catenin (CA-β-cat) in Gli1⁺ cells (induced at postnatal days [P] 21–22 and harvested at P42) on cementogenesis: micro–computed tomography images (left panel), DMP1 and BSP immunostainings (green color) in the Gli1^Lin tracing background (red color) (middle panels) with cellular cementum in the upper panel and acellular cementum in the lower panel, and quantification of cellular cementum area (upper; n = 4; ***P < 0.001) and acellular cementum thickness (lower; n = 4; P = 0.07) in the right panels. (B) The impact of conditional removal of β-catenin in Gli1⁺ cells (Gli1^CreERT2/+; β-catenin^flox/flox; R26R^tdTomato/+ (cKO)) with tamoxifen induced at P21–22 and harvested at P56: DMP1 and BSP immunostainings in the Gli1^Lin tracing background with cellular cementum in the left panels and acellular cementum in the middle panels, as well as quantification of cellular cementum area (upper; n = 4; ***P < 0.001) and acellular cementum thickness (lower; n = 4; *P < 0.05) in the right panels. AC, acellular cementum; ALB, alveolar bone; CC, cellular cementum; D, dentin; PDL, periodontal ligament. Scale bars indicate 50 µm.

To further support the above hypothesis, we next generated a loss-of-function model with the genotype of Gli1^CreERT2/+; β-catenin^flox/flox; R26R^tdTomato/+ (cKO mice). Tamoxifen was administrated to both cKO and CTR mice for 2 d (at P21–22) and sacrificed at P56. Immunostaining data showed that both cellular (Fig. 4B, left) and acellular cementum (Fig. 4B, middle; Appendix Fig. 6) masses of the first mandibular molar distal roots were drastically reduced in the cKO mice with few tdTomato⁺ cementocytes in cKO mice. Moreover, tdTomato⁺/OSX⁺ cementoblasts were sharply reduced in cKO mice (Appendix Fig. 7), suggesting that downregulation of the Wnt activity impairs the differentiation of Gli1⁺ progenitor cells to OSX⁺ cementoblasts. Quantification data confirmed a significant decrease in the cellular cementum area (P < 0.001; n = 4) and acellular cementum thickness (P < 0.05; n = 4) in cKO mice.

Finally, we blocked sclerostin, a potent antagonist of Wnt signaling (Li et al. 2009; Ren et al. 2015), using the Scl-Ab in the Gli1^Lin line. Specifically, Gli1^CreERT2/+; R26R^tdTomato/+ mice were first induced with tamoxifen in young adults (P42), and experimental group animals were intraperitoneally injected with Scl-Ab at 25 mg/kg (twice a week) (Ren et al. 2015), starting from P70 and lasting for 4 wk (Fig. 5A). Systemic administration of Scl-Ab greatly increased the cellular cementum volume and collagen fibers (Appendix Fig. 8). Cell lineage tracing data showed a sharp increase of Gli1-derived cementocytes (Fig. 5B, left panel) and cementoblasts lining acellular cementum surface (Fig. 5B, right panel). The quantitative data demonstrated a 4-fold increase in both the tdTomato⁺ cell number and the ratio of tdTomato⁺/DAPI⁺ cells in cellular cementum of the Scl-Ab-treated group (Fig. 5C, left and middle panels) plus a 3-fold increase in the tdTomato⁺ cells lining the acellular cementum surface (Fig. 5C, right panel). Together, the above findings support a vital role of Wnt/β-catenin in control of Gli1⁺ mesenchymal cells during cementogenesis.

Figure 5.

Wnt/β-catenin signaling defines the cell fate of Gli1⁺ progenitors during cementogenesis. (A) Schematic diagram of intraperitoneal injections with tamoxifen (one time at postnatal day [P] 42) and Scl-Ab at 25 mg/kg (twice a week; starting from P70 and harvested at P98 in the Gli1^Lin tracing lines [Gli1^CreERT2/+; R26R^tdTomato/+]). (B) The impact of Scl-Ab on cementogenesis: DMP1 immunostaining in the Gli1^Lin tracing background with cellular cementum in the left panels; BSP immunostaining in the Gli1^Lin tracing background with acellular cementum in the right panels. (C) Quantification of tdTomato⁺ cells (n = 4; **P < 0.01) in cellular cementum in the left panel; the ratio of tdTomato⁺/DAPI⁺ cells (n = 4; **P < 0.01) in cellular cementum in the middle panel; quantification of tdTomato⁺ cells (n = 4; **P < 0.01) lining acellular cementum in the right panel. (D) Working hypothesis: Gli1⁺-mesenchymal progenitors are a vital cell source for cementoblasts, and the differentiation process is positively controlled by Wnt/β-catenin signaling. AC, a cellular cementum; CC, cellular cementum; PDL, periodontal ligament. Scale bars indicate 50 µm.

Discussion

Accumulating new evidence (Huang et al. 2009; Cao et al. 2015; Wang and Feng 2017; Xie et al. 2019) strongly supports the “classic theory” that the origin of cementum is from dental mesenchymal progenitors in PDL (Paynter and Pudy 1958; Cho and Garant 1988; Ten Cate 1997; Chai et al. 2000; Diekwisch 2001). However, there is also evidence backing up the epithelium cell origin (Orban 1952; Thomas 1995; Bosshardt 2005). In other words, the cell origin of cementum remains unsolved (Yamamoto et al. 2016; Zhao et al. 2016). In this study, we selected a well-documented marker of mesenchymal stem/progenitor cells, Gli1 (Zhao et al. 2015; Shi et al. 2017; Men et al. 2020), to define the cell origin of cementoblasts/cementocytes using the cell lineage tracing approach. Our key findings include 1) documenting a progressive decrease of the Gli1⁺ progenitor cell number during rapid expansion of cementum during postnatal development, 2) observation of a positive relationship between Gli1-derived acellular and cellular cementoblast cell number and cementum growth, 3) demonstration of a vital role of β-catenin in control of the Gli1⁺ PDL progenitor cells during cementum growth by studying gain or loss of function of β-catenin, and 4) systemic administration of Scl-Ab drastically accelerates cementogenesis, including Gli1 lineage cells (such as PDL, cementoblasts, and cementocytes).

Previous studies have demonstrated Gli1⁺ cells in vivo are negative for periodontium differentiation markers, and no Gli1 expression is detected within terminally differentiated cells, including cementoblasts and cementocytes (Men et al. 2020). In the present study, the X-gal stain images showed that the Gli1-lacZ signal is restricted in PDL progenitor cells but not in cementoblasts or cementocytes (i.e., a true progenitor marker), which is consistent with a previous report (Men et al. 2020), and that there is a progressive decrease of Gli1⁺ PDL progenitor cells during the development of cementum (Fig. 1). On the other hand, the cell lineage tracing study showed a progressive increase of Gli1 lineage cells in PDL, cementoblasts, and cementocytes, indicating a positive association between Gli1 lineage cells and cementum growth. Interestingly, we noticed that only some cementoblasts lining the acellular cementum were tdTomato⁺ while most of cells around the cellular cementum were Gli1 lineage cells. In theory, the acellular cementum is formed prior to the cellular cementum. Our speculation is that the timing of tamoxifen injection at P21 targets most cellular cementum progenitor cells but a few acellular cementum progenitor cells. Furthermore, partial ablation of Gli1⁺ cells by inducing expression of DTA leads to a sharp reduction of Gli1 lineage cells (PDL, cementoblasts, and cementocytes) and cementum volume (Fig. 3). This information supports the notion that cementogenesis needs only a limited number of Gli1⁺ PDL progenitors, although they are highly active to maintain normal cementum growth.

Similar to the Gli1 expression pattern, β-catenin expression levels were gradually decreased during cementum growth as well (Appendix Fig. 3). The loss or gain of function of β-catenin in Gli1⁺ PDL progenitor cells displayed drastic changes of cementogenesis (including both cell number and cementum volume; Fig. 4A, B) through affecting the differentiation of Gli1⁺ PDL progenitor cells (Appendix Figs. 5, 7). Importantly, systemic administration of Scl-Ab at the adult age results in a great increase in both Gli1 lineage cells (such as PDL, cementoblasts, and cementocytes) and cementum volume (Fig. 5B, C; Appendix Fig. 8). These in vivo data support a vital role of Wnt/β-catenin signaling in regulation of Gli1⁺ PDL progenitor cells.

In conclusion, our studies support a vital role of the Gli1⁺ progenitor cells in contribution to both types of cementum, in which the canonical Wnt/β-catenin signaling positively regulates the differentiation of Gli1⁺ progenitors to cementoblasts (Fig. 5D). Successfully increasing Gli1^Line PDL and cementoblasts/cementocytes plus a sharp increase in cementum masses by applying Scl-Ab at the adult stage provides a promising approach for repair and regeneration of cementum in future.

Author Contributions

X. Xie, J. Wang, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; C. Xu, contributed to data acquisition, critically revised the manuscript; H. Zhao, contributed to data analysis and interpretation, critically revised the manuscript; J.Q. Feng, contributed to conception, design, data analysis, and interpretation, drafted and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.