Male New Zealand white rabbits weighing 3.5 to 4 kg were used in the study. Approval was obtained from the Institutional Animal Care and Use Committee at the authors’ institute before the study (permit number: 2015121201).
Preparation of rabbit adipose-derived mesenchymal stem cells (AMSCs).
Adipose tissue collected from New Zealand white rabbits was immersed in 2% PSA (Thermo Fisher Scientific, Waltham, MA) in PBS (Sigma-Aldrich, St Louis, MO) and washed twice in PBS by centrifugation. The blood vessels in the adipose tissue were cleaned to remove blood from the tissue. Then, cut them into small pieces and placed in a 0.075% collagenase type I solution for 60 minutes at 37°C. The collagenase activity was neutralized with 10% FBS (Hyclone, Logan, UT) in low-glucose DMEM (Gibco-BRL, Carlsbad, CA) and centrifuged at 1200 rpm for 10 minutes. Aspirated the supernatant and dissolved with PBS. Filter excess tissue with 100um nylon mash. The filtrate was centrifuged at 1200 rpm for 10 minutes and removed the supernatant. The cells were reconstituted with 20% FBS, 100 U/mL penicillin-streptomycin (Hyclone) in low-glucose DMEM, and cultured at 37°C in 5% CO2 incubator.
Flow cytometric analysis of AMSCs.
After 14-day cultivation, the AMSCs (passage 3) were harvested using 0.25% trypsin/EDTA. A 100-µL cell suspension with a cell density of 5 × 105/mL was transferred into an Eppendorf tube and then incubated with the following specific fluorescein isothiocyanate (FITC) conjugated monoclonal antibodies, including CD9 (1:200, AbD Serotec Ltd., Oxford, UK), CD29 (1:200, MilliporeSigma, Burlington, MA), CD44 (1:200, AbD Serotec Ltd.), CD73 (1:200, eBioScience, San Diego, CA), CD90 (1:200, BioLegend, San Diego, CA), CD105 (1:200, Biorbyt, Cambridge, UK), CD45 (1:200, Thermo Fisher Scientific, Waltham, MA) and CD34 (1:200, GeneTex, Irvine, CA) for 20 min at 4°C. Then the stained samples were assessed by a flow cytometer (Beckman coulter, Brea, CA) and analyzed by FlowJo software. CD90 used as a marker for a variety of stem cells.
Preparation of endothelial-differentiated AMSCs impregnated alginate beads (vascularized bone construct).
200 mg of sodium alginate (SA) was dissolved in 10 mL of PBS, resulting in 2 % sodium alginate solution. The SA solution was heated on a hot plate with stirring thoroughly. The homogeneous SA solution was sterilized using a 0.22 µm filter. 10.2 mL of 1 M CaCl2 stock solution was added to 89.8 mL of double-distilled water in 100 mL volumetric flask, resulting in 102 mM CaCl2 solution. The CaCl2 solution was sterilized using a 0.22 µm filter. The ASCs (5 x 105 cells/mL) were suspended in SA solution in a 10 mL conical tube. The mixture of ASCs-SA was dripped into 1 mL of pre-warmed CaCl2 solution. After incubation at 37° C for 5 min, CaCl2 solution was discarded, and the beads were cultured in 24 well plate with a conventional Endothelial Cell Media 2 (PromoCell, Heidelberg, Germany).
The success of endothelial cell differentiation was confirmed by real-time polymerase chain reaction (PCR) and immunofluorescence staining analysis. For real-time PCR analysis, the samples were retrieved at 3, 7, and 14 days. The total cellular RNA of the cells was extracted with RNeasy Mini Kit (QIAGEN, Hilden, Germany) and reverse-transcribed into cDNA using M-MLV Reverse Transcriptase (Promega, Madison, WI). EZtime Real-Time PCR Premix (Yeastern, Taipei, Taiwan) real-time PCR were used to amplify and simultaneously quantify targeted genes on oryctolagus cuniculus CD31 (Forward: TAAAATCGCCGCAGAGTGGG; Reverse: AGTTCCATTTGATTGGCAGCTC) and Von Willebrand Factor (vWF) (Forward: CCGGCGATGTGTGGACC; Reverse: CCCATGCACATACAGGGACAA). Each Q-PCR was performed in triplicate for PCR yield validation. Data were analyzed by the 2 ΔΔCt methods, with normalization by the Ct of the housekeeping gene GAPDH (Forward: GGCAAAGTGGATGTTGTCGC; Reverse: TTCCCGTTCTCAGCCTTGAC). For immunohistochemistry (IHC) staining analysis, the samples were dehydrated and embedded with paraffin after 14 days of induction. These blocks were sectioned with the microtome at 2 µm thickness. Sections were stained with hematoxylin and eosin. For antigen retrieval, the specimens were deparaffinized and treated with 0.01 M citrate buffer (Sigma), pH 6.0, in a pressure cooker for 1 min. For blocking of endogenous peroxidase, sections were incubated in 3% H2O2 for 10 min. The antibody against platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) (Bioss, Woburn, MA) and vWF (Bioss) was used at 4°C overnight. Then, specimens were incubated in secondary antibody conjugated with Alexa Fluor® 488 (Abcam, Cambridge, MA) for 30 min. Hoechst 33342 (Thermo Fisher Scientific) was used to counter staining.
Preparation and scanning electron microscopy (SEM) analysis of engineered periosteum-mimetic cell sheet.
One percent of chitosan solution and collagen solution were prepared using 1% of acetic acid, respectively and equal volume of them was mixed uniformly. Eight hundred microliter of the mixture was injected into a mold (length: 8.5 cm, width: 3 cm, and height: 0.1 cm). The collagen/chitosan membranes were obtained after lyophilization process. The collagen/chitosan membranes were then washed three times at 10 cm of culture dishes. To make the periosteum-mimetic cell sheet (AMSCs seeded collagen/chitosan membrane), 5x105 cells/mL of AMSCs were seeded in the collagen/chitosan membrane evenly at 37°C and cultivated with medium composed of low-glucose DMEM medium containing 10% of FBS and 1% of antibiotics for 7 days in a 5% CO2 incubator. The specimens were dehydrated through a graded series of ethanol solutions, beginning with a 50% solution and progressing through 70%, 95% and 100% solutions. The specimens were dried in an HCP-2 critical-point dryer (Hitachi, Tokyo, Japan) and were sputter-coated using an IB.3 ion coater (EiKo, Tokyo, Japan). Finally, the samples were visualized using a field emission SEM (Hitachi S-5000) to determine the matrix formation of AMSCs on collagen/chitosan membrane at 7 days. Fluorescence staining was used to observe the distribution of AMSCs in collagen/chitosan membrane.
Preparation of biomimetic vascularized bone-periosteum construct (VBPC).
The endothelial-differentiated AMSCs impregnated alginate beads were washed in a saline buffer first, and then were wrapped with periosteum-mimetic cell sheet to form biomimetic VBPC for further in vivo animal experiment.
In vivo animal experiment.
Twenty-four male New Zealand white rabbits were used and were divided into four groups according to the experimental materials and survived for 12 weeks for final analysis. Figure 1A demonstrated the schematic diagram of the experimental design and groups. Group 1: Acellular alginate beads wrapped with cell sheet (Acellular alginate-sheet construct, n = 6); Group 2: Endothelial-differentiated AMSCs impregnated alginate beads (VBC, n = 6); Group 3: Non-differentiated AMSCs impregnated alginate beads wrapped with cell sheet (Non-vascularized AMSCs-alginate-periosteum construct, n = 6); Group 4: Endothelial-differentiated AMSCs impregnated alginate beads wrapped with cell sheet (VBPC, n = 6). In order to determine whether the new bone formation and neovascularization were initiated during the early stage or not, one additional rabbit was operated in each experimental group and was sacrificed at 4 weeks. Under anesthesia with an intramuscular injection of Zoletil (10 mg/kgw) (Virbac Laboratories, Carros, France), the animals underwent intertransverse fusion at the L4-L5 level with different bone graft materials by aseptic manner. The bilateral L4 and L5 transverse processes were exposed and were decorticated by electric burr. The bone graft material was then placed on each side between the transverse processes (Fig. 1B and C). The fascia and skin were closed layer-by-layer with absorbable sutures. After operation, all animals received 200 mg cefamezine per day for 3 days. The animals were allowed unlimited activity without brace application.
Radiographic examination and bone volume-total volume (BV/TV) analysis.
All the animals underwent 2-mm thin-cut computed tomography (CT) scanning of the lumbosacral spine at 12 weeks. The bone volume (BV) was examined in a defined total volume (TV) by Image J software with the BoneJ plugin. The result of bone volume–total volume ratio (BV/TV) of each group was calculated for new bone formation. Briefly, datasets were rearranged with software, AlignStacks plugin. A computed cylinder (12 mm diameter, 20 mm height) was set to the region of interest (ROI), which contained the scaffold. As a first step, the grey-value CT data was converted into binaries. The rearranged image sequence was shortened up to the height of 20 mm. Then, the bone volume in this total cylinder (12 mm diameter, 20 mm height) was computed with the BoneJ plugin. To gain information about bone growth rates into the different sectors of the scaffold, the total cylinder was subdivided. Horizontal slices of 1 mm thickness provided information about bone distribution from the surface at the cortical bone to the internal parts of the cylinder, while hollow cylinders around a central core cylinder gave insight into bone growth from the periphery to the scaffolds’ center. The L4-5 intertransverse fusion areas were collected separately for statistical analysis.
Biomechanical analysis.
The L4-L5 fusion segment was tested for torsional biomechanical strength at weeks 12 post surgery. The L3, L3-4 disc, L5-6 disc, and L6 were embedded along the longitudinal axis in cylindrically shaped epoxy blocks. The length of the L4-L5 non-embedded portion of each specimen was kept identical. The potted samples were then mounted on a Material Testing System machine (GT-7054-A1) (GOTECH Testing Machines Inc., Taichung, Taiwan) for rotational torque assessments. Each individual specimen was tested until ultimate failure in external rotation along its longitudinal axes at 1 degree per second. The maximum torque values were obtained from the torque-rotation angle curve. The results of maximum torque value were presented as means ± standard deviation (SD).
Histology evaluation.
The enbloc spine specimens were fixed in 10% neutral buffered formaldehyde, decalcified, dehydrated through alcohol gradients, cleared and embedded in paraffin blocks. Tissue blocks of the intertransverse fusion areas were sectioned and stained with hematoxylin and eosin (H&E) and Masson’s trichrome methods, and visualized using standard light microscopy. For immunohistochemistry staining, the slide firstly was blocked with normal goat serum for 45 min. Mouse anti-rabbit monoclonal primary antibodies of CD 31 diluted 1:100 (Novus Biologicals, Littleton, CO) were applied at 4°C for 24 h, followed by incubation with HRP-conjugated anti-mouse secondary antibody diluted 1:250. Diaminobenzidine (Sigma-Aldrich) was used as the substrate to develop brown color in the presence of CD31. The slides were dehydrated before being cover slipped.
Statistical analysis
Statistical analyses were performed using the Statistical Package for the Social Sciences for Windows (SPSS, version 12.0; IBM, Armonk, NY, USA). Numerical data are expressed as mean value ± standard deviation (SD). Statistical analysis was performed by analysis of variance (ANOVA) to determine statistical significance. For the histomorphometry data and mechanical strength comparison between groups, the maximal values were analyzed using two-tailed Student’s test. A statistical significant difference was set at p < 0.05.