EphrinB2 overexpression enhances osteogenic differentiation of dental pulp stem cells partially through ephrinB2-mediated reverse signaling

Background Alveolar bone loss is a frequent occurrence. Dental pulp stem cells (DPSCs) which have invasive accessibility and high osteogenic potential is a promising source for cell-based bone regeneration. EphrinB2 is involved in bone homeostasis and osteogenesis. The aim of this study was to investigate the effect and mechanism of ephrinB2 overexpression on osteogenic differentiation of DPSCs and bone defect repair. Methods EphrinB2 expression was analyzed during osteogenic induction of human DPSCs (hDPSCs). Endogenous ephrinB2 expression in hDPSCs was then upregulated using EfnB2 lentiviral vectors. The effect of ephrinB2 overexpression on osteogenic differentiation capacity of hDPSCs was investigated in vitro, and activation of ephrinB2-EphB4 bidirectional signaling in ephrinB2-overexpressing hDPSCs was detected. In vivo, a canine alveolar bone defect model was established and canine DPSCs (cDPSCs) were cultured, characterized, EfnB2-tranfected, and combined with a PuraMatrix scaffold. Micro-CT analysis was performed to evaluate the therapeutic effect of ephrinB2-overexpressing cDPSCs on bone defect repair. Results EphrinB2 was upregulated after osteogenic induction of hDPSCs. EphrinB2 overexpression enhanced osteogenic differentiation capacity of hDPSCs in vitro. Moreover, p-ephrinB2 instead of p-EphB4 was upregulated by ephrinB2 overexpression, and activation of ephrinB2-mediated reverse signaling promoted osteogenic differentiation of hDPSCs. In a canine bone defect model, ephrinB2 overexpression in cDPSCs significantly improved trabecular bone volume per tissue volume (BV/TV) and trabecular thickness, as demonstrated by radiographic analysis. Conclusions EphrinB2 overexpression enhanced osteogenic potential of DPSCs partially via upregulation of ephrinB2-mediated reverse signaling and effectively promoted alveolar bone defect repair. Electronic supplementary material The online version of this article (10.1186/s13287-019-1540-2) contains supplementary material, which is available to authorized users.

easily and invasively obtained from the discarded or removed teeth, which makes DPSCs a promising source for autologous cell therapy. Although the natural function of DPSCs is to differentiate into odontoblasts producing reparative dentin [12], growing evidence has proved the ability of DPSCs for generating bone-like tissues and repairing bone loss [13][14][15][16][17]. The osteogenic differentiation capability of DPSCs is crucial for their application in bone regeneration and has been well documented by several studies [18][19][20][21]. In addition, it has been reported that DPSCs exhibit higher osteogenic potential along with decreased adipogenic potential compared to BMSCs [22].
Ephrin is the ligand for the tyrosine kinase receptor Eph, and the ephrin/Eph interaction plays a key role in many biological processes such as angiogenesis [23], axon guidance [24], cell migration [25,26], and skeletal patterning [27]. Ephrin ligands are divided into ephrinA ligands (A1-A5), which are membrane-anchored proteins, and ephrinB ligands (B1-B3), which are transmembrane proteins. In general, ephrinB ligands preferentially bind to EphB receptors, with a few exceptions [28,29]. The ephrinB/EphB interaction leads to activation of bidirectional signaling: the forward signaling is mediated by EphB receptors and the reverse signaling is mediated by ephrinB ligands. Both forward and reverse signaling can activate downstream signaling cascades and regulate biological processes.
EphrinB2 is one of three ephrinB ligands, and its involvement in bone homeostasis and osteogenesis was firstly verified by Zhao et al. [30]. They found that ephrinB2 expressed on osteoclasts could stimulate EphB4-mediated forward signaling in osteoblasts, resulting in enhanced osteoblast differentiation [30]. Similarly, exogenous recombinant ephrinB2-Fc has been demonstrated to promote osteoblast differentiation [31,32]. In fact, ephrinB2 and their receptors are co-expressed on osteoblasts [30], and the important role of endogenous ephrinB2 expressed on osteoblasts for regulating differentiation and promoting bone mineralization has also been shown [33,34]. Excluding osteoblasts, ephrinB2 and its receptors are also co-expressed in BMSCs [35] and DPSCs [36]. Previous studies have reported that exogenous ephrinB2-Fc treatment can upregulate osteogenic gene expression and promote mineralized nodule formation in BMSCs [37][38][39] and DPSCs [36], while the role of endogenous ephrinB2 is unclear.
In this study, we aimed to investigate the effect of endogenous ephrinB2 overexpression on osteogenic differentiation of DPSCs. We observed that ephrinB2 overexpression enhanced osteogenic differentiation of DPSCs partially through ephrinB2-mediated reverse signaling. Furthermore, we found that transplantation of ephrinB2-overexpressing DPSCs promoted repairing alveolar defect in a beagle model.

Ethics statement
Collection of human third molars was carried out after receiving approval from the Ethics Committee of Xuzhou Medical University (20161108) and written informed consent from volunteers. The dog experiments in this study were approved by the Experimental Animal Ethics Committee of Xuzhou Medical University (20161108). All experimental procedures were carried out in accordance with relevant guidelines and regulations. Dogs were looked after by professional breeders, and no cases of animal cruelty occurred.

Primary cell culture and identification
Human DPSCs (hDPSCs) were obtained from extracted third molars of healthy volunteers (18-25 years old), and canine DPSCs (cDPSCs) were obtained from sound anterior teeth of six beagles (15-month-old; 10-15 kg; male). Briefly, pulp tissues were isolated from pulp cavities, washed several times, minced, and digested with 3 mg/ml collagenase type I (Gibco, Grand Island, NY, USA) and 4 mg/ml dispase (Gibco, Grand Island, NY, USA) for 1 h at 37°C. The human pulp samples from diverse individuals were mixed and pooled to decrease individual differences, while each dog's DPSCs were cultured separately to avoid the immunological rejection after transplantation. After digestion, cells and remaining tissues were suspended in growth medium containing α-minimum essential medium (α-MEM; Gibco, Beijing, China), 20% fetal bovine serum (FBS; Gibco, South America), 100 U/ml penicillin, and 100 μg/ml streptomycin (Vicmed, Xuzhou, Jiangsu, China) at 37°C and 5% CO 2 . Cells below passage six were used in experiments.
Self-renewal ability was confirmed by a colonyformation assay. Briefly, 400 cells were dispersedly seeded in a 6-cm dish and cultured for 10 days. Then, cell colonies were fixed and stained with crystal violet.

Cell proliferation assay
To assess the influence of EfnB2 transfection on proliferation of hDPSCs, a cell proliferation assay was performed. Related cells (5000 cells/well) were planted into 96-well plates. On days 0, 2, 4, 6, and 8, the medium in each well was changed with 100 μl α-MEM supplemented with 10% Cell Counting Kit-8 (CCK-8; Vicmed, Xuzhou, Jiangsu, China). One hour later, absorbance at 450 nm was measured.

Cell migration assay
To assess the influence of EfnB2 transfection on migration capability of hDPSCs, related cells were plated in the upper chamber of a transwell-24-well permeable support with 0.8-μm pore polyester membrane (Corning, NY, USA) at a density of 5 × 10 4 cells per well. α-MEM was added into the upper chamber, while α-MEM supplemented with 10% FBS was in the lower chamber. The cells migrating to the lower surface of membrane were fixed, stained, and counted after 6 and 9 h.

Stimulation with EphB4-Fc
Recombinant EphB4-Fc chimeras (R&D Systems, Wiesbaden, Germany) were used to stimulate ephrinB2mediated reverse signaling, and IgG-Fc (R&D Systems, Wiesbaden, Germany) was used as the negative control. hDPSCs were cultured in osteogenic medium supplemented with 2 or 4 μg/ml EphB4-Fc or 4 μg/ml IgG-Fc. Alizarin Red S staining was performed on day 14, and osteogenic gene transcription was measured on day 7.

Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Total RNA was isolated using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol and 1 μg RNA was reverse transcribed into cDNA by HiScript Q RT SuperMix for qRT-PCR (Vazyme, Nanjing, Jiangsu, China) in a 20 μl reaction volume. qRT-PCR was performed with UltraSYBR Mixture (Cwbio, Beijing, China) on an ABI7500 quantitative PCR instrument (Applied Biosystems, Darmstadt, Germany). A total of 20 μl reaction system was used: 2 μl cDNA, 100 nM forward and reverse primers, and 10 μl 1× Ultra-SYBR Mixture. Primers are listed in Table 1. The cycle conditions were as follows: initial denaturation at 95°C for 10 min, 40-cycle amplification of 95°C for 15 s and 60°C for 1 min, and melt curve analysis of 95°C for 15 s, 60°C for 1 min, 95°C for 15 s, and 60°C for 15 s. Relative expression was calculated by the comparative cycle threshold method (ΔΔCT) and normalized to β-actin.

Endogenous RhoA activity assay
Active GTP-RhoA was captured using the RhoA Pulldown Activation Assay Biochem Kit (bead pull-down format) (Cytoskeleton, Inc., Japan). Briefly, cell lysates were incubated with GST-rhotekin-RBD beads for 1 h at 4°C. The protein/beads complexes were washed and the bound proteins were resuspended. GTP-RhoA and total RhoA were detected by western blotting with a RhoAspecific antibody.

Cell growth in PuraMatrix
To deliver cells into defect areas, cells were encapsulated in PuraMatrix Peptide Hydrogel (Corning, Bedford, USA). PuraMatrix is a type I self-assembling peptide (SAPs), which can self-assemble into a 3D structured hydrogel under certain physiological conditions. The proliferation of cDPSCs in 0.5%, 0.25%, or 0.125% Pura-Matrix was measured. cDPSCs were suspended in varying dilutions of PuraMatrix in sucrose and added into 96-well plates. Then gelation was induced by careful addition of 100 μl growth medium onto the gel. Medium was changed twice in the next 1 h to equilibrate pH. On days 1, 3, 5, and 7, 10 μl CCK-8 regent was added into 100 μl growth medium and absorbance at 450 nm was measured 1 h later. To assess proliferation of cDPSCs at different densities in 0.25% PuraMatrix, cDPSCs (0.25, 0.5, 1, 2 or 4 × 10 6 cells/ml) encapsulated with 0.25% PuraMatrix were seeded into 96-well plates. Cell growth was measured on days 1, 3, 5, and 7.

Alveolar bone defect model establishment and cell transplantation
The six beagle dogs whose cDPSCs had been isolated before were used to establish bone defect models. All surgical procedures were performed under general anesthesia, which was induced with propofol (5-7 mg/ kg, i.v.) and maintained by isoflurane inhalation (1.5-2% isoflurane/O 2 to effect). Three months after extraction of bilateral mandibular third premolars, horizontal incisions were made between the second and fourth premolars, and mucoperiosteal flaps were elevated. Bilateral four-wall critical-sized alveolar bone defects (4 × 2 × 5 mm, length × width × depth) were created mesial to the fourth premolars and distal to the second premolars with a 1-2-mm distance between defects and premolars. There were four defects for each dog, which were randomly assigned into four groups (n = 3 per group per time point): NC group (without any treatment), PuraMatrix group, Vector-cDPSCs + PuraMatrix group, and EfnB2-cDPSCs + PuraMatrix group. For the latter two groups, cells encapsulated with PuraMatrix were cultured in osteogenic medium for 7 days before transplantation in vivo. The wounds were sutured without stress. At the 4th and 8th week after the surgery, the mandibles were fixed by arterial perfusion of 10% formalin and the dogs were euthanized by an overdose of anesthetic. The Abbreviations: ALP alkaline phosphatase, RUNX2 runt-related transcription factor 2, BMP2 bone morphogenetic protein 2, COL1 collagen type I, OCN osteocalcin, DSPP dentin sialophosphoprotein, DMP1 dentin matrix protein 1, For forward, Rev reverse bone segments containing defect areas were separated, fixed, and scanned by micro-CT.

Micro-CT analysis
For micro-CT evaluation, samples were scanned by a micro-CT scanner (Scanco Medical AG, Brüttisellen, Zurich, Switzerland) at 80 kV, 116 μA. The slice thickness was 25 μm. TRI/3D-BON (Ratoc System Engineering, Tokyo, Japan) was used to perform 3-D structural analysis. The original defect areas, whose borders were visually recognizable, were defined as the region of interest. Trabecular bone volume per tissue volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp), connectivity density (Conn-Den), and structure model index (SMI) were measured.

Statistical analysis
All experiments were repeated at least in triplicate. Collected data were analyzed using SPSS 19.0 (IBM Corp, Armonk, NY, USA), and results were displayed as mean ± standard deviation. Comparison between two groups was analyzed with two-tailed Student's t test and differences among more than two groups were determined by a one-way ANOVA followed by Bonferroni's post hoc test. A value of p < 0.05 was accepted as statistically significant.

EphrinB2 expression in hDPSCs during osteogenic induction
We investigated ephrinB2 expression level during osteogenic induction of hDPSCs. Results showed that ephrinB2 and its phosphorylated form (p-ephrinB2) were upregulated on days 1, 3, and 5 of induction and on days 1 and 3 of induction, respectively (Fig. 1d). EphB4 is a receptor of ephrinB2, and the EphB4-ephrinB2 interaction specifically stimulates osteogenesis, so we measured EphB4 and p-EphB4 expression simultaneously. EphB4 was upregulated on days 1, 3, and 5 of osteogenic induction compared to day 0, while p-EphB4 expression showed no significant change during osteogenic induction of hDPSCs (Fig. 1d).

Recombinant EfnB2-lentivirus transfection and its influence on proliferation and migration of hDPSCs
EphrinB2-Fc has been reported to stimulate osteogenic differentiation of hDPSCs, while the influence of endogenous ephrinB2 overexpression on biological functions of hDPSCs is unclear. To clarify this issue, we constructed ephrinB2-overexpressing hDPSCs (EfnB2-hDPSCs) via recombinant EfnB2-lentivirus transfection, and hDPSCs infected with corresponding control lentivirus were treated as a control (Vector-hDPSCs). Three days after infection, green fluorescence was visible in both EfnB2-hDPSCs and Vector-hDPSCs (Fig. 2a). In comparison with uninfected hDPSCs and Vector-hDPSCs, ephrinB2 mRNA and protein were significantly upregulated in EfnB2-hDPSCs (Fig. 2b, c), which indicated successful establishment of ephrinB2-overexpressed hDPSCs. The consequence of EfnB2 transfection on proliferation and migration of hDPSCs were then analyzed. Results showed that EfnB2-hDPSCs proliferated more slowly than uninfected hDPSCs and Vector-hDPSCs (Fig. 2d), while there was no significant difference in migration capability among uninfected hDPSCs, Vector-hDPSCs, and EfnB2-hDPSCs (Fig. 2e).
EphrinB2 overexpression promoted osteogenic differentiation of hDPSCs partially via enhancement of ephrinB2-mediated reverse signaling The ephrinB2-EphB4 interaction plays a critical role in osteogenesis. Therefore, we hypothesized that ephrinB2 overexpression might elevate the ephrinB2-EphB4 interaction, thereby stimulating osteogenic differentiation of hDPSCs. First, we compared p-ephrinB2 and p-EphB4 in EfnB2-hDPSCs with those in Vector-hDPSCs. Western blot analysis showed that the p-ephrinB2 level was higher in EfnB2-hDPSCs at 0, 2, and 6 h, along with a continuously higher level of ephrinB2 in EfnB2-DPSCs (Fig. 4a). However, EphB4 mRNA and protein expression levels were downregulated in EfnB2-hDPSCs, which was surprising ( Fig. 4a and Additional file 1: Figure S1A, B). May be due to downregulation of EphB4, p-EphB4 was not elevated in EfnB2-hDPSCs (Fig. 4a). We also detected the protein expression level of other EphB receptors, EphB1 and EphB2, which was not affected by ephrinB2 overexpression (Additional file 1: Figure S1C). These results indicated that ephrinB2 overexpression in hDPSCs enhanced ephrinB2-mediated reverse signaling instead of EphB4-mediated forward signaling.
To further test the role of EphB4-mediated forward in the osteogenic differerentiation of ephrinB2-overexpression hDPSCs, we detected active GTP-RhoA that is the downstream target of EphB4-mediated forward signaling in the process of osteogenesis. Results showed that active GTP-RhoA expression was not affected by ephrinB2 overexpression. Protein expression levels were normalized to that of β-actin. Data are shown as mean ± SD. Assays were repeated three times. *p < 0.05 vs. sample on day 0 of osteogenic induction Next, the stimulating effect of ephrinB2-mediated reverse signaling on osteogenic differentiation of hDPSCs was explored. EphB4-Fc only interacts with the ephrinB2 ligand; therefore, it is commonly used to selectively activate ephrinB2-mediated reverse signaling. In our study, western blot analysis showed that ephrinB2-mediated reverse signaling was activated by EphB4-Fc (Fig. 4b). hDPSCs were then treated with 2 or 4 μg/ml EphB4-Fc, and their osteogenic differentiation was tested. Results revealed that there were more calcium deposits in EphB4-Fc treated hDPSCs compared with that in IgG-Fc-treated hDPSCs and uninduced hDPSCs (Fig. 4c). In addition, EphB4-Fc treatment also increased RUNX2, BMP2, and OCN transcription (Fig. 4e).

Culture, characterization, and transfection of cDPSCs and encapsulation of cDPSCs with PuraMatrix
The studies above demonstrated that ephrinB2overexpressing DPSCs possessed higher osteogenic potential in vitro. To determine the effect of ephrinB2overexpressing DPSCs in vivo, a beagle dog experiment was performed. cDPSCs were isolated from canine anterior teeth, and their colony-forming abilities and multidifferentiation potentials were demonstrated (Fig. 5a-c). Immunofluorescence analysis revealed that they were positive for CD73, CD90, CD105, and STRO-1, and negative for CD45 (Fig. 5d). cDPSCs were then infected with EfnB2 lentiviral vectors, and the high transfection efficiency was confirmed by green fluorescence expression (Fig. 5e) and ephrinB2 upregulation (Fig. 5f). Similar to EfnB2-hDPSCs, ephrinB2-overexpressing cDPSCs (EfnB2-cDPSCs) possessed elevated osteogenic potential relative to uninfected cDPSCs and control vector-infected cDPSCs (Vector-cDPSCs), as indicated by Alizarin Red S staining (Fig. 5g).
To effectively transfer cells into the defective regions, cDPSCs were encapsulated in PuraMatrix. Cell proliferation assays revealed that cDPSCs proliferated more slowly in 0.5% PuraMatrix (Fig. 5h). PuraMatrix at 0.125% dilution is very loose and easily disrupted. Therefore, 0.25% was the appropriate PuraMatrix concentration. The proliferation of cDPSCs at different densities hour. Data are shown as mean ± SD. Assays were repeated three times. *p < 0.05 and **p < 0.01 vs. hDPSCs; #p < 0.05 and ##p < 0.01 vs. Vector-hDPSCs in 0.25% PuraMatrix revealed that 4 × 10 6 cells/ml is the optimal cell density, which could guarantee sufficient cell numbers and satisfactory proliferation (Fig. 5i).

EphrinB2 overexpression cDPSCs promoted defect repair in a canine alveolar bone defect model
A canine alveolar bone defect model was constructed (Fig. 6a-d) and the effect of ephrinB2 overexpression cDPSCs on bone damage repair was observed at the 4th and 8th week from the NC, PuraMatrix, Vector-cDPSCs + PuraMatrix, and EfnB2-cDPSCs + PuraMatrix groups. The 2-D images of representative sagittal slices and 3-D reconstruction images showed that the new mineralized tissues were denser with less lacunas and cavities in the EfnB2-cDPSCs + PuraMatrix group compared with the other groups (Fig. 6e). At the 4th week after the operation, analysis revealed that the BV/TV of the EfnB2-cDPSCs + PuraMatrix group was 1.67-fold higher and 3.22-fold higher than that of the Vector-cDPSCs + Pura-Matrix group and NC group, respectively. In addition, the EfnB2-cDPSCs + PuraMatrix group had a higher Tb.Th and lower SMI compared with the other groups ( Fig. 7). At the 8th week, the EfnB2-cDPSCs + PuraMatrix group still had the highest BV/TV (1.30-fold higher than the Vector-cDPSCs + PuraMatrix group; 1.59-fold higher than the NC group). EfnB2-cDPSCs still resulted in a higher Tb.Th compared to Vector-cDPSCs and NC (Fig. 7). Micro-CT analysis demonstrated that there were more and denser mineralized tissues in the EfnB2-cDPSCs + PuraMatrix group compared with the other groups.

Discussion
Successful and sufficient osteogenesis is essential when DPSCs are applied to bone tissue engineering and regeneration. While studies have proven DPSCs enhancing bone regeneration in vitro and in vivo, large volume bone formation by DPSC-based therapy has not yet achieved [12]. It is urgent to find a way to upregulate DPSC osteogenic capacity for broader clinical application. To the best of our knowledge, this is the first study providing evidence that ephrinB2 overexpression offers a novel strategy to enhance osteogenic capacity of DPSCs, thereby promoting DPSC-based bone regeneration. EphrinB2 expressed on osteoclasts or exogenous ephrinB2-Fc have been verified to stimulate osteogenesis of osteoblasts and BMSCs, mainly via interaction with the EphB4 receptor [30,[37][38][39][40]. Endogenous ephrinB2 expressed on osteoblasts also plays pivotal roles in bone formation and mineralization [33,34,41,42]. However, the expression and role of endogenous ephrinB2 in DPSCs under osteogenic differentiation has not been investigated yet. In this study, we observed that ephrinB2, p-ephrinB2 and its receptor EphB4 in hDPSCs were upregulated after osteogenic induction of hDPSCs, which is consistent with a previous report [36]. Meanwhile, p-EphB4 expression was steady during this osteogenesis process. These results proved that ephrinB2 and its receptor EphB4 were co-expressed on hDPSCs and suggested the involvement of ephrinB2 in osteogenic differentiation of hDPSCs. To elucidate the osteogenic effect of endogenous ephrinB2, we upregulated ephrinB2 expression in hDPSCs via lentiviral EfnB2 transfection, and the results revealed that ephrinB2 overexpression significantly increased mineral deposition and upregulated osteogenic genes instead of odontogenic genes in hDPSCs.
Previous studies have indicated that exogenous ephrinB2-Fc enhances osteoblast differentiation via stimulating EphB4-mediated forward signaling [30], and endogenous ephrinB2 might act in a paracrine or autocrine manner on EphB4 in osteoblasts to promote osteogenesis [33]. Thus, we speculated that ephrinB2 overexpressed on hDPSCs might bind to and activate the EphB4 receptor, thereby accelerating osteogenesis of hDPSCs. Contrary to this, there was no upregulation of phosphorylation of EphB4 and EphB4 expression decreased at both mRNA and protein levels in ephrinB2-overexpressing hDPSCs. The negative feedback regulation of the Eph receptor expression by the ephrin ligand has also been reported by previous studies, which revealed that lack of ephrinB1 expression led to post-transcriptional upregulation of the EphB receptor expression through relief of endocytosis and degradation [43]. Furthermore, we found that GTP-RhoA, which is a downstream target of EphB4-mediated forward signaling in the process of osteogenesis, was not affected by ephrinB2 overexpression. These results excluded the contribution of EphB4-mediated forward signaling in the osteogenesis of ephrinB2-overexpressing hDPSCs.
In this study, we identified elevated activation of ephrinB2 reverse signaling in ephrinB2 overexpression hDPSCs. Although the role of ephrinB2 reverse signaling in osteogenesis is unclear, ephrinB1-mediated reverse signaling has been proven to stimulate osteogenic differentiation of BMSCs and osteoblasts by influencing PDZ binding motif (TAZ) transactivation [44,45]. To further assess the effect of ephrinB2-mediated reverse signaling on osteogenesis, we used EphB4-Fc to specifically activate ephrinB2 signaling, since EphB4 only binds with the ephrinB2 ligand instead of other ligands [46,47]. We found EphB4-Fc treatment resulted in more calcium nodule formation in hDPSCs together with an increase in osteogenic gene transcription, which indicated the contribution of ephrinB2-mediated reverse signaling to osteogenic differentiation of hDPSCs. Taken together, these findings suggested that ephrinB2 overexpression promoted osteogenic differentiation of hDPSCs partially through activation of ephrinB2mediated reverse signaling.
Although the ephrinB2/EphB4 interaction is mostly studied, EphB1 and EphB2 have also been reported as likely candidate receptors for ephrinB2 in calvarial bone formation [31]. We detected EphB1 and EphB2 expression patterns in ephrinB2-overexpressing hDPSCs and found that EphB1 and EphB2 expression were not affected by ephrinB2 overexpression. This may partially account for the phosphorylation of ephrinB2 in the absence of EphB4. However, the osteogenic function of EphB1 and EphB2 in DPSCs has not been reported before and whether ephrinB2 overexpression stimulated osteogenesis partially through EphB1 or EphB2 mediated forward signaling requires further research.
To explore the impact of ephrinB2 overexpression DPSCs on bone regeneration in vivo, a canine alveolar bone defect model was established and EfnB2-modified cDPSCs encapsulated with PuraMatrix hydrogel were transplanted into defects. We found that ephrinB2 overexpression cDPSCs accelerated bone defect repair, as indicated by an increased BV/TV and Tb.Th. It has been reported that gene-modified stem cells, such as osteoprotegerin (OPG)-modified periodontal ligament stem cells, promote bone defect repair in vivo [48,49]. There were various methods to evaluate new bone formation, such as radiography, micro-CT, and histological examination. We observed greater and denser new mineralized tissues radiographically by ephrinB2 overexpression cDPSCs. Further studies are needed to histologically verify bone matrix formation instead of dentin matrix formation and vascularization of the newborn tissues [50].

Conclusions
In conclusion, this study demonstrated that ephrinB2 was upregulated in hDPSCs under osteogenic differentiation. EphrinB2 overexpression enhanced osteogenic potential of hDPSCs partially via ephrinB2-mediated reverse signaling in vitro, and ephrinB2-modified cDPSCs accelerated bone regeneration in a canine bone defect model. Thus, EphrinB2 signaling might be a potential target to enhance osteogenesis of DPSCs.