Biomaterials 35 (2014) 6859e6870

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

A comparison of bone regeneration with human mesenchymal stem cells and muscle-derived stem cells and the critical role of BMP Xueqin Gao a, Arvydas Usas a, Ying Tang a, c, Aiping Lu a, Jian Tan b, Johannes Schneppendahl a, Adam M. Kozemchak d, Bing Wang c, James H. Cummins a, Rocky S. Tuan b, Johnny Huard a, * a

Stem Cell Research Center, University of Pittsburgh, Pittsburgh, PA, United States Center for Cellular and Molecular Engineering, University of Pittsburgh, Pittsburgh, PA, United States Molecular Therapy Laboratory, Department of Orthopaedic Surgery, University of Pittsburgh, Pittsburgh, PA, United States d Neuroscience Program, University of Michigan Class of 2013, Pittsburgh Tissue Engineering Initiative Summer Internship, United States b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 April 2014 Accepted 27 April 2014 Available online 21 May 2014

Adult multipotent stem cells have been isolated from a variety of human tissues including human skeletal muscle, which represent an easily accessible source of stem cells. It has been shown that human skeletal muscle-derived stem cells (hMDSCs) are muscle-derived mesenchymal stem cells capable of multipotent differentiation. Although hMDSCs can undergo osteogenic differentiation and form bone when genetically modified to express BMP2; it is still unclear whether hMDSCs are as efficient as human bone marrow mesenchymal stem cells (hBMMSCs) for bone regeneration. The current study aimed to address this question by performing a parallel comparison between hMDSCs and hBMMSCs to evaluate their osteogenic and bone regeneration capacities. Our results demonstrated that hMDSCs and hBMMSCs had similar osteogenic-related gene expression profiles and had similar osteogenic differentiation capacities in vitro when transduced to express BMP2. Both the untransduced hMDSCs and hBMMSCs formed very negligible amounts of bone in the critical sized bone defect model when using a fibrin sealant scaffold; however, when genetically modified with lenti-BMP2, both populations successfully regenerated bone in the defect area. No significant differences were found in the newly formed bone volumes and bone defect coverage between the hMDSC and hBMMSC groups. Although both cell types formed mature bone tissue by 6 weeks post-implantation, the newly formed bone in the hMDSCs group underwent quicker remodelling than the hBMMSCs group. In conclusion, our results demonstrated that hMDSCs are as efficient as hBMMSCs in terms of their bone regeneration capacity; however, both cell types required genetic modification with BMP in order to regenerate bone in vivo. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Human muscle-derived stem cells Human bone marrow mesenchymal stem cells BMP2 Lentivirus Calvarial defect Bone tissue engineering

1. Introduction Adult mesenchymal stem cells have been isolated from a number of human tissue sources including bone marrow, adipose tissue, dental pulp, and umbilical cord tissue. Although all adult derived stem cells have been shown to be capable of multipotent differentiation, there are advantages and disadvantages of using each of the populations. Among the various cell types, bone marrow and adipose-derived stem cells are two of the most practical for clinical application because of their availability; however, healthy bone marrow cells are not always readily available. The

* Corresponding author. E-mail address: [email protected] (J. Huard). http://dx.doi.org/10.1016/j.biomaterials.2014.04.113 0142-9612/Ó 2014 Elsevier Ltd. All rights reserved.

abundance of adipose-derived stem cells makes them highly attractive [1e3]; however, they are less efficient for bone regeneration after freezing [4]. Hence, identifying other tissue sources for acquiring adult stem cells is highly desirable. Different populations of human muscle-derived cells possess a multi-lineage differentiation capacity in vitro and are capable of forming bone and cartilage in vivo [5,6] and skeletal muscle is easily accessible through a minimally invasive needle biopsy procedure. Human musclederived stem cells (hMDSCs) isolated by the preplate technique have been shown to be capable of improving the functionality of myocardial infarcted cardiac muscle more effectively than myoblasts, and have been shown capable of effectively treating stress urinary incontinence in human patients [7,8]. hMDSCs display a similar marker profile as human bone marrow mesenchymal stem cells (hBMMSCs), with more than 95% of the cells expressing CD73,

6860

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

CD90, CD105, CD44, and being negative for CD45. Moreover, a high percentage of hMDSCs express CD56 and CD146. These hMDSCs exhibit myogenic, osteogenic, chondrogenic, and adipogenic capacities in vitro and are considered to be MSCs of muscle origin. These cells were also shown to be capable of enhancing the healing of a critical size calvarial bone defect created in mice when transduced with lenti-BMP2 [9]; however, it has never been determined if hMDSCs are as efficient as bone marrow MSCs in terms of their ability to promote bone repair. Consequently, we conducted a parallel comparison study between these two human cell populations in terms of their osteogenic differentiation capacities in vitro and their regeneration capacities in vivo utilizing a criticalsize calvarial defect model. Many different scaffolds have been used for promoting the osteogenesis of bone marrow MSCs in vivo, including collagen type I, alginate hydrogel [10,11], gelatin beads [12], hydroxyapatite [13,14], small intestine submucosa, and akermanite bioceramics [15,16]. In the current study, we utilized fibrin sealant, which is the natural product of blood clot formation and is completely bioresorbable. Upon activation by thrombin, it forms a clot-like gel instantly and has been successfully used as scaffold for bone repair [9,17e19]. It has also been used as a cell delivery vehicle to repair nerve and articular cartilage [20,21] and exhibits no adverse side effects on the transplanted cells or host tissue. Fibrin glue (Tisseel, BAXTER) is FDA approved and is routinely used in clinic; therefore, this scaffold was used to compare the bone regeneration capacities of both hBMMSCs and hMDSCs in vivo. BMP2 effectively promotes the osteogenic differentiation of hBMMSCs and hMDSCs in vitro and can mediate bone repair in vivo; therefore, we used a lenti-BMP2 viral vector to transduce both populations of cells and compared their osteogenic and bone regenerative potentials. In the current study, we performed a parallel systematic comparison between transduced and nontransduced hMDSCs and hBMMSCs in terms of their osteogenic gene expression profiles, in vitro osteogenic potential, and in vivo bone regeneration capacity in a mouse critical size calvarial defect model using fibrin sealant as a scaffold. 2. Material and methods

hBMMSCs were isolated from bone marrow obtained from the femoral heads of four patients who had undergone total hip arthroplasty from an 81 y/o female (81F), 66 y/o female (66F), 68 y/o male (68M), and a 52 y/o male (52M). Briefly, as described previously [23], trabecular bone was cored out using a curette or rongeur and flushed with rinsing medium containing a-MEM, 1% antibioticeantimycotic (Invitrogen, CA, USA) using 18-gauge hypodermic needles. The bone chips were then minced with scissors and the flushed medium was passed through a 40 mm mesh cell strainer to remove debris and centrifuged for 5 min at 300G. Pellets were washed twice with rinsing medium and resuspended in mesenchymal stem cells (MSC) proliferation medium containing a-MEM (Invitrogen, CA, USA), 10% foetal bovine serum (FBS, Invitrogen, CA, USA), 1% antibioticeantimycotic (Invitrogen), and 1 ng/ml FGF2 (RayBiotech, Inc., GA, USA) and plated on two 150 cm2 tissue culture flasks. On day 4, the cells were washed with PBS and fresh proliferation medium was added. Proliferation medium was changed every 3e4 days. Once the cell cultures reached 70e80% confluency, the cells were dissociated with 0.25% trypsin containing 1 mM EDTA (Invitrogen, CA, USA), frozen, and stored in liquid nitrogen for future studies. 2.2. Construction of the lentiviral-BMP2 construct A lenti-viral construct encoding human BMP2 gene under the control of the human cytomegalovirus (CMV) promoter was constructed in our laboratory (in collaboration with Dr. Bing Wang Laboratory). An SV40 early promoter was inserted down-stream of the BMP2 expression cassette to drive a higher level of constitutive expression of an antibiotic resistance gene to allow us to use blasticidin to select for the BMP2 transduced cells. Lenti-BMP2 virus was packaged using 293T cells (ATCC). 2.3. Cell transduction hMDSCs were transduced with lenti-viral-BMP2 construct in the presence of polybrene (8 mg/ml) at passage 8 for 16 h and the hBMMSCs were transduced with the same virus at passage 2. Twenty-four hours after transduction, both the hMDSCs and hBMMSCs were selected with 10 mg/ml blasticidin for 24 h. After selection, the cells were expanded in their respective proliferation media. BMP2 secretion levels were measured using ELISA (R&D system) at different passages post-transduction. 2.4. Analysis of osteogenic-related genes and early osteogenic markers Semi-quantitative RT-PCR was performed on the hMDSCs and hBMMSCs to assess their respective gene expression profile before and after lenti-BMP2 transduction. Total RNA was extracted from both non-transduced and lenti-BMP2 transduced cells using Trizol reagent (Invitrogen). cDNA was synthesized using Superscript II reverse transcriptase (Invitrogen). PCR reactions were performed using a Promega GOtaq Kit. Primers were designed using Primer 3 [24,25] and are shown in Table 1. Alkaline phosphatase (ALP, SigmaeAldrich) staining was performed on both, non-transduced and lenti-BMP2 transduced hMDSCs and hBMMSCs. 2.5. In vitro osteogenic differentiation using 3D pellet cultures

The use of human tissues was approved by the Institutional Review Board (University of Pittsburgh and University of Washington), and all animal experiments and procedures were approved by Institutional Animal Care and Use Committee of the University of Pittsburgh. 2.1. Cell isolation Four populations of hMDSCs were isolated, via a modified preplate technique as previously described [22], from skeletal muscle biopsies purchased from the National Disease Research Interchange (NDRI) from a 23 y/o male (23M), a 31y/o female (31F), a 21 y/o male (21M), and a 76 y/o female (76F). The late adhering (PP6) cells were grown and maintained in proliferation medium that contained high glucose DMEM (Invitrogen) supplemented with 20% FBS, 1% chicken embryo extract, and 1% penicillin/streptomycin.

Four populations of hMDSCs and hBMMSCs were subjected to pellet culture before and after lenti-BMP2 transduction using a protocol previously described [9]. Four to eight replicate pellets were prepared for each population and cultured in osteogenic medium utilizing the same conditions. At 3 and 4 weeks after initiating the osteogenic cultures, the pellets were scanned with a microCT (Viva CT 40, Scanco Medical, Switzerland) to detect mineralization using 21 voxel size and medium resolution. The mineralized pellet volumes were evaluated using the following parameters: Gauss Sigma 0.8, Gauss support 1.0, and threshold 122. After 4 weeks, the cell pellets were fixed in 4% neutral buffered formaldehyde for 2 h at RT, embedded in NEG 50 freezing medium, snap frozen in liquid nitrogen, and stored at
80  C until cryosectioned at 8 mm. Von Kossa staining was performed using an online protocol (IHC world) to verify the calcification of the pellets. Osteocalcin immunohistochemistry using mouse anti-human osteocalcin (R &D system, USA) primary

Table 1 Primer information. Gene name

Accession number

Primers (50 e30 )

Product size (bp)

Annealing temperature ( C)

SOX9

NM_000346.3

249

55

BMPR1B

NM_001203.2

108

55

BMPR2

NM_001204.6

231

55

COX-2

NM_000963.2

225

55

GAPDH

NM_002046.4

F:GCTCAGCAAGACGCTGGGCA R:CCGGAGGAGGAGTGTGGCGA F:AGGCCTCCCTCTGCTGGTCC R:TTCGCCACGCCACTTTCCCA F:TCCCATGCTGCCACAACCCA R:TGGCCGTTCCCTCCTGTCCA F:GCGAGGGCCAGCTTTCACCA R:CCTGCCCCACAGCAAACCGT F: GCCTTCCGTGTCCCCACTGC R:CAATGCCAGCCCCAGCGTCA

211

55

F: Forward; R: Reverse.

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

6861

Fig. 1. Gene expression and ALP staining of hBMMSCs and hMDSCs, before and after lenti-BMP2 transduction. A) Gene expression analysis by semi-quantitative RT-PCR. B) Quantification of the expression of osteogenic genes by hMDSCs and hBMMSCs. C) ALP staining. D) Quantification of ALP positive cells indicated that their abundance was significantly higher in the hBMMSC81F population and significantly lower in the hBMMSC 66F population after transduction. No difference was found in other populations after lenti-BMP2 transduction. Scale bars in C represent 200 mm. **p < 0.01, ***p < 0.001.

antibody was also performed as previously reported [9]. DAB colour reaction was used to reveal osteogenic differentiation of the cells as shown in brown. Pellet culture experiments were repeated three times for all cell populations. 2.6. Creation of critical size calvarial bone defects Critical size calvarial bone defects were created using an established protocol [19]. Briefly, CD-1 nude mice (Charles River) were anesthetized using 2% isoflurane. An incision was made just off the middle line of the skull in the scalp and the right parietal bone was exposed. After removal of the periosteum, a 5 mm bone defect was created using a 5 mm diameter trephine (Fine Science Tools, USA). The bone was

carefully removed and the defect area was rinsed with normal saline. The cultured cells were resuspended in 20 ml of PBS and mixed with 15 ml of thrombin immediately before their implantation into the defect area. Following cell implantation, 15 ml fibrin sealant was put on top of the cells and allowed 1e2 min to solidify. The wound was closed with suture and the mice were allowed to recover in an oxygen chamber. 2.7. In vivo bone regeneration using non-transduced hBMMSCs and hMDSCs Eight week-old male CD-1 nude mice (Charles River) were divided into 2 groups (N ¼ 8). One group received 1.5  106 hBMMSCs from 2 donors (81F and 52M) with

6862

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

Fig. 2. MicroCT imaging of mineralization of in vitro 3D pellet cultures at 3 and 4 weeks after osteogenic differentiation. A) Representative 3D microCT images of two populations of hBMMSCs and hMDSCs. BeC) Quantification of mineralized volumes of the hBMMSC pellets. DeE) Quantification of the mineralized volumes of hMDSC pellets. Transduction with lenti-BMP2 enhanced mineralization of the pellets with the exception of donor hMDSCs 21M. *p < 0.05, **p < 0.01, ***p < 0.001 compared to the untransduced cells. each of the donor cell populations were implanted into 4 mice. The other group received 1.5  106 hMDSCs from two donors (76F and 21M). hBMMSCs at passage 5 and hMDSCs at passage 13 were implanted into the defect area using fibrin sealant as the scaffold (Tisseel, Baxter). MicroCT scans (Viva CT-40, Scanco Medical, Switzerland) were performed at days 1, 14, 28, and 42 post-implantation (PI) using 30 mm voxel size and medium resolution. Bone volume was quantified using Scanco evaluation software according to the guidelines of the American Society of Bone and Mineral Research [26]. Bone defect coverage was measured using Image J software (NIH). 2.8. In vivo bone regeneration using BMP2 transduced hBMMSCs and hMDSCs Thirty-two male CD-1 nude mice at 8 weeks of age were divided into two groups (N ¼ 16): Group 1 received 4 populations of lenti-BMP2 transduced hBMMSCs (N ¼ 4 for each population at passage 3e4 after blasticidin selection); Group 2 received 4 populations of lenti-BMP2 transduced hMDSCs at passage 5e6 after blasticidin selection (N ¼ 4 for each population). 5 mm critical size calvarial defects were created in the right parietal bone of the mice as stated above and 1.5  106 cells were mixed with thrombin and implanted in the defect, with fibrin sealant then being added onto the cells. The fibrin was allowed to solidify and the wound was then closed with sutures. MicroCT scans (Scanco Medical) were performed at days 1, 14, 28, and 42 PI using 30 mm voxel size and medium resolution as stated above. 2.9. Histology Herovici’s staining was used to identify collagen type I as previously described [27e29]. H&E staining was used to reveal the newly regenerated bone and bone

marrow. Tartrate resistant acid phosphatase (TRAP) staining was performed on the decalcified paraffin sections using a 387A kit (Sigma, Milwaukee, USA). TRAP positive osteoclasts on the bone surface were normalized to bone area (excluding bone marrow area) and expressed as cell number/mm2. 2.10. Statistical analysis One way analysis of variance or Student’s t-test was used to analyse bone volume and defect coverage and TRAP positive cells between the different hBMMSC and hMDSC populations and between the two groups. For the data that had high standard deviations, such as the in vitro pellet cultures of the hMDSC (76F) and hBMMSC (52M), we used the Wilcoxon Rank sum non-parametric test. A value of p < 0.05 was considered statistically significant.

3. Results 3.1. Osteogenic gene expression profiles of hMDSCs and hBMMSCs It took about 10 days to isolate and expand the hBMMSCs and 2 weeks to isolate the hMDSCs. The hBMMSCs could be used up to passage 10 and the hMDSCs could be cultured up to passage 20 without any noticeable change in their phenotype. The BMP2 secretion levels of the 4 populations of hBMMSCs at the time of

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

implantation were: 29.9, 12.7, 14.1, 14.1 ng/million/24 h for donors 81F, 66F, 68M, and 52M respectively. The BMP2 secretion levels of the 4 populations of hMDSCs were: 2.7, 1.3, 0.5, and 0.35 ng/million/ 24 h for donors 23M, 30F, 21M, and 76F respectively. Both hMDSCs and hBMMSCs expressed the osteogenic-related genes BMPR2, BMPR1b, COX-2, and SOX-9 (Fig. 1A). Quantification of the relative gene expression levels indicated that lenti-BMP2 transduction increased COX-2 expression in the hBMMSC group (Fig. 1B). ALP staining indicated that only a very small percentage of hMDSCs and hBMMSCs expressed ALP (Fig. 1C). Quantification of the number of positive cells after lenti-BMP2 transduction showed that hBMMSCs 81F exhibited higher ALP positivity while hBMMSCs 66F displayed a lower ALP positivity. All other populations had no significant changes (Fig. 1D). 3.2. In vitro osteogenic capacities of hMDSCs and hBMMSCs MicroCT evaluation of the 3D pellet cultures showed that all 4 populations of hBMMSCs underwent mineralization by 3e4 weeks after initiating the osteogenic cultures; in contrast, the nontransduced cells exhibited only very minimal mineralization (Fig. 2A showing two populations). Lenti-BMP2 transduction greatly enhanced the pellets’ mineralization volumes (Fig. 2B, C). Similarly, the 4 populations of hMDSCs cultured in osteogenic medium also exhibited mineralization as detected by microCT at 3 and 4 weeks, although the mineralized pellet volumes varied more among the different hMDSC populations. Of the four populations of hMDSCs, three populations of cells (76F, 31F, and 23M) displayed significant increases in mineralization volume after lenti-BMP2 transduction at the 3 and 4-week time points (Fig. 2D, showing 1

6863

donor); however, one population of untransduced hMDSCs (21M) actually displayed a significantly higher mineralized pellet volumes at both the 3- and 4-week time points than the lenti-BMP2 transduced cells (Fig. 2E). This phenomenon could have been caused by the different growth rates between the untransduced and lentiBMP2 transduced hMDSCs, since we observed that this population of untransduced cells grew much faster than the lenti-BMP2 transduced cells (data not shown). Von Kossa staining demonstrated that all the hBMMSCs pellets underwent mineralization. Notably, the mineralization of the untransduced cells was located at the periphery of the pellets, while the lenti-BMP2 transduced cells mineralized at both the edge and center of the pellets. Similarly, the untransduced hMDSCs mineralized at the periphery of the pellets while the lenti-BMP2 transduced cells mineralized both centrally and peripherally. For the 21M population, both the untransduced and transduced cells demonstrated extensive mineralization throughout their pellets. As a result of this high level of mineralization, some parts of the pellets peeled off especially in those areas that were highly mineralized (Fig. 3A). Immunohistochemical staining of osteocalcin indicated that the cells in the pellets of the hBMMSCs expressed the early mineralization marker osteocalcin (brown staining) as did the hMDSCs(Fig. 3B). 3.3. In vivo bone regeneration capacities of untransduced hMDSCs and hBMMSCs We tested the capacity of the hMDSCs and hBMMSCs to regenerate bone, using fibrin sealant as a scaffold, in a critical size

Fig. 3. Von Kossa staining and osteocalcin immunohistochemistry of hBMMSCs and hMDSCs pellet cultures. A) Von Kossa staining. Untransduced hBMMSCs mineralized mainly at the periphery of the pellets and lenti-BMP2 transduction resulted in both peripheral and central mineralization of the pellets. hMDSCs tended to mineralize throughout the entire pellet. LentiBMP2 transduction enhanced the extent of mineralization. B) Osteocalcin immunohistochemistry. Osteocalcin positive cells stained dark brown. Osteocalcin was identified in the same location as the Von Kossa positive cells in both the hBMMSCs and hMDSCs pellets. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

6864

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

calvarial defect model. After implantation of the non-transduced hBMMSCs or hMDSCs into the calvarial defects, microCT analyses revealed no obvious bone formation in either the hMDSCs or hBMMSCs transplantation groups (Fig. 4A). Quantification of the bone defect area demonstrated less than 30% healing of the defect in all groups with no significant difference between groups (Fig. 4B). Herovici’s staining indicated no bone matrix collagen

type I (red in the web version) formation in the defect area; however, collagen type 3 (blue in the web version) staining, which stains fibrotic tissues, was present in the defect area of both the hBMMSCs and hMDSCs groups (Fig. 5A). H&E staining displayed no bone formation, but did demonstrate a layer of fibrotic tissue in both the hBMMSCs and hMDSCs implantation groups (Fig. 5B).

Fig. 4. Comparison of in vivo bone regeneration by untransduced hBMMSCs and hMDSCs. A) MicroCT 3D images of the calvarial bone defect at different times points post-implantation of the stem cells. B) Quantification of the defect healing percentage. The defect closure ranged from 15 to 26% on average for both the hBMMSCs and hMDSCs. No statistical differences were found.

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

6865

Fig. 5. Histological analysis of the regenerated tissues formed by the untransduced hBMMSCs and hMDSCs. A) Herovici’s staining. The defect area is shown between the two black arrows at 20 magnification. Both the hBMMSCs and hMDSCs transplantation groups stained positively for collagen type III (blue) in the defect area while the host bone stained positively for collagen type I (red) matrix. B) High magnification (200) showing the defect area was mainly filled with collagen type III (blue) in both populations of hBMMSCs and hMDSCs while the host bones stained positively for collagen type I (red). C) H&E staining further indicated the formation of fibrotic tissue in the defect area while the host bone showed a dense bone matrix. COL3: Collagen type III; HB: Host bone; Br: Brain; COL1: Collagen type I; FT: Fibrotic tissue. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4. In vivo bone regeneration capacities of BMP-2 transduced hMDSCs and hBMMSCs All 4 populations of lenti-BMP2 transduced hBMMSCs were capable of forming new bone in the calvarial bone defect site as early as 2 weeks post-implantation, as shown by 3D reconstruction of the microCT scans. Some animals showed complete healing (81F) at 4 and 6 weeks (Fig. 6A). Similarly, the 4 populations of hMDSCs also displayed significant new bone formation in the defect area with some of the animals showing complete defect healing (31F) (Fig. 6B); however, both cell types demonstrated variability between the different populations. The quantification of the newly regenerated bone volumes is shown in Fig. 6C. There were no statistical differences in the newly regenerated bone volumes or bone defect coverage among the 4 populations of hBMMSCs or the different populations of hMDSCs (Fig. 6B). There were also no differences between the hMDSC and hBMMSC groups when the data was combined and compared (Fig. 6D). Quantification of the bone defect healing areas revealed 50e90% healing in the 4 populations of hBMMSCs (Fig. 6E). The defect healing of the lenti-BMP2 transduced hMDSCs ranged from 60 to 80% (Fig. 6E). There were no statistical differences among the 4 populations in terms of their percentage of total defect healing. When the data from the 4 populations from both the hBMMSCs and hMDSCs were combined and compared, there was no statistical difference between two groups (Fig. 6F). 3.5. Functional bone tissue formation by BMP2 transduced hBMMSCs and hMDSCs Herovici’s staining revealed the formation of bone matrix which consisted of collagen type I (red) in the defect site (the area between the two black arrows), that resembled normal host bone on the contralateral side of the skull in both the hBMMSC and hMDSCs groups (Fig. 7A). The newly regenerated bone in the defect area was found to be fully integrated with the host bone (Fig. 7B). Bone matrix (shown in red) and typical cancellous bone could be seen at high magnification (Fig. 7C). Collagen type III is detected as dark blue and bone marrow stains light blue. H&E staining revealed the formation of cancellous bone and bone marrow. We found myeloid

cells (blue arrows), erythrocytes (green arrows), and megakaryocytes in the newly formed bone marrow (enlarged in inset boxes) and blood vessels in all the defect sites implanted with the lentiBMP2 transduced hBMMSCs and hMDSCs transplanted populations (Fig. 7D showing two populations). 3.6. Remodelling of the newly formed bone To assess the bone remodelling of the newly formed bone, osteoclast formation was examined by TRAP staining. Violet red TRAP positive osteoclasts (green arrows) were found on the bone surface of the newly formed bone, similar to the contralateral side of the native host bone in all the transplanted groups transduced with lenti-BMP2 (Fig. 8A). Quantification of the TRAP positive osteoclasts in the bone area indicated variability between the different populations of hBMMSCs; however, the 4 populations of hMDSCs displayed similar trends (Fig. 8B). When the 4 populations were combined and compared between the hBMMSCs and the hMDSCs, we found significantly more TRAP positive osteoclasts in the new bone areas in the hMDSCs group than in the hBMMSCs group (Fig. 8C). 4. Discussion In the current study, we performed a parallel comparison of the osteogenic capacity of hMDSCs and hBMMSCs both in vitro and in vivo using a critical size bone defect model and bioresorbable fibrin sealant scaffold. Our results demonstrated that both cell types had similar osteogenic-related gene expression profiles and exhibited equivalent osteogenic differentiation capacities in vitro and bone regeneration capacities in vivo. hBMMSCs have been extensively studied for their bone regeneration capacities and have become the “gold standard” adult stem cells for bone repair. Autologous bone marrow stem cells have been shown to be effective for treating non-union bone fractures [30e 34] and have been used to promote bone elongation [35]. Furthermore, genetic modification of hBMMSCs with osteogenic genes like BMP has been shown to greatly increase their osteogenic potential of the and effectively promote bone defect repair in different animal models [36e39]. Recently, comparison studies of

6866

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

Fig. 6. Comparison of bone regeneration capacity of lenti-BMP2 transduced hBMMSC and hMDSCs. AeB) Representative microCT images of 4 populations of lenti-BMP2 transduced hBMMSCs and 4 populations of lenti-BMP2 transduced hMDSCs at various time points post-implantation. C) Quantification of newly formed bone volume by microCT for all populations of both hBMMSCs and hMDSCs groups. D) Combined bone volume of hBMMSCs and hMDSCs indicated no statistical differences. E) Quantification of defect coverage (%) for both groups. N ¼ 4 for each population. F) Combined bone defect coverage showing no statistical differences between the hBMMSCs and hMDSCs groups.

the hBMMSCs with other adult stem cells have been performed. A comparative study of hBMMSCs and human adipose-derived stem cells in a rat spine fusion model using ex vivo adenoviral based gene transfer of BMP2 showed that both cell populations induced abundant bone formation and promoted similar posterolateral spinal fusion [40]. hBMMSCs have also been shown to be more efficient at promoting functional bone fracture healing than mesenchymal stem cells derived from human embryonic stem cells [41]. However, until now, there have been no reports that compared human bone marrow derived stem cells with other sources of adult stem cells for bone repair using a critical size calvarial bone defect model. Human muscle-derived multipotent progenitor cells (MPCs) isolated from traumatized muscle tissues have been shown to

express the mesenchymal stem cell markers CD73, CD90 andCD105 and were negative for CD45. MPCs have been shown to be capable of differentiating into osteoblasts, chondrocytes and adipocytes, but not myoblasts [42,43]. Although they appeared to have a limited lineage commitment compared to bone marrow derived MSCs, they are remain a promising cell source for bone regeneration due to their availability and abundance following orthopaedic injuries [44] and are putative osteoprogenitor cells that initiate ectopic bone formation in heterotopic ossification [45]. Recently, it has been shown that hMDSCs isolated via the preplate technique also display a mesenchymal stem cell marker profile which is highly positive for CD73, CD90, CD105, CD44, CD56, and CD146, and negative for CD45 and are capable of differentiating into osteogenic, chondrogenic, adipogenic and myogenic lineages, and are

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

6867

Fig. 7. Histological analysis of the newly formed bone in the defect area by the lenti-BMP2 transduced hBMMSCs and hMDSCs. A) Herovici’s staining. The newly formed bone in both groups contained similar collagen type I positive bone matrix (red) (20 magnifications) as host bone. The region between the two black arrows indicates the defect area. B) Herovici’s staining indicated the whole defect area in detail. C). Herovici’s staining at high magnification indicated the structure of the bone matrix (red) and bone marrow (light blue). D) H&E staining displayed the formation of bone and functional bone marrow. Both the hBMMSCs and hMDSCs groups formed new functional bone as indicated by the dense bone matrix (red) and functional bone marrow was demonstrated by the appearance of blood vessels (yellow arrows) and three lineages of bone marrow cells including erythrocytes (green arrows), myeloid cells (blue arrows), and megakaryocytes (blue box). Inserts are the enlarged megakaryocytes in the blue boxes. COL1: Collagen type 1; COL3: collagen type 3; BM: Bone marrow; B: Bone; Br: Brain. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

considered to be a muscle-derived mesenchymal stem cell population [9]. Indeed, hMDSCs, like hBMMSCs, also expressed BMPR1b, BMPR2, COX-2, and SOX9 which allow them to be responsive to BMP2 signalling through the BMP2-BMPR2 and BMPR1b pathways to activate the osteogenic differentiation processes [46]. We found that BMP2 secretion levels were much higher in the lenti-BMP2 transduced hBMMSCs than in the lenti-BMP2 transduced hMDSCs; however, this difference was not reflected in the in vitro or in vivo osteogenic potential of the cells. Both cell types formed comparable sized pellets that underwent similar

mineralization in vitro and formed comparable mature bone in vivo; however, neither the hBMMSCs nor the hMDSCs underwent osteogenic differentiation after lenti-BMP2 transduction in vitro without being grown in osteogenic medium in pellet cultures. Both non-transduced hBMMSCs and hMDSCs formed negligible amounts of bone after implantation into critical size calvarial bone defects in mice. This indicated that although both stem cell types possess the ability to differentiate toward an osteogenic lineage in vitro, sufficient osteoinductive signalling is required to initiate the bone regeneration processes in vivo. It has been shown that hBMMSCs

6868

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

Fig. 8. TRAP staining revealed osteoclasts mediated bone remodelling. A) Representative TRAP staining images of the newly regenerated bone in the defect area of two populations of hBMMSCs and hMDSCs. TRAP positive osteoclasts stain violet red (green arrows) on the bone surface seen in the newly regenerated bone of 4 populations of hBMMSCs and hMDSCs, inserts in each picture showing TRAP positive osteoclasts. B). Quantification of TRAP positive osteoclasts for each cell population. C). Comparison of hBMMSCs and hMDSCs groups indicated more TRAP positive osteoclasts in the hMDSCs group than the hBMMSCs group. *p < 0.05 compared to hBMMSCs group. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

could facilitate bone regeneration and repair without gene modification or stimulation with BMPs [47,48]; however, the enhanced bone repair capacity of the hBMMSCs in these studies was probably due to the osteogenic inductive effect that the scaffolds had on the hBMMSCs. The fibrin sealant scaffold used in the current study has no bone inductive effects on the transplanted cells and only served to keep the cells at the defect site. Fibrin sealant is resorbed in approximately 7 days post-transplantation, as can be observed at early time points of several of our previous studies [9,19].

In contrast, the current study demonstrated that both hMDSCs and hBMMSCs could efficiently promote critical size bone defect healing when genetically modified with lenti-BMP2. Although the secretion levels of BMP2 were much lower in the lenti-BMP2 transduced hMDSCs than in the lenti-BMP2 transduced hBMMSCs, they formed similar amounts of new bone. This finding cannot be explained by the age differences of the patients from which the hMDSCs and hBMMSCs cells were isolated because we found that the hMDSCs and hBMMSCs isolated from the older

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

donor were not inferior to those of the younger donor (81F hBMMSC versus 52M, 66F, 68M hMDSCs; 76F hMDSC versus 21M, 23M, 31F hMDSCs). Importantly, we found that all the populations of lenti-BMP2 transduced hBMMSCs and hMDSCs formed mature bone and bone marrow as shown by the presence of typical cancellous (trabecular) bone structure which is composed of a highly organized dense collagen type I matrix and is vascularized and populated by at least three haematopoietic cell lineages (erythrocytes, myeloid cells, and megakaryocytes) in the newly formed bone marrow, which was very similar to the native contralateral host bone (data not shown). In the past, most research in the field of bone tissue engineering has focused on the formation of mineralized tissue and not on the formation of truly mature bone; however, recently, Scotti et al. were able to establish a mature haematopoietic stem cell niche using in vitro cartilage based engineering in combination with in vivo implantation of hBMMSCs. In their study, they seeded the stem cells in a collagen scaffold, stimulated them with different factors for 5 weeks until the cells differentiated toward a cartilaginous lineage, and subsequently implanted the scaffold and cells subcutaneously into host mice. Five to 12 weeks after implantation they found that the implants attracted host cells to form vascularized bone and mature bone marrow [49]. In our study, both the lenti-BMP2 transduced hMDSCs and hBMMSCs demonstrated similar capacities to form functional bone by 6 weeks post-implantation, and the newly formed bone marrow was populated by at least three different haematopoietic lineages. These results were very encouraging because our method omitted the necessity of the in vitro culture process, did not require the use of a complicated scaffold, and took a shorter time to form mature bone which actually integrated with the host bone. We also found that the newly formed bone in the hMDSCs group underwent faster remodelling (a measure of bone maturation) than the hBMMSCs group, as indicated by the appearance of a greater number of TRAP positive osteoclasts in the newly formed bone created by the hMDSCs. This is important because the faster bone remodelling occurs, the quicker functional integration can occur with the host bone. Limitations of the current study included the fact that we could not obtain bone marrow and skeletal muscle biopsies from the same donors to perform side-by-side comparative studies of the isolated hBMMSCs and hMDSCs. Also, the skull defect model is not an easy model to determine biomechanical testing post-treatment, and therefore additional studies using a long bone defect are required in the future. 5. Conclusion In summary, we found that hMDSCs were as efficient as hBMMSCs in terms of their in vitro osteogenesis and in vivo bone regeneration capacity, and all the populations studied formed mature bone and bone marrow by 6 weeks when genetically modified with a lenti-BMP2 viral vector. Genetic modification or stimulation with BMP of both cell populations was a key requirement for the cells to efficiently regenerate bone. Therefore, skeletal muscle is a potential reliable alternative source of adult mesenchymal stem cells (hMDSCs) for applications in bone tissue engineering and bone defect repair. Acknowledgements We highly appreciate Deanna Rhodes from Histology core of McGowan Institute For Regenerative Medicine from University of Pittsburgh Medical Center for performing Herovici’s staining and

6869

H&E staining. We thank Nick Oyster for his assistance performing the microCT scanning and evaluation of non-transduced cells in vivo experiments. This project was funded by NIH grant 5RO1DE13420-09 to Dr. Johnny Huard and Commonwealth of Pennsylvania Department of Health funding (SAP4100050913) to Dr. Rocky S Tuan. References [1] Behr B, Tang C, Germann G, Longaker MT, Quarto N. Locally applied vascular endothelial growth factor A increases the osteogenic healing capacity of human adipose-derived stem cells by promoting osteogenic and endothelial differentiation. Stem Cells 2011;29(2):286e96. [2] Levi B, Nelson ER, Li S, James AW, Hyun JS, Montoro DT, et al. Dura mater stimulates human adipose-derived stromal cells to undergo bone formation in mouse calvarial defects. Stem Cells 2011;29(8):1241e55. [3] Cowan CM, Shi YY, Aalami OO, Chou YF, Mari C, Thomas R, et al. Adiposederived adult stromal cells heal critical-size mouse calvarial defects. Nat Biotechnol 2004;22(5):560e7. [4] James AW, Levi B, Nelson ER, Peng M, Commons GW, Lee M, et al. Deleterious effects of freezing on osteogenic differentiation of human adipose-derived stromal cells in vitro and in vivo. Stem Cells Dev 2011;20(3):427e39. [5] Mastrogiacomo M, Derubeis AR, Cancedda R. Bone and cartilage formation by skeletal muscle derived cells. J Cell Physiol 2005;204(2):594e603. [6] Musgrave DS, Pruchnic R, Bosch P, Ziran BH, Whalen J, Huard J. Human skeletal muscle cells in ex vivo gene therapy to deliver bone morphogenetic protein-2. J Bone Jt Surg Br 2002;84(1):120e7. [7] Okada M, Payne TR, Drowley L, Jankowski RJ, Momoi N, Beckman S, et al. Human skeletal muscle cells with a slow adhesion rate after isolation and an enhanced stress resistance improve function of ischemic hearts. Mol Ther 2012;20(1):138e45. [8] Smaldone MC, Chancellor MB. Muscle derived stem cell therapy for stress urinary incontinence. World J Urol 2008;26(4):327e32. [9] Gao X, Usas A, Lu A, Tang Y, Wang B, Chen CW, et al. BMP2 is superior to BMP4 for promoting human muscle-derived stem cell-mediated bone regeneration in a critical-sized calvarial defect model. Cell Transplant 2013;22(12):2393e 408. [10] Schneider RK, Puellen A, Kramann R, Raupach K, Bornemann J, Knuechel R, et al. The osteogenic differentiation of adult bone marrow and perinatal umbilical mesenchymal stem cells and matrix remodelling in threedimensional collagen scaffolds. Biomaterials 2010;31(3):467e80. [11] Chang SC, Chung HY, Tai CL, Chen PK, Lin TM, Jeng LB. Repair of large cranial defects by hBMP-2 expressing bone marrow stromal cells: comparison between alginate and collagen type I systems. J Biomed Mater Res A 2010;94(2): 433e41. [12] Yang Y, Hallgrimsson B, Putnins EE. Craniofacial defect regeneration using engineered bone marrow mesenchymal stromal cells. J Biomed Mater Res A 2011;99(1):74e85. [13] Zannettino AC, Paton S, Itescu S, Gronthos S. Comparative assessment of the osteoconductive properties of different biomaterials in vivo seeded with human or ovine mesenchymal stem/stromal cells. Tissue Eng Part A 2010;16(12):3579e87. [14] Chang SH, Tung KY, Wang YJ, Tsao YP, Ni TS, Liu HK. Fabrication of vascularized bone grafts of predetermined shape with hydroxyapatite-collagen gel beads and autogenous mesenchymal stem cell composites. Plast Reconstr Surg 2010;125(5):1393e402. [15] Kim KS, Lee JY, Kang YM, Kim ES, Kim GH, Rhee SD, et al. Small intestine submucosa sponge for in vivo support of tissue-engineered bone formation in the presence of rat bone marrow stem cells. Biomaterials 2010;31(6):1104e 13. [16] Huang Y, Jin X, Zhang X, Sun H, Tu J, Tang T, et al. In vitro and in vivo evaluation of akermanite bioceramics for bone regeneration. Biomaterials 2009;30(28):5041e8. [17] Hale BW, Goodrich LR, Frisbie DD, McIlwraith CW, Kisiday JD. Effect of scaffold dilution on migration of mesenchymal stem cells from fibrin hydrogels. Am J Vet Res 2012;73(2):313e8. [18] Kretlow JD, Spicer PP, Jansen JA, Vacanti CA, Kasper FK, Mikos AG. Uncultured marrow mononuclear cells delivered within fibrin glue hydrogels to porous scaffolds enhance bone regeneration within critical-sized rat cranial defects. Tissue Eng Part A 2010;16(12):3555e68. [19] Usas A, Ho AM, Cooper GM, Olshanski A, Peng H, Huard J. Bone regeneration mediated by BMP4-expressing muscle-derived stem cells is affected by delivery system. Tissue Eng Part A 2009;15(2):285e93. [20] Zhao Z, Wang Y, Peng J, Ren Z, Zhang L, Guo Q, et al. Improvement in nerve regeneration through a decellularized nerve graft by supplementation with bone marrow stromal cells in fibrin. Cell Transplant 2014;23(1):97e110. [21] Ahmed TA, Giulivi A, Griffith M, Hincke M. Fibrin glues in combination with mesenchymal stem cells to develop a tissue-engineered cartilage substitute. Tissue Eng Part A 2011;17(3e4):323e35. [22] Gharaibeh B, Lu A, Tebbets J, Zheng B, Feduska J, Crisan M, et al. Isolation of a slowly adhering cell fraction containing stem cells from murine skeletal muscle by the preplate technique. Nat Protoc 2008;3(9):1501e9.

6870

X. Gao et al. / Biomaterials 35 (2014) 6859e6870

[23] Caterson EJ, Nesti LJ, Danielson KG, Tuan RS. Human marrow-derived mesenchymal progenitor cells: isolation, culture expansion, and analysis of differentiation. Mol Biotechnol 2002;20(3):245e56. [24] Koressaar T, Remm M. Enhancements and modifications of primer design program Primer3. Bioinformatics 2007;23(10):1289e91. [25] Untergasser A, Cutcutache I, Koressaar T, Ye J, Faircloth BC, Remm M, et al. Primer3 e new capabilities and interfaces. Nucleic Acids Res 2012;40(15):e115. [26] Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Muller R. Guidelines for assessment of bone microstructure in rodents using microcomputed tomography. J Bone Miner Res 2010;25(7):1468e86. [27] Levame M, Meyer F. Herovici’s picropolychromium. Application to the identification of type I and III collagens. Pathol Biol (Paris) 1987;35(8):1183e8. [28] Pieraggi MT, Julian M, Bouissou H, Stocker S, Grimaud JA. Dermal aging. Immunofluorescence study of collagens I and III and fibronectin. Ann Pathol 1984;4(3):185e94. [29] Turner NJ, Pezzone MA, Brown BN, Badylak SF. Quantitative multispectral imaging of Herovici’s polychrome for the assessment of collagen content and tissue remodelling. J Tissue Eng Regen Med 2013;7(2):139e48. [30] Connolly JF, Guse R, Tiedeman J, Dehne R. Autologous marrow injection as a substitute for operative grafting of tibial nonunions. Clin Orthop Relat Res 1991;266:259e70. [31] Tiedeman JJ, Garvin KL, Kile TA, Connolly JF. The role of a composite, demineralized bone matrix and bone marrow in the treatment of osseous defects. Orthopedics 1995;18(12):1153e8. [32] Hernigou P, Poignard A, Beaujean F, Rouard H. Percutaneous autologous bonemarrow grafting for nonunions. Influence of the number and concentration of progenitor cells. J Bone Jt Surg Am 2005;87(7):1430e7. [33] Quarto R, Mastrogiacomo M, Cancedda R, Kutepov SM, Mukhachev V, Lavroukov A, et al. Repair of large bone defects with the use of autologous bone marrow stromal cells. N Engl J Med 2001;344(5):385e6. [34] Marcacci M, Kon E, Moukhachev V, Lavroukov A, Kutepov S, Quarto R, et al. Stem cells associated with macroporous bioceramics for long bone repair: 6to 7-year outcome of a pilot clinical study. Tissue Eng 2007;13(5):947e55. [35] Kitoh H, Kitakoji T, Tsuchiya H, Mitsuyama H, Nakamura H, Katoh M, et al. Transplantation of marrow-derived mesenchymal stem cells and platelet-rich plasma during distraction osteogenesis e a preliminary result of three cases. Bone 2004;35(4):892e8. [36] Srouji S, Ben-David D, Fromigue O, Vaudin P, Kuhn G, Muller R, et al. Lentiviral-mediated integrin alpha5 expression in human adult mesenchymal stromal cells promotes bone repair in mouse cranial and long-bone defects. Hum Gene Ther 2012;23(2):167e72. [37] Kleinschmidt K, Ploeger F, Nickel J, Glockenmeier J, Kunz P, Richter W. Enhanced reconstruction of long bone architecture by a growth factor mutant combining positive features of GDF-5 and BMP-2. Biomaterials 2013;34(24):5926e36.

[38] Kanczler JM, Ginty PJ, White L, Clarke NM, Howdle SM, Shakesheff KM, et al. The effect of the delivery of vascular endothelial growth factor and bone morphogenic protein-2 to osteoprogenitor cell populations on bone formation. Biomaterials 2010;31(6):1242e50. [39] Meinel L, Hofmann S, Betz O, Fajardo R, Merkle HP, Langer R, et al. Osteogenesis by human mesenchymal stem cells cultured on silk biomaterials: comparison of adenovirus mediated gene transfer and protein delivery of BMP-2. Biomaterials 2006;27(28):4993e5002. [40] Miyazaki M, Zuk PA, Zou J, Yoon SH, Wei F, Morishita Y, et al. Comparison of human mesenchymal stem cells derived from adipose tissue and bone marrow for ex vivo gene therapy in rat spinal fusion model. Spine 2008;33(8): 863e9. [41] Undale A, Fraser D, Hefferan T, Kopher RA, Herrick J, Evans GL, et al. Induction of fracture repair by mesenchymal cells derived from human embryonic stem cells or bone marrow. J Orthop Res 2011;29(12):1804e11. [42] Nesti LJ, Jackson WM, Shanti RM, Koehler SM, Aragon AB, Bailey JR, et al. Differentiation potential of multipotent progenitor cells derived from wartraumatized muscle tissue. J Bone Jt Surg Am 2008;90(11):2390e8. [43] Jackson WM, Aragon AB, Djouad F, Song Y, Koehler SM, Nesti LJ, et al. Mesenchymal progenitor cells derived from traumatized human muscle. J Tissue Eng Regen Med 2009;3(2):129e38. [44] Jackson WM, Lozito TP, Djouad F, Kuhn NZ, Nesti LJ, Tuan RS. Differentiation and regeneration potential of mesenchymal progenitor cells derived from traumatized muscle tissue. J Cell Mol Med 2011;15(11):2377e88. [45] Jackson WM, Aragon AB, Bulken-Hoover JD, Nesti LJ, Tuan RS. Putative heterotopic ossification progenitor cells derived from traumatized muscle. J Orthop Res 2009;27(12):1645e51. [46] Lavery K, Swain P, Falb D, Alaoui-Ismaili MH. BMP-2/4 and BMP-6/7 differentially utilize cell surface receptors to induce osteoblastic differentiation of human bone marrow-derived mesenchymal stem cells. J Biol Chem 2008;283(30):20948e58. [47] Ko EK, Jeong SI, Rim NG, Lee YM, Shin H, Lee BK. In vitro osteogenic differentiation of human mesenchymal stem cells and in vivo bone formation in composite nanofiber meshes. Tissue Eng Part A 2008;14(12):2105e19. [48] Degano IR, Vilalta M, Bago JR, Matthies AM, Hubbell JA, Dimitriou H, et al. Bioluminescence imaging of calvarial bone repair using bone marrow and adipose tissue-derived mesenchymal stem cells. Biomaterials 2008;29(4): 427e37. [49] Scotti C, Piccinini E, Takizawa H, Todorov A, Bourgine P, Papadimitropoulos A, et al. Engineering of a functional bone organ through endochondral ossification. Proc Natl Acad Sci U S A 2013;110(10):3997e4002.