Effect of CD146 on Bone Regeneration by Transplantation of Stem Cells from Human Exfoliated Deciduous Teeth into Mouse Skull Defect

Stem cells from human exfoliated deciduous teeth (SHED) possess bone regeneration ability and may have therapeutic applications. CD146, a cell adhesion protein expressed by vascular endothelial cells, is involved in the osteoblastic differentiation of stem cells. However, the effect of CD146 on SHED-mediated bone regeneration in vivo remains unknown. Hence, in this study we aimed to establish ecient conditions for SHED transplantation. SHED were isolated from the pulp of an extracted deciduous tooth and cultured, and CD146-positive (CD146+) and CD146-negative (CD146−) populations were sorted. Heterogeneous populations of SHED and CD146+ and CD146– cells were transplanted into bone defects generated in the skulls of individual immunodecient mice. Micro-computed tomography was performed immediately post-transplantation and at 4- and 8-weeks thereafter to evaluate bone regeneration. Histological and immunohistochemical assessments were also performed at 8 weeks after transplantation. Micro-computed tomography revealed bone regeneration upon transplantation with CD146+ and heterogeneous populations of SHED, particularly at 8 weeks after transplantation, with signicantly higher bone regeneration observed following transplantation with CD146+ cells. Furthermore, histological and immunohistochemical assessments revealed that CD146+ cells promoted bone regeneration and angiogenesis. Therefore, transplantation of CD146+ SHED into bone defects may serve as a useful strategy for bone regeneration. groups, along with the presence of a large number of CD31+ blood vessels (**p <0.01, *p < 0.05). VEGF promotes macrophage recruitment, angiogenesis, and osteoblast differentiation in bone defects 27,28 . CD31 is an adhesion molecule expressed in platelets and vascular endothelial cells and is a marker of angiogenesis 29 . BMP-2 is a bone morphogenetic protein that induces the differentiation of MSCs into osteoblasts and promotes angiogenesis via VEGF expression 30,31 . Our results indicated activation of angiogenesis in the CD146+ group. Interestingly, CD146 functions as a co-receptor for VEGFR-2 in endothelial cells to enhance VEGF signaling and angiogenesis via nuclear factor-κB 32,33 .Therefore, CD146 may activate VEGF and the nuclear factor-κB signaling pathway to promote angiogenesis and bone regeneration.


Introduction
Investigations of bone regenerative therapy using mesenchymal stem cells (MSCs) have been actively conducted in recent years. [1][2][3][4] Bone regeneration was observed following transplantation of bone marrow mesenchymal stem cells (BMSCs) into the jaw defects of a beagle dog 5,6 . However, bone marrow puncture, which is performed to collect BMSCs, causes pain and gait disturbance in animals 7,8 . Therefore, we focused on stem cells from human exfoliated deciduous teeth (SHED) to develop a non-invasive and e cient bone regeneration treatment strategy. SHED have high proliferative capacity and, like BMSCs, can differentiate into osteoblasts [9][10][11] . Thus, SHED may serve as a useful resource for bone regeneration.
CD146 is a cell adhesion molecule expressed in vascular endothelial cells and smooth muscle cells and is involved in angiogenesis and the osteoblastic differentiation of stem cells 12 . CD146+ cells isolated from a heterogeneous cell population of MSCs exhibit higher bone regeneration capacity than CD146cells [13][14][15] . Thus, CD146 may promote the bone regeneration capacity of MSCs. However, in vivo promotion of bone regeneration by CD146 using SHED has not been reported.
Therefore, in this study, we investigated the effect of CD146 on bone regeneration in vivo and optimized the transplantation conditions for SHED. Results 2.1. Isolation of CD146+ and CD146-cells SHED collected from four patients were cultured, and the cells were sorted to isolate CD146+ and CD146-cells. The sorted CD146+ cells accounted for 83.5%, 74.1%, 89.1%, and 87.9% of the heterogeneous SHED population.

3D evaluation of regenerated bone by μCT
At t 0 , no regenerated bone was observed in the bone defects of all groups. At t 1 and t 2 , shrinkage of the bone defect was observed in the control group; shrinkage of the bone defect with newly regenerated bone was observed at the center of the bone defect in the other groups (Fig. 1a).

H&E and MT staining
Newly formed bone was clearly observed in the CD146+ group, whereas insigni cant bone formation was observed in the SHED group. Large blank areas were observed in the control and CD146-groups (Fig.  2a).
Mature bone was stained as red, whereas collagen bers and osteoids were stained as blue in all groups.
Only a small amount of mature bone was observed in the control and CD146-groups, whereas in the SHED and CD146+ groups, mature bone was widely observed (Fig. 2b). The area ratio of mature bone in each group was as follows: control group, 3.065%; SHED group, 6.654%; CD146+ group, 11.759%; and CD146-group, 6.098%. The area ratio of mature bone in the CD146+ group was signi cantly larger than that in the other groups (**p < 0.01, *p < 0.05) (Fig. 2c).

Immunohistochemistry
Faint brown VEGF-A staining was observed throughout the transplant site in the control and CD146groups. In contrast, in the SHED and CD146+ group, dark brown staining was observed around and at the center of the transplant site (Fig. 3a). The ratio of the area of the VEGF-stained region to the area of the transplant site was as follows in the different groups: control group, 1.008%; SHED group, 2.662%; CD146+ group, 6.977%; and CD146-group, 1.158%. For the ratio of the VEGF-stained area, the CD146+ group showed signi cantly higher values compared to the other groups (**p < 0.01, *p < 0.05). The SHED group showed signi cantly higher values than the control group (*p < 0.05) (Fig. 3b).

Fluorescence immunohistochemistry
BMP-2 was weakly expressed in the upper part of the transplant site in the control group. In contrast, BMP-2 was expressed in the lower part of the transplant site in the SHED group. In the CD146+ group, the entire transplant site exhibited BMP-2 expression, whereas no expression was observed in the CD146group (Fig. 4c).

Discussion
Human dental pulp stem cells and SHED are obtained non-invasively, reducing the physical burden on patients, and these cells have the same regenerative ability as BMSCs [17][18][19][20] . However, transplantation of BMSCs does not result in signi cantly higher bone regeneration compared to autogenous bone graft 21,22 .
Hence, we focused on CD146 to determine cell transplantation conditions that could improve the bone regeneration ability of SHED. CD146 contributes to the bone differentiation and regeneration of MSCs derived from various tissues in vitro, but its effect in vivo remains unclear 14,15,23,24 . In this study, we investigated the role of CD146 in SHED in promoting bone regeneration in vivo.
We observed remarkable bone regeneration in the SHED and CD146+ groups, and the CD146+ group showed signi cantly higher regeneration than the SHED group. CD146+ cells isolated from MSCs have been reported to promote bone formation [25][26][27] . This study revealed that CD146+ cells from SHED also contribute to bone regeneration.
In this study, CD146+ cells comprised approximately 74.1-89.1% of the total heterogeneous cell population in SHED. This indicates that a su cient number of CD146+ cells can be isolated from a heterogeneous cell population of SHED. Thus, our approach may be useful for future clinical applications.
The CD146+ group also exhibited signi cantly higher expression of VEGF and BMP-2 compared with the other groups, along with the presence of a large number of CD31+ blood vessels (**p <0.01, *p < 0.05).
VEGF promotes macrophage recruitment, angiogenesis, and osteoblast differentiation in bone defects 27,28 . CD31 is an adhesion molecule expressed in platelets and vascular endothelial cells and is a marker of angiogenesis 29 . BMP-2 is a bone morphogenetic protein that induces the differentiation of MSCs into osteoblasts and promotes angiogenesis via VEGF expression 30,31 . Our results indicated activation of angiogenesis in the CD146+ group. Interestingly, CD146 functions as a co-receptor for VEGFR-2 in endothelial cells to enhance VEGF signaling and angiogenesis via nuclear factor-κB 32,33 .Therefore, CD146 may activate VEGF and the nuclear factor-κB signaling pathway to promote angiogenesis and bone regeneration.
Additionally, VEGF and BMP-2 interact in bone regeneration via MSCs 34,35 . The association of BMP with the BMP receptor results in phosphorylation of Smad1, Smad5, Smad8 and regulates the expression of target genes such as VEGF and Runx2 36 . In contrast, binding of VEGF to the VEGF receptor activates the mitogen-activated protein kinase/extracellular signal-regulated kinase pathway, culminating in increased BMP-2 expression 34,35 . Thus, VEGF and BMP-2 interact and contribute to bone regeneration in MSCs. In this study, advanced bone regeneration was promoted in the CD146 + group through this mechanism.
However, the detailed effect of CD146+ cells in SHED on bone regeneration remains unclear. Therefore, it is necessary to determine the signal transduction pathways related to CD146+ cells in SHED, VEGF, and BMP-2. In addition, studies in humans are warranted to con rm the ndings observed in the animal model for future clinical applications.
In conclusion, we demonstrated that CD146+ cells present in SHED are more useful for in vivo bone regeneration than a heterogeneous population of SHED. Additionally, CD146 and VEGF may be intricately involved in bone regeneration by SHED, and further studies are required to determine their precise roles. Our study demonstrates the immense potential for the development and clinical application of SHED transplantation in bone regeneration therapy.

Cell isolation and culture
Pulp tissue was collected from deciduous teeth extracted from patients being treated at the Department of Orthodontics, Hiroshima University Hospital. SHED were isolated and cultured using previously described methods 16 . The guidelines pertaining to epidemiological research at Hiroshima University Hospital were strictly followed (Approval No. E-20-2). Informed consent was obtained from all participants.

Fluorescence-activated cell sorting
A heterogeneous population of the 3rd passage cells obtained from SHED was sorted to isolate CD146+ and CD146-cells using a FACS Aria II cell sorter (BD Biosciences, San Jose, CA, USA). The cells were stained with PE-conjugated Mouse Anti-Human CD146 (BD Pharmingen, San Jose, CA, USA) or PEconjugated Mouse IgG1, κ Isotype control (BD Biosciences). The numbers of CD146+ and CD146-cells in SHED were analyzed using FlowJo software (TreeStar, Ashland, OR, USA).

SHED transplantation into a mouse model of bone defect
CD146+ cells, CD146-cells, and the heterogeneous SHED population were collected from the same patient and isolated by cell sorting. As SHED were of human origin, 6-week-old immunode cient mice (BALB/c-nu; Japan Charles River International Laboratories, Inc., Yokohama, Japan) were used to avoid immunogenic and graft rejections. Non-uorescent alfalfa-free solid food (D10001; AIN-76A; Research Diet, EPS Masuzo, New Brunswick, NJ, USA) was administered for one week before initiating the experiment. The cells were seeded with an atelocollagen sponge (Mighty ® ; ⌀ 5.0 × 1.5 mm; Koken, Tokyo, Japan) and transplanted into mouse skull defects (diameter, 5.0 mm) under general anesthesia in accordance with a previously described study 17 . Anaesthesia consisted of midazolam (4 mg/kg; Sandoz), medetomidine (0.3 mg/kg; Orion Corp.) and butorphanol (5 mg/kg; Meiji Seika Pharma Co., Ltd.). The following groups were de ned according to the implanted materials: (a) CD146+: CD146+ cells immediately after atelocollagen transplantation (t 0 ), as well as at 4 (t 1 ) and 8 weeks after transplantation (t 2 ). The CT image resolution was 512 × 512 pixels, and the slice width was 35 μm. The transplant site was indicated in 3D data using ZedView (Lexi, Tokyo, Japan). The volume of regenerated bone was measured using Rapidform (Inus Technologies, Seoul, Korea) and FreeForm (SensAble Technologies, Wilmington, MA, USA).

Histological evaluation of regenerated bone
Eight weeks after atelocollagen transplantation, the immunode cient mice were euthanized and the parietal bone was extracted. The tissue specimens were decalci ed using 14% ethylenediaminetetraacetic acid, embedded in para n, and sectioned (thickness, 7 μm) along the sagittal plane. Staining was performed as described by Hiraki et al 17 . After staining, tissues were imaged and observed using a BZ-X800 uorescence microscope (Keyence, Osaka, Japan).

Hematoxylin and eosin (H&E) and Masson's trichrome (MT) staining
Tissue sections were depara nized and dehydrated, and then subjected to H&E staining and MT staining. MT staining was performed to detect mature bone. A section with atelocollagen (diameter, 5.0 mm) was used as the center of the transplant site. Using the BZ-II image analysis application (Keyence), the ratio of the area of mature bone to the area of the transplant site was calculated.

Immunohistochemistry
Immunostaining for vascular endothelial growth factor (VEGF) and CD31 was performed to examine angiogenesis using a slice from the center of the transplant site. The ratio of the area of the VEGF-Astained region to the area of the transplant site and number of CD31+ blood vessels was calculated using the BZ-II image analysis application.

Statistical analysis
All data are presented as the mean ± standard deviation. Signi cant differences among groups were analyzed using the Bonferroni method in BellCurve® for Excel (SSRI; Tokyo, Japan). Results with p < 0.05 and p < 0.01 were considered statistically signi cant.

Declarations Data Availability
The data that support the ndings of this study are available from the corresponding author, upon reasonable request. gave critical comments on the draft of the manuscript. All authors reviewed the manuscript.

Competing interests
The authors declare no competing interests. Figure 1 Regenerated bone evaluation by micro-computed tomography. No notable difference was detected between groups at t0. In the control-and CD146-group, a small reduction in the bone defect and regenerated bone was observed. In the SHED-and CD146+ group, several regenerated bones were found in the center of the bone defect (a). The SHED and CD146+ groups exhibited signi cantly higher regenerated bone mass compared to the control-and CD146-groups in the t0-t1 and t1-t2 periods.

Figure 2
Histological evaluation of regenerated bone. H&E staining showed calci ed-like tissue in the CD146+ group but not in the control group (a). Masson's Trichrome staining revealed intensely stained sites showing mature bone in the SHED and CD146+ groups compared to in the control and CD146-groups (b). The proportion of areas stained red by Masson's Trichrome staining was signi cantly higher in the CD146+ group than in all other groups (c). (n = 5 for each group, ** p < 0.01, * p < 0.05). Scale bars = 500 μm.

Figure 3
Immunohistochemical analysis of VEGF expression. VEGF immunohistochemical staining was extensive and stronger in the SHED-and CD146+ groups than in the control and CD146-groups. In the CD146+ group, many stained sites were found around atelocollagen and in the lower central region (a). The proportion of VEGF-stained area was signi cantly higher in the CD146+ group than in all other groups (b). (n = 5 for each group, ** p < 0.01, * p < 0.05). Scale bars = 500 μm. Analysis of CD31 and BMP-2 expression. Blood vessels stained using the anti-CD31 antibody were observed in all groups. Several large blood vessels were observed in the CD146+ group (a). The number of CD31+ blood vessels was high in the CD146+ group (b). Representative section depicting immunostaining with BMP-2 alone (upper panel), and co-staining with BMP-2 and DAPI (lower panel). A few sites were stained with BMP-2 in the control and CD146-groups. In the SHED group, only the lower center of the transplant site exhibited BMP-2 expression, whereas the entire transplant site in the CD146+ group showed BMP-2 expression (c). (n = 5 for each group, * p < 0.05). Scale bars = 500 μm.