Axin2⁺-Mesenchymal PDL Cells, Instead of K14⁺ Epithelial Cells, Play a Key Role in Rapid Cementum Growth

Abstract

To date, attempts to regenerate functional periodontal tissues (including cementum) are largely unsuccessful due to a lack of full understanding about the cellular origin (epithelial or mesenchymal cells) essential for root cementum growth. To address this issue, we first identified a rapid cementum growth window from the ages of postnatal day 28 (P28) to P56. Next, we showed that expression patterns of Axin2 and β-catenin within cementum-forming periodontal ligament (PDL) cells are negatively associated with rapid cementum growth. Furthermore, cell lineage tracing studies revealed that the Axin2⁺-mesenchymal PDL cells and their progeny rapidly expand and directly contribute to postnatal acellular and cellular cementum growth. In contrast, the number of K14⁺ epithelial cells, which were initially active at early stages of development, was reduced during rapid cementum formation from P28 to P56. The in vivo cell ablation of these Axin2⁺ cells using Axin2^CreERT2/+; R26R^DTA/+ mice led to severe cementum hypoplasia, whereas constitutive activation of β-catenin in the Axin2⁺ cells resulted in an acceleration in cellular cementogenesis plus a transition from acellular cementum to cellular cementum. Thus, we conclude that Axin2⁺-mesenchymal PDL cells, instead of K14⁺ epithelial cells, significantly contribute to rapid cementum growth.

Introduction

Periodontal tissues consist of 4 defined structures: gingiva, cementum, alveolar bone, and periodontal ligament (PDL). Cementum, a thin layer of mineralized tissue, covers the tooth root surface and serves to attach the tooth to alveolar bone via collagen fibers of PDL. Based on morphological, histological, and functional differences, there are 2 major types of cementum: 1) acellular (primary) cementum, located from the cervical margin to the apical third, anchors principal fibers of PDL to the root surface and is critical for tooth attachment, and 2) cellular (secondary) cementum, located from the middle to apical third, plays an essential role in adaption to occlusal loads and repair (Nanci and Ten Cate 2013; Foster 2017).

Periodontitis is a local inflammatory disease featured by the destruction of periodontal tissues. The ultimate goal for periodontal treatment is the regeneration of the lost periodontal tissues. In periodontal tissue engineering and regeneration, cementum regeneration remains a critical and challenging phase (Chen and Jin 2010; Vaquette et al. 2018). Thus, it is necessary to get a better understanding of the cell sources and developmental processes of cementum formation, as well as gene regulations.

Osterix (OSX), a key transcription factor downstream of Runx2, is essential for both bone and cementum formation as deletion of Osx leads to a lack of bone (Baek et al. 2009) and a sharp reduction in cementum formation (Cao et al. 2012). OSX controls bone and cementum formation via a similar mechanism: inhibition of cell proliferation and acceleration of cell differentiation. Additional mechanistic studies suggest that the key role of OSX in controlling cementogenesis is to maintain a proper level of Wnt/β-catenin via direct upregulation of DKK1 (a potent Wnt antagonist) (Cao et al. 2015).

The canonical Wnt/β-catenin signaling plays a broad role in development and maintenance of many organs and tissues. In addition, Wnt/β-catenin signaling has been implicated in the control over various types of stem/progenitor cells (Nusse 2008; Visweswaran et al. 2015; Nusse and Clevers 2017). Activation of Wnt/β-catenin signaling causes a complex series of events that lead to the prevention of β-catenin degradation and its consequent stabilization and accumulation in the cytoplasm. Stabilized β-catenin then translocates to the nucleus, where it stimulates downstream gene expression through Tcf/Lef (Komiya and Habas 2008). Wnt responsive genes, such as Axin2, are expressed in a wide variety of stem/progenitor cells (Ontiveros et al. 2008; Lim et al. 2016; Maruyama et al. 2016; Ransom et al. 2016; Usami et al. 2019). Using a lineage-tracing study, a population of Wnt-responsive cells (Axin2-expressing cells) has been identified in PDL, which becomes activated in response to tooth extraction and forms new alveolar bone (Yuan et al. 2018).

In comparison with other craniofacial organs, studies of cementogenesis fall behind with many fundamental questions largely unsolved, including the cellular origin of root cementum. One school of thought is that dental follicle cells (mesenchymal origin) directly differentiate into cementoblasts after invading the gaps between the ruptured Hertwig’s epithelial root sheath (HERS) and then attaching to the root dentin (Nanci and Ten Cate 2013; Wang and Feng 2017). Another hypothesis is that the epithelial cells from HERS form cementoblasts via an epithelial-mesenchymal transition process (Orban 1952; Thomas 1995; Bosshardt 2005; Wang and Feng 2017). Both theories were further supported by recent evidence (Sonoyama et al. 2007; Huang et al. 2009; Cao et al. 2012; Chen et al. 2014; Cao et al. 2015).

The goal of this study was to investigate whether Axin2⁺-mesenchymal PDL cells directly contribute to postnatal cementogenesis using multiple mouse lines (Axin2^lacZ/+; Axin2^CreERT2/+; K14^Cre/+; β-catenin^flox(Ex3)/+; R26R^DTA/+; R26R^tdTomato/+) with different combinations in the following 4 approaches: 1) Axin2 lineage tracing lines (lacZ or tdTomato reporter) combined with immunostaining assays to reveal a close association between the temporal and spatial expression pattern of Axin2 and β-catenin⁺ PDL cells, OSX⁺ cementoblasts, or DMP1⁺ cementocytes; 2) ablation of Axin2⁺ cells (loss of function); 3) constitutive activation of Axin2⁺ cells by β-catenin stabilization (gain of function); and 4) K14 tracing line to show K14⁺ cells during the postnatal cementogenesis (up to the age of postnatal day 56 [P56]). Our findings demonstrate that Axin2⁺-mesenchymal PDL cells, as key progenitor cell sources, play a vital role in postnatal cementogenesis.

Materials and Methods

Breeding 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.

Axin2^lacZ/+ (stock number: 009120), Axin2^CreERT2/+ (stock number: 018867), K14^Cre/+ (stock number: 004782), R26R^tdTomato/+ (stock number: 007905), and R26R^DTA/+ (stock number: 006331) mice were purchased from Jackson Laboratory and housed in a temperature-controlled environment with 12-h light/dark cycles. To trace Axin2-expressing Wnt-responsive cells during cementum formation, Axin2^CreERT2/+ mice were crossed with R26R^tdTomato/+ reporter mice. A single intraperitoneal injection of tamoxifen (75 mg/kg body weight; T5648, Sigma-Aldrich) was administered at the age of P28. Animals were sacrificed at P31, P35, P42, and P56, separately, corresponding to 3, 7, 14, and 28 d after the tamoxifen induction.

To generate triple transgenic mice to conditionally ablate Axin2-lineage cells, Axin2^CreERT2/+; R26R^tdTomato/+ mice were crossed with R26R^DTA/+ mice. Tamoxifen was administrated once daily for 3 consecutive days starting from P28, and the mice were harvested at P35.

To constitutively activate β-catenin in Axin2-lineage cells, Axin2^CreERT2/+; R26R^tdTomato/+ mice were crossed with β-catenin^flox(Ex3)/+ mice (Harada et al. 1999). Tamoxifen was administrated once daily for 3 consecutive days starting from P28, and then the animals were sacrificed at P56.

Tissue Preparation, Histology, and Immunostaining

Mandibles were fixed in freshly prepared 4% paraformaldehyde and decalcified in 10% ethylenediaminetetraacetic acid (EDTA). Samples for histological staining were embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) or Masson’s trichrome stain as previously reported (Wang, Muir, et al. 2017). Samples for cell lineage tracing were dehydrated with 30% sucrose and embedded in OCT compound (Sakura Tissue-Tek). Ten-μm-thick sections were prepared with a Leica cryostat equipped with Cryojane as previously described (Jiang et al. 2005; Jing et al. 2016).

Immunostaining was then carried out with anti–β-catenin mouse monoclonal antibody (PY489, DSHB; 2 µg/mL), anti-DMP1 rabbit polyclonal antibody (generously provided by Dr. Chunlin Qin, Texas A&M University College of Dentistry; 1:400), anti-osterix rabbit polyclonal antibody (ab22552, Abcam; 1:400), anti–β-galactosidase chicken polyclonal antibody (ab9361, Abcam; 1:200), or anti-Axin2 rabbit polyclonal antibody (ab227158, Abcam; 1:100).

PEGASOS Tissue Clearing

PEGASOS tissue-clearing procedures were performed as previously reported (Jing et al. 2018; Yi et al. 2019). Fluorescent images of mouse mandibles were acquired with a SP8 Leica confocal microscope. Three-dimensional reconstruction was performed with IMARIS 9.0 (Bitplane).

Micro–Computed Tomography (µCT) Analysis

Mandibles were dissected and analyzed by a µCT 35 imaging system (Scanco Medical) as previously reported (Wang, Massoudi, et al. 2017).

Histomorphometric Analyses

Histomorphometric measurements were performed using ImageJ software (National Institutes of Health). To quantify the number of tdTomato⁺ cells and the total number of cells (counterstained with 4′,6-diamidino-2-phenylindole [DAPI]) in cellular cementum or along acellular cementum, tissue sections were photographed using a SP5 Leica confocal microscope. 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. Acellular cementum thickness was measured on mesial surfaces of the first mandibular molar distal roots 300 µm from the cementoenamel junction (Foster et al. 2015). Four animals per group were used for quantitative analyses (n = 4), with at least 5 comparable sections from each mouse.

Statistical Analysis

The results were expressed as the mean ± standard deviation, and the level of significance was determined by using 1-way analysis of variance (ANOVA) in combination with Dunnett’s test using SPSS 17.0 software (SPSS, Inc.). A value of P < 0.05 was considered statistically significant.

Results

Temporal and Spatial Expression Patterns of Axin2 and β-Catenin Are Negatively Associated with Rapid Cementum Growth

It is known that cementogenesis, in contrast to the rest of tooth components, starts late and experiences lifelong deposition (Gupta et al. 2014). We first analyzed serial mesiodistal sections of the first murine mandibular molars from P28 to P84, reflecting puberty to adulthood, using H&E and Masson’s trichrome stain (Fig. 1a). At P28, there was a small mass of cellular cementum lining the first molar root surface. From P28 to P42, there was a gradual and steady increase in cellular cementum mass. However, from P42 to P56, cellular cementum rapidly expanded. Histomorphometric analyses showed a drastic difference between this time window (P42 to P56) compared to the other 2 developmental periods (P28 to P42 and P56 to P84), which was statistically significant (Fig. 1b; n = 4). To be specific, the growth rate of cellular cementum from P42 to P56 was ~0.025 mm²/wk, while the growth rate was only ~0.011 mm²/wk from P28 to P42 or ~0.012 mm²/wk from P56 to P84. Furthermore, we used DMP1 immunofluorescence to measure the thickness of acellular cementum at different stages. Our data revealed a prominent difference of the growth rate of acellular cementum between the period of P42 to P56 and the other 2 periods (Appendix Fig. 1; n = 4). The growth rate of acellular cementum from P42 to P56 was ~2.97 µm/wk, whereas it was only ~0.89 µm/wk from P28 to P42 or ~0.35 µm/wk from P56 to P84.

Figure 1.

Temporal and spatial expression patterns of Axin2 and β-catenin are negatively associated with rapid cementum growth. (a) H&E (upper panel) and Masson’s trichrome (lower panel) staining demonstrated expanded cellular cementum in the first mandibular molar distal roots from postnatal day 28 (P28) to P84. (b) Histomorphometric analyses showed a drastic difference between the time window of P42 to P56 compared to the other 2 developmental periods (P28 to P42 and P56 to P84), which is statistically significant (n = 4; ^aP < 0.05, ^bP < 0.001, ^cP < 0.01 versus P28 group; ^dP < 0.001, ^eP < 0.01 versus P42 group; ^fP < 0.05 versus P56 group). (c) Axin2-lacZ⁺ cells were visualized by β-galactosidase (β-gal) immunofluorescence (yellow arrows). (d) Quantification of Axin2-lacZ⁺ PDL cells in close proximity to cellular cementum (n = 4; ^aP < 0.05, ^bP < 0.01 versus P28 group; ^cP < 0.05 versus P42 group). (e) Immunohistochemistry showed gradually decreased expression of β-catenin (red arrows) during cellular cementum growth. (f) Quantification of β-catenin⁺ PDL cells in close proximity to cellular cementum (n = 4; ^aP < 0.05, ^bP < 0.01 versus P28 group; ^cP < 0.001 versus P42 group). ALB, alveolar bone; CC, cellular cementum; D, dentin; DP, dental pulp; PDL, periodontal ligament. Scale bars indicate 50 µm.

Next, we examined expression profiles of Axin2 during cementum growth at P28, P42, and P56 using Axin2-lacZ reporter mice. In Axin2-lacZ mice, the lacZ cassette is inserted in frame with the ATG start codon of the Axin2 gene, and thus β-galactosidase activity reflects Axin2 gene expression (Lustig et al. 2002). Interestingly, Axin2-lacZ level in the PDL adjacent to both cellular cementum (Fig. 1c, d; n = 4) and acellular cementum (Appendix Fig. 2a; n = 4) gradually decreased from P28 to P56 qualitatively and quantitatively, indicating a negative correlation between Axin2 expression and cementum growth throughout puberty to adulthood. Similarly, expression profiles of β-catenin also gradually decreased during cellular and acellular cementum growth (Fig. 1e, f and Appendix Fig. 2b; n = 4). Axin2 immunohistochemistry data further confirmed the decreased expression of Axin2 in periodontium from P28 to P84 (Appendix Fig. 2c). These data support the notion that a low level of Wnt signaling activity is required for cementum growth.

Axin2⁺ PDL Cells Are Primary Progenitor Cell Sources for Cementum Formation

To trace the fate of the Axin2⁺ PDL cells, we crossed Axin2^CreERT2/+ mice with R26R^tdTomato/+ mice. Without tamoxifen administration, tdTomato signal was undetectable in the periodontium of Axin2^CreERT2/+; R26R^tdTomato/+ mice at P24, indicating there was no “leaky” expression of the tdTomato reporter in periodontium (Fig. 2a). Then, 1 dose of tamoxifen was administrated at P28 followed by 4 chasing timelines: 3, 7, 14, and 28 d, separately. For a better view of the relationship between Axin2-lineage tdTomato⁺ cells and cementocytes, we performed DMP1 (a marker for cementocytes) immunostaining. Analyses at day 3 showed few relatively small Axin2⁺ cementoblasts lining the root surface with very few Axin2⁺ cementocytes (Fig. 2b, left panel). After 7 d of chasing, there were much more enlarged tdTomato⁺ cementoblasts along the cellular cementum surface, with some of these red cells becoming DMP1⁺ cementocytes (Fig. 2b, left-middle panel). By 28 d of chasing, there were even more but smaller tdTomato⁺ cementoblasts plus additional tdTomato⁺ cementocytes (Fig. 2b, right panel). The histomorphometric analyses showed that these changes were statistically significant regarding the number of tdTomato⁺ cells/mm² (Fig. 2c; n = 4) and the ratio of tdTomato⁺/DAPI⁺ cementocytes in the cellular cementum of the first mandibular molar distal roots (Fig. 2d; n = 4). Notably, after 28 d of chasing, most cementocytes were tdTomato⁺ with the ratio of tdTomato⁺/DAPI⁺ cells reaching ~70% (Fig. 2d; n = 4). Similarly, there was a gradual increase of Axin2-lineage cells lining acellular cementum over time, correlated with acellular cementum growth (Appendix Fig. 3a–d), which was statistically significant between P31 and P56 (Appendix Fig. 3e; n = 4). But in view of the limitations of the analyses depending on tissue sectioning methods, we applied the PEGASOS tissue-clearing technique (Jing et al. 2018; Jing et al. 2019) to further confirm the contribution of Axin2-lineage cells during cementum growth in a 3-dimensional view. Both 3-dimensional images and movies of cleared mandibles revealed that tdTomato⁺ cells were drastically increased in cellular cementum and acellular cementum from P31 (Fig. 2e, left; Appendix Movie 1) to P56 (Fig. 2e, right; Appendix Movie 2).

Figure 2.

Lineage tracing of Axin2⁺ cells and their progeny in cellular cementum formation. (a) No-tamoxifen control. (b) Axin2^CreERT2/+; R26R^tdTomato/+ mice were administrated 1 dose of tamoxifen at postnatal day 28 (P28) and followed by 4 chasing timelines: 3, 7, 14, and 28 d, separately. DMP1 immunostaining was performed for a better view of cementum. (c) The number of tdTomato⁺ cells/mm² in the cellular cementum of the first mandibular molar distal roots (n = 4; ^aP < 0.05, ^bP < 0.01 versus P31 group; ^cP < 0.05, ^dP < 0.05 versus P35 group). (d) The ratio of tdTomato⁺/DAPI⁺ cells in the cellular cementum of the first mandibular molar distal roots (n = 4; ^aP < 0.01, ^bP < 0.001 versus P31 group; ^cP < 0.01, ^dP < 0.01 versus P35 group). (e) Three-dimensional images of cleared mandibles revealed significantly increased tdTomato⁺ cells in cellular cementum (yellow arrows) or along acellular cementum (white arrows) from P31 to P56. (f) Immunofluorescence showed some tdTomato⁺ cells expressing osterix (OSX) (white arrows). CC, cellular cementum; D, dentin; DP, dental pulp; PDL, periodontal ligament. Scale bars indicate 50 µm.

Furthermore, immunostaining revealed tdTomato⁺ cells expressing OSX, a transcription factor in mesenchymal cells essential for controlling cementum formation (Cao et al. 2012), especially at age P56 (after 28 d of chasing) (Fig. 2f). Together, these findings suggest that Axin2⁺ PDL cells are the major progenitor cell sources for root cementum growth.

Ablation of Axin2-Lineage Cells Leads to Severe Cementum Hypoplasia

To further support the notion that Axin2⁺ PDL cells are the major progenitor cell sources for root cementum growth, we performed cell ablation assays using Axin2^CreERT2/+; R26R^DTA/+; R26R^tdTomato/+ mice with tamoxifen administrated from P28 (once a day for 3 consecutive days). Mice were harvested at P35 (Fig. 3a, right). Immunostaining images showed the DMP1⁺–extracellular matrix (ECM) masses in the DTA-ablated cellular cementum (Fig. 3b, right) and acellular cementum (Appendix Fig. 4a, c, right) were notably reduced, in which there were a lack of tdTomato⁺ cementocytes with an extremely low level of DMP1 compared to the control (CTR) mice. Quantitative data showed a significant decrease in the cellular cementum area of the first mandibular molar distal roots by ~70% (Fig. 3b; n = 4). Moreover, there was a striking reduction in the acellular cementum thickness of the second mandibular molar mesial roots by ~50% (Appendix Fig. 4d; n = 4), although there was no statistically significant difference for the first mandibular molar distal roots due to a large variation (Appendix Fig. 4b; n = 4). OSX immunostaining images further revealed a lack of tdTomato⁺/OSX⁺ cementoblasts and few tdTomato⁺ cementocytes in the DTA-ablated mice (Fig. 3c, right). Similarly, β-catenin immunohistochemistry data demonstrated a lack of β-catenin⁺ PDL cells along either cellular cementum (Appendix Fig. 4e, left) or acellular cementum (Appendix Fig. 4e, right) in Axin2-DTA mice. In sum, the data support that Axin2⁺ PDL progenitor cells are essential for both cellular and acellular cementum growth.

Figure 3.

Ablation of Axin2-lineage cells leads to severe cementum hypoplasia. (a) Schematic diagram of Axin2^CreERT2/+; R26R^DTA/+; R26R^tdTomato/+ mice with tamoxifen administration from postnatal day 28 (P28) (once a day for 3 consecutive days) and harvest at P35. (b) The DMP1 immunostaining showed severely defective cellular cementum in Axin2-DTA mice characterized by a lack of tdTomato⁺ cementocytes and an extremely low level of DMP1 expression. Quantitative data based on DMP1⁺–extracellular matrix areas showed a significant decrease in the cellular cementum area of the first mandibular molar distal roots by ~70% (n = 4; *P < 0.05). (c) Osterix (OSX) immunostaining revealed a lack of OSX⁺ cementoblasts on the surface of Axin2-DTA cellular cementum, whereas there were many OSX⁺ cementoblasts (white arrows) in CTR. CC, cellular cementum; D, dentin; PDL, periodontal ligament. Scale bars indicate 50 µm.

Constitutive Activation of β-Catenin in the Axin2⁺ PDL Cells Accelerates Root Cementum Growth, Leading to Cementum Hyperplasia

Because Axin2 is a key target of Wnt/β-catenin signaling, we next generated a gain-of-function model containing Axin2^CreERT2/+; β-catenin^flox(Ex3)/+; R26R^tdTomato/+ (CA-β-cat mouse), constitutively activating β-catenin in the Axin2⁺ PDL cells. Micro–computed tomography (CT) imaging analyses showed a cementum hyperplasia phenotype characterized by excessive buildup of mineralized cementum over root surface at P56 (Fig. 4a, lower panel). Confocal immunostaining images showed great expansion of cellular cementum with DMP1 expression in CA-β-cat tdTomato⁺ cementocytes (Fig. 4b, right). The quantitative data showed that the cellular cementum area of the first mandibular molar mesial roots was significantly enlarged by close to 2-fold and the acellular cementum thickness was notably increased by ~2.5-fold (Fig. 4c; n = 4). Interestingly, the CA-β-cat mice exhibited a transition from acellular cementum to cellular cementum at the cervical region, characterized by many tdTomato⁺ cells trapped in the DMP1⁺ mineral matrix (Fig. 4d, right). This information indicates that both cellular and acellular cementoblasts originate from the same Axin2⁺ progenitor cells but are under different regulation mechanisms during cementum growth.

Figure 4.

Constitutive activation of β-catenin in Axin2-lineage cells contributes to cementum hyperplasia. (a) Representative micro–computed tomography images showed notably expanded cellular cementum in the CA-β-cat mice. (b) Confocal immunostaining images showed great expansion of cellular cementum with DMP1 expression in the CA-β-cat tdTomato⁺ cementocytes. (c) Histomorphometric analyses suggested the cellular cementum area of the first mandibular molar mesial roots was significantly enlarged by close to 2-fold (n = 4; ***P < 0.001). Meanwhile, acellular cementum thickness was notably increased by ~2.5-fold (n = 4; *P < 0.05). (d) The CA-β-cat mice exhibited a transition from acellular cementum to cellular cementum at the cervical region, featured by many tdTomato⁺ cells trapped in the DMP1⁺ mineral matrix. Scale bars indicate 50 µm.

HERS Cells Play No Apparent Role in Rapid Cementum Growth from P28 to P56

It is widely believed that HERS cells play a role in cementogenesis, especially in acellular cementum growth (Huang et al. 2009). To address whether HERS cells truly contribute to cementum growth, we performed cell lineage tracing using noninducible K14^Cre/+; R26R^tdTomato/+ mice labeling the precursors of the epithelial cells of tooth germ and their progeny. At early stages from P3 to P21, there were K14⁺ epithelial cells along the cementum surface, suggesting these cells may play a role in the initiation of cementogenesis (Appendix Fig. 5). However, there were few tdTomato⁺ cells on the acellular cementum surface at P28 (Fig. 5a, right upper panel). By P56, we observed only 2 tdTomato⁺ cells along the acellular cementum surface in 1 jawbone section (Fig. 5a, right lower panel), whereas the K14⁺ gingival epithelial cells remained highly active as they were at P28 (Fig. 5a, left panel). Thus, the cell-tracing data indicate the notion that K14⁺ HERS cells may play a role in the initiation of cementogenesis but have no apparent role in rapid cementum growth from P28 to P56.

Figure 5.

Axin2⁺-mesenchymal periodontal ligament (PDL) cells are responsible for cementum growth. (a) The representative K14^Cre/+; R26R^tdTomato/+ tracing data displayed strong red fluorescence in gingival epithelial cells but very few tdTomato⁺ cells in PDL at ages of postnatal day 28 (P28) (upper panel) and P56 (lower panel). DMP1 immunostaining showed green fluorescence in bone and cementum. (b) Schematic diagram depicting essential roles of Axin2⁺ PDL progenitor cells in cementum formation: Axin2⁺-mesenchymal PDL cells are key cell sources for rapid cementum growth. Ablation of these cells led to hypoplastic cementum formation, while constitutive activation of β-catenin in these cells resulted in cementum hyperplasia. Scale bars indicate 250 µm.

Discussion

Although the cellular origin for cementum growth has been debated for many decades (Slavkin 1976; Slavkin et al. 1989; Thomas 1995; Bosshardt and Nanci 2004; Bosshardt 2005; Huang et al. 2009), none of the theories are conclusive partly due to a limitation of early research approaches, including simple gene expression comparison, cell culture or organ culture, a pure gene knockout, or overexpression without a side-by-side cell lineage tracing study. In this study, we revisited this critical issue (i.e., cementoblasts/cementocytes originated from mesenchymal cells or epithelial cells) using multiple mouse lines (Axin2^lacZ/+; Axin2^CreERT2/+; K14^Cre/+; β-catenin^flox(Ex3)/+; R26R^DTA/+; R26R^tdTomato/+) with different combinations plus comprehensive imaging approaches. Our key findings are as follows: 1) the temporal and spatial expression patterns of Axin2 and β-catenin in PDL cells directly concur with postnatal cementum growth, which reach peak levels at P28 right before rapid cementum growth; 2) ablation of Axin2⁺ PDL cells (loss of function using the compound mouse line of Axin2^CreERT2/+; R26R^DTA/+; R26R^tdTomato/+) leads to a sharp reduction in both cellular and acellular cementum growth; 3) activation of Axin2⁺ PDL cells by constitutive stabilization of β-catenin (gain of function) results in acceleration of cellular and acellular cementum growth in concert, including a transition from acellular cementum to cellular cementum; and 4) there are few K14⁺ HERS cells at P28 and P56. These findings support the notion that Axin2⁺-mesenchymal PDL cells, instead of K14⁺ HERS cells, are key progenitor cell sources that play a crucial role in rapid cementum growth, and both cellular and acellular cementum share the same progenitor cell sources.

It has been documented extensively that Wnt signaling is essential for periodontal development and homeostasis. The current study has extended our understanding about Wnt signaling in cementum formation. Here, we demonstrated that both cellular and acellular cementum expanded rapidly from P42 to P56, while the peak level of Wnt signaling is at P28, followed by a gradual reduction (Fig. 1 and Appendix Fig. 2). On the other hand, OSX, an essential transcription factor for cementum growth (Cao et al. 2012; Cao et al. 2015), remains at a relatively stable level from P28 to P56 (Fig. 2f). These data agree with Bae et al. (2017) and our previous report (Cao et al. 2012) that appropriate regulation of Wnt signaling is required for cellular and acellular cementum-specific apposition patterns, and they indicate a requirement for a low Wnt activity in differentiated cementoblasts to preserve characteristics of acellular cementum. Thus, we propose that Wnt signaling is an initial driving force for cementogenesis, while maintaining a low level of Wnt signaling is essential for cementum growth and mineralization.

One of the key findings in this study is that we demonstrated that Axin2⁺-mesenchymal PDL cells, instead of K14⁺ epithelial cells, are directly responsible for rapid cementum growth. The supporting evidence includes the following: 1) the Axin2⁺-mesenchymal PDL cells give rise to the majority of cementoblasts and cementocytes (Fig. 2b–e and Appendix Fig. 3); 2) the Axin2-lineage cells, to a large extent, express OSX (Fig. 2f), the key mesenchymal transcription factor for cementum formation (Cao et al. 2012; Cao et al. 2015); 3) ablation of Axin2⁺ PDL cells using Axin2^CreERT2/+; R26R^DTA/+ leads to an arrest of both cellular and acellular cementum growth (Fig. 3 and Appendix Fig. 4); 4) activation of Axin2⁺ PDL cells by constitutive stabilization of β-catenin using Axin2^CreERT2/+; β-catenin^flox(Ex3)/+ shows not only a great expansion of cellular cementum with numerous tdTomato⁺ cementocytes but also a transition of tdTomato⁺ cementoblasts into cementocytes in the acellular cementum (Fig. 4); and 5) the K14^Cre/+; R26R^tdTomato/+ mice display very few K14⁺ cells in PDL during the rapid cementum growth (P28 to P56) (Fig. 5a). Previously, Huang et al. (2009) reported K14⁺ HERS cells in PDL and thus proposed roles of HERS in cementum growth. Here we followed the fate of K14⁺ cells during the early stages (P3, P7, P14, and P21; Appendix Fig. 5) and the late stages (P28 and P56; Fig. 5a). We observed an overall decreasing trend in the number of K14⁺ PDL cells as cementum grew, whereas K14⁺ gingival epithelial cells were highly active during all developmental stages. This information suggests that K14⁺ epithelial cells may contribute to the initiation of cementogenesis but play no apparent role in rapid cementum growth.

Together, our comprehensive in vivo studies identified the Axin2⁺ PDL progenitor cells, which are directly responsible for postnatal cementum growth (Fig. 5b). More important, these findings, coupled with K14^Cre/+; R26R^tdTomato tracing data, suggest that Axin2⁺ mesenchymal PDL cells play a key role in rapid cementum growth, whereas K14⁺ HERS cells may play a role in the initiation of cementogenesis but not in rapid cementum growth.

Author Contributions

X. Xie, J. Wang, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; K. Wang, C. Li, S. Zhang, D. Jing, C. Xu, contributed to data acquisition, critically revised the manuscript; X. Wang, 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.