Combining Bone Collagen Matrix with hUC-MSCs for Application to Alveolar Process Cleft in a Rabbit Model

Most materials used clinically for filling severe bone defects either cannot induce bone re-generation or exhibit low bone conversion, therefore, their therapeutic effects are limited. Human umbilical cord mesenchymal stem cells (hUC-MSCs) exhibit good osteoinduction. However, the mechanism by which combining a heterogeneous bone collagen matrix with hUC-MSCs to repair the bone defects of alveolar process clefts remains unclear. A rabbit alveolar process cleft model was established by removing the bone tissue from the left maxillary bone. Forty-eight young Japanese white rabbits (JWRs) were divided into normal, control, material and MSCs groups. An equal volume of a bone collagen matrix alone or combined with hUC-MSCs was implanted in the defect. X-ray, micro-focus computerized tomography (micro-CT), blood analysis, histochemical staining and TUNEL were used to detect the newly formed bone in the defect area at 3 and 6 months after the surgery. The bone formation rate obtained from the skull tissue in MSCs group was significantly higher than that in control group at 3 months (P < 0.01) and 6 months (P < 0.05) after the surgery. The apoptosis rate in the MSCs group was significantly higher at 3 months after the surgery (P < 0.05) and lower at 6 months after the surgery (P < 0.01) than those in the normal group. Combining bone collagen matrix with hUC-MSCs promoted the new bone regeneration in the rabbit alveolar process cleft model through promoting osteoblasts formations and chondrocyte growth, and inducing type I collagen formation and BMP-2 generation.


Introduction
An alveolar process cleft is a common clinical diagnosis of alveolar bone defects caused by birth defects, trauma, or inflammation, which severely affects a patient's facial and oral functions. It also greatly hinders the patient's subsequent denture repair. The goal of alveolar bone reconstruction is to generate physiologically functioning bone and eventually restore facial morphology and occlusal function. Currently, autogenous bone transplantation remains the clinical gold standard and, commonly autogenous bone includes the skull [1], cancellous bone [2], ilium [3], etc. In addition, dozens of bone substitutes such as allogeneic [4], alloplastic [5], and tissue-engineered [6] bones have been applied in clinical practice.
The ideal bone repair material must exhibit numerous properties. On the one hand, it should have a wide range of readily available sources, and exhibit high bio-compatibility and safety. On the other hand, it must also exhibit certain plasticity and bio-degradability. In addition, it must exhibit good bone conductivity and inductance. The collagen matrix used in this study was prepared by refining bovine cancellous bone through a series of de-cellularizaisation and degreasing processes. Its main components are hydroxyapatite and collagen. The material not only exhibited greatly reduced immunogenicity, but also maintained the natural bone structure. The material has a suitable pore size for facilitating the growth of cells and blood vessels. Because the bone re-generation environment is complex, it is difficult to meet bone re-generation requirements using only a bone collagen matrix. Studies have shown that human umbilical cord mesenchymal stem cells (hUC-MSCs) play an important role in inducing bone regeneration [7][8][9]. HUC-MSCs have the advantages of wide availability, rapid proliferation and low immunogenicity [9]. Although the collagen matrix used in this study have been used in clinical bone repair research, the effects of combining bone collagen matrix with hUC-MSCs on the alveolar process cleft remain unclear.
Developing a suitable animal model of an alveolar bone defect according to the clinical characteristics of oral bone grafting surgery is the basis and focus of studying alveolar bone repair and evaluating the osteogenic ability of bone grafting materials. Although adult monkey [10], beagle [11], miniature pig [12], rabbit [13], rat [14] and other experimental animals had been used to develop bone defect models in different parts of the animals' jaws to evaluate of the effects of tissue-engineered graft on bone repair ability, the animal models of the alveolar process cleft used to estimate the ability of tissue-engineered materials to induce regeneration were mainly established using adult rats [15][16][17][18], and hard to reflect status of human congenital cleft lip and palate. And most of the models have a small area of bone damage, and easy to repair. In our previous study, we had established an animal models of the alveolar process cleft with young Japanese white rabbits (JWRs) [19]. The larger the rabbit grows, the larger the area of the bone defects of alveolar process clefts. Moreover, our alveolar process cleft model lacked bone on three sides and was difficult to induce bone regeneration [19]. If the ability of tissue-engineered materials to induce regeneration is weak, the defect area will be larger as the rabbit grows up. It is unclear whether combining bone collagen matrix with hUC-MSCs can substantially induce bone regeneration at the defect site of the alveolar process clefts established by young rabbits.
In this study, the animal models of the alveolar process cleft with young rabbits [19] were used to analyze the ability of combining bone collagen matrix with hUC-MSCs to induce bone regeneration. This study may provide a new approach for bone repair of congenital cleft palate.

Isolation and Culture of hUC-MSCs
Human umbilical cords were obtained from normal pregnant women with term delivery in Haidian Maternal and Child Health Hospital (Beijing, China). The hUC-MSCs were obtained by tissue mass culture. The study was approved by Ethics Committee of Research Institute for Family Planning (No. 2015-16). The informed consent was got from all participants. The blood vessels of the human umbilical cord were removed, and then the cord was cut into approximately 1 mm 3 of tissue. Until tissue block was attached to the bottom of the cell culture dish, α-Minimal Essential Medium (α-MEM) culture medium containing 10% fetal bovine serum, 100 IU/mL penicillin, and 10 mg/ mL streptomycin was added to the dishes and cultured in a carbon dioxide incubator. The cell growth was monitored, and the tissue was removed when the cells had radially covered the surface of the culture plate. The surface antigens of within 5 passages cells were detected using a human MSC assay kit (BD Biosciences Franklin Lakes, NJ, USA). The surface antigens include CD11b, CD19, CD45, CD34, CD44, CD73, CD105, CD90 and HLA-DR. These antibody sources were mouse.

Preparation of Implant Materials
The bone collagen matrix is a heterogeneous bone matrix prepared from bovine cancellous bone refined in a series of processes, thereby retaining its natural three-dimensional porous structure. The main components of the bone collagen matrix are hydroxyapatite and collagen. The collagen membrane covers the injury site implanted with the bone collagen matrix. The collagen membrane was approximately 0.8 mm thick. Before use, the collagen membrane was cut into small pieces the same size as the defect area. The bone collagen matrix and collagen membrane were both provided by Yantai Zhenghai Bio-technology Co., Ltd (Shandong, China). The hUC-MSCs within five passages were harvested and suspended in phosphate buffer solution (PBS) at the concentration of 10 7 cells/ mL.

Groups and Treatment
In this study, 48 female JWRs (bodyweight: 2000 ± 300 g, about 2-month-old) were used. The JWRs were purchased from Huafukang Bio-technology Co., Ltd (Beijing, China). All the animals were kept in the animal room at the National Research Institute for Family Planning and were provided with clean water and fresh food. The indoor conditions were as follows: temperature (24 ± 1 ℃); air humidity (55% ± 5%); noise (less than 60 dB); lighting time (12 h). The room always was kept clean, dry and ventilated. The experimental design and implementation were approved by the local research and ethics committee.
The model of alveolar process cleft was established according to our previous method [19]. Briefly, the rabbits were anesthetized by an intravenous injection of serazine hydrochloride into the ear margin (concentration: 1-2 mg/ kg; Fig. 1Cc). The rabbit's mouth was opened with a pair of tweezers and disinfected by 75% alcohol (Fig. 1Cd). An incision was created on the left maxilla along the mucosa of the alveolar bone surface towards the gingival margin to expose the alveolar bone (Fig. 1Ce). Bone tissue 1.00 ± 0.20 cm long × 0.50 ± 0.05 cm wide × 0.40 ± 0.04 cm high was removed from the alveolar process with a rongeur to form the alveolar process cleft (Fig. 2Af). JWRs were randomly assigned to each of four groups: normal, control, material, and MSCs. The flowchart of the different groups was showed in Fig. 1D. Rabbits in the normal group were fed normally without any treatment. In the control group, the bone defects of the alveolar process cleft were filled with nothing. In the material group, the bone defects of the alveolar process cleft were filled with the same volume of the bone collagen matrix. In the MSCs group, bone collagen matrix was inoculated with 10 7 cells/mL hUC-MSCs suspension and cultured in a carbon dioxide incubator for 0.5 h. The bone defects of the alveolar process cleft were filled with combination of bone collagen matrix with hUC-MSCs. And then the collagen membrane was used to directly cover the injury in control, material and MSCs group to prevent the influence of soft tissue such as muscle on the repair of bone tissue injury, and the muscles and skin at the injured site were sutured. Conventional anti-inflammatory therapy (i.e. penicillin potassium: 4000 IU/kg/days) was then given to all the rabbits for 1 week to prevent postoperative infection. The rabbits were removed from each group at 3 and 6 months after the surgery. The rabbits were euthanised by intravenous injection of an overdose of serazine hydrochloride in the ear margin. The skull was removed, and partial skulls were separately placed in 4% NaOH and 95% ethanol for 24 h, and then taken out and placed in an oven to dry. Finally, the appearance of the skull tissue was photographed and marked. Partial fresh skull tissue was first fixed in 4% paraformaldehyde for 24 h, de-calcified with 10% ethylenediaminetetraacetic acid (EDTA) for 1 month, and finally embedded in paraffin. The paraffin sections (6 μm) were prepared using a rotary microtome (Leica RM2245, Leica, GmbH, Germany). For histology staining, the paraffin sections were first de-paraffinized in xylene and then re-hydrated in graded alcohol solutions to pure water.

X-ray Analysis
X-ray analysis was performed using a SOFTEX® M-60 X-ray machine (Kanagawa, Japan) operated at 80 kV and 125 mA on the tissue samples prepared from the surgical sites in each group at 3 and 6 months after the surgery at the Beijing Ornamental Animal Hospital. The exposure time was 40 millisecond (ms).

Blood Biochemical Analysis
Three rabbits were randomly selected from each group at 3 and 6 months after the surgery. Approximately 3.5 mL of blood was collected by ear-vein sampling. Some of the blood was used for direct detection and the remainder for serum separation. The blood routine, liver function, renal function and serum bone Gla protein (BGP) of the rabbits were all measured. The routine blood tests were performed using an LH 750 automated haematology analyser (Beckman Coulter, USA). The blood biochemistry test was performed using a DXC 800 automated biochemical analyser (Beckman Coulter, USA).

Bone Formation Rate Analysis
Three rabbit skull models were randomly made in each group at 3 and 6 months after surgery, respectively. The lateral view of the surgical side of the rabbit was obtained by photographing. Then, the actual bone defect area and the total bone defect area were measured by Image J, and the difference value between the two was the area of new osteogenesis. The percentage of osteogenic area to defect area is the bone formation rate.

Micro-focus Computed Tomography (Micro-CT) Analysis
Three rabbits were randomly selected from each group at 3 and 6 months after the surgery. A skull model was prepared. The general appearance of the skull was recorded laterally and vertically. The bone re-generation in the material transfer area were scanned using Siemens Inveon micro-CT scanner system and data were analyzed by Inveon™ Research Workplace 4.2 (Siemens, Erlangen, Germany). The repair of the maxillary region in each group was observed stereoscopically. In order to analyse the the quality of the newly formed bone in the defect area, three 1 mm 3 areas were randomly selected from the centre of the cubic bonedefect area in each group to record the bone trabeculae and mineral density.

Hematoxylin and Eosin (HE) Staining
HE staining was used to observe the tissue morphology. The nucleus and other areas were stained blue and red by the hematoxylin and eosin respectively. The slides were observed using TE2000-U inverted phase contrast microscope (Nikon, Tokyo, Japan) and scanned by Pannoramic MIDI scanner (3DHISTECH, Budapest, Hungary).

Periodic Acid-Schif (PAS) Staining
PAS staining was used to assess the glycogen concentration and was performed using a commercial kit (Senbeijia (e-f) Incising alveolar process mucosa and removing the bone tissue of alveolar process; (g) Measuring the size of the bone that has been removed; (h) Filling bone collagen matrix without/with hUC-MSCs into bone defects; (i) Covering the injury site with the collagen membrane; (j) suturing the muscles and skin. D The flowchart of the different groups. Rabbits were fed normally without any treatment in normal group Biological Technology Co. Ltd, NanJing, Jiangsu, China) according to the manufacturer's instructions. Briefly, the sections were incubated in the dark first with periodic acid solution and then with the Schiff reagent for 5 and then 20 min at room temperature, respectively. The sections were then counterstained with Lillie-Mayer's hematoxylin and were observed using Leica DMIL LED microscope (Leica, Fig. 2 The general appearance of the skull. A The general appearance of the skull 3 months after surgery. B the general appearance of the skull 6 months after surgery. The soft tissues of the skulls were eliminated by soaking them in 4% NaOH and 95% ethanol for 24 h, respectively. The red box shows the postoperative appearance of the transplanted area. The rabbits were randomly assigned to four groups: normal group, control group, material group and MSCs group. In the normal group, rabbits were fed normally without any treatment. In the control group, only collagen membrane was used to cover injury site. In material group, bone collagen matrix was implanted in injury site and collagen membrane covered the material. In MSCs group, combining bone collagen matrix with hUC-MSCs was implanted in injury site and collagen membrane covered the material. C and D The bone formation rate at 3 months and 6 months after the surgery, respectively. *P < 0.05, **P < 0.01 Wetzlar, Germany). The cartilage structure can be dyed either deep purple or red.

Sirius Red Staining
Sirius red staining was used to detect different collagen fibers and was performed using a commercial kit (Senbeijia Biological Technology Co. Ltd, Nanjing, Jiangsu, China) according to the manufacturer's instructions. Briefly, the sections were incubated with sirius red for 1 h at room temperature and counter-stained with Lillie Mayer's hematoxylin. The slides were observed by a polarized light microscope (Putuo XS-18C, Shanghai, China). Sirius red can dye type I collagen bright orange. Image J software was used to calculate the relative percentage of the type I collagen staining area under different fields in each group.

Bone-Specific Alkaline Phosphatase (ALP) Assay
Bone-specific ALP is an osteoblasts phenotypic marker, which can directly reflect the activity or function of osteoblasts. ALP calcium-cobalt staining was used to detect the bone-specific ALP content by a commercial kit (KeyGEN BioTECH Co.Ltd, NanJing, Jiangsu, China) used according to the manufacturer's instructions. Briefly, the sections were incubated with ALP solution for 5 min and then with cobalt nitrate solution for 2 min at room temperature. The sections were then counter stained with eosin. The slides were observed by the Leica DMIL LED microscope (Leica, Wetzlar, Germany). The osteoblasts can be dyed black.

Immunohistochemical Staining for Bone Morphogenetic Protein 2 (BMP-2)
After the sections were de-waxed, re-hydrated and subjected to heat-induced epitope retrieval, they were incubated with rabbit anti-BMP-2 polyclonal antibody (1:1000; Abcam, ab6285, Cambridge, UK), and then incubated with HRPconjugated goat anti-rabbit IgG (1:5000; Jackson Immu-noResearch Laboratories, West Grove, Pennsylvania, USA). Normal rabbit serum was used as blocking solution. The sections were developed with diaminobenzidine tetrahydrochloride (DAB) and counterstained with hematoxylin. Samples were viewed at Leica DMIL LED microscope (Leica, Wetzlar, Germany). To evaluate the specificity of the antibodies, negative control staining was performed by replacing the primary antibody with normal rabbit serum (Fig. S1). The methods of image analysis and signal quantification referred to many other researchers [20,21]. Briefly, to avoid possible bias in light intensity during image capture, the image brightness was normalized to background level through white balance. All images are photographed at the same conditions by using the same microscope and camera (Leica DMIL LED, Wetzlar, Germany). The mean optical densities (MOD) were measured by Image J software. 24-bit RGB images were first converted into 8-bit grayscale images. The threshold optical density discriminating negatively from positively stained cells was set after evaluating several fields of the negative control slides at which the best discrimination between cell nuclei and cytoplasm, or background. The areas of BMP-2 signals and the entire field were measured. The percentage of MOD of BMP-2 signals were expressed as ratio of area of BMP-2 signals vs total area. Three fields were randomly selected for each section. Three different sections were chosen for the same animal. There are at least three animals in each group.

Immunofluorescence for Ki67
The sections were incubated with mouse anti Ki-67 monoclonal anti-body (1:200; Abcam, ab15580, Cambridge, UK), and then incubated with fluorescein isothiocyanate (FITC)conjugated goat anti-mouse IgG (1:300; Jackson Immu-noResearch Laboratories, West Grove, Pennsylvania, USA). Goat serum was used as blocking solution. Nuclei were stained with Hoechst 33258. Sections were viewed under a laser-scanning confocal microscope (ZEISS LSM 710 META, Oberkochen, Germany). Three different sections were chosen for the same animal. There are at least three animals in each group. The proliferative cells labeled by Ki67 were counted in at least three different optical fields selected in a random manner, and counted at least 100 cells for each section. If the total number of Ki67 positive cells for each section did not reach 100, more fields would be randomly selected until the total number of Ki67 positive cells reached 100. All the cells were counted from the same fields. The percentage of proliferative cells for each section were expressed as ratio of the total number of Ki67 positive cells vs the total number of all cells. Each sample was observed at least three sections and the average value was used for the statistical analysis. Each group has at least three samples for proliferation analysis.

Detection of Apoptosis Assay by TdT-Mediated dUTP Nick-end Labelling (TUNEL)
TUNEL assays were prepared using an in-situ Cell Death Detection Kit, Fluorescein (Roche, Mannheim, Germany) according to the manufacturer's instructions. To correlate the TUNEL assay results with the nuclear morphology, the sections were counterstained with hematoxylin. The number of apoptotic cells was counted in 3 randomly selected optical fields (magnification × 400). The samples were by Leica DMIL LED microscope (Leica, Wetzlar, Germany). At least 100 randomly selected cells in each sample were evaluated for apoptosis in the different optical fields (magnification × 400). The results were expressed as the ratio of TUNEL-positive cells to the total number of cells. Each sample was observed at least three times. The TUNEL analysis was performed 3 times for each animal, and the average value was used for the statistical analysis.

Statistical Analysis
The results are presented as mean ± standard deviation (SD), and P < 0.05 is considered statistically significant. The data from the blood, HE staining, Sirius red staining, PAS staining, ALP staining immunohistochemical tests and TUNEL were statistically analyzed using one-way analysis of variance (ANOVA). The difference of the bone trabeculae and mineral density between injury side and normal side was analyzed using Student's t-test.

HUC-MSCs Express Specific Surface Antigens
The hUC-MSCs used in this study were previously identified by our laboratory [22]. Before the experiment, the surface antigen of hUC-MSCs was identified again. The expression of positive markers for MSC such as CD73, CD105, CD90, and CD44 of the hUC-MSC were detected almost above 95%. The expression of negative markers for MSC such as CD34, CD11b, CD19, CD45, and HLA-DR were all below 5% (Fig. S2). These facts are consistent with the characteristics of MSCs [23].

Morphometric Analysis of Alveolar Process Tissues
The maxilla in the normal group is smooth and without any defect ( Fig. 2A). At 3 months after the surgery, little repair was found in the bone defect site in the control or material group, while partial repair was found in the bone defect site in the MSCs group ( Fig. 2A). At 6 months after the surgery, there was slightly repair in the bone defect site in the control or material group. In MSCs group, the degree of bone repair in the bone defect site was close to the normal bone observed from the external appearance (Fig. 2B). The bone formation rate in MSCs group was significantly higher than that in control group at 3 months (P < 0.01) and 6 months (P < 0.05) after the surgery, respectively ( Fig. 2C and D). Although bone collagen matrix alone did not significantly increase the bone defect area, it induced slight bone regeneration.
To evaluate the degree of bone repair in the alveolar process cleft, X-ray analysis was used for the preliminary analysis in which bone density was positively correlated with brightness ( Fig. 3). In the normal group, the interior of the maxilla had a uniform bone matrix with high bone density and brightness. Compared with the brightness (an indicator of bone density) of the normal group, those of the control, material, and MSC groups were the lowest, mid-range, and highest at 3 months after the surgery. At 6 months after the surgery, the brightness of the MSCs group remained higher than those of the other groups. Therefore, the osteogenic effect of combining the bone collagen matrix with the hUC-MSCs may be observably stronger than that obtained for only the bone collagen matrix.
To further analyze the absorption of the bone materials and the bone formation of the alveolar process cleft, micro-CT scans were conducted to detect the skull tissues from Fig. 3 Bone repair in the alveolar process cleft was detected by X-ray analysis. a-d X-ray analysis of the skull 3 months after surgery; e-h X-ray analysis of the skull 6 months after surgery. The red box is the surgical site. After the rabbits were anesthetized, the left maxilla was scanned using a SOFTEX® M-60 X-ray machine different perspectives at 3 (Fig. 4A) and 6 months (Fig. 4B) after the surgery. The normal maxilla has a smooth appearance, no defect, and high brightness. The internal structure of normal maxilla is uniform bone matrix without defect and scar structure. The injury and normal sides were compared in each group. The micro-CT scans were consistent with the visual inspection. Compared with the control or material group, the MSCs groups had more visible new bone formation in the bone defect site at 3 months after surgery (Fig. 4A) and were close to normal bone morphology at Fig. 4 The bone formation of the alveolar process cleft was detected by micro-CT system. A CT images from different angles 3 months after surgery. B CT images from different angles 6 months after surgery. The red box is the surgical site. The surgical site of the skull was scanned using a micro-CT system for small-animal imaging. Two-and three-dimensional image processing, the bone trabeculae and bone density analysis were performed using the scanning results. The injury side of each group was compared with that of the control group and the difference was denoted by *. The difference between the injury side and the normal side was expressed as #. The difference between the material group and the MSCs group on the injury side was expressed by *. *P < 0.05, **P < 0.01, ## P < 0.01, ## P < 0.01 6 months after surgery (Fig. 4B). The percentage of the bone density and trabecular bone were then calculated for each group (Fig. 4C). At 3 months after the surgery (Fig. 4Ca), the percentage of the trabecular bone in the MSCs group (60.916 ± 2.072%) was the highest, followed by the material group (52.647 ± 2.857%) and the control group was almost zero. The results obtained at 6 months after the surgery were like those obtained at 3 months after the surgery ( Fig. 4Cc; MSCs group: 73.338 ± 2.132%, material group: 61.180 ± 4.241%, control group: 0). The results showed that the percentage of bone density in the control group (0.466 ± 0.110%) was the lowest and that the material (53.013 ± 2.002%) and MSCs groups (64.337 ± 2.011%) were negligibly different 3 months after the surgery (Fig. 4Cb). The percentage of the bone density in the MSCs group (82.936 ± 2.112%) was significantly higher than that in the other two groups (material group: 43.858 ± 0.522%, control group: 1.29 ± 0.522%) at 6 months after the surgery (Fig. 4Cd). Therefore, the osteogenic ability and the quality of the newly formed bone of combining the bone collagen matrix with the hUC-MSCs was higher than that of only the bone collagen matrix.

Microscopic Structure of Bone Defect Site Detected by HE Staining
The results of HE staining showed that the maxillary bone tissue in the normal group was uniform bone matrix, without trabecular bone and vacuolated structure, and had a few cells ( Fig. 5Aa1 and Ae1). HE staining showed that no significant bone repair was observed in the control group ( Fig. 5Ab1 and Af1) at both 3 and 6 months after the surgery compared to the normal group. In the material group, a few bone fibers, some bone marrow and trabeculae, and numerous cavitation structures were observed in the damaged area at 3 and 6 months after the surgery (Fig. 5Ab1 and Af1). In the MSCs group, no cavitation was observed in the damaged area at 3 months after the surgery, and numerous bone trabeculae and fibrous tissues were observed (Fig. 5Ad1). At 6 months after the surgery, the bone trabeculae in the damaged area had connected to form new bone tissue (Fig. 5Ah1). The results of HE staining showed that the trabeculae forming new bone showed irregular filamentous or reticular structure, and the surface of the trabeculae was surrounded by osteoblasts ( Fig. 5Ad1 and Ah1). Existing bone is a large, regular, dense matrix of bone and contains a small number of bone cells (Fig. 5Aa1 and Ae1). All these facts indicate that combining bone collagen matrix with hUC-MSCs can induce bone regeneration better than bone collagen matrix alone.

Measurement of Serum Bone Gla Protein (BGP)
BGP is a bone-specific-protein synthesized by osteoblasts and is incorporated into the bone matrix. The serum BGP results showed that the BGP concentration in the control and material groups was approximately equal to that in normal group and significantly enhanced in the MSCs groups at 3 months after the surgery (P < 0.05; Fig. 5B1). At 6 months after the surgery, the BGP concentration was similar in normal, control and material groups and had an upward trend in MSCs group (Fig. 5B2). These results show that the osteogenic ability of combining the bone collagen matrix with the hUC-MSCs was higher than that of only the bone collagen matrix.

Effects of Bone Collagen Matrix Combined with hUC-MSCs on Osteoblasts, Collagen, and Bone-Formation-Associated Saccharides
ALP staining was used to detect the black-stained osteoblasts. The results showed that there was a small amount of black staining in the maxillary bone in the normal group, suggesting that there were a few osteoblasts in normal bone tissue ( Fig. 6Aa and Ae). There was no obvious black staining in the control group ( Fig. 6Ab and Af). In the material group, black appeared at the edge of the trabecular bone at 3 months after the surgery (Fig. 6Ac), and was also observed at 6 months after the surgery (Fig. 6Ag). In the MSCs group, there was obvious black staining at the edge of the trabecular bone at 3 months after the surgery (Fig. 6Ad), and remained at 6 months after the surgery (Fig. 6Ah). The percentages of ALP-stained signals in the material group were approximately equal to that in the normal group at 3 months after the surgery, and higher than that in the normal group at 6 months after the surgery (P < 0.05). In the MSC group, the percentages of ALP-stained signals were significantly higher than that in the normal group (P < 0.01; Fig. 6B1 and B2). The results show that combining the bone collagen matrix with the hUC-MSCs can promote osteoblasts formations better than the bone collagen matrix alone.
Sirius red staining is used to analyze the collagen distribution and can dye type I collagen bright yellow/orange. Bright yellow/orange was distributed evenly and widely in the maxillary bone tissue in the normal group, revealing the uniform distribution of type I collagen in the normal maxillary bone (Fig. 7Aa and Ae). In the control group, no obvious type I collagen was observed at 3 months after the surgery and only a small amount of type I collagen was present at 6 months after the surgery (Fig. 7Ab and Af). In the material group, a small amount of collagen type I was observed at 3 months after the surgery and some collagen type I was present at 6 months after the surgery in the bone defect area after the implantation of the bone collagen . e-f showed the results of HE staining 6 months after surgery (original magnification, × 10), and represented the normal group, control group, material group and MSCs group, respectively. e1-f1 were the amplification of the black frame in e-f (original magnification, × 50). Bar = 500 μm. BM bone marrow (green arrow); FT: fibrous tissue (blue arrow); BT bone trabecular (red arrow); CM collagen materials (orange arrow); NB new bone (black arrow); CS cavitation structure (yellow arrow). B Measurement of serum bone BGP. B1 Serum BGP levels 3 months after surgery; B2 Serum BGP levels 6 months after surgery. The injury side of each group was compared with that of the control group and the difference was denoted by *. The difference between the injury side and the normal side was expressed as #. **P < 0.01 matrix alone (Fig. 7Ac and Ag). In the MSCs group, a modest number of collagen type I was observed at 3 months after the surgery and a large amount of collagen type I was found at 6 months after the surgery in the bone defect area after the implantation of the bone collagen matrix with the hUC-MSCs (Fig. 7Ad and Ah). The statistical results showed that the ratio of collagen type I in the normal group was significantly higher than that in the control or material group at 3 and 6 months after the surgery (P < 0.01). In the MSCs group, the collagen type I ratio in the bone defect area was dramatically diminished as compared with normal maxillary bone (P < 0.01), but higher than that in the control or material group (P < 0.05) at 3 months after the surgery. At 6 months after the surgery, the ratio of collagen type I is above the normal level (P < 0.01) and more than that in the control or material group (P < 0.01; Fig. 7B and C). These results show that combining the bone collagen matrix with the hUC-MSCs could induce type I collagen formation more effectively than the bone collagen matrix alone.
Saccharide is a constituent of cartilage-matrix constituent, and PAS staining was used to assay the saccharide content. The saccharide is dyed red or fuchsia. The maxillary bone tissue in the normal group was uniformly purple with only a few fuchsia areas, suggesting that the normal maxillary bone might have a small amount of chondrocytes ( Fig. 8Aa and Ae). The control group did not exhibit any obvious red or fuchsia areas at either 3 or 6 months after the surgery (Fig. 8Ab and Af). In the material group, the trabecular Fig. 6 The osteoblasts in the alveolar process cleft were analyzed by ALP staining. A The photographs of ALP staining. a-d showed the results of ALP staining 3 months after surgery (original magnification, × 400), and represented the normal group, control group, material group and MSCs group, respectively. e-f showed the results of ALP staining 6 months after surgery (original magnification, × 400), and represented the normal group, control group, material group and MSCs group, respectively. Osteoblasts appeared black. Red arrows indicate positive staining. OB osteoblasts; PB primary bone; BM bone marrow; FT fibrous tissue; BT bone trabecular; NB new bone; CS cavitation structure. Bar = 20 μm. B1 Percentage of ALP-stained signals at 3 months after surgery. B2 Percentage of ALP-stained signals at 6 months after surgery. Compared with the normal group, *P < 0.05; **P < 0.01 bone border appeared fuchsia at 3 months after the surgery (Fig. 8Ac) and was reduced at 6 months after the surgery (Fig. 8Ag). In the MSCs group, the trabecular bone showed large fuchsia areas at 3 months after the surgery (Fig. 8Ad). While at 6 months after the surgery, only the edges of the new bone tissue were fuchsia (Fig. 8Ah). The percentages of PAS-stained signals at 3 and 6 months after the surgery in the material group were approximately equal to those in the normal group. In the MSC group, the percentage of PAS-stained signals was significantly higher than those in the normal group at 3 months (P < 0.05) and 6 months (P < 0.01) after the surgery, respectively (Fig. 8B1 and B2). These results show that combining the bone collagen matrix combined with hUC-MSCs can promote chondrocyte growth better than only the bone collagen matrix.

Effect of Combining Bone Collagen Matrix with hUC-MSCs on BMP-2 Expression
The BMP-2 expression in tissues was detected by immunochemistry. The positive signals of BMP-2 appear brown. The BMP-2 positive signals were mainly located in osteocytes in the normal maxillary bone. There was a small amount of brown in the normal group ( Fig. 9Aa and Ae), implying that the normal maxillary bone expressed low levels of BMP-2. In the control group, only cavitation Fig. 7 The distributions of collagen were analyzed by sirius red staining. A The photographs of sirius red staining. a-d showed the results of sirius red staining 3 months after surgery (original magnification, × 40), and represented the normal group, control group, material group and MSCs group, respectively. e-f showed the results of sirius red staining 6 months after surgery (original magnification, × 40), and represented the normal group, control group, material group and MSCs group, respectively. Collagen of type I appeared yellow/orange under polarized light microscopy. PB primary bone; FT fibrous tissue; BT bone trabecular; NB new bone; CM collagen materials; CS cavitation structure. Bar = 1000 μm. B Percentage of type I collagen 3 months after surgery. C Percentage of type I collagen 6 months after surgery. Compared with the normal group, *P < 0.05; **P < 0.01 structure was observed and no obvious BMP-2 positive staining was found at either 3 or 6 months after the surgery ( Fig. 9Ab and Af). In the material group, BMP-2 was mainly expressed in the osteocytes and osteoclasts at the edge of the trabecular bone at either 3 or 6 months after the surgery (Fig. 9Ac and Ag). In the MSCs group, the BMP-2 positive signals were mainly concentrated at the edge of the trabecular bone at 3 months after the surgery (Fig. 9Ad), and in the area of new bone formation at 6 months after the surgery (Fig. 9Ah). The percentage of BMP-2 positive signals in the material group was approximately equal to that in the normal group at 3 and 6 months after the surgery. In the MSC group, the percentage of BMP-2 positive signals was significantly higher than those in the normal group at 3 months (P < 0.05) and 6 months (P < 0.01) after the surgery (Fig. 9B1 and B2). The results show that the ability of the bone collagen matrix combined with the hUC-MSCs to induce BMP-2 generation was better than that of the bone collagen matrix alone.

Proliferation and Apoptosis Analysis of Bone Defect Site
TUNEL assays were used to detect apoptosis in the bone defect center area of each group, and the apoptotic nuclei were stained brown (Fig. 10A). Ki67 immunohistochemistry was used to detect the proliferation of cells in the bone defect center area of each group, and the proliferation cells were labeled with fluorescent green (Fig. 10B). The cells in the normal group were mainly distributed in the cancellous bone region inside the maxilla, while the cells in the dense bone region were less. The apoptotic and proliferative cells were present in the maxillary bone in the normal group (Fig. 10Aa, Ae, Ba1 and Be1). Although there was slightly repair in the bone defect site in the control or material group, the images from the proliferation and apoptosis staining did not observed visible brown staining and green fluorescence at 3 and 6 months after the surgery (Fig. 10Ab, Ac, Af, Ag, Bb1, Bc1, Bf1 and Bg1), suggesting there was no obvious signals of proliferation and apoptosis in the control or material group. There was visible brown staining and green fluorescence in the bone growth area in MSCs group at 3 and 6 months after the surgery (Fig. 10Ad, Ah, Bd1 and Bh1). The TUNEL staining results showed that the apoptosis rate in the MSCs group was significantly higher at 3 months after the surgery (P < 0.05) and lower at 6 months after the surgery (P < 0.01) than those in the normal group (Fig. 10C1). The Ki67 immunohistochemical results showed that the proliferation rate in MSCs group have an increasing trend at 3 months after the surgery and a decreasing trend at 6 months after the surgery as compared with the normal group, but there was no statistical difference (Fig. 10C2). All these facts show that combining the bone collagen matrix with the hUC-MSCs effectively promote bone regeneration and repair by regulating cell proliferation and apoptosis.

Assessment of Postoperative Health Status of Rabbits
Blood routine (Tables 1, 2), liver function (Tables 3,  Table 4) and renal function (Tables 5, Table 6) were used to estimate the health status of the rabbits at 3 and 6 months after the surgery. The blood routine results showed that the percentage of lymphocyte (LYM) and neutrophil granulocyte (NEUT) in the MSCs group was significantly lower (P < 0.01) and higher (P < 0.05), respectively, than that in normal or control group at 3 months after the surgery, and both returned to normal level at 6 months after the surgery. The percentage of LYM in the material group was significantly decreased (P < 0.05) and the percentage of NEUT was dramatically increased (P < 0.05) compared with the normal group at 6 months after the surgery. The C-reactive protein (CRP) concentration in the MSCs group had a downward trend as compared with the normal, control and materials group at 3 months after the surgery, and returned to normal level at 6 months after the surgery. CRP concentration in the control group was higher than that in the normal group at 3 months and approximately equal to that in the normal group at 6 months after the surgery. The liver function results showed the ALP concentration in the control, material and MSCs groups had an upward trend at 3 months after the surgery, and was significantly decreased in these group as compared with the normal group at 6 months after the surgery (P < 0.05). Total bilirubin (TBIL) and bilirubin direct (DBIL) in the MSCs group was significantly higher than that in the control group (P < 0.05), and indirect bilirubin (IBIL) was dramatically increased in the MSCs group as compared with the normal group at 3 months after the surgery (P < 0.05). TBIL, DBIL and IBIL were back to normal level at 6 months after the surgery. Renal function results showed that the blood urea nitrogen (BUN) and creatinine (CR) concentration showed an increase trend in the control, material and MSCs groups compared with the normal group at 3 months after the surgery, and returned to normal level at 6 months after the surgery. There was no significant difference in uric acid (UA) concentration among all groups at both 3 and 6 months after the surgery.

Discussion
Bone induction refers to the induction of connective tissue adjacent to the bone graft area by bone growth factors or seed cells in the bone material. The formation of new bone is facilitated by affecting undifferentiated bone progenitor cells, promoting their differentiation and proliferation and eventually becoming osteoblast. Recently, MSCs have been used to induce bone regeneration in bone defect of orofacial clefts [15,16,24]. Human dental pulp stem cells are used to repair maxillary alveolar defects in rats [16]. Bone-marrow-derived mesenchymal stem/stromal cells (BM-MSCs) loaded in hydroxyl apatite/collagen promote local osteogenesis in alveolar cleft in rat and human [15,24]. Although hUC-MSCs have also been used treat various diseases [25,26], there are few reports about the use of hUC-MSCs in the treatment of alveolar process cleft.
HUC-MSCs are characterized by easy extraction, multidirectional differentiation, short proliferation time, low immunogenicity, and long survival time after transplantation and have, therefore, become the preferred seed cells for transplantation [27,28]. In addition, because hUC-MSCs are obtained from neonatal umbilical cords. Neonatal umbilical cords are medical waste and no harm to the donor. And few ethical implications arise from using hUC-MSCs [29]. Moreover, the hUC-MSC stem cells obtained from umbilical cords show rapid self-renewal [30]. HUC-MSCs also express low levels of major histocompatibility complex II and costimulatory molecules, thereby reducing the possibility of rejection [31]. Other MSCs, such as BM-MSCs, have obvious osteogenic ability and low immunogenicity. However, their sources are limited. Furthermore, it is difficult to obtain such cells and easy to harm donors when the cells are collected [32,33]. Therefore, compared with other MSCs, hUC-MSCs have more extensive application prospects.
The bone collagen matrix used in this study was a bone matrix prepared from de-greased and de-cellularised bovine cancellous bone. The decalcified and de-proteinised heterogeneous bone matrix was mainly composed of type I collagen, which is insoluble and highly hinged and exhibits good bone-guiding activity, toughness and strength. This material preserves the natural structure of the bone and exhibits the right-sized pores wherein cells and blood vessels can grow. Owing to the removal of antigens, the antigenicity of the material is very weak, and there were no obvious signs of rejection after implantation. Because the heterogeneous bone matrix is more widely derived than allogeneic bone, and can be bio-degraded and absorbed in a shorter time, it meets the requirements of the ideal carrier.
Studies have shown that the collagen biomaterials alone exhibited no obvious effective function in bone repair [34]. However, combining collagen bio-materials with a bone collagen matrix can slow the degradation of hUC-MSCs, thereby prolonging the bone repair time. Using collagen scaffold alone to repair rabbit alveolar cleft had no significant effect. Collagen scaffold material combined with hUC-MSCs can significantly repair rabbit alveolar cleft, and the defective jaw can be repaired to be close to the normal jaw. and was difficult to induce bone regeneration. Therefore, it is more likely to be useful to evaluate the bone regeneration capacity of biomaterials.
In order to evaluate the quality of the newly formed bone in the defect area, the bone trabeculae and mineral density in the centre of the cubic bone-defect area were detect by micro-CT analysis. Although bone collagen matrix alone did not significantly increase the bone defect area, it induced partial bone regeneration. The percentage of bone trabeculae and bone density is visibly higher in MSCs group than material group, but significantly less than normal bone. These facts showed that combining bone collagen matrix with hUC-MSCs could enhance the quality of the newly formed bone in the defect area, but it doesn't reach the normal level.
The repair of bone defect generally involves in osteogenesis, the collagen distribution and chondrogenesis [35]. Here, combining bone collagen matrix with hUC- MSCs   Fig. 10 Detection of cell apoptosis and proliferative assay. A Cell apoptosis detected by TUNEL (original magnification, × 400). B Cell proliferation detected by Ki67 immunofluorescence (original magnification, × 400). a-d and a1-d1 showed the results 3 months after surgery, and represented the normal group, control group, material group and MSCs group, respectively. e-h and e1-h1 showed the results 6 months after surgery, and represented the normal group, control group, material group and MSCs group, respectively. C1 The percentage of apoptotic cells in each group at 3 or 6 months after surgery. C2 The percentage of proliferative cells in each group at 3 or 6 months after surgery. TUNEL-positive cells appeared brown and Ki67-positive cells appeared green. Compared with the normal group, *P < 0.05; **P < 0.01 ◂ Table 1 Blood routine test results 3 months after surgery contributed to the repair of the bone defect in the alveolar process cleft, likely through promoting the production of osteoblasts, chondrocytes, type I collagen and BMP-2. BMP-2 pathway is activated at the onset of cortical bone repair, and mainly expressed by osteoblasts and osteoclast to regulates osteoblast differentiation and bone formation [36,37]. In the process of the bone defect repair, bone collagen matrix combined with hUC-MSCs is probably through increasing in the number of osteoblasts, and then enhancing BMP-2 expression to induce bone regeneration.
In the detection of apoptosis and proliferation, our experimental results showed that no obvious signals of proliferation and apoptosis were observed in the control or material group, although slight bone formation was found. It may be caused by the following reasons: firstly, although new bone could be distinguished from existing bone, there was no unclear boundary between the new bone and the existing bone. In order to avoid the effect of unclear boundary on the experimental results, the central area of bone defects in each group was chosen for apoptosis and proliferation analysis. In this study, about 2-month-old rabbits were used for alveolar process cleft repair experiment. The larger the rabbit grows, the larger the area of the area of bone defects of alveolar process clefts. At 3 months, bone tissue could be seen around the center of the breach. In the same field of vision, no bone tissue could be seen around the center of the breach at 6 months. Secondly, even if bone tissue can be seen in the control or material group at 3 months after the surgery, the images from the proliferation and apoptosis staining did not observed visible brown staining and green fluorescence, implying that no obvious cell apoptosis and proliferation occur in the control or material group. The results also showed that the apoptosis rate of the MSCs group was significantly higher than that of the normal group at 3 months after surgery and significantly lower than that of the normal group at 3 months after surgery. The trend was similar with cell proliferation. Usually, apoptosis is a ubiquitous phenomenon for the natural, developmental cell death and takes place in proliferating cell populations and is involved in tissue morphogenesis [38][39][40]. Therefore, we speculated that the increase of apoptosis in MSCs group at 3 months after surgery was due to the increase of proliferative cells. At 6 months, the cell proliferation level was downward and the corresponding apoptosis rate was decreased. Blood routine, liver and renal function can be used to estimate the health of the body. The blood routine results showed that the percentage of LYM was significantly decreased and the percentage of NEUT was dramatically increased in the material group at 6 months after the surgery and in the MSCs group at 3 months after the surgery. CRP concentration in the MSCs group is lower than that in the normal, control and materials group at 3 months after the surgery, and regained to similar level to other groups at 6 months after the surgery. NEUT/LYM ratio is considered as a clinical cellular inflammation marker [41]. CRP be supposed to be an early indicator of infectious or inflammatory conditions [42]. It was reported that hUC-MSCs played a role in bone tissue repair by inhibiting inflammation [22]. Thus it can be seen that that combining bone collagen matrix with hUC-MSCs may more effectively inhibit the inflammatory reaction than bone collagen matrix alone. The liver function results found that the ALP concentrations in the control, material and MSCs groups showed an increase tendency at 3 months and was significantly decreased at 6 months after the surgery compared with the normal group. TBIL in the MSCs group was markedly increased compared with the control group (P < 0.05) at 3 months after the surgery, and back to normal level at 6 months after the surgery. Serum ALP in normal people mainly comes from liver and bone, and may be considered as the biomarkers of liver injury or bone damage [43,44]. Total bilirubin (TBIL) is the sum of DBIL and IBIL, and the biomarker of liver injury [45]. The results from liver function suggest that bone collagen matrix combining with hUC-MSCs may induce liver injury at 3 months after the surgery and completely recover well at 6 months after the surgery. All these facts show that bone collagen matrix combining with hUC-MSCs may not affect the health status of the body over a long period of time.
Recent studies have shown that MSCs exhibit a short survival time and poor cell differentiation activity in animal models [46,47]. Furthermore, the conditioned media (CM) of MSCs can also induce inflammatory repair and tissue regeneration [47][48][49]. Therefore, it is believed that implanted MSCs repair tissues mainly by changing the local microenvironment rather than directly transforming into target cells because the cells involved in tissue repair are mainly cells of the receptor itself, which are homed by the paracrine regulation of MSCs and further proliferate and differentiate [46].

Conclusions
In the rabbit alveolar process cleft model, combining bone collagen matrix with hUC-MSCs could repaired the damaged bone tissue and promoted the new bone regeneration via improving osteogenesis, the collagen distribution and chondrogenesis. The study shows that the combining of hUC-MSCs with a bone collagen matrix is a promising strategy for repairing injured bone tissue in regenerative medicine.

Informed Consent
There are no human subjects in this article and informed consent is not applicable.