Repair of Canine Mandibular Bone Defects with Bone Marrow Stromal Cells and Coral

Abstract

Tissue engineering has become a new approach for repairing bone defects. Previous studies indicated that coral scaffolds had been utilized with bone marrow stromal cells (BMSCs) in a variety of approaches for bony reconstruction. In these applications, the degradation rate of the material did not match the rate at which bone was regenerated. In this study, a previously established 30 mm long mandibular segmental defect was repaired with engineered bone using green fluorescent protein–labeled osteogenic BMSCs seeded on porous coral (n = 12). Defects treated with coral alone (n = 12) were used as an experimental control. In the BMSCs/coral group, new bone formation was observed from 4 weeks postoperation, and bony-union was achieved after 32 postoperative weeks. The residual coral volume of the BMSCs/coral grafts at 12 weeks (20–30%) was significantly higher than that at 32 weeks (10–15%, p < 0.05), which was detected by microcomputed tomography and histological examination. The engineered bone with BMSCs/coral achieved satisfactory biomechanical properties at 32 weeks postoperation, which was very close to that of the contralateral edentulous mandible. More importantly, immunostaining demonstrated that the implanted BMSCs differentiated into osteoblast-like cells. In contrast, minimal bone formation with almost solely fibrous connection was observed in the group treated with coral alone. Based on these results, we conclude that engineered bone from osteogenically induced BMSCs and biodegradable coral can successfully repair the critical-sized segmental mandibular defects in canines and the seeding cells could be used for bony restoration.

Introduction

Tissue engineering has been shown to be an effective approach in bone regeneration. The successful repair of bone defects has been demonstrated in many large mammals, such as sheep and canine.^1–5 Meanwhile, a variety of biomaterials have been utilized for bone engineering. It is clear that the degradation rate of the material could affect the bone regeneration process. In a previous study,⁶ we demonstrated that porous β-tricalcium phosphate (β-TCP) scaffolds loaded with osteogenically induced bone marrow stromal cells (BMSCs) could repair critical-sized mandibular defects in a canine model. However, the degradation rate of β-TCP was slow;30–40% remained at 32 weeks postoperation, a time after which new bone has already formed, indicating that this material may not be optimal for bone engineering in repairing mandibular bone defects.

Porous coral is a natural and biodegradable ceramic material that can be considered a fast-degrading material, as coral alone can degrade nearly completely in vivo after 3 months.⁷ Coral has been successfully used for repairing goat and sheep segmental limb defects when combined with BMSCs.^8–10 However, different data regarding coral degradation in areas of newly formed bone complicates the use of this material in bone tissue engineering applications. Petite et al.⁸ observed that less than 2% of the coral remained after 16 weeks of bone regeneration in a sheep critical-size metatarsus defect model, whereas Geiger et al.¹⁰ found that 6% of the original scaffold remained after 16 weeks in a rabbit radial critical-size defect model. Therefore, the exact long-term degradation rate of coral in engineered bone is still unclear. Since the scaffold residual volume change with time could represent the synchronization of the scaffold resorption and bone formation processes, it is necessary to address the issue for future clinical application.

In this study, porous coral seeded with in vitro osteogenically induced, green fluorescent protein (GFP)–labeled BMSCs were implanted to repair a 30 mm long critical-sized segmental mandibular defect in a canine model. The residual coral volume for the BMSCs/coral grafts at 12 weeks (20–30%) was significantly higher than that at 32 weeks (10–15%), as measured by microcomputed tomography (CT) and histomorphometry analyses. A satisfactory repair was, in most cases, achieved at 32 weeks postimplantation, suggesting that BMSCs along with porous coral could be potentially used for the clinical repair of mandibular bone defects. Moreover, the implanted BMSCs differentiated into osteoblast-like cells in vivo, demonstrating that the seeded cells could be the source of bony restoration.

Materials and Methods

Animals

A total of 24 healthy adult mongrel dogs, aged 16 months old with an average weight of 19.7 kg, were used in this study. All dogs were prepared by extraction of the upper and lower right premolar and first molar teeth 8 weeks before segmental mandibular resection. The dogs were kept on a soft diet during the study and were allowed complete gingival healing before mandibular resection surgery. The experimental protocol was approved by the Animal Care and Experiment Committee of Shanghai Jiao Tong University School of Medicine.

Cell culture, induction, and GFP labeling

After intravenous anesthesia with 5% sodium pentobarbital (0.5 mL/kg), 3 mL of bone marrow aspirates were harvested from the iliac crests of each dog and transferred into a preheparinized centrifuge tube. Mononuclear cells were separated by percoll (1.073 g/mL; Sigma, St. Louis, MO) gradient centrifugation¹¹ and were plated in 100 mm dishes (Falcon, Franklin Lakes, NJ) at a density of 1 × 10⁵ cells/cm². Cells were cultured in osteogenic induction media (Dulbecco's modified Eagle's medium; Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (Gibco), 10⁻⁸ mol/L dexamethasone, 10 mmol/L β-phosphoglycerol, and 50 μmol/L L-2-ascorbic acid (all from Sigma) at 37°C with 5% CO₂. The culture medium was changed after 48 h and then every 3 days. When BMSCs reached 80–90% confluence, cells were detached with 0.25% trypsin/ethylenediaminetetraacetic acid (Gibco) and subcultured at a density of 1 × 10⁵ cells/cm² in 100 mm dishes.

To trace the in vivo implanted cells, BMSCs were retrovirally labeled with GFP, as described.^12,13 After selection in G418 (300–400 μg/mL), these labeled cells were also osteogenically induced and observed under an inverted fluorescence microscope (Nikon, Tokyo, Japan).

Preparation of BMSCs/coral constructs

Natural corals were obtained from Hainan (Sanya, China) and used as the scaffold. The corals were molded into cuboids (30 × 15 × 10 mm) with a 3 mm diameter tunnel mimicking the normal structure of mandible (Fig. 1A), and they were sterilized by ⁶⁰Co irradiation before use. From analysis by micro-CT (μCT 80; Basserdorf, Zurich, Switzerland), the coral exhibited a reasonably homogenous structure with a volume porosity of 57 ± 6.5%, mean pore diameter of 195 ± 75 μm (range, 140–260 μm), and pore wall thickness of 98 ± 22 μm (range, 72–120 μm), as previously described.⁹ The initial compression strength of the coral was 6.46 ± 0.93 MPa. Scanning electron microscopy showed interconnected three-dimensional networks of porous coral (Fig. 2A), in which all the pores were interconnected (open porosity). For cell seeding, osteogenically induced BMSCs at passage 2 were detached from culture dishes, centrifuged to remove supernatant, and then resuspended in culture media at a density of 2 × 10⁷ cells/mL. Cells in suspension were slowly injected into the coral cuboids using a syringe (nearly 3 mL/cuboid). After being incubated for 4 h to allow cell attachment, 30 mL of osteogenic induction medium were then added to cover the constructs. The BMSCs/coral constructs were subsequently cultured for 7 days in vitro before implantation.

FIG. 1.

(A) Gross view of coral cuboids. The cuboids are 30 mm long, 15 mm high, and 10 mm thick, with a tunnel of 3 mm in diameter through each cuboid. (B, C) Comparing light and fluorescence microscopic images of GFP-labeled BMSCs, it was confirmed that over 90% cells expressed green fluorescence (scale bars: 50 μm). (D) Scanning electron microscopy evaluation of the three-dimensional porous structures of coral. (E) Scanning electron microscopy showed BMSCs attach on coral with abundant matrix after 7 days of culture. GFP, green fluorescent protein; BMSCs, bone marrow stromal cells. Color images available online at www.liebertonline.com/ten.

FIG. 2.

Radiographs of treated defects taken at different time points postoperation. (A) In the BMSCs/coral group, a few calluses were formed at 4 weeks, (C) and more calluses were observed at 12 weeks. (E) At 26 weeks, the radiopacity was highly increased and bony-union was achieved. (G) The radiopacity was close to that of the contralateral normal mandible with smooth, remodeled bony contour at 32 weeks. (B) In the coral group, the scaffolds were rigidly implanted. (D) No callus was observed at 4 weeks. (F) The scaffold mostly degraded at 12 weeks. (H) Nearly complete radiolucency and minimal callus formation at the cutting ends were observed at 32 weeks.

In a parallel experiment, 3 × 3 × 3 mm cuboids were prepared and seeded with BMSCs at an identical cell density. After 7 days of incubation, the constructs were fixed in 2% formalin for 2 h and then examined using scanning electron microscopy (SEM; Philips SEM XL-30, Amsterdam, The Netherlands).

Surgical procedure

The canine mandibular defects were made in the right hemimandible as previously described.⁶ The defects were filled either with osteogenically induced BMSCs/coral constructs (experimental group, n = 12) or with coral cuboids alone (scaffold group, n = 12). Meanwhile, the left lower premolar and first molar teeth were extracted, and the edentulous left hemimandibles served as normal control (n = 24). All animals were kept on a soft diet during the study. To give the newly formed tissue mechanical stimulation, the titanium plates were removed from both groups at 22 weeks postoperation.

Radiographic and micro-CT analyses

Mandibular radiograms were obtained under general anesthesia at 4, 12, 26, and 32 weeks postoperation. For micro-CT analysis, all animals were sacrificed at 12 and 32 weeks postoperation, and both sides of the mandibles were harvested. A micro-CT system (μCT-80; Bassersdorf, Zurich, Switzerland) was used to visualize cross-sectional images of the mandibular specimens. The CT settings were used as follows: pixel matrix, 1024 × 1024; voxel size, 50 μm; and slice thickness, 50 μm. The bone tissue was segmented from marrow and soft tissue using a global thresholding procedure.¹⁴ The threshold was set at 25% of the maximal gray scale value, and the density of bone volume (DBV, mg HA/cm³) was calculated. Also, we used the global thresholding approach to analyze the residual material volume (RMV, %) in the specimens of the experimental group, whereas the threshold was set at 46.5% of the maximal gray scale value to segment residual coral from new bone tissue.

Biomechanical analysis

For biomechanical analysis, mandibles from the experimental group were harvested at 12 and 32 weeks postoperation. The left side mandible of each animal was also collected as normal control. Each mandible was placed in a three-point bending system, providing an unsupported length of 30 mm. Load was applied in the occlusal-to-inferior direction to the midpoint of the unsupported length, approximately in the middle of the repaired defect. The measurement was conducted in a hydraulic materials testing machine (Shimadzu AG-5KN, Kyoto, Japan) and stopped after bone fracture. Deformation was measured using the strain gauge at a loading rate of 0.5 mm/min. The bending stress was calculated with the following equation: bending stress = 3LF/(2WT²), wherein L, F, W, and T are the test span (mm), the bending strength load (KN), the width (mm), and the thickness of the specimen (mm), respectively. Young's modulus was derived from stress–strain curve as the slope in the linear region.

Histological evaluation

Tissues from the original defect area of both groups were harvested at 12 and 32 weeks postoperation. Samples were fixed in freshly made 4% paraformaldehyde at 4°C overnight. They were dehydrated through an ethanol gradient of 70%, 80%, 90%, 95%, 100% and 100% ethanol with 12 h for each step at 4°C. Then the samples were defatted in xylene twice at 4°C for 12 h (removing lipids from the tissue to facilitate penetration of the embedding medium) before embedding in methyl methacrylate resin. After hardening, 100–120-μm-thick sections were cut under cooling water with a sawing microtome Leica SP1600 (Leica Microsystems, Wetzlar, Germany). The sections glued onto a plastic support were polished to 50 ± 5 μm thickness under cooling water. They were finally stained with Van Gieson's Picro-fuchsine staining. The undecalcified sections from the experimental group at 32 weeks postoperation were also observed using anti-GFP polyclonal antibody immunostaining (BD, Franklin Lakes, NJ).

The two most central sections per defect of BMSCs/coral group at 12 and 32 weeks were histomorphometrically analyzed using Image Pro Plus 4.0 software (Media Cybernetics, Bethesda, MD). The images were taken with a Nikon E600 microscope (Nikon) and Pixera 600CL CCD (Pixera, San Jose, CA) at a magnification of 100 × . The BV (%) and the RMV (%) were expressed as the percentage of newly formed bone tissue and residual material quantities on the implant stained sections, respectively.

Statistical analysis

Data of DBV from micro-CT examination were analyzed using a one-way analysis of variance test. The differences between experimental group (n = 6 at 12 and 32 weeks, respectively), scaffold group (n = 6 at 12 and 32 weeks, respectively), and normal group (n = 6 at 12 and 32 weeks, respectively) were assessed using Student-Newman-Kewls test. Data of BV and RMV from histomorphometrical and micro-CT examination (experimental group, n = 6 at 12 and 32 weeks, respectively) were analyzed by paired Student's t-test. Data from biomechanical examination were analyzed by one-way analysis of variance test. The differences between experimental group (n = 6 at 12 and 32 weeks, respectively) and normal group (n = 6 at 12 and 32 weeks, respectively) were assessed using Student-Newman-Kewls test. A p-value of less than 0.05 was considered statistically significant.

Results

Growth and fluorescence of GFP-labeled osteogenic BMSCs on coral

After being retrovirally labeled with GFP and cultivated in the presence of dexamethasone and β-phosphoglycerol for two passages, BMSCs grew well on the plate (Fig. 1B) and showed a typical osteogenic phenotype, including calcium deposition and positive alkaline phosphatase staining (data not shown).

To assess the efficiency of GFP labeling technique, BMSCs were also observed under an inverted fluorescence microscope. Comparing Figure 1B with C, it was confirmed that over 90% cells expressed green fluorescence and that implanted BMSCs were mostly GFP labeled.

To evaluate the compatibility of the scaffold, BMSCs were seeded on the coral scaffolds and incubated in vitro for 7 days, and cell-scaffold complexes were subjected to SEM examination. As shown in Figure 1D and E, the pores were filled by cells together with abundant extracellular matrices after 7 days of incubation, indicating that the scaffold has good biocompatibility.

Radiographic analysis

After a critical-sized segmental defect was created, the defect was filled with either GFP-labeled osteogenic BMSCs/coral constructs (experimental group, n = 12) or coral cuboid alone (scaffold group, n = 12). All animals survived the surgical procedure; no dog was excluded from the study.

To serially examine the new bone formation and the development of bone union within the defects, X-ray images were taken at 4, 12, 26, and 32 weeks postoperation. Representative photographs of each group are shown in Figure 2. In the experimental group implanted with BMSCs/coral constructs, a few calluses were observed around the implants at 4 weeks postoperation (Fig. 2A), and more calluses were observed at 12 weeks (Fig. 2C). At 26 weeks postoperation, the volume and radiopacity of newly formed bone were highly increased and bony-union was achieved (Fig. 2E). At 32 weeks, the bone contour was remodeled smoothly (Fig. 2G), and the radiopacity was close to that of the contralateral edentulous mandible. On the contrary, although the scaffold was rigidly implanted in the coral alone group (Fig. 2B), callus could neither be observed at 4 weeks (Fig. 2D) nor at 12 weeks (Fig. 2F), and the scaffold was mostly degraded at 12 weeks. At 32 weeks, nearly complete radiolucency and minimal callus formation at the cutting ends were observed (Fig. 2H). Radiographic union was observed in all of the experiment (BMSCs/coral) group (12/12) but not in the scaffold group (0/12).

Gross and micro-CT analyses

Correlated with the radiographic analysis, gross views of the mandibles repaired with BMSCs/coral constructs showed that bones achieved bony-union at either 12 or 32 weeks postoperation (Fig. 3A, B). The shape of repaired mandibles was very close to that of normal edentulous mandibles at 12 and 32 weeks postoperation, respectively (Fig. 3E, F). On the contrary, no bony-union was observed when the defects were implanted with coral alone; obvious soft tissue was detected in the original defect area at either 12 or 32 weeks postoperation (Fig. 3C, D).

FIG. 3.

Gross view of repaired mandibles at 12 and 32 weeks postoperation. (A, B) In the BMSCs/coral group, bony-union with red-white colored cortex and some notches are observed at 12 and 32 weeks. (C, D) In the coral group, soft tissues are detected in the defect at 12 and 32 weeks. (E, F) Contralateral edentulous native mandible at 12 and 32 weeks. Arrows indicate the junction sites. Color images available online at www.liebertonline.com/ten.

To detect the inner structure of repaired mandibles, micro-CT images were taken at 12 and 32 weeks postoperation. Representative images from each group are shown in Figure 4. In the experimental group treated with BMSCs/coral constructs, the scaffolds were mostly degraded at 12 weeks from the section images (Fig. 4A). Porous structure occurred and the scaffold continuously degraded at 32 weeks for BMSCs/coral grafts (Fig. 4B), which were different from the normal mandibular canal structure (Fig. 4C). However, in the group treated with coral alone, nonunions were observed and the scaffolds almost completely degraded at both 12 and 32 weeks (Fig. 4D).

FIG. 4.

Microcomputed tomography analysis at 12 and 32 weeks postoperation (scale bars: 5 mm). (A) In BMSCs/coral group, the scaffold mostly degraded at 12 weeks. (B) In BMSCs/coral group, porous structure occurred and the scaffold continuously degraded at 32 weeks. (C) In the normal group, the mandibular canal structure was observed. (D) In the coral alone group, the scaffold almost completely degraded and nonunion was observed at 32 weeks. (E) Both at 12 and 32 weeks, the BMSCs/coral grafts showed significantly higher bone density than the normal group (p < 0.05) and scaffold group (p < 0.01). (F) The BMSCs/coral grafts at 12 weeks showed significantly higher residual material volume than that at 32 weeks (p < 0.01). Color images available online at www.liebertonline.com/ten.

To quantify the calcification and residual material in the repaired mandibles, the bone density (Fig. 4E) and the RMV (Fig. 4F) were also measured using micro-CT at 12 and 32 weeks postoperation. In terms of the DBV, at 12 weeks postoperation, the BMSCs/coral grafts showed relatively high bone density of 562.76 ± 85.68 mg HA/cm³, which was significantly higher than the contralateral normal group (474.04 ± 86.85 mg HA/cm³, p < 0.05) and the scaffold group (90.95 ± 20.5 mg HA/cm³, p < 0.01). At 32 weeks postoperation, the BMSCs/coral grafts also showed significantly higher bone density (554.3 ± 59.43 mg HA/cm³) than the contralateral normal group (469.36 ± 67.74 mg HA/cm³, p < 0.05) and the scaffold group (47.51 ± 6.41 mg HA/cm³, p < 0.01). In terms of the RMV, the data from BMSCs/coral grafts at 12 weeks (21.44 ± 4.33%) were significantly higher than those at 32 weeks (14.46 ± 2.94%, p < 0.01).

Biomechanical analysis

To evaluate the biomechanical property of the repaired mandibles, three-point bending test was performed at 12 and 32 weeks postoperation. As shown in Table 1, the bending load strength and bending stress in the BMSCs/coral group at 12 weeks were 1.74 ± 0.37 KN and 33.43 ± 9.47 MPa, respectively (73.50 ± 11.69% and 81.67 ± 19.50% of the contralateral normal edentulous mandible), whereas at 32 weeks these data were 1.52 ± 0.42 KN and 54.40 ± 9.88 MPa, respectively (86.37 ± 12.03% and 90.80 ± 18.81% of the normal group). Compared with the normal mandible, tissue-engineered segmental bone revealed similarities with respect to bending load strength, bending displacement, and bending stress at either 12 or 32 weeks postoperation (p > 0.05), except for the Young's modulus at 12 weeks (p < 0.05). Further, the bending stress of the engineered bone at 32 weeks was significantly higher than that at 12 weeks (p < 0.05). Since there was no bony-union in the scaffold group that was treated with coral alone, three-point bending test could not be applied in this group.

Table 1.

Biomechanical Analyses at 12 and 32 Weeks Postoperation

Parameters	Experimental group (12 weeks)	Normal group (12 weeks)	Experimental group (32 weeks)	Normal group (32 weeks)
Bending strength load (KN)	1.74 ± 0.37	2.38 ± 0.55	1.52 ± 0.42	1.84 ± 0.79
Bending displacement (mm)	1.82 ± 0.39	2.55 ± 0.52	1.89 ± 0.76	2.64 ± 0.56
Bending stress (MPa)	33.43 ± 9.47	42.25 ± 5.59	54.40 ± 9.88	61.18 ± 13.01
Young's modulus (MPa)	507.58 ± 194.91	239.63 ± 71.82	433.95 ± 159.51	292.88 ± 113.61

Data present mean value ± standard deviation (experimental and normal group n = 12, each time point n = 6, respectively). No significant differences were observed among bone marrow stromal cells/coral and normal control group in terms of most biomechanical parameters tested at either 12 or 32 weeks postoperation (p > 0.05), except for the Young's modulus at 12 weeks (p < 0.05). Further, the bending stress of engineered bone at 32 weeks was significantly high than at 12 weeks (p < 0.05).

Histological examination

To further confirm the above findings, histology of repaired defects at 12 and 32 weeks postoperation was performed using undecalcified tissue sections stained with Van Gieson's Picro-fuchsine stain. Representative photographs from each group are presented in Figure 5. In the defects treated with BMSCs/coral constructs, osteon formation was observed together with some undegraded coral particles at 12 and 32 weeks; the residual coral at 12 weeks was greater than that at 32 weeks (Fig. 5A1, B1, B2). In addition, bony-union was observed at both 12 and 32 weeks for the BMSCs/coral group (Fig. 5A2, B2); however, in the defects treated with coral alone, mostly soft tissue and some residual coral were observed at 12 weeks. At 32 weeks, fibrous tissue with minimal new bone formation at the cut end was observed (data not shown).

FIG. 5.

Van Gieson's Picro-fuchsine staining of repaired mandibles at 12 and 32 weeks postoperation. The macroscopic view of the repaired area in each group is shown. (A) BMSCs/coral group at 12 weeks. (B) BMSCs/coral group at 32 weeks. Detailed views of the inside area (solid frame, A1, B1) and interfacial area (dot frame, A2, B2) were taken at high magnification (scale bars: 100 μm). (A1) Osteon formation was observed (arrowhead) together with some nondegraded coral particles (arrow) at inside area of 12 weeks. (B1, B2) The osteon remained (arrowhead) and the coral residue (arrow) decreased at both the inside and interfacial areas after 32 weeks. (A2, B2) Bony-union occurred at the interfacial area after both 12 and 32 weeks. Color images available online at www.liebertonline.com/ten.

To further quantify the calcification and material residue of repaired mandibles, histomorphometry was used in the BMSCs/coral group at 12 and 32 weeks postoperation. The BV of BMSCs/coral group at 32 weeks (60.66 ± 8.09%) was significantly higher than that at 12 weeks (46.74 ± 5.17%, p < 0.05), whereas the RMV of BMSCs/coral group at 12 weeks (24.17 ± 5.44%) was significantly higher than that at 32 weeks (11.63 ± 3.29%, p < 0.01).

To trace the long-term turnover of GFP-labeled BMSCs in the repaired mandibles, anti-GFP polyclonal antibody immunostaining was applied in the BMSCs/coral group at 32 weeks postoperation. At the central area of the scaffold/bone, GFP-immunopositive cells differentiated into osteoblast-like cells on the surface of the newly formed bone (Fig. 6A1, A2). At the interfacial area, no GFP-immunopositive cells were observed in the pores (Fig. 6B1, B2).

FIG. 6.

Anti-GFP polyclonal antibody immunostaining of BMSCs/coral group at 32 weeks postoperation. The macroscopical view of repaired area was shown, and detail view of inside area (solid frame, A1, A2) and interfacial area (dot frame, B1, B2) were taken at high magnification (scale bars: 100 μm). (A1, A2) At the inside area, GFP-immunopositive cells differentiated into osteoblast-like cells on the surface of the newly formed bone; arrows show the GFP-immunopositive cells. (B1, B2) At the interfacial area, no GFP-immunopositive cells were observed in the pores. Color images available online at www.liebertonline.com/ten.

Discussion

The current study demonstrates that critical-sized segmental defects of the canine mandible can be repaired using GFP-labeled osteogenic BMSCs with biodegradable coral scaffolds. The residual coral volume of BMSCs/coral grafts after 12 weeks was significantly higher than that at 32 weeks (p < 0.05), which indicated that the scaffold gradually degraded in concert with the speed of new bone formation. Moreover, the implanted BMSCs differentiated into osteoblast-like cells. Meanwhile, the scaffold alone is not sufficient enough to repair such a sizable defect in this model.

With a microstructure of interconnecting pores that is similar to cancellous bone, coral could be degraded within a few weeks at a rate commensurate with the new bone formation in bone engineering applications.^7,9,15–17 The degradation of coral involves a combination of physicochemical dissolution and osteoclast-like cell-mediated degradation. Carbonic anhydrase, an enzyme present in osteoclasts, is believed to play a key role in the resorption of coral.^18,19 In this study, we also found multinucleated giant cells around some nondegraded residual scaffold of BMSCs/coral grafts at 12 weeks (data not shown). These osteoclast-like cells likely originated from mononuclear myeloid precursors brought to the coral scaffold through new vessel formation or migration from the microenvironment.²⁰ In previous reports, the amount of coral remaining within newly formed bone in vivo was not clear, as the data differed from each other.^8–10 Thus, the exact long-term degradation rate of coral in engineered bone repair of load-bearing defects becomes an attractive issue for study. In the current study, the RMVs were precisely measured using micro-CT, and the BMSCs/coral grafts at 12 weeks showed significantly higher RMV (21.44 ± 4.33%) than at 32 weeks (14.46 ± 2.94%) by micro-CT analysis, which was in agreement with histomorphometrical findings. Further, the bending stress of engineered bone at 32 weeks was significantly higher than at 12 weeks (p < 0.05). These data demonstrated that the degradation of coral was a gradual process after initial engineered bone formation at 12 weeks in the load-bearing mandibular defect, which finally achieved synchronized scaffold resorption and bone formation to obtain high quality bone restoration at 32 weeks.

Poorly resorbable scaffolds prevent extensive bone formation by failing to leave space for newly formed bone. In both this study and a previous study carried out by our group,⁹ porous coral has shown a satisfactory degradation rate that could match the rate of new bone formation in the repair process, which was better than undegradable pyrolized bovine bone that has been used in mandibular reconstruction.²¹ Compared with resorbable β-TCP that we previously used in the same model,⁶ coral scaffolds degraded faster and left less residual scaffold at 32 weeks postoperation. Meanwhile, although β-TCP scaffolds with high porosity and large pore size were favorable for degradation as well as for better nutrient exchange and cell penetration, the mechanical properties of the scaffold were accordingly less due to this porosity, and we found that one β-TCP scaffold without seeded cells was fractured during the experiment. On the other hand, due to the higher initial compressive strength of coral (6.46 ± 0.93 MPa), no scaffolds were fractured in this study. Therefore, coral may be more suitable than β-TCP for clinical application. In addition, coralline hydroxyapatite (CHA) scaffolds have an architecture similar to that of coral and are also resorbable, which when seeded with BMSCs degrade well and were used to successfully repair a segmental metatarsal bone defect in sheep.²² In contrast, in our experience in using CHA to repair canine segmental mandibular defects, the residual scaffold volume was much greater than even β-TCP at 32 weeks (data not shown). The different degradation rate of CHA may be correlated with the different amount of mechanical load placed on the material. Increasing evidence indicates that load bearing is an important functional influence on bone regeneration and scaffold degradation^23,24; higher load leads to less residual material.

In addition, although BMSCs, whose osteogenic potential has been demonstrated both in vitro and in vivo,^25–28 have proved to be a major cell source for bone engineering, the role of these implanted BMSCs during engineered bone formation was still unclear. Three major points may be contributing to engineered bone formation: (1) The implanted BMSCs differentiated into osteogenic cells and formed new bone. (2) Well-known “creeping substitution,” as occurred in autogenous bone transplantation and which includes both resorptive and appositional processes and native cells, could migrate into the scaffold to generate new tissues and achieve bony-union.²⁹ (3) The implanted BMSCs can secrete growth factors to recruit native cells to migrate into the defect site.^30–32 Endogenous bone marrow and BMSCs also participated in recruitment and differentiation into osteogenic cells during bone defect restoration.³³ In rodent tibial defects, Ito et al.³⁴ have demonstrated that the implanted BMSCs survived 8 weeks and differentiated into osteoblast-like cells; however, in repairing segmental bone defect of large mammals and over longer time periods, the turnover of implanted BMSCs was still unknown. Therefore, in this study, the implanted BMSCs were retrovirally labeled with GFP by gene transfection, and it was then possible to trace the implanted BMSCs and further illustrate the differentiation and distribution of BMSCs after implantation. As shown in the results, the GFP-labeled cells survived in the repaired mandible even at 32 weeks postoperation, which provides convincing evidence that the implanted BMSCs can differentiate into osteoblasts, and the implanted BMSCs are the vital cell source to repair mandibular defects even in large animals. Further, the GFP-labeled cells were not found at the defect/bone interface area, indicating that the repair process might be also combined with endogenous mesenchymal stem cells through local recruitment to the junction site and the strong spontaneous repair capability of bone.

Conclusion

This study demonstrates that bone substitutes constructed with porous coral scaffolds loaded with GFP-labeled osteogenic BMSCs could repair critical-sized mandibular defects in a large mammal, and the implanted BMSCs survived and differentiated into osteoblast-like cells after 32 weeks. The satisfactory degradation rate suggests that porous coral might be a feasible bone engineering scaffold for the clinical reconstruction of load-bearing mandibular defects.

Footnotes

Acknowledgments

This work was supported by Major State Basic Research Development Program of China (2005CB522700), National High Technology Research and Development Program of China (2006AA02A123), and Research Development Foundation of Shanghai educational committee (08YZ49). We thank Drs. Deli Liu, Yulai Weng, Guangdong Zhou, Lian Zhu, Ming Wang, Li Zhao, Gang Cai, Wanyao Xia, and Jinjun Chen for their contributions to this paper. We also appreciate the technical assistance from Drs. Demin Ying, Lijuan Zong, Juanjuan Wu, and Bin Zhong.

Disclosure Statement

No competing financial interests exist.

References

Kon

, Muraglia

, Corsi

, Bianco

, Maracacci

, Martin

, Boyde

, Ruspantini

, Chistolini

, Rocca

, Giardino

, Cancedda

, Quarto

Autologous bone marrow stromal cells loaded onto porous hydroxyapatite ceramic accelerate bone repair in critical-size defects of sheep long bones. J Biomed Mater Res, 49:328. 2000.

Shang

, Wang

, Liu

, Shi

, Cui

, Cao

Tissue-engineered bone repair of sheep cranial defects with autologous bone marrow stromal cells. J Craniofac Surg, 12:586. 2001.

Arinzeh

T.L.

, Peter

S.J.

, Archambault

M.P.

, van den Bos

, Gordon

, Kraus

, Smith

, Kadiyala

Allogeneic mesenchymal stem cells regenerate bone in a critical-sized canine segmental defect. J Bone Joint Surg Am, 85A:1927. 2003.

Abukawa

, Shin

, Williams

W.B.

, Vacanti

J.P.

, Kaban

L.B.

, Troulis

M.J.

Reconstruction of mandibular defects with autologous tissue-engineered bone. J Oral Maxillofac Surg, 62:601. 2004.

Viateau

, Guillemin

, Bousson

, Oudina

, Hannouche

, Sedel

, Logeart-Avramoglou

, Petite

Long-bone critical-size defects treated with tissue-engineered grafts: a study on sheep. J Orthop Res, 25:741. 2007.

Yuan

, Cui

, Zhang

W.J.

, Liu

, Cao

Repair of canine mandibular bone defects with bone marrow stromal cells and porous beta-tricalcium phosphate. Biomaterials, 28:1005. 2007.

Sciadini

M.F.

, Dawson

J.M.

, Johnson

K.D.

Evaluation of bovine-derived bone protein with a natural coral carrier as a bone-graft substitute in a canine segmental defect model. J Orthop Res, 15:844. 1997.

Petite

, Viateau

, Bensaid

, Meunier

, de Pollak

, Bourguignon

, Oudina

, Sedel

, Guillemin

Tissue-engineered bone regeneration. Nat Biotechnol, 18:959. 2000.

Zhu

, Liu

, Cui

, Cao

Review article: tissue-engineered bone repair of goat-femur defects with osteogenically induced bone marrow stromal cells. Tissue Eng, 12:423. 2006.

10.

Geiger

, Lorenz

, Xu

, Szalay

, Kasten

, Claes

, Augat

, Richter

VEGF producing bone marrow stromal cells (BMSC) enhance vascularization and resorption of a natural coral bone substitute. Bone, 41:516. 2007.

11.

Guo

, Yang

, Liu

, Li

, Hou

, Tang

P.H.

, Mao

Biological features of mesenchymal stem cells from human bone marrow. Chin Med J (Engl), 114:950. 2001.

12.

Takayama

, Tahara

, Thomson

A.W.

Transduction of dendritic cell progenitors with a retroviral vector encoding viral interleukin-10 and enhanced green fluorescent protein allows purification of potentially tolerogenic antigen-presenting cells. Transplantation, 68:1903. 1999.

13.

Zhou

, Liu

, Cui

, Wang

, Liu

, Cao

Repair of porcine articular osteochondral defects in non-weightbearing areas with autologous bone marrow stromal cells. Tissue Eng, 12:3209. 2006.

14.

Müller

, Rüegsegger

Micro-tomographic imaging for the nondestructive evaluation of trabecular bone architecture. Stud Health Technol Inform, 40:61. 1997.

15.

Vuola

, Taurio

, Göransson

, Asko-Seljavaara

Compressive strength of calcium carbonate and hydroxyapatite implants after bone-marrow-induced osteogenesis. Biomaterials, 19:223. 1998.

16.

Jaber

, Klein

B.Y.

, Leichter

, Ben-Bassat

, Gal

, Liebergall

, Segal

Tibial implant mineralization in rats is inversely related to serum osteogenic capacity. Bone, 16:81. 1995.

17.

Roudier

, Bouchon

, Rouvillain

J.L.

, Amedee

, Bareille

, Rouais

, Fricain

J.C.

, Dupuy

, Kien

, Jeandot

The resorption of bone-implanted corals varies with porosity but also with the host reaction. J Biomed Mater Res, 29:909. 1995.

18.

Guillemin

, Patat

J.L.

, Fournie

, Chetail

The use of coral as a bone graft substitute. J Biomed Mater Res, 21:557. 1987.

19.

Leize

E.M.

, Hemmerle

, Voegel

J.C.

, Leize

Characterization and histological analyses of a coral-collagen composite used for bone-replacement graft material: a report of clinical cases. J Mater Sci Mater Med, 10:47. 1999.

20.

Trojani

, Boukhechba

, Scimeca

J.C.

, Vandenbos

, Michiels

J.F.

, Daculsi

, Boileau

, Weiss

, Carle

G.F.

, Rochet

Ectopic bone formation using an injectable biphasic calcium phosphate/Si-HPMC hydrogel composite loaded with undifferentiated bone marrow stromal cells. Biomaterials, 27:3256. 2006.

21.

Schliephake

, Knebel

J.W.

, Aufderheide

, Tauscher

Use of cultivated osteoprogenitor cells to increase bone formation in segmental mandibular defects: an experimental pilot study in sheep. Int J Oral Maxillofac Surg, 30:531. 2001.

22.

Bensaïd

, Oudina

, Viateau

, Potier

, Bousson

, Blanchat

, Sedel

, Guillemin

, Petite

De novo reconstruction of functional bone by tissue engineering in the metatarsal sheep model. Tissue Eng, 11:814. 2005.

23.

Lanyon

L.E.

Using functional loading to influence bone mass and architecture: objectives, mechanisms, and relationship with estrogen of the mechanically adaptive process in bone. Bone, 18:37s. 1996.

24.

Handschel

, Wiesmann

H.P.

, Stratmann

, Kleinheinz

, Meyer

, Joos

TCP is hardly resorbed and not osteoconductive in a nonloading calvarial model. Biomaterials, 23:1689. 2002.

25.

Maniatopoulos

, Sodek

, Melcher

A.H.

Bone formation in vitro by stromal cells obtained from bone marrow of young adult rats. Cell Tissue Res, 254:317. 1988.

26.

Muraglia

, Martin

, Cancedda

, Quarto

A nude mouse model for human bone formation in unloaded conditions. Bone, 22:131s. 1998.

27.

Bruder

S.P.

, Kurth

A.A.

, Shea

, Hayes

W.C.

, Jaiswal

, Kadiyala

Bone regeneration by implantation of purified, culture-expanded human mesenchymal stem cells. J Orthop Res, 16:155. 1998.

28.

Bruder

S.P.

, Kraus

K.H.

, Goldberg

V.M.

, Kadiyala

The effect of implants loaded with autologous mesenchymal stem cells on the healing of canine segmental bone defects. J Bone Joint Surg Am, 80:985. 1998.

29.

Burchardt

The biology of bone graft repair. Clin Orthop Relat Res, 174:28. 1983.

30.

Rickard

D.J.

, Sullivan

T.A.

, Shencker

B.J.

, Leboy

P.S.

, Kazhdan

Induction of rapid osteoblast differentiation in rat bone marrow stromal cell cultures by dexamethasone and BMP-2. Dev Biol, 161:218. 1994.

31.

Harris

S.E.

, Sabatini

M.A.

, Harris

M.A.

, Feng

J.Q.

, Wozney

, Mundy

G.R.

Expression of bone morphogenetic protein messenger RNA in prolonged cultures of fetal rat calvarial cells. J Bone Miner Res, 9:389. 1994.

32.

Hoshino

, Egi

, Terai

, Namikawa

, Takaoka

Repair of long intercalated rib defects using porous beta-tricalcium phosphate cylinders containing recombinant human bone morphogenetic protein-2 in dogs. Biomaterials, 27:4934. 2006.

33.

Bianco

, Gehron Robey

Marrow stromal stem cells. J Clin Invest, 105:1663. 2000.

34.

Ito

, Tanaka

, Fujimoto

, Yasunaga

, Ishida

, Agung

, Ochi

Bone formation using novel interconnected porous calcium hydroxyapatite ceramic hybridized with cultured marrow stromal stem cells derived from Green rat. J Biomed Mater Res A, 69:454. 2004.