BMP9‐initiated osteogenic/odontogenic differentiation of mouse tooth germ mesenchymal cells (TGMCS) requires Wnt/β‐catenin signalling activity

Abstract Teeth arise from the tooth germ through sequential and reciprocal interactions between immature epithelium and mesenchyme during development. However, the detailed mechanism underlying tooth development from tooth germ mesenchymal cells (TGMCs) remains to be fully understood. Here, we investigate the role of Wnt/β‐catenin signalling in BMP9‐induced osteogenic/odontogenic differentiation of TGMCs. We first established the reversibly immortalized TGMCs (iTGMCs) derived from young mouse mandibular molar tooth germs using a retroviral vector expressing SV40 T antigen flanked with the FRT sites. We demonstrated that BMP9 effectively induced expression of osteogenic markers alkaline phosphatase, collagen A1 and osteocalcin in iTGMCs, as well as in vitro matrix mineralization, which could be remarkably blunted by knocking down β‐catenin expression. In vivo implantation assay revealed that while BMP9‐stimulated iTGMCs induced robust formation of ectopic bone, knocking down β‐catenin expression in iTGMCs remarkably diminished BMP9‐initiated osteogenic/odontogenic differentiation potential of these cells. Taken together, these discoveries strongly demonstrate that reversibly immortalized iTGMCs retained osteogenic/odontogenic ability upon BMP9 stimulation, but this process required the participation of canonical Wnt signalling both in vitro and in vivo. Therefore, BMP9 has a potential to be applied as an efficacious bio‐factor in osteo/odontogenic regeneration and tooth engineering. Furthermore, the iTGMCs may serve as an important resource for translational studies in tooth tissue engineering.


| INTRODUC TI ON
Tooth is a highly mineralized organ arising from the tooth germ, which is induced by sequential and reciprocal interactions between the immature epithelium and the underlying cranial neural-crestderived mesenchyme in the developing embryo. 1 Most of the dental tissues are formed by dental ectomesenchyme, which is formed by condensed cells originated from cranial neural-crest-cells during embryonic development. Following tooth germ formation, cells in tooth germ differentiate into multiple lineages of cells forming tooth tissues, including ameloblast, odontoblast, pulp cells and periodontal ligament cells. These cells secrete extracellular matrix proteins for mineral deposition and then form hard tissues of tooth, such as enamel, dentin and cementum. 2 Tooth germ mesenchyme determines both tooth identity and the ability to reprogram nondental epithelium to tooth fate, as a result of transmission of the odontogenic potential from epithelium to the mesenchyme in mouse tooth between embryonic days 11 and 12, which was demonstrated by the classic tissue recombination studies. [3][4][5] While regeneration of fully functional teeth utilizing autologous bioengineered tooth germ transplantation 6 may provide great provision for improving clinical outcomes in cases involved with tooth loss, harnessing the natural regenerative potential of cranial neuralcrest-derived mesenchyme cells in tooth-forming tissue may provide an alternative solution to reconstruct defects of tooth hard tissue and maintain pulp vitality. 7 For this reason, the stem cells discovered in adult tooth, including the mesenchymal stem cells in dental papilla, pulp and periodontal ligament have attracted intensive attention during the past decades for they are capable of differentiating into cells forming dentin/cementum and periodontal ligament. However, it is unclear whether tooth germ mesenchymal cells (TGMCs) possess osteo/odontoblastic differentiation potential.

Numerous signalling pathways have been implicated in regulat-
ing the fate and lineage determination of tooth mesenchyme, including Wnt/β-catenin pathway and BMP signalling. 8,9 Nonetheless, the detailed mechanism underlying tooth development from TGMCs remains to be fully understood. We have demonstrated in a previous study that BMP9 (aka, growth and differentiation factor 2, or GDF2) possesses great potential to induce osteogenic, adipogenic and, to a lesser extent, chondrogenic differentiation of mesenchymal stem cells (MSCs), 10,11 and that BMP9 and canonical Wnt signalling act synergistically to induce osteo/odontoblastic differentiation of stem cells harvested from apical papilla of mature lower incisor teeth, which is considered as a subpopulation of MSC-like cells. 12,13 In this study, we investigated the role of Wnt/β-catenin signalling in BMP9-induced osteogenic/odontogenic differentiation of TGMCs.
To overcome the technical challenge of maintaining primary dental germ cells, we first established the reversibly immortalized TGMCs (iTGMCs) derived from young mouse mandibular molar tooth germs.
We demonstrated that BMP9 effectively induced the expression of osteogenic markers and in vitro matrix mineralization, which could be effectively blunted by β-catenin knockdown. In vivo implantation of BMP9-stimulated iTGMCs induced robust formation of ectopic bone, knocking down β-catenin expression in iTGMCs significantly diminished BMP9-induced ectopic bone formation. Taken together, these discoveries demonstrate the osteogenic/odontogenic ability of engineered iTGMCs upon BMP9 stimulation, but this process requires the participation of canonical Wnt signalling. Therefore, BMP9 has the potential to be explored as an efficacious bio-factor for osteo/odontogenic tissue engineering and tooth engineering.

| Cell culture
The HEK-293 cell line (ATCC) was maintained in complete Dulbecco's Modified Eagle's Medium (DMEM) containing 10% foetal bovine serum (FBS, Invitrogen), 100 units of penicillin and 100 mg of streptomycin at 37°C in 5% CO 2 . The recently engineered 293pTP line was used for adenovirus amplification. 14 Both cell lines were maintained in complete DMEM. Unless indicated otherwise, all chemicals were purchased from Sigma-Aldrich or Fisher Scientific.

| Isolation and immortalization of mouse TGMCs from late bell stage molar tooth germs
All animal studies were conducted following the guidelines approved by Institutional Animal Care and Use Committee (IACUC). Mandibles retained osteogenic/odontogenic ability upon BMP9 stimulation, but this process required the participation of canonical Wnt signalling both in vitro and in vivo. Therefore, BMP9 has a potential to be applied as an efficacious bio-factor in osteo/odontogenic regeneration and tooth engineering. Furthermore, the iTGMCs may serve as an important resource for translational studies in tooth tissue engineering.

| Establishment of reversibly immortalized tooth germ mesenchymal cells (iTGMCs)
The use of the retroviral vector SSR #41 to express SV40 T antigen flanked with the FRT sites have been described previously. [15][16][17][18] Briefly, the SSR #41 vector and pCL-Ampho packaging vector were co-transfected into293 Phoenix Ampho (293PA) cells to produce the packaged retrovirus. Exponentially growing TGMCs were infected with the SSR #41 retrovirus and subjected to hygromycin B selection (0.4mg/mL) for 3-5 days twice in complete DMEM at 37°C, yielding the stably immortalized tooth germ mesenchymal cells, designated as the iTGMC line.

| Generation and amplification of recombinant adenoviruses expressing BMP9, Wnt3A, Flippase (FLP) and GFP
AdEasy technology was used to generate recombinant adenoviruses as described previously. 19 Briefly, the coding regions of human BMP9, mouse Wnt3A and FLP recombinase were cloned into an adenoviral shuttle vector, respectively, then these vectors were used to generate recombinant adenoviruses in HEK-293 or 293pTP cells. The resulting adenoviruses were designated as Ad-BMP9, Ad-Wnt3A and Ad-FLP, both of which also express GFP to monitor infection efficiency. Analogous adenovirus expressing only GFP (Ad-GFP) was used as control. Polybrene (4-8 mg/ mL) was added to enhance infection efficiency as previously reported. 20

| Construction of multiplex siRNA expression system targeting mouse β-catenin and the generation of stable iTGMC-KD and iTGMC-Ctrl lines
The single vector-based multiplex expression of siRNAs targeting the coding regions of mouse β-catenin (NM_007614.3) were constructed by using the BSG Versatile shotgun cloning strategy as recently reported. 21 Briefly, the 3 siRNA sites targeting βcatenin were driven by the opposing U6 and H1 promoters into pSEB361-BSG destination vectors, yielding pBSG361-simBC.
Vector containing the scrambled sites which do not target any significant known human and rodent transcripts was also constructed as a control (eg, pBSG361-siControl). To generate retrovirus supernatants for infecting subconfluent iTGMCs, the resultant retroviral transfer vectors pSEB361-simBC and pSEB361-siControl were co-transfected with retroviral packaging plasmids into the293PA cells as described in a previous study. 18 Cells were subjected to blasticidin S selection (final concentration at 3 µg/mL) for 5 days after 36-48 hours post viral infection.
The resultant stable lines were designated as iTGMC-KD, and iTGMC-Ctrl, respectively.

| RNA isolation and quantitative RT-PCR (qPCR)
Total RNA was isolated by using TRIZOL Reagents (Invitrogen), and cDNA templates were generated by reverse transcription reactions with hexamer and MMuLV reverse transcriptase (New England Biolabs). QPCR was carried out as described. 22 The primers used in this study are show in Table 1. SYBR Green-based qPCR analysis was carried out by using the thermo cycler CXF-Connect (Bio-Rad) as described elsewhere. [23][24][25][26] All qPCR reactions were done in triplicate. All samples were normalized by the expression level of GAPDH.

| Immunofluorescence staining
Immunofluorescence staining was performed as previously described. 10 Briefly, cells were first infected with Ad-Wnt3A or Ad-GFP for 48 hours or directly fixed with methanol, permeabilized with 1% NP-40, and blocked with 10% BSA, followed by incubating with primary antibodies. Concentrations of primary antibodies used are shown in Table 2 Stains without primary antibodies, or with control IgG, were used as negative controls.

| Alkaline phosphatase (ALP) activity assay
Alkaline phosphatase activity was assessed using the Great ESCAPEe SEAP Chemiluminescence assay kit (BD Clontech) and histochemical staining assay (using a mixture of 0.1 mg/mL napthol AS-MX phosphate and 0.6 mg/mL Fast Blue BB salt) as described. 27 Each assay condition was performed in triplicate and the results were repeated in at least 3 independent experiments.
ALP activity was normalized by total cellular protein concentrations among the samples.

| Matrix mineralization assay (Alizarin Red S staining)
The iTGMCs were seeded in 24-well cell culture plates and infected with Ad-BMP9 or Ad-GFP. Infected cells were cultured in the presence of ascorbic acid (50 mg/mL) and β-glycerophosphate (10 mM).
Mineralized matrix nodules were stained by Alizarin Red S staining 14 days after infection as described. 28 Briefly, cells were fixed with 2.5% (v/v) glutaraldehyde at room temperature for 10 minutes and washed with PBS before incubated with 2% Alizarin Red S (Sigma-Aldrich) for 20 minutes. Excess dye was washed off by PBS after incubation. The staining of calcium mineral deposits was recorded under bright-field microscopy.

| Oil Red O staining assay
The iTGMCs were seeded in 24-well cell culture plates, infected with

| Subcutaneous iTGMCs cell implantation, ectopic bone formation and micro-computed tomography (μCT) analysis
All animal studies were conducted under the guidelines approved by the Institutional Animal Care and Use Committee (IACUC). The thermo-responsive PPCNg was used as a scaffold for the delivery of BMP9-stimulated iTGMCs for bone formation as described. 29 The iTGMCs-mediated ectopic bone formation was carried out as previously described. 19 Briefly, iTGMC-Ctrl and iTGMC-KD were infected with Ad-BMP9 or Ad-GFP for 20 hours, collected and resuspended in for 15μ micro-CT imaging. All specimens were imaged using the vivaCT 40 preclinical imaging system (Scanco Medical AG) and all image data and 3D volumetric data were analysed using μCT V6.1 software. 13

| Statistical analysis
All quantitative assays were performed in triplicate and/or repeated 3 times. Analysis results were expressed as mean ± SD. Statistical significances were determined by one-way analysis of variance or the student's t test. A value of P < .05 was considered statistically significant.

| The SV40 T antigen-immortalized TGMCs (iTGMCs) exhibit long-term proliferative activity in vitro
To establish reversibly immortalized mouse TGMCs that retain longterm proliferation capability, tooth germ mesenchyme was dissected from the mouse mandibular molar teeth, and the TGMCs were isolated and cultured ( Figure 1A). The adherent cells were infected with retroviral SSR#41 vector that expresses SV40 T antigen and selection marker hygromycin ( Figure 1B). After hygromycin selection for 3 days, the surviving cells were replated. The resultant pooled immortalized TGMCs were designated as iTGMCs ( Figure 1C).
The proliferative activities were compared between the iTGMCs and the primary TGMCs. We found that when seeded at a similar initial cell density, the iTGMCs reached confluence faster than the primary TGMCs as demonstrated by crystal violet staining assay ( Figure 1D-a). Quantitative assessment indicated that iTGMCs grown significantly higher numbers of cells at each time point than that of TGMCs ( Figure 1D-b). Taken together, these results demonstrate that the iTGMCs can be stably maintained in culture condition and exhibit a high proliferative rate.

| FLP recombinase can reverse SV40 T antigenmediated immortalization of iTGMCs which also express mesenchymal markers
Immunofluorescence staining was used to assess whether the es- To test whether the SV40 T antigen-mediated immortalization could be effectively reversed in iTGMCs, a recombinant adenoviral vector Ad-FLP was constructed as an effective tool to deliver FLP into iTGMCs. 33,34 We found that the Ad-FLP infected iTGMCs grew at a significantly decreased rate ( Figure 2B). The efficient removal of SV40 T antigen by FLP was further confirmed by qPCR analysis in the Ad-FLP infected iTGMCs ( Figure 2C). Collectively, these results convincingly indicate that the immortalization phenotype of the iTGMC could be effectively reversed by FLP recombinase and the proliferative activity be diminished accordingly.

| The iTGMCs are not tumorigenic in xenograft cell implantation assay in athymic nude mice
As SV40 T antigen is capable of stimulating cell proliferation, 14 we

| BMP9 effectively induces osteogenic/ odontogenic activity of iTGMCs, which requires the participation of canonical Wnt signalling
To test whether the iTGMCs have the osteogenic/odontogenic potential when induced by BMP9, we found that early osteogenic A similar dose-dependent effect was found in the relative ALP activity in iTGMCs when induced by BMP9 ( Figure 3B). Alizarin red staining further demonstrated the late-stage mineralization in BMPstimulated iTGMCs ( Figure 3C).
To further test whether the BMP9-induced osteogenic/odontogenic activity of iTGMCs requires the participation of canonical Wnt signalling as we found in immortalized stem cells of dental apical papilla (iSCAPs), 12 the endogenous β-catenin expression was knockdown by the multiplex siRNAs targeting mouse β-catenin.
QPCR analysis showed that the knockdown efficiency was about 70% ( Figure 4A). We found that cytoplasmic/nuclear accumulation Furthermore, similar results as in mesenchymal stem cell (MSC) 10,11,19,27,35 were found in iTGMCs. Oil Red O staining implied that BMP9 was capable of inducing adipogenic differentiation in iT-GMCs ( Figure 3D). Stronger adipogenic differentiation was found in iTGMC-Ctrl than in iTGMC-KD in a dose-dependent fashion. These results suggest that this effect may also be affected by knocking down of β-catenin expression ( Figure 6C).

| BMP9-induced ectopic bone formation from iTGMCs can be attenuated by β-catenin knockdown
We tested the in vivo effect of BMP9 on ectopic bone formation in iTGMCs via previously established stem cell implantation assay.

| D ISCUSS I ON
Applying the TGMCs isolated from the tooth germ mesenchyme of mouse mandibular molar teeth, we demonstrated that these Even though the search for leading transcriptional regulators of odontogenesis is far from over, the importance of Bmp family members has been recognized by many investigators. Several members in Bmp family have proven to be key participants in inducing terminal differentiation of odontoblast in vitro. [36][37][38] Evidence also shows that Wnt signalling is associated with dentinogenesis. Loss of Wnt signalling would lead to compromised odontoblast maturation and reduced dentin formation, whereas over-activation of Wnt signalling through constitutive stabilization of β-catenin results in excessive dentin and cementum deposition. 39,40 BMP9 (aka, GDF2) is found by us as one of the least characterized yet the most potent osteogenic BMPs. 11, 35 We have recently demonstrated that BMP9 is capable of efficiently inducing bone, cartilage and, to a lesser extent, adipocytes differentiation of iSCAPs. 13 We also found that canonical Wnt signalling acted synergistically on BMP9-induced osteo/odontoblastic differentiation of stem cells of dental apical papilla. 12 It was reported that BMP9-induced osteogenic differentiation in mesenchymal stem cell-like C3H10T1/2 cells

| CON CLUS IONS
In summary, we demonstrated that primary TGMCs can be reversibly immortalized by SV40 T antigen. These iTGMCs exhibit high proliferative activity and maintain long-term cell proliferation under ex vivo culture conditions, which can be reversed by introducing FLP recombinase. We further demonstrated that the iTGMCs retained osteogenic/odontogenic ability upon BMP9 stimulation, but this process required the participation of canonical Wnt signalling both in vitro and in vivo. Therefore, BMP9 has a potential to be applied as an efficacious bio-factor in osteo/odontogenic tissue engineering and tooth engineering. Furthermore, the iTGMCs may serve as an important seed cell resource for translational studies in tooth tissue engineering. were not involved in the study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

CO N FLI C T O F I NTE R E S T
The authors declare no competing conflicts of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supporting the findings of this study are available from the corresponding authors upon request.