Cytotherapy, 2014; 0: 1e13
Adipose-derived stem cells alleviate osteoporosis by enchancing osteogenesis and inhibiting adipogenesis in a rabbit model
XINHAI YE1,*, PENG ZHANG2,*, SHAOBO XUE3, YIPIN XU1, JIAN TAN1 & GUANGPENG LIU1 1
Department of Plastic and Reconstructive Surgery, Shanghai Tenth People’s Hospital, Tongji University, Shanghai, PR China, 2Department of Orthopedic Surgery, Provincial Hospital Affiliated to Shandong University, Jinan, PR China, and 3The Central Laboratory of Shanghai Tenth People’s Hospital, Tongji University, Shanghai, PR China
Abstract Background aims. Osteoporosis (OP) is characterized by a reduction in bone quality, which is associated with inadequacies in bone marrow mesenchymal stromal cells (BMSCs). As an alternative cell source to BMSCs, adipose-derived stem cells (ASCs) have been investigated for bone repair because of their osteogenic potential and self-renewal capability. Nevertheless, whether autologous ASCs can be used to promote bone regeneration under osteoporotic conditions has not been elucidated. Methods. The OP rabbit model was established by means of bilateral ovariectomy (OVX). Both BMSCs and ASCs were harvested from OVX rabbits and expanded in vitro. The effects of osteogenic-induced ASCs on the in vitro adipogenic and osteogenic capabilities of BMSCs were evaluated. Autologous ASCs were then encapsulated by calcium alginate gel and transplanted into the distal femurs of OVX rabbits (n ¼ 12). Hydrogel without loading cells was injected into the contralateral femurs as a control. Animals were killed for investigation at 12 weeks after transplantation. Results. Osteogenicinduced ASCs were able to promote osteogenesis and inhibit adipogenesis of osteoporotic BMSCs through activation of the bone morphogenetic protein 2/bone morphogenetic protein receptor type IB signal pathway. Local bone mineral density began to increase at 8 weeks after ASC transplantation (P < 0.05). At 12 weeks, microecomputed tomography and histological evaluation revealed more new bone formation in the cell-treated femurs than in the control group (P < 0.05). Conclusions. This study demonstrated that ASCs could stimulate proliferation and osteogenic differentiation of BMSCs in vitro and enhance bone regeneration in vivo, which suggests that autologous osteogenic-induced ASCs might be useful to alleviate OP temporally. Key Words: adipose-derived stem cells, alginate, bone regeneration, osteoporosis, rabbit, tissue engineering
Introduction Osteoporosis (OP) is a systemic skeletal disease characterized by low bone mass and microarchitectural deterioration of bone tissue, which leads to increased bone fragility and susceptibility to fracture (1,2). With the increase of life expectancy, the prevalence of OP has exceeded that of osteoarthritis or of hard- and soft-tissue healing problems in the world’s aging populations (3,4). Current treatment of OP mainly includes pharmacological drugs, physical activity and adaption of nutrition, aiming at preventing the progression of OP by promoting bone forming and/or decreasing bone resorbing (2). Although the detailed pathologic mechanism of OP remains unknown, some studies indicated that its
occurrence is correlated to the osteogenic deficiency of bone marrow mesenchymal stromal cells (BMSCs) (5,6). Bone regeneration is a complex process involving the intimate relationship between the activities of osteogenic and adipogenic progenitor cells, which are both derived from BMSCs (7). The osteoprogenitor cell number, proliferating capability and differentiating potential of BMSCs decrease with aging or under osteoporotic conditions (8e12). The balance between osteogenesis and adipogenesis of BMSCs is then disrupted, and the latter takes predominance. The increase of fat proportion in bone marrow subsequently induces apoptosis of osteoblasts and proliferation of osteoclasts, which results in further bone resorption and overall bone loss (13,14).
*These authors contributed equally to this work. Correspondence: Guangpeng Liu, MD, Department of Plastic and Reconstructive Surgery, Shanghai Tenth People’s Hospital, Tongji University, 301# Middle Yanchang Road, Shanghai, PR China, 200072. E-mail: [email protected] (Received 14 March 2014; accepted 28 July 2014) http://dx.doi.org/10.1016/j.jcyt.2014.07.009 ISSN 1465-3249 Copyright Ó 2014, International Society for Cellular Therapy. Published by Elsevier Inc. All rights reserved.
2
X. Ye et al.
Therefore, it is speculated that OP might be ameliorated by supplementing osteo-progenitor cells to stimulate osteogenesis and inhibit adipogenesis, and the stem cell transplantation approach becomes attractive. During the past decade, adipose-derived stem cells (ASCs) have been described as an alternative source to BMSCs because of their easy access, multilineage potential and self-renewal capability (15). When induced in vitro into osteoblasts, ASCs could upregulate alkaline phosphatase (ALP) activity, produce osteogenic proteins and deposit mineralized extracellular matrix (ECM) (16e18). In combination with proper scaffolds, osteo-differentiated ASCs were able to form osteoid and enhance bone repair in vivo (19e22). The OP-ameliorating potential of ASCs has also been demonstrated in a syngenic murine model, and young ASCs were shown more effective in restoring bone mineral density (BMD) than were aged cells (23). Nevertheless, the young and aged ASCs were not derived from OP mice. The mechanism of ASCs to enhance BMSC osteogenic potential and the bone microstructure changes after ASC transplantation remained to be elucidated. Moreover, mice cannot represent the gold standard in OP animal models because they lack the Haversian system and do not achieve true skeletal maturity (24,25). The aim of our present study was to investigate whether autologous ASCs could be applied to treat OP in larger animal models. We hypothesized that osteo-differentiated ASCs could reverse the imbalance between osteogenesis and adipogenesis in bone marrow and promote bone regeneration when transplanted into osteoporotic sites. An OP rabbit model was established by means of a bilateral ovariectomy (OVX) procedure in our study. Compared with rodents, rabbits have a moderate body size with sufficient adipose tissue for harvesting ASCs. In addition, they are the smallest animals with Haversian structures in cortical bone and can display seasonal estrogen-deficiency bone loss similar to that in larger mammals (26). Three main questions were sought to be addressed in this study. First, could ASCs derived from OVX rabbits be served as osteoprogenitor cells? Second, what effects did ASCs have on the proliferation and differentiation capabilities of BMSCs under osteoporotic conditions? Finally, could autologous osteogenic-differentiated ASCs be used to alleviate OP in the rabbit model?
Methods Ethical approval All the animals enrolled in this study were cared for and processed in accordance with protocols approved
by the Laboratory Animal Care and Use Committee and the Research Ethics Committee of Tongji University (2012e0083). Rabbit OP model A total of 20, 12-month-old, skeletally mature female New Zealand White rabbits were enrolled and randomized into 2 groups. Group 1 (n ¼ 13) underwent bilateral OVX and group 2 (n ¼ 7) was subjected to sham surgery. Each animal was housed in one cage and fed with standard chow (containing 0.8% calcium and 0.5% phosphate). Serum estrogen levels and BMD values were evaluated before and 8 months after surgery. The estrogen level was determined with the use of an electrochemical immunoassay kit (Roche, Mannheim, Germany). BMD of the rabbit distal femurs (10 mm from the distal end) was measured with the use of dual-energy X-ray absorptiometry (DXA) (Hologic Discovery A, Bedford, MA, USA). In addition, at 8 months after surgery, 1 animal in each group was randomly killed and the distal femurs were harvested for microecomputed tomography (micro-CT) measurement (uCT-80, Scanco Medical, Bassersdorf, Switzerland) and histological examination (hematoxylin and eosin staining) to confirm the osetopenic status. Isolation, in vitro expansion and characterization of rabbit ASCs At 8 months after OVX, rabbit ASCs were isolated and expanded in vitro as described previously (27). Briefly, approximately 5 grams of subcutaneous adipose tissue in the groin area were harvested, washed 3 times with 0.1 M phosphate-buffered saline (PBS, pH 7.4) and treated with 0.075% type I collagenase (Washington Biochemical Corp, Lakewood, NJ, USA) at 37 C for 30 min. Enzymatic activity was neutralized with low-glucose Dulbecco’s Modified Eagle’s Medium (LG-DMEM, Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (HyClone, Logan, UT, USA). After being centrifuged at 1200g for 10 min, the yielding ASCs were resuspended in the basic culture medium (containing LG-DMEM, 10% fetal bovine serum, 100 mg/mL streptomycin and 100 U/mL penicillin; both from Sigma Aldrich, St Louis, MO, USA), and plated at 4 104 cells/cm2 in F100-mm culture dishes (Falcon, B&D Bioscience, San Jose, CA, USA). The medium was changed twice per week, and cells were passaged on reaching 80e90% confluence by the use of 0.25% trypsin/ethylenediaminetetra-acetic acid solution (from Sigma) and seeded at a density of 1 104 cells per cm2. The phenotypic characterization of ASCs of passage 3 (P3) from both sham-surgery and OVX
Effect of adipose-derived stem cells on osteoporosis rabbits was performed with the use of a FACScan cytometer (Coulter Epics Altra, Becton Dickson, San Jose, CA, USA) with specific fluorescein isothiocyanateeconjugated monoclonal antibodies, including CD34, CD45, CD73 and CD90 (all from B&D Bioscience). The multilineage potential of ASCs to differentiate into osteoblasts, adipocytes and chondrocytes was also confirmed in monolayer culture or in pellet culture, as previously described (16e18). Osteogenic differentiation of ASCs ASCs of P3 from both OVX and sham-surgery rabbits were released and seeded into 12-well plates (Falcon) at a density of 10,000 cells per well. The culture medium was replaced with osteogenic medium (OM, consisting of the basic culture medium supplemented with 10 mmol/L b-glycerophosphate, 0.1 mmol/L dexamethasone, 10 mmol/L vitamin D3 and 50 mmol/L L-ascorbic acid; all from Sigma), and was changed twice per week thereafter. Cells incubated in the basic culture medium served as the control groups. ASC number and osteogenic differentiation were determined at days of 3, 6, 9, 12, 15 and 18 in culture through the use of quantitative assays of DNA and calcium (Ca2þ) as previously described (28e30). At each time point, both the osteo-induced and noninduced ASCs were homogenized. DNA content in the cell lysate was quantified spectrofluorometrically with the use of Hoechst 33258 dye (Sigma), and the cell number was obtained by means of correlating with a standard curve, which was generated by lysing serial dilutions of a known concentration of ASCs. The calcium content was determined with the use of a calcium quantification kit (Diagnostic Chemicals, Charlottetown, PEI, Canada). A Varioskan multimode detection reader (Thermo Electron, Waltham, MA, USA) was used for the absorbance/fluorescence measurements, and all measurements were conducted in triplicate. Osteogenic differentiation of ASCs was further confirmed by alizarin red S staining and collagen type I immunohistochemical staining at day 14 according to the standard procedures. Fabrication of ASC/calcium alginate gel composites Sodium alginate solution (1.5 wt%) was prepared by dissolving the sodium alginate powder (Sigma) in PBS and sterilized by autoclaving before use. After ASCs were osteogenically induced in vitro for 2 weeks, ASCs of P3 were released and resuspended in the sterile sodium alginate solution, with a cell concentration of 10 106/mL. The 0.5-mL cell suspension was then mixed with 0.1 mL of 0.1 mol/L
3
CaCl2 solution (Sigma). On contacting Ca2þ, alginate polymerized and encapsulated the cells in suspension. All the cell/calcium alginate gel (ASC/ alginate) composites were transferred into 12-well plates (Falcon) and subcultured in OM for 1 week. To observe the spatial distribution of ASCs inside the gel, we performed scanning electron microscopy (SEM, Jeol, Tokyo, Japan) 3 h after cell encapsulation. At days 1 and 7 in culture, ASC/alginate composites were stained by the Live-Dead Cell Double Staining Kit (BioVision, Milpitas, CA, USA) and subjected to confocal laser scanning microscopy (Leica Microsystem, Mannheim, Germany) examination. Calcein TM, a cell-permeable green fluorescent dye, was used to stain live cells. Dead cells could be easily stained with the use of propidium iodide, a cellenon-permeable red fluorescent dye. Cell number and osteogenic differentiation of ASCs inside the gel were determined at days of 1, 4 and 7 in culture with the use of quantitative assays of DNA, ALP and osteocalcin (OCN), as previously reported (28e30). Briefly, at each time point, ASC/ alginate composites were homogenized and digested in proteinase K (Sigma) at 56 C for 16 h. The homogenates were then subjected to 3 freeze-thawsonicate cycles (30 min at 80 C, 30 min at room temperature, 30 min of sonication) for complete extraction of DNA from the cell cytoplasm. DNA content in the lysate was quantified spectrofluorometrically with the use of Hoechst 33258 dye (Sigma), and ALP activity was measured with the use of an alkaline phosphatase assay (Sigma). OCN concentration in the medium was determined with the use of an OCN enzyme-linked immunoassay (Invitrogen, Carlsbad, CA, USA). The amounts of OCN and ALP produced by each sample were divided by the total cell number of that sample, which was derived from DNA measurement, thereby allowing statistical comparisons to be made between different groups. Effects of ASC conditioned medium on in vitro proliferation and differentiation of BMSCs After being osteogenically induced for 7 days, ASC/ alginate composites were washed with PBS 3 times, and OM was replaced with the fresh basic culture medium. Forty-eight hours later, the conditioned medium from ASC/alginate composites (osteogenic differentiated ASC-CM) was collected, centrifuged at 300g for 5 min, filtered through a 0.22-mm syringe filter and stored at 20 C before use. Samples of bone marrow (2e3 mL) were aspirated from the distal femur of OVX rabbits. BMSCs were isolated, expanded in vitro and characterized as previously reported (31). Seeded into 6-well plates (Falcon)
4
X. Ye et al.
at a density of 106 cells per well, third-passaged BMSCs were induced into osteoblasts and adipocytes in the presence of lineage-specific induction factors. For osteogenic induction, the mixture of osteogenically differentiated ASC-CM and OM (1:1) was used, and OM alone served as the control. For adipogenic differentiation, equal volumes of osteogenically differentiated ASC-CM and adipogenic induction medium (consisting of the basic culture medium plus 0.5 mmol/L isobutyl-methylxanthine, 10 mmol/L insulin and 200 mmol/L indomethacin; all from Sigma) were prepared, and adipogenic induction medium was used as the control medium. Medium was changed twice per week, and cell number was determined at days 1, 7 and 14 of induction as described above. To evaluate the adipogenic and osteogenic differentiation capabilities of BMSCs, total RNA was isolated at days 7 and 14 with the use of Trizol reagent (Invitrogen), and the expression levels of adipogenic- and osteogenic-related genes were measured by means of reverse transcriptaseepolymerase chain reaction (RT-PCR). The primer sequences used for RT-PCR are listed in Table I, including peroxisome proliferator-activated receptor gamma 2 (PPAR-g2), leptin, runt-related transcription factor 2 (RUNX2), ALP and OCN, with a house-keeping gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) used as the internal control. Alizarin red S staining and oil red O staining were also performed at day 14. Deposited calcium and accumulated intracellular lipid were extracted, and the absorbance of supernatants was measured in spectrophotometer for quantification evaluation (7). To explore the underlying molecular mechanisms involved in ASC-regulated osteogenesis, at day 14 after ASC-CM treatment, the bone morphogenetic
protein (BMP) signaling pathways of BMSCs (including BMP-2, BMP receptor type IA [BMPRIA] and BMPR-IB) were examined from the gene and protein expression levels by means of RT-PCR and Western blot methods. The results (band density) were normalized to the expression of GAPDH and semi-quantitatively analyzed by means of image acquisition and analysis software (LabWorks, Upland, CA, USA). Primer sequences for RT-PCR are listed in Table I; antibodies used for Western blot were purchased from Santa Cruz Biotechnology Inc (Shanghai, China). Transplantation of ASC/alginate composites in vivo Rabbit ASCs of P3 were resuspended in the sterile sodium alginate solution, loaded into a 1-mL sterile syringe and mixed with CaCl2 solution to obtain the ASC/alginate composites as described above. Autologous ASC/alginate composites (0.6 mL in volume with approximately 5 106 cells) were injected into the medial condyle cancellous space of the rabbit left distal femurs with the use of a bone marrow aspiration needle connected to the 1-mL syringe (n ¼ 12), and 600 mL of calcium alginate gel without loading cells was injected into the right femoral condyles as the control group (n ¼ 12). BMD micro-CT and histological evaluation BMD measurements (2-dimensional, mg/cm2) of the rabbit distal femurs were performed with the use of DXA at 0, 4, 8 and 12 weeks after cell transplantation, covering a distance of 10 mm starting from the distal end, as described previously (31). All animals were
Table I. Sequences of primers and RT-PCR conditions.
Gene PPAR-g2 Leptin RUNX2 ALP OCN BMP-2 BMPR-IA BMPR-IB GAPDH
Primers (F ¼ forward; R ¼ reverse) F: 50 - CCAGCTAGCCAAAGTCACCAT -30 R: 50 - GTCTCGGAGCCATACAGGATT -30 F: 50 - GTGAAGGTCGGTGTGAACGGATTT -30 R: 50 - CACAGTCTTCTGAGTGGCAGTGAT -30 F: 50 -CCTTCCACTCTCAGTAAGAAGA-30 R: 50 -TAAGTAAAGGTGGCTGGATAGA-30 F: 50 - CTCCATTGTCCACAGGAAATGC -30 R: 50 - TGTGACTGGTGACAGCAGTCTT -30 F: 50 - GAGGAAGTGGGCAGGAGAATG -30 R: 50 - GTAGTAGAAAGGGGACAGGAC -30 F: 50 - CTCGAATTCTGTAACAGATGAGATGCTCCA -30 R: 50 - CGTGGATCCACCAAAGGGGCACGATTCCC -30 F: 50 - CTGAGGCTGAAGGTGATAGC -30 R: 50 - AGTATTTTGCTTCTGGGGAC -30 F: 50 - CTGAAAAGTAAGAACATCCTGGT -30 R: 50 - GGTATGTCAACTTCATTGGTATCAC -30 F: 50 -ATCCCATCACCATCTTCCAG-30 R: 50 -CCATCACGCCACAGTTTCC-30
Amplicon size (bp)
Annealing temperature ( C)
354
55
555
63
149
58
180
57
294
50
209
55
296
50
271
57
383
60
Effect of adipose-derived stem cells on osteoporosis
5
Figure 1. Serum estrogen level (A) and BMD (B) were measured before and 8 months after surgery. Both values in the OVX group were significantly lower than those in the sham-surgery group (*P < 0.05). (C) Micro-CT 3D reconstruction of the distal femurs. The trabecular bone structure of OVX rabbit was much looser than that of the control rabbit (scale bar, 1 mm). (D) Hematoxylin and eosin staining confirmed the reduced bone structure of OVX rabbits, and more adipocytes could be found within the bone marrow (scale bar, 200 mm).
killed at 12 weeks with the use of CO2 euthanasia, and distal femurs were dissected for micro-CT scanning (70 kV, 114 mA, 20-mm pixel size). The region of interest (ROI) was set as a cylinder (2.5 mm in diameter and 3.0 mm in height) at the central part of the medial condyle, that is, the site of cell administration, and was 3-dimensionally (3D) reconstructed. The quantitative morphometric analysis of the ROI was performed automatically with the use of microCT auxiliary software (Scanco Medical, Bassersdorf, Switzerland), including parameters of BMD (3D, mg HA:hydroxyapatite/cm3), bone volume/total volume (BV/TV, %), connectivity density (Conn.D., 1/mm3), structure model index (SMI), trabecular number (Tb.N, 1/mm), trabecular thickness (Tb.Th, mm), trabecular separation (Tb.Sp, mm) and bone surface area/bone volume (BS/BV, 1/mm). After being scanned, samples were fixed in buffered 10% formalin in PBS for 48 h and decalcified in 15% formic acid solution for 3e4 weeks. The decalcified tissues were then dehydrated through an ethanol series and embedded in paraffin. Tissue sections of 5-mm thickness were stained with hematoxylin and eosin, and images were taken with the use of an optical microscope (IX70, Olympus, Tokyo, Japan). The histomorphometry analysis was performed as described previously (29). Briefly, the bone area and total area within the sections were quantified from one photomicrograph by semiautomatically drawing a line around the perimeter of these regions with a software tool, and the total number of pixels was determined within the selected area with the use of the Image-Pro Plus 6.0 software
system (Media Cybernetics, Silver Spring, MD, USA). The trabecular bone thickness was measured directly, and the amount of trabecular bone was displayed as a percentage of the bone area versus the total area within the implant section. For each sample, 10 sections were used for analysis, and 4 images from every section were chosen. Statistical analysis All data collected are presented as mean standard deviation (SD). One-way analysis of variance and the Student-Newman-Keuls test were used to determine possible significant differences (P < 0.05) between groups. Results Characterization of OP rabbit model All the animals survived during the observation period, and no sign of infection or other surgical complications was found. The average body weight of OVX rabbits increased from 2.84 0.45 kg before surgery to 3.28 0.43 kg at 8 months after surgery (n ¼ 13), and that of the sham-surgery rabbits increased from 2.83 0.37 kg to 3.44 0.23 kg (n ¼ 7). No significant difference was found regarding the body weight changes between the 2 groups (P > 0.05). However, a significant reduction of serum estrogen level and BMD value was observed in the OVX group compared with those of the control group (P < 0.05, Figure 1A,B). Compared with their respective baselines before surgery, the
6
X. Ye et al.
Table II. Phenotypic characterization of ASCs. CD markers CD34 CD45 CD73 CD90
ASCs from OVX rabbits (1.57 (2.76 (91.88 (93.26
0.34)% 0.57)% 7.38)% 4.34)%
ASCs from sham-surgery rabbits
P value
0.145 0.487 0.262 0.439
(3.53 (2.33 (97.84 (90.81
0.28)% 0.32)% 6.77)% 8.28)%
Data are expressed as mean SD; n ¼ 6.
estrogen level decreased by 44% and BMD was reduced by 24% at 8 months after OVX, whereas those in the sham-surgery group remained at similar levels. As observed by micro-CT and hematoxylin and eosin evaluation, the trabecular bone structure of OVX rabbit became much looser and thinner than that of the normal rabbit (Figure 1C,D). Hematoxylin and eosin staining also showed more adipocytes inside the bone marrow of OVX rabbits, indicating the replacement of red marrow by adipose-rich yellow marrow as the result of estrogen depletion (31).
Phenotypic characterization and osteogenic differentiation of ASCs from OVX rabbits ASCs could be isolated from adipose tissue of OVX rabbits without difficulty. The flow cytometry analysis revealed that ASCs of P3 from both groups were positive for CD73 and CD90 and negative for CD34 and CD45 (Table II), exhibiting their MSC properties. To confirm their multilineage capability, ASCs of P3 from OVX rabbits were differentiated into chondrocytes, osteoblasts and adipocytes under lineage-specific induction conditions. Chondrogenic differentiation was confirmed by immunohistochemical staining of collagen type II. Osteogenic differentiation was detected by means of von Kossa staining for calcified nodule formation. Adipogenic differentiation was characterized by means of oil red O staining of intracellular lipid droplet accumulation (Supplementary Figure S1). ASCs of P3 from both OVX and sham-surgery groups appeared to have similar fibroblast-like morphology. After induced in OM for 14 days, osteogenic differentiation of ASCs from sham-surgery rabbits was confirmed by means of positive alizarin red staining and collagen type I immunohistochemical staining, compared with the noninduced cells. Similar positive results were observed for osteo-induced ASCs of OVX rabbits, though the alizarin red staining was less intensive than that of their counterpart cells (Figure 2A). Cell number and osteogenic differentiation were quantified by measuring DNA and Ca2þ at days of 3, 6, 9, 12, 15 and 18 in culture. As shown in Figure 2B, the cell number at day 18 in the non-induced sham
group increased by 20.9-fold compared with the initial cell-seeding number, followed by the non-induced OVX group (19.1-fold), the osteo-induced OVX group (14.7-fold) and the osteo-induced sham group (13.3-fold). Cell number in the osteo-induced groups became stabilized from day 9 and less than that in the non-induced groups from day 12 (P < 0.05). However, no difference was found between groups under the same culture condition at each time point checked (P > 0.05). The increase of Ca2þ was obvious in the osteo-induced OVX group over the time course but lower than the osteo-induced sham group at days 15 and 18 (P < 0.05, Figure 2C). ASCs encapsulation with calcium alginate gel After being osteo-induced for 2 weeks, ASCs from OVX rabbits were encapsulated with the calcium alginate gel (Figure 3A,B). Day 14 was chosen because effective osteogenic differentiation could be observed at this stage of the differentiation process (Figure 2A), but more extracellular calcification, which could influence the cell detaching and harvesting, had not occurred to the full extent. SEM imaging showed the interconnected porous structure of the scaffold before cell loading and the aggregates of spherical ASCs in the gel pores after cell encapsulation (Figure 3C,D). Live/dead double staining demonstrated that most ASCs remained viable inside the scaffold. However, fewer viable cells were found at day 7 in culture compared with day 1 (Figure 3E,F). Cell number and osteogenic differentiation within the gel were quantified by measuring DNA, ALP and OCN at days of 1, 4 and 7 in subculture. As shown in Figure 3G, the cell numbers in both OVX and shamsurgery groups decreased from day 1 to day 4 and remained similar thereafter. No significant difference was found between the 2 groups (P > 0.05). ALP activity and OCN content continued to increase from day 1 to day 7. Compared with those in the control group, no significant differences of ALP activity were observed at each time point checked, but the OCN secretion was lower at day 7 (P < 0.05, Figure 3H,I). Effects of osteogenically differentiated ASC-CM on proliferation and differentiation of BMSCs BMSCs harvested from OVX rabbits could be induced into osteoblasts, chondrocytes and adipocytes in vitro (Supplementary Figure S1) but exhibited decreased colony-forming capabilities and osteogenic potential compared with the counterpart cells from normal rabbits (Supplementary Figure S2). To observe the effects of ASC-secreted cytokines and growth factors on the proliferation and differentiation of BMSCs, conditioned medium from osteo-induced ASC/alginate
Effect of adipose-derived stem cells on osteoporosis
7
Figure 2. (A) As observed by light microscopy, ASCs of P3 from both the sham surgery and OVX rabbits exhibited similar spindle-shaped morphology. After induced in OM for 14 days, osteogenic differentiation was confirmed by positive alizarin red staining and positive immunohistochemical staining of collagen type I (scale bars, 500 mm for light microscopy observation and alizarin red staining; 200 mm for collagen type I staining). In vitro proliferation and osteogenic differentiation of osteo-induced and non-induced ASCs were evaluated by means of quantitative assays of DNA (B) and Ca2þ (C). *P < 0.05; n ¼ 6.
composites (osteogenically differentiated ASC-CM) was prepared. BMSCs were cultured in adipo- and osteo-promoting medium with or without osteogenically differentiated ASC-CM for 14 days, respectively. The conditioned medium could slow down BMSC proliferation under adipogenic conditions at day 14 but increased that of osteo-induced cells at days 7 and 14 (P < 0.05, Figure 4A). Before adding osteogenically differentiated ASC-CM, BMSCs showed high expression of adipogenetic gene markers (PPAR-g2 and leptin) and low expression of osteogenic markers (RUNX2, OCN and ALP). With the addition of conditioned medium, the reversed gene expression was observed (Figure 4B). At day 14, the histochemical stainings (alizarin red S and oil red O) and their corresponding quantitative measurements further confirmed that the presence of osteogenically differentiated ASC-
CM promoted osteogenesis and inhibited adipogenesis of BMSCs from OVX rabbits (P < 0.05, Figure 4C). Also at day 14, the gene and protein expression levels of BMP-related signaling pathways were evaluated to determine the underlying molecular mechanisms of ASC-promoting osteogenesis. For BMSCs treated with osteogenically differentiated ASC-CM, the expression of BMP-2 and BMPR-IB, which can promote osteogenesis, was significantly elevated compared with the control cells; the expression of BMPR-IA, which induces adipogenic differentiation, was downregulated (P < 0.05, Figure 5). BMD, micro-CT and histological evaluation The covering range of DXA measurement and the ROI setting for micro-CT analysis are illustrated in
8
X. Ye et al.
Figure 3. (A) ASCs were suspended in sodium alginate solution (red) and mixed with CaCl2 solution. (B) On contacting Ca2þ, the sodium alginate solution solidified within 5 min and the cells were encapsulated inside the gel. SEM images show the interconnected porous structure of the calcium alginate gel (C) and the spheroid ASCs distributing in the gel pores 3 h after cell encapsulation (D) (magnification 300). Live/dead double staining was performed at days 1 (E) and 7 (F) in culture. Most of the cells in the scaffold were viable (green), but the dead cells (red) increased with time (scale bar, 200 mm). Cell number and osteogenic differentiation of ASCs within the gels were evaluated by means of DNA assay (G), ALP activity (H) and OCN level (I), respectively. *P < 0.05, n ¼ 6.
Figure 6A and Figure 6B, respectively. As shown in Figure 6C, autologous ASC transplantation enhanced bone mass and decreased bone loss from 8 weeks, and BMD reached 88% of the normal rabbits’ value at 12 weeks, much higher than that of the alginate-injected group (P < 0.05). Micro-CT scanning showed denser trabecular bone structure at the ASC/alginate-injected area than in the alginate-treated group (Figure 6D). In addition to 3D-BMD, which was significantly increased by 14.8% over the control, other bone quantitative indices of the ASC/alginate group, including BV/ TV, Conn.D., SMI, Tb. N, Tb.Th, Tb.Sp and BS/BV, all differed significantly from those of the alginatetreated group (P < 0.05, Table III). Histological examination confirmed the findings of micro-CT measurement, showing more cancellous bone formation at the ASC/alginate-injected areas than in the alginate-treated group. No evident inflammatory response or necrosis was observed in
either group (Figure 7A). Histomorphometric analysis revealed that both the amount of trabecular bone and the average bone thickness at the injection sites in the ASC/alginate group were significantly higher than those of the control group (19.6% 4.9% versus 14.7% 3.8% and 175 31 mm versus 136 24 mm, P < 0.05, Figure 7B,C). Discussion Primary OP is generally categorized into two types, namely postmenopausal and senile OP. Postmenopausal OP is the result of significant decrease in estrogen levels and senile OP is mainly associated with aging (32). Estrogen plays an important role in the process of bone regeneration and remodeling. It can stimulate osteoblastic differentiation and bone matrix synthesis, maintain osteocyte viability and inhibit osteoclastic apoptosis (33,34). Therefore, estrogenic
Effect of adipose-derived stem cells on osteoporosis
9
Figure 4. (A) By addition of osteogenically differentiated ASC-CM, the proliferation rate of adipo-induced BMSCs was slowed down at day 14, but that of the osteo-induced cells increased at days 7 and 14 (*P < 0.05, n ¼ 5). (B) RT-PCR results demonstrate the downregulated expression of adipo-related genes (PPAR-g2 and leptin) and upelevated expression of osteo-related genes (RUNX2, ALP and OCN) with the presence of osteogenically differentiated ASC-CM. (C) After ASC-CM treatment for 14 days, alizarin red S staining showed more ECM mineralization of osteo-induced BMSCs, and oil red O staining showed less lipid droplet formation of adipo-induced BMSCs. Upper panels indicate quantitative results of staining intensities observed. *P < 0.05, n ¼ 5; magnification 100.
deprivation (OVX) has been the most commonly used experimental model of postmenopausal OP in animals. In our study, we established the OVX rabbit model to mimic estrogen-deficiency osteopenia.
Compared with the sham-surgery group, the estrogen level was decreased by 43% at 8 months after OVX (Figure 1A). BMD value was reduced by 30%, which was approximately 2.8 SD of that in the control group
Figure 5. (A) RT-PCR (left) and Western blot (right) evaluation of BMP-related signaling pathways. As shown by their corresponding quantitative analyses (B and C), expression levels of BMP-2 and BMPR-IB in BMSCs were increased, whereas BMPR-IA was decreased after treatment of osteogenically differentiated ASC-CM, which indicates that ASCs promoted osteogenesis and inhibited adipogenesis of BMSCs through activation of BMP-2 and BMPR-IB signaling pathway. *P < 0.05, n ¼ 5.
10
X. Ye et al.
Figure 6. (A) DXA measurement of the distal femurs covered a distance of 10 mm starting from the distal end (scale bar, 5 mm). (B) For micro-CT quantitative measurement, the ROI was set as a gray cylinder (2.5 mm in diameter and 3.0 mm in height) at the central part of the medial condyle (scale bar, 2 mm). (C) BMD value began to increase at 8 weeks after cell transplantation and was significantly higher than that in the alginate-injected group at 12 weeks (*P < 0.05, n ¼ 12). (D) At 12 weeks after transplantation, micro-CT 2-dimensional scanning (upper) showed dense trabecular bone at the injected area of ASC/alginate-treated femur and sparse bone structure in the alginate-treated femur (square areas). As shown by 3D observation (lower), the trabeculae of the control group appeared looser and thinner than that of the experimental group (scale bars, 1 mm).
(Figure 1B). According to World Health Organization criteria, OP can be defined as a reduction of >2.5 SD of the normal BMD (2,35). These results, combined with the micro-CT and histological findings, confirm the successful establishment of the OP rabbit model. BMSC transplantation represents a potential strategy to treat OP in animal models (31,36). Nevertheless, the acquisition of BMSCs from bone marrow is an invasive, painful procedure, and the osteoprogenitor number and differentiation potential of BMSCs decrease with age (9e13). Compared with BMSCs, ASCs are easier to harvest, and their self-renewal potential and osteogenic differentiation are less affected by aging and estrogen deficiency (37,38). It has been reported that allogeneic or human ASCs can prevent bone loss in OVX mice (23,39). However, whether autologous ASC transplantation could augment bone formation under osteoporotic conditions has not yet been explored. In this study, we obtained ASCs from OVX rabbits and evaluated their suitability as a source of osteogenic precursor cells. ASCs from osteoporotic rabbits exhibited similar cell morphology, phenotypic characteristics and growth rate as those from sham-surgery animals. When osto-induced, these cells demonstrated positive alizarin red staining and collagen type I staining and an increased trend of calcium content.
However, their ECM mineralization extent (alizarin red staining and Ca2þ level) was lower than that of the normal ASCs after being induced in OM over 14 days (Figure 2). It might reflect the hormone-driven changes in ECM composition, which can influence the osteogenic differentiation of ASCs (38). Because of several advantageous features, such as injectability, high porosity, biocompatibility and biodegradation, calcium cross-linked sodium alginate hydrogels have been used as MSC carriers in several in vivo animal studies (31,40). In the present study, homogeneous calcium alginate gels were formulated as an injectable scaffold for osteoinduced ASCs. Confocal laser scanning microscopy results revealed that most ASCs remained viable after encapsulation, but dead cells increased over time (Figure 3E,F). Quantitative analyses showed that these cells retained the capability of osteogenic differentiation within the hydrogel, though their number decreased from day 1 to day 4 (Figure 3GeI). It has been reported that the encapsulation process may result in cell death at early time points (30), and we found that the cell number remained the same from day 4 to day 7. Another possible reason may be that the cells had differentiated into osteoblasts before encapsulation. Osteogenic differentiation of osteoblasts is usually accompanied with a stage of growth arrest (Figure 2B).
13.07 1.29 15.46 1.35
Adipose-derived stem cells alleviate osteoporosis by enchancing osteogenesis and inhibiting adipogenesis in a rabbit model
XINHAI YE1,*, PENG ZHANG2,*, SHAOBO XUE3, YIPIN XU1, JIAN TAN1 & GUANGPENG LIU1 1
Department of Plastic and Reconstructive Surgery, Shanghai Tenth People’s Hospital, Tongji University, Shanghai, PR China, 2Department of Orthopedic Surgery, Provincial Hospital Affiliated to Shandong University, Jinan, PR China, and 3The Central Laboratory of Shanghai Tenth People’s Hospital, Tongji University, Shanghai, PR China
Abstract Background aims. Osteoporosis (OP) is characterized by a reduction in bone quality, which is associated with inadequacies in bone marrow mesenchymal stromal cells (BMSCs). As an alternative cell source to BMSCs, adipose-derived stem cells (ASCs) have been investigated for bone repair because of their osteogenic potential and self-renewal capability. Nevertheless, whether autologous ASCs can be used to promote bone regeneration under osteoporotic conditions has not been elucidated. Methods. The OP rabbit model was established by means of bilateral ovariectomy (OVX). Both BMSCs and ASCs were harvested from OVX rabbits and expanded in vitro. The effects of osteogenic-induced ASCs on the in vitro adipogenic and osteogenic capabilities of BMSCs were evaluated. Autologous ASCs were then encapsulated by calcium alginate gel and transplanted into the distal femurs of OVX rabbits (n ¼ 12). Hydrogel without loading cells was injected into the contralateral femurs as a control. Animals were killed for investigation at 12 weeks after transplantation. Results. Osteogenicinduced ASCs were able to promote osteogenesis and inhibit adipogenesis of osteoporotic BMSCs through activation of the bone morphogenetic protein 2/bone morphogenetic protein receptor type IB signal pathway. Local bone mineral density began to increase at 8 weeks after ASC transplantation (P < 0.05). At 12 weeks, microecomputed tomography and histological evaluation revealed more new bone formation in the cell-treated femurs than in the control group (P < 0.05). Conclusions. This study demonstrated that ASCs could stimulate proliferation and osteogenic differentiation of BMSCs in vitro and enhance bone regeneration in vivo, which suggests that autologous osteogenic-induced ASCs might be useful to alleviate OP temporally. Key Words: adipose-derived stem cells, alginate, bone regeneration, osteoporosis, rabbit, tissue engineering
Introduction Osteoporosis (OP) is a systemic skeletal disease characterized by low bone mass and microarchitectural deterioration of bone tissue, which leads to increased bone fragility and susceptibility to fracture (1,2). With the increase of life expectancy, the prevalence of OP has exceeded that of osteoarthritis or of hard- and soft-tissue healing problems in the world’s aging populations (3,4). Current treatment of OP mainly includes pharmacological drugs, physical activity and adaption of nutrition, aiming at preventing the progression of OP by promoting bone forming and/or decreasing bone resorbing (2). Although the detailed pathologic mechanism of OP remains unknown, some studies indicated that its
occurrence is correlated to the osteogenic deficiency of bone marrow mesenchymal stromal cells (BMSCs) (5,6). Bone regeneration is a complex process involving the intimate relationship between the activities of osteogenic and adipogenic progenitor cells, which are both derived from BMSCs (7). The osteoprogenitor cell number, proliferating capability and differentiating potential of BMSCs decrease with aging or under osteoporotic conditions (8e12). The balance between osteogenesis and adipogenesis of BMSCs is then disrupted, and the latter takes predominance. The increase of fat proportion in bone marrow subsequently induces apoptosis of osteoblasts and proliferation of osteoclasts, which results in further bone resorption and overall bone loss (13,14).
*These authors contributed equally to this work. Correspondence: Guangpeng Liu, MD, Department of Plastic and Reconstructive Surgery, Shanghai Tenth People’s Hospital, Tongji University, 301# Middle Yanchang Road, Shanghai, PR China, 200072. E-mail: [email protected] (Received 14 March 2014; accepted 28 July 2014) http://dx.doi.org/10.1016/j.jcyt.2014.07.009 ISSN 1465-3249 Copyright Ó 2014, International Society for Cellular Therapy. Published by Elsevier Inc. All rights reserved.
2
X. Ye et al.
Therefore, it is speculated that OP might be ameliorated by supplementing osteo-progenitor cells to stimulate osteogenesis and inhibit adipogenesis, and the stem cell transplantation approach becomes attractive. During the past decade, adipose-derived stem cells (ASCs) have been described as an alternative source to BMSCs because of their easy access, multilineage potential and self-renewal capability (15). When induced in vitro into osteoblasts, ASCs could upregulate alkaline phosphatase (ALP) activity, produce osteogenic proteins and deposit mineralized extracellular matrix (ECM) (16e18). In combination with proper scaffolds, osteo-differentiated ASCs were able to form osteoid and enhance bone repair in vivo (19e22). The OP-ameliorating potential of ASCs has also been demonstrated in a syngenic murine model, and young ASCs were shown more effective in restoring bone mineral density (BMD) than were aged cells (23). Nevertheless, the young and aged ASCs were not derived from OP mice. The mechanism of ASCs to enhance BMSC osteogenic potential and the bone microstructure changes after ASC transplantation remained to be elucidated. Moreover, mice cannot represent the gold standard in OP animal models because they lack the Haversian system and do not achieve true skeletal maturity (24,25). The aim of our present study was to investigate whether autologous ASCs could be applied to treat OP in larger animal models. We hypothesized that osteo-differentiated ASCs could reverse the imbalance between osteogenesis and adipogenesis in bone marrow and promote bone regeneration when transplanted into osteoporotic sites. An OP rabbit model was established by means of a bilateral ovariectomy (OVX) procedure in our study. Compared with rodents, rabbits have a moderate body size with sufficient adipose tissue for harvesting ASCs. In addition, they are the smallest animals with Haversian structures in cortical bone and can display seasonal estrogen-deficiency bone loss similar to that in larger mammals (26). Three main questions were sought to be addressed in this study. First, could ASCs derived from OVX rabbits be served as osteoprogenitor cells? Second, what effects did ASCs have on the proliferation and differentiation capabilities of BMSCs under osteoporotic conditions? Finally, could autologous osteogenic-differentiated ASCs be used to alleviate OP in the rabbit model?
Methods Ethical approval All the animals enrolled in this study were cared for and processed in accordance with protocols approved
by the Laboratory Animal Care and Use Committee and the Research Ethics Committee of Tongji University (2012e0083). Rabbit OP model A total of 20, 12-month-old, skeletally mature female New Zealand White rabbits were enrolled and randomized into 2 groups. Group 1 (n ¼ 13) underwent bilateral OVX and group 2 (n ¼ 7) was subjected to sham surgery. Each animal was housed in one cage and fed with standard chow (containing 0.8% calcium and 0.5% phosphate). Serum estrogen levels and BMD values were evaluated before and 8 months after surgery. The estrogen level was determined with the use of an electrochemical immunoassay kit (Roche, Mannheim, Germany). BMD of the rabbit distal femurs (10 mm from the distal end) was measured with the use of dual-energy X-ray absorptiometry (DXA) (Hologic Discovery A, Bedford, MA, USA). In addition, at 8 months after surgery, 1 animal in each group was randomly killed and the distal femurs were harvested for microecomputed tomography (micro-CT) measurement (uCT-80, Scanco Medical, Bassersdorf, Switzerland) and histological examination (hematoxylin and eosin staining) to confirm the osetopenic status. Isolation, in vitro expansion and characterization of rabbit ASCs At 8 months after OVX, rabbit ASCs were isolated and expanded in vitro as described previously (27). Briefly, approximately 5 grams of subcutaneous adipose tissue in the groin area were harvested, washed 3 times with 0.1 M phosphate-buffered saline (PBS, pH 7.4) and treated with 0.075% type I collagenase (Washington Biochemical Corp, Lakewood, NJ, USA) at 37 C for 30 min. Enzymatic activity was neutralized with low-glucose Dulbecco’s Modified Eagle’s Medium (LG-DMEM, Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (HyClone, Logan, UT, USA). After being centrifuged at 1200g for 10 min, the yielding ASCs were resuspended in the basic culture medium (containing LG-DMEM, 10% fetal bovine serum, 100 mg/mL streptomycin and 100 U/mL penicillin; both from Sigma Aldrich, St Louis, MO, USA), and plated at 4 104 cells/cm2 in F100-mm culture dishes (Falcon, B&D Bioscience, San Jose, CA, USA). The medium was changed twice per week, and cells were passaged on reaching 80e90% confluence by the use of 0.25% trypsin/ethylenediaminetetra-acetic acid solution (from Sigma) and seeded at a density of 1 104 cells per cm2. The phenotypic characterization of ASCs of passage 3 (P3) from both sham-surgery and OVX
Effect of adipose-derived stem cells on osteoporosis rabbits was performed with the use of a FACScan cytometer (Coulter Epics Altra, Becton Dickson, San Jose, CA, USA) with specific fluorescein isothiocyanateeconjugated monoclonal antibodies, including CD34, CD45, CD73 and CD90 (all from B&D Bioscience). The multilineage potential of ASCs to differentiate into osteoblasts, adipocytes and chondrocytes was also confirmed in monolayer culture or in pellet culture, as previously described (16e18). Osteogenic differentiation of ASCs ASCs of P3 from both OVX and sham-surgery rabbits were released and seeded into 12-well plates (Falcon) at a density of 10,000 cells per well. The culture medium was replaced with osteogenic medium (OM, consisting of the basic culture medium supplemented with 10 mmol/L b-glycerophosphate, 0.1 mmol/L dexamethasone, 10 mmol/L vitamin D3 and 50 mmol/L L-ascorbic acid; all from Sigma), and was changed twice per week thereafter. Cells incubated in the basic culture medium served as the control groups. ASC number and osteogenic differentiation were determined at days of 3, 6, 9, 12, 15 and 18 in culture through the use of quantitative assays of DNA and calcium (Ca2þ) as previously described (28e30). At each time point, both the osteo-induced and noninduced ASCs were homogenized. DNA content in the cell lysate was quantified spectrofluorometrically with the use of Hoechst 33258 dye (Sigma), and the cell number was obtained by means of correlating with a standard curve, which was generated by lysing serial dilutions of a known concentration of ASCs. The calcium content was determined with the use of a calcium quantification kit (Diagnostic Chemicals, Charlottetown, PEI, Canada). A Varioskan multimode detection reader (Thermo Electron, Waltham, MA, USA) was used for the absorbance/fluorescence measurements, and all measurements were conducted in triplicate. Osteogenic differentiation of ASCs was further confirmed by alizarin red S staining and collagen type I immunohistochemical staining at day 14 according to the standard procedures. Fabrication of ASC/calcium alginate gel composites Sodium alginate solution (1.5 wt%) was prepared by dissolving the sodium alginate powder (Sigma) in PBS and sterilized by autoclaving before use. After ASCs were osteogenically induced in vitro for 2 weeks, ASCs of P3 were released and resuspended in the sterile sodium alginate solution, with a cell concentration of 10 106/mL. The 0.5-mL cell suspension was then mixed with 0.1 mL of 0.1 mol/L
3
CaCl2 solution (Sigma). On contacting Ca2þ, alginate polymerized and encapsulated the cells in suspension. All the cell/calcium alginate gel (ASC/ alginate) composites were transferred into 12-well plates (Falcon) and subcultured in OM for 1 week. To observe the spatial distribution of ASCs inside the gel, we performed scanning electron microscopy (SEM, Jeol, Tokyo, Japan) 3 h after cell encapsulation. At days 1 and 7 in culture, ASC/alginate composites were stained by the Live-Dead Cell Double Staining Kit (BioVision, Milpitas, CA, USA) and subjected to confocal laser scanning microscopy (Leica Microsystem, Mannheim, Germany) examination. Calcein TM, a cell-permeable green fluorescent dye, was used to stain live cells. Dead cells could be easily stained with the use of propidium iodide, a cellenon-permeable red fluorescent dye. Cell number and osteogenic differentiation of ASCs inside the gel were determined at days of 1, 4 and 7 in culture with the use of quantitative assays of DNA, ALP and osteocalcin (OCN), as previously reported (28e30). Briefly, at each time point, ASC/ alginate composites were homogenized and digested in proteinase K (Sigma) at 56 C for 16 h. The homogenates were then subjected to 3 freeze-thawsonicate cycles (30 min at 80 C, 30 min at room temperature, 30 min of sonication) for complete extraction of DNA from the cell cytoplasm. DNA content in the lysate was quantified spectrofluorometrically with the use of Hoechst 33258 dye (Sigma), and ALP activity was measured with the use of an alkaline phosphatase assay (Sigma). OCN concentration in the medium was determined with the use of an OCN enzyme-linked immunoassay (Invitrogen, Carlsbad, CA, USA). The amounts of OCN and ALP produced by each sample were divided by the total cell number of that sample, which was derived from DNA measurement, thereby allowing statistical comparisons to be made between different groups. Effects of ASC conditioned medium on in vitro proliferation and differentiation of BMSCs After being osteogenically induced for 7 days, ASC/ alginate composites were washed with PBS 3 times, and OM was replaced with the fresh basic culture medium. Forty-eight hours later, the conditioned medium from ASC/alginate composites (osteogenic differentiated ASC-CM) was collected, centrifuged at 300g for 5 min, filtered through a 0.22-mm syringe filter and stored at 20 C before use. Samples of bone marrow (2e3 mL) were aspirated from the distal femur of OVX rabbits. BMSCs were isolated, expanded in vitro and characterized as previously reported (31). Seeded into 6-well plates (Falcon)
4
X. Ye et al.
at a density of 106 cells per well, third-passaged BMSCs were induced into osteoblasts and adipocytes in the presence of lineage-specific induction factors. For osteogenic induction, the mixture of osteogenically differentiated ASC-CM and OM (1:1) was used, and OM alone served as the control. For adipogenic differentiation, equal volumes of osteogenically differentiated ASC-CM and adipogenic induction medium (consisting of the basic culture medium plus 0.5 mmol/L isobutyl-methylxanthine, 10 mmol/L insulin and 200 mmol/L indomethacin; all from Sigma) were prepared, and adipogenic induction medium was used as the control medium. Medium was changed twice per week, and cell number was determined at days 1, 7 and 14 of induction as described above. To evaluate the adipogenic and osteogenic differentiation capabilities of BMSCs, total RNA was isolated at days 7 and 14 with the use of Trizol reagent (Invitrogen), and the expression levels of adipogenic- and osteogenic-related genes were measured by means of reverse transcriptaseepolymerase chain reaction (RT-PCR). The primer sequences used for RT-PCR are listed in Table I, including peroxisome proliferator-activated receptor gamma 2 (PPAR-g2), leptin, runt-related transcription factor 2 (RUNX2), ALP and OCN, with a house-keeping gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) used as the internal control. Alizarin red S staining and oil red O staining were also performed at day 14. Deposited calcium and accumulated intracellular lipid were extracted, and the absorbance of supernatants was measured in spectrophotometer for quantification evaluation (7). To explore the underlying molecular mechanisms involved in ASC-regulated osteogenesis, at day 14 after ASC-CM treatment, the bone morphogenetic
protein (BMP) signaling pathways of BMSCs (including BMP-2, BMP receptor type IA [BMPRIA] and BMPR-IB) were examined from the gene and protein expression levels by means of RT-PCR and Western blot methods. The results (band density) were normalized to the expression of GAPDH and semi-quantitatively analyzed by means of image acquisition and analysis software (LabWorks, Upland, CA, USA). Primer sequences for RT-PCR are listed in Table I; antibodies used for Western blot were purchased from Santa Cruz Biotechnology Inc (Shanghai, China). Transplantation of ASC/alginate composites in vivo Rabbit ASCs of P3 were resuspended in the sterile sodium alginate solution, loaded into a 1-mL sterile syringe and mixed with CaCl2 solution to obtain the ASC/alginate composites as described above. Autologous ASC/alginate composites (0.6 mL in volume with approximately 5 106 cells) were injected into the medial condyle cancellous space of the rabbit left distal femurs with the use of a bone marrow aspiration needle connected to the 1-mL syringe (n ¼ 12), and 600 mL of calcium alginate gel without loading cells was injected into the right femoral condyles as the control group (n ¼ 12). BMD micro-CT and histological evaluation BMD measurements (2-dimensional, mg/cm2) of the rabbit distal femurs were performed with the use of DXA at 0, 4, 8 and 12 weeks after cell transplantation, covering a distance of 10 mm starting from the distal end, as described previously (31). All animals were
Table I. Sequences of primers and RT-PCR conditions.
Gene PPAR-g2 Leptin RUNX2 ALP OCN BMP-2 BMPR-IA BMPR-IB GAPDH
Primers (F ¼ forward; R ¼ reverse) F: 50 - CCAGCTAGCCAAAGTCACCAT -30 R: 50 - GTCTCGGAGCCATACAGGATT -30 F: 50 - GTGAAGGTCGGTGTGAACGGATTT -30 R: 50 - CACAGTCTTCTGAGTGGCAGTGAT -30 F: 50 -CCTTCCACTCTCAGTAAGAAGA-30 R: 50 -TAAGTAAAGGTGGCTGGATAGA-30 F: 50 - CTCCATTGTCCACAGGAAATGC -30 R: 50 - TGTGACTGGTGACAGCAGTCTT -30 F: 50 - GAGGAAGTGGGCAGGAGAATG -30 R: 50 - GTAGTAGAAAGGGGACAGGAC -30 F: 50 - CTCGAATTCTGTAACAGATGAGATGCTCCA -30 R: 50 - CGTGGATCCACCAAAGGGGCACGATTCCC -30 F: 50 - CTGAGGCTGAAGGTGATAGC -30 R: 50 - AGTATTTTGCTTCTGGGGAC -30 F: 50 - CTGAAAAGTAAGAACATCCTGGT -30 R: 50 - GGTATGTCAACTTCATTGGTATCAC -30 F: 50 -ATCCCATCACCATCTTCCAG-30 R: 50 -CCATCACGCCACAGTTTCC-30
Amplicon size (bp)
Annealing temperature ( C)
354
55
555
63
149
58
180
57
294
50
209
55
296
50
271
57
383
60
Effect of adipose-derived stem cells on osteoporosis
5
Figure 1. Serum estrogen level (A) and BMD (B) were measured before and 8 months after surgery. Both values in the OVX group were significantly lower than those in the sham-surgery group (*P < 0.05). (C) Micro-CT 3D reconstruction of the distal femurs. The trabecular bone structure of OVX rabbit was much looser than that of the control rabbit (scale bar, 1 mm). (D) Hematoxylin and eosin staining confirmed the reduced bone structure of OVX rabbits, and more adipocytes could be found within the bone marrow (scale bar, 200 mm).
killed at 12 weeks with the use of CO2 euthanasia, and distal femurs were dissected for micro-CT scanning (70 kV, 114 mA, 20-mm pixel size). The region of interest (ROI) was set as a cylinder (2.5 mm in diameter and 3.0 mm in height) at the central part of the medial condyle, that is, the site of cell administration, and was 3-dimensionally (3D) reconstructed. The quantitative morphometric analysis of the ROI was performed automatically with the use of microCT auxiliary software (Scanco Medical, Bassersdorf, Switzerland), including parameters of BMD (3D, mg HA:hydroxyapatite/cm3), bone volume/total volume (BV/TV, %), connectivity density (Conn.D., 1/mm3), structure model index (SMI), trabecular number (Tb.N, 1/mm), trabecular thickness (Tb.Th, mm), trabecular separation (Tb.Sp, mm) and bone surface area/bone volume (BS/BV, 1/mm). After being scanned, samples were fixed in buffered 10% formalin in PBS for 48 h and decalcified in 15% formic acid solution for 3e4 weeks. The decalcified tissues were then dehydrated through an ethanol series and embedded in paraffin. Tissue sections of 5-mm thickness were stained with hematoxylin and eosin, and images were taken with the use of an optical microscope (IX70, Olympus, Tokyo, Japan). The histomorphometry analysis was performed as described previously (29). Briefly, the bone area and total area within the sections were quantified from one photomicrograph by semiautomatically drawing a line around the perimeter of these regions with a software tool, and the total number of pixels was determined within the selected area with the use of the Image-Pro Plus 6.0 software
system (Media Cybernetics, Silver Spring, MD, USA). The trabecular bone thickness was measured directly, and the amount of trabecular bone was displayed as a percentage of the bone area versus the total area within the implant section. For each sample, 10 sections were used for analysis, and 4 images from every section were chosen. Statistical analysis All data collected are presented as mean standard deviation (SD). One-way analysis of variance and the Student-Newman-Keuls test were used to determine possible significant differences (P < 0.05) between groups. Results Characterization of OP rabbit model All the animals survived during the observation period, and no sign of infection or other surgical complications was found. The average body weight of OVX rabbits increased from 2.84 0.45 kg before surgery to 3.28 0.43 kg at 8 months after surgery (n ¼ 13), and that of the sham-surgery rabbits increased from 2.83 0.37 kg to 3.44 0.23 kg (n ¼ 7). No significant difference was found regarding the body weight changes between the 2 groups (P > 0.05). However, a significant reduction of serum estrogen level and BMD value was observed in the OVX group compared with those of the control group (P < 0.05, Figure 1A,B). Compared with their respective baselines before surgery, the
6
X. Ye et al.
Table II. Phenotypic characterization of ASCs. CD markers CD34 CD45 CD73 CD90
ASCs from OVX rabbits (1.57 (2.76 (91.88 (93.26
0.34)% 0.57)% 7.38)% 4.34)%
ASCs from sham-surgery rabbits
P value
0.145 0.487 0.262 0.439
(3.53 (2.33 (97.84 (90.81
0.28)% 0.32)% 6.77)% 8.28)%
Data are expressed as mean SD; n ¼ 6.
estrogen level decreased by 44% and BMD was reduced by 24% at 8 months after OVX, whereas those in the sham-surgery group remained at similar levels. As observed by micro-CT and hematoxylin and eosin evaluation, the trabecular bone structure of OVX rabbit became much looser and thinner than that of the normal rabbit (Figure 1C,D). Hematoxylin and eosin staining also showed more adipocytes inside the bone marrow of OVX rabbits, indicating the replacement of red marrow by adipose-rich yellow marrow as the result of estrogen depletion (31).
Phenotypic characterization and osteogenic differentiation of ASCs from OVX rabbits ASCs could be isolated from adipose tissue of OVX rabbits without difficulty. The flow cytometry analysis revealed that ASCs of P3 from both groups were positive for CD73 and CD90 and negative for CD34 and CD45 (Table II), exhibiting their MSC properties. To confirm their multilineage capability, ASCs of P3 from OVX rabbits were differentiated into chondrocytes, osteoblasts and adipocytes under lineage-specific induction conditions. Chondrogenic differentiation was confirmed by immunohistochemical staining of collagen type II. Osteogenic differentiation was detected by means of von Kossa staining for calcified nodule formation. Adipogenic differentiation was characterized by means of oil red O staining of intracellular lipid droplet accumulation (Supplementary Figure S1). ASCs of P3 from both OVX and sham-surgery groups appeared to have similar fibroblast-like morphology. After induced in OM for 14 days, osteogenic differentiation of ASCs from sham-surgery rabbits was confirmed by means of positive alizarin red staining and collagen type I immunohistochemical staining, compared with the noninduced cells. Similar positive results were observed for osteo-induced ASCs of OVX rabbits, though the alizarin red staining was less intensive than that of their counterpart cells (Figure 2A). Cell number and osteogenic differentiation were quantified by measuring DNA and Ca2þ at days of 3, 6, 9, 12, 15 and 18 in culture. As shown in Figure 2B, the cell number at day 18 in the non-induced sham
group increased by 20.9-fold compared with the initial cell-seeding number, followed by the non-induced OVX group (19.1-fold), the osteo-induced OVX group (14.7-fold) and the osteo-induced sham group (13.3-fold). Cell number in the osteo-induced groups became stabilized from day 9 and less than that in the non-induced groups from day 12 (P < 0.05). However, no difference was found between groups under the same culture condition at each time point checked (P > 0.05). The increase of Ca2þ was obvious in the osteo-induced OVX group over the time course but lower than the osteo-induced sham group at days 15 and 18 (P < 0.05, Figure 2C). ASCs encapsulation with calcium alginate gel After being osteo-induced for 2 weeks, ASCs from OVX rabbits were encapsulated with the calcium alginate gel (Figure 3A,B). Day 14 was chosen because effective osteogenic differentiation could be observed at this stage of the differentiation process (Figure 2A), but more extracellular calcification, which could influence the cell detaching and harvesting, had not occurred to the full extent. SEM imaging showed the interconnected porous structure of the scaffold before cell loading and the aggregates of spherical ASCs in the gel pores after cell encapsulation (Figure 3C,D). Live/dead double staining demonstrated that most ASCs remained viable inside the scaffold. However, fewer viable cells were found at day 7 in culture compared with day 1 (Figure 3E,F). Cell number and osteogenic differentiation within the gel were quantified by measuring DNA, ALP and OCN at days of 1, 4 and 7 in subculture. As shown in Figure 3G, the cell numbers in both OVX and shamsurgery groups decreased from day 1 to day 4 and remained similar thereafter. No significant difference was found between the 2 groups (P > 0.05). ALP activity and OCN content continued to increase from day 1 to day 7. Compared with those in the control group, no significant differences of ALP activity were observed at each time point checked, but the OCN secretion was lower at day 7 (P < 0.05, Figure 3H,I). Effects of osteogenically differentiated ASC-CM on proliferation and differentiation of BMSCs BMSCs harvested from OVX rabbits could be induced into osteoblasts, chondrocytes and adipocytes in vitro (Supplementary Figure S1) but exhibited decreased colony-forming capabilities and osteogenic potential compared with the counterpart cells from normal rabbits (Supplementary Figure S2). To observe the effects of ASC-secreted cytokines and growth factors on the proliferation and differentiation of BMSCs, conditioned medium from osteo-induced ASC/alginate
Effect of adipose-derived stem cells on osteoporosis
7
Figure 2. (A) As observed by light microscopy, ASCs of P3 from both the sham surgery and OVX rabbits exhibited similar spindle-shaped morphology. After induced in OM for 14 days, osteogenic differentiation was confirmed by positive alizarin red staining and positive immunohistochemical staining of collagen type I (scale bars, 500 mm for light microscopy observation and alizarin red staining; 200 mm for collagen type I staining). In vitro proliferation and osteogenic differentiation of osteo-induced and non-induced ASCs were evaluated by means of quantitative assays of DNA (B) and Ca2þ (C). *P < 0.05; n ¼ 6.
composites (osteogenically differentiated ASC-CM) was prepared. BMSCs were cultured in adipo- and osteo-promoting medium with or without osteogenically differentiated ASC-CM for 14 days, respectively. The conditioned medium could slow down BMSC proliferation under adipogenic conditions at day 14 but increased that of osteo-induced cells at days 7 and 14 (P < 0.05, Figure 4A). Before adding osteogenically differentiated ASC-CM, BMSCs showed high expression of adipogenetic gene markers (PPAR-g2 and leptin) and low expression of osteogenic markers (RUNX2, OCN and ALP). With the addition of conditioned medium, the reversed gene expression was observed (Figure 4B). At day 14, the histochemical stainings (alizarin red S and oil red O) and their corresponding quantitative measurements further confirmed that the presence of osteogenically differentiated ASC-
CM promoted osteogenesis and inhibited adipogenesis of BMSCs from OVX rabbits (P < 0.05, Figure 4C). Also at day 14, the gene and protein expression levels of BMP-related signaling pathways were evaluated to determine the underlying molecular mechanisms of ASC-promoting osteogenesis. For BMSCs treated with osteogenically differentiated ASC-CM, the expression of BMP-2 and BMPR-IB, which can promote osteogenesis, was significantly elevated compared with the control cells; the expression of BMPR-IA, which induces adipogenic differentiation, was downregulated (P < 0.05, Figure 5). BMD, micro-CT and histological evaluation The covering range of DXA measurement and the ROI setting for micro-CT analysis are illustrated in
8
X. Ye et al.
Figure 3. (A) ASCs were suspended in sodium alginate solution (red) and mixed with CaCl2 solution. (B) On contacting Ca2þ, the sodium alginate solution solidified within 5 min and the cells were encapsulated inside the gel. SEM images show the interconnected porous structure of the calcium alginate gel (C) and the spheroid ASCs distributing in the gel pores 3 h after cell encapsulation (D) (magnification 300). Live/dead double staining was performed at days 1 (E) and 7 (F) in culture. Most of the cells in the scaffold were viable (green), but the dead cells (red) increased with time (scale bar, 200 mm). Cell number and osteogenic differentiation of ASCs within the gels were evaluated by means of DNA assay (G), ALP activity (H) and OCN level (I), respectively. *P < 0.05, n ¼ 6.
Figure 6A and Figure 6B, respectively. As shown in Figure 6C, autologous ASC transplantation enhanced bone mass and decreased bone loss from 8 weeks, and BMD reached 88% of the normal rabbits’ value at 12 weeks, much higher than that of the alginate-injected group (P < 0.05). Micro-CT scanning showed denser trabecular bone structure at the ASC/alginate-injected area than in the alginate-treated group (Figure 6D). In addition to 3D-BMD, which was significantly increased by 14.8% over the control, other bone quantitative indices of the ASC/alginate group, including BV/ TV, Conn.D., SMI, Tb. N, Tb.Th, Tb.Sp and BS/BV, all differed significantly from those of the alginatetreated group (P < 0.05, Table III). Histological examination confirmed the findings of micro-CT measurement, showing more cancellous bone formation at the ASC/alginate-injected areas than in the alginate-treated group. No evident inflammatory response or necrosis was observed in
either group (Figure 7A). Histomorphometric analysis revealed that both the amount of trabecular bone and the average bone thickness at the injection sites in the ASC/alginate group were significantly higher than those of the control group (19.6% 4.9% versus 14.7% 3.8% and 175 31 mm versus 136 24 mm, P < 0.05, Figure 7B,C). Discussion Primary OP is generally categorized into two types, namely postmenopausal and senile OP. Postmenopausal OP is the result of significant decrease in estrogen levels and senile OP is mainly associated with aging (32). Estrogen plays an important role in the process of bone regeneration and remodeling. It can stimulate osteoblastic differentiation and bone matrix synthesis, maintain osteocyte viability and inhibit osteoclastic apoptosis (33,34). Therefore, estrogenic
Effect of adipose-derived stem cells on osteoporosis
9
Figure 4. (A) By addition of osteogenically differentiated ASC-CM, the proliferation rate of adipo-induced BMSCs was slowed down at day 14, but that of the osteo-induced cells increased at days 7 and 14 (*P < 0.05, n ¼ 5). (B) RT-PCR results demonstrate the downregulated expression of adipo-related genes (PPAR-g2 and leptin) and upelevated expression of osteo-related genes (RUNX2, ALP and OCN) with the presence of osteogenically differentiated ASC-CM. (C) After ASC-CM treatment for 14 days, alizarin red S staining showed more ECM mineralization of osteo-induced BMSCs, and oil red O staining showed less lipid droplet formation of adipo-induced BMSCs. Upper panels indicate quantitative results of staining intensities observed. *P < 0.05, n ¼ 5; magnification 100.
deprivation (OVX) has been the most commonly used experimental model of postmenopausal OP in animals. In our study, we established the OVX rabbit model to mimic estrogen-deficiency osteopenia.
Compared with the sham-surgery group, the estrogen level was decreased by 43% at 8 months after OVX (Figure 1A). BMD value was reduced by 30%, which was approximately 2.8 SD of that in the control group
Figure 5. (A) RT-PCR (left) and Western blot (right) evaluation of BMP-related signaling pathways. As shown by their corresponding quantitative analyses (B and C), expression levels of BMP-2 and BMPR-IB in BMSCs were increased, whereas BMPR-IA was decreased after treatment of osteogenically differentiated ASC-CM, which indicates that ASCs promoted osteogenesis and inhibited adipogenesis of BMSCs through activation of BMP-2 and BMPR-IB signaling pathway. *P < 0.05, n ¼ 5.
10
X. Ye et al.
Figure 6. (A) DXA measurement of the distal femurs covered a distance of 10 mm starting from the distal end (scale bar, 5 mm). (B) For micro-CT quantitative measurement, the ROI was set as a gray cylinder (2.5 mm in diameter and 3.0 mm in height) at the central part of the medial condyle (scale bar, 2 mm). (C) BMD value began to increase at 8 weeks after cell transplantation and was significantly higher than that in the alginate-injected group at 12 weeks (*P < 0.05, n ¼ 12). (D) At 12 weeks after transplantation, micro-CT 2-dimensional scanning (upper) showed dense trabecular bone at the injected area of ASC/alginate-treated femur and sparse bone structure in the alginate-treated femur (square areas). As shown by 3D observation (lower), the trabeculae of the control group appeared looser and thinner than that of the experimental group (scale bars, 1 mm).
(Figure 1B). According to World Health Organization criteria, OP can be defined as a reduction of >2.5 SD of the normal BMD (2,35). These results, combined with the micro-CT and histological findings, confirm the successful establishment of the OP rabbit model. BMSC transplantation represents a potential strategy to treat OP in animal models (31,36). Nevertheless, the acquisition of BMSCs from bone marrow is an invasive, painful procedure, and the osteoprogenitor number and differentiation potential of BMSCs decrease with age (9e13). Compared with BMSCs, ASCs are easier to harvest, and their self-renewal potential and osteogenic differentiation are less affected by aging and estrogen deficiency (37,38). It has been reported that allogeneic or human ASCs can prevent bone loss in OVX mice (23,39). However, whether autologous ASC transplantation could augment bone formation under osteoporotic conditions has not yet been explored. In this study, we obtained ASCs from OVX rabbits and evaluated their suitability as a source of osteogenic precursor cells. ASCs from osteoporotic rabbits exhibited similar cell morphology, phenotypic characteristics and growth rate as those from sham-surgery animals. When osto-induced, these cells demonstrated positive alizarin red staining and collagen type I staining and an increased trend of calcium content.
However, their ECM mineralization extent (alizarin red staining and Ca2þ level) was lower than that of the normal ASCs after being induced in OM over 14 days (Figure 2). It might reflect the hormone-driven changes in ECM composition, which can influence the osteogenic differentiation of ASCs (38). Because of several advantageous features, such as injectability, high porosity, biocompatibility and biodegradation, calcium cross-linked sodium alginate hydrogels have been used as MSC carriers in several in vivo animal studies (31,40). In the present study, homogeneous calcium alginate gels were formulated as an injectable scaffold for osteoinduced ASCs. Confocal laser scanning microscopy results revealed that most ASCs remained viable after encapsulation, but dead cells increased over time (Figure 3E,F). Quantitative analyses showed that these cells retained the capability of osteogenic differentiation within the hydrogel, though their number decreased from day 1 to day 4 (Figure 3GeI). It has been reported that the encapsulation process may result in cell death at early time points (30), and we found that the cell number remained the same from day 4 to day 7. Another possible reason may be that the cells had differentiated into osteoblasts before encapsulation. Osteogenic differentiation of osteoblasts is usually accompanied with a stage of growth arrest (Figure 2B).
13.07 1.29 15.46 1.35
