Biomaterials 35 (2014) 2680e2691

Contents lists available at ScienceDirect

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

The effect of type II collagen on MSC osteogenic differentiation and bone defect repair Li-Hsuan Chiu a,1, Wen-Fu T. Lai a, b, c, Shwu-Fen Chang a, Chin-Chean Wong d, Cheng-Yu Fan e, Chia-Lang Fang f, Yu-Hui Tsai a, c, * a

Graduate Institute of Medical Sciences, Taipei Medical University, Taipei 11031, Taiwan Graduate Institute of Clinical Medicine, Taipei Medical University, Taipei 11031, Taiwan Center for Nano Biomedicine Research, Taipei Medical University, Taipei 11031, Taiwan d Department of Orthopaedic Surgery, Wanfang Hospital, Taipei Medical University, Taipei 11031, Taiwan e Department of Orthopaedic Surgery, Taipei Medical University Hospital, Taipei 11031, Taiwan f Department of Pathology, School of Medicine, College of Medicine, Taipei Medical University, Taipei 11031, Taiwan b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 October 2013 Accepted 8 December 2013 Available online 8 January 2014

The function of type II collagen in cartilage is well documented and its importance for long bone development has been implicated. However, the involvement of type II collagen in bone marrow derived mesenchymal stem cell (BMSC) osteogenesis has not been well investigated. This study elucidated the pivotal role of type II collagen in BMSC osteogenesis and its potential application to bone healing. Type II collagen-coated surface was found to accelerate calcium deposition, and the interaction of osteogenic medium-induced BMSCs with type II collagen-coated surface was mainly mediated through integrin a2b1. Exogenous type II collagen directly activated FAK-JNK signaling and resulted in the phosphorylation of RUNX2. In a segmental defect model in rats, type II collagen-HA/TCP-implanted rats showed significant callus formation at the reunion site, and a higher SFI (sciatic function index) scoring as comparing to other groups were also observed at 7, 14, and 21 day post-surgery. Collectively, type II collagen serves as a better modulator during early osteogenic differentiation of BMSCs by facilitating RUNX2 activation through integrin a2b1-FAK-JNK signaling axis, and enhance bone defect repair through an endochondral ossification-like process. These results advance our understanding about the cartilaginous ECM-BMSC interaction, and provide perspective for bone defect repair strategies. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Mesenchymal stem cell Type II collagen Osteogenesis Cell adhesion Integrin signaling Bone healing

1. Introduction Osteogenesis involves the differentiation of bone marrow derived mesenchymal stem cells (BMSCs), which is an important process contributing to normal growth and wound healing of the bone tissue. The lost of function of BMSCs may result in delayed fracture healing and further pathological changes of the bone tissue [1,2]. Means to enhance the osteogenic differentiation of BMSCs could contribute to the healing of bone tissue after trauma,

orthopedic surgery, or dental procedure. Type II collagen, a cartilaginous ECM molecule mainly present in the cartilage and developing bone, has been implicated to play important roles in both fracture healing and long bone development [1,3]. Transgenic mice bearing partially deleted type II collagen gene showed a temporary impairment of callus remodeling and fracture healing [4]. On the other hand, mice carrying a partially deleted type II collagen gene present phenotypes of chondrodysplasia, with characteristics of dwarfism, thick limbs, and delayed mineralization of bone [5].

Abbreviations: BMSC, bone marrow derived mesenchymal stem cell; ECM, extracellular matrix; VLA-2, very late antigen-2 (integrin a2b1 complex); FAK, focal adhesion kinase; MEK, mitogen-activated protein kinase; ERK1/2, extracellular signal-regulated kinase 1/2; JNK, c-Jun N-terminal Kinase; RUNX2(Cbfa1), core binding factor alpha1; SOX9, SRY-related high mobility group-box gene9; AGN, aggrecan gene; COL2A1, type-II collagen alpha1 chain; ITGA2, integrin alpha2; ITGB1, integrin beta 1; HA/TCP, hydroxyapatite/tricalcium phosphate. * Corresponding author. Graduate Institute of Medical Sciences, Taipei Medical University, 250, Wu-Xing Street, Taipei 11031, Taiwan. Tel.: þ886 2 27361661x3417/3070; fax: þ886 2 23778620. E-mail addresses: [email protected], [email protected] (Y.-H. Tsai). 1 Current affiliation: Department of Obstetrics and Gynecology, Taipei Medical University Hospital, Taipei 11031, Taiwan. 0142-9612/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biomaterials.2013.12.005

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

Furthermore, transgenic mice with inactivated type II collagen gene expressed phenotypes that lack of endochondral bones and epiphyseal growth plates [6]. Some in vitro studies also imply the role of type II collagen matrix in the osteogenic differentiation of BMSCs. A proper chondrogenic induction of mesenchymal stem cells before implantation resulted in a more successful bone formation, in which a persisting type II collagen expression of the implant was observed [7]. Similarly, the chondrogenic pre-induction of b-TCP/BMSC composites which exhibited a significant production of type II collagen could enhance full bone formation, even including marrow organization [8]. At the cellular level, it is suggested that skeletal progenitor cells displayed type II collagen expression is sufficient to drive ectopic ossicle formation with myelosupportive stroma and adipogenesis [9]. Furthermore, it have been demonstrated that exogenous type II collagen promotes BMSC osteogenesis and inhibits adipogenesis, providing a clue that type II collagen itself may play an important role in cell fate commitment during the early stage of BMSC differentiation [10]. These results implied that type II collagen matrix is crucial to the differentiation of BMSCs not only during the early stage of embryonic bone development but also the fracture healing process. Consequently, it is hypothesized that through an endochondral ossification-like process, type II collagen is an important modulator for osteogenesis of BMSCs. As is known, type I collagen enhances osteogenic differentiation and facilitates cell attachment of BMSCs [11]. When osteoblasts or BMSCs bind to type I collagen, ERK1/2 signaling pathway is activated to trigger the osteogenic differentiation of the cell [12,13]. Thus type I collagen has been widely applied for bone regeneration, such as combined with hydroxyapatite or calcium phosphates as a bone filling material [14e16]. On the other hand, the signaling pathway that type II collagen modulates BMSC osteogenic differentiation, or its possible application in bone regeneration has not been well elucidated. Herein, the molecular mechanisms of type II collagen for modulating BMSC osteogenic differentiation, and its possible role in bone defect repair, were carefully evaluated in this present study. 2. Materials and methods 2.1. Ethics statement The protocols and informed consent form for BMSC isolation were approved by the Taipei Medical University Joint Institutional Review Board (TMUH-03-08-12). The specimen donor was provided with the IRB-approved formal consent form describing sufficient information for one to make an informed decision about his/ her participation in this study. The formal consent form was signed by the subject before specimen collection. 2.2. Reagents Antibodies against human antigens CD105, CD73, CD44, CD29, CD90w, CD34, CD45, and CD14 were purchased from BD Biosciences (San Jose, CA, USA). Antibodies against human STRO-1 and RUNX2 were obtained from R&D Systems (Minneapolis, MN, USA). Functional blocking antibodies against integrins a1b1, a2b1, a5b1, and avb3, and antibody against VLA-2 complex used in FACS analysis were purchased from Millipore (Billerica, MA, USA). Antibodies against pFAK397, pFAK576/577, pFAK925, pERK1/2, pJNK, ERK1/2, JNK, and b-actin were obtained from Cell Signaling Technology (Danvers, MA, USA). Antibody against RUNX2 was purchased from MBL International Corporation (Woburn, MA, USA). Antibody against phosphoserine was purchased from Abcam (Cambridge, CB4 0FW, UK). Protein A-agarose immunoprecipitation reagent was purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). Alexa FluorÒ 594 phalloidin was from Invitrogen (Carlsbad, CA, USA). DAPI (D9542) was purchased from SigmaeAldrich (St. Louis, MO, USA). Dulbecco’s Modified Eagle Medium low glucose (DMEM/LG), Dulbecco’s Modified Eagle Medium high glucose (DMEM/HG), Dulbecco’s Modified Eagle Medium Nutrient Mixture F-12 (DMEM/F12), fetal bovine serum and other cell culture-related supplies were from Invitrogen (Carlsbad, CA, USA). Percoll solution was from GE Healthcare Bio-Sciences (Piscataway, NJ, USA). Type II collagen of chicken sternal cartilage (c9301) and type I collagen of rat tail tendon (c3867) were obtained from SigmaeAldrich (St. Louis, MO, USA). Alizarin Red S (A5533) was purchased from SigmaeAldrich (St. Louis, MO, USA). RT-PCR related reagents include TRIzolÒ and

2681

SuperScriptÒ III RT system were from Invitrogen (Carlsbad, CA, USA). SYBR Green I qPCR system was obtained from Roche Applied Science (Indianapolis, IN, USA). The porous biphasic HA/TCP scaffolds (Sinbone HT, hydroxyapatite/tricalcium phosphate ¼ 60/40 in weight %) used in the in vitro study were obtained from Purzer Pharmaceutical Co., Ltd. (Taipei, Taiwan). The biphasic HA/TCP bone substitutes (BoneGraft, hydroxyapatite/tricalcium phosphate ¼ 60/40 in weight %, granule size: 63e250 um) used in the animal study were obtained from Biotech One Inc (Taipei, Taiwan). 2.3. Mesenchymal stem cell (BMSC) isolation, cultivation and storage Bone marrow aspirates were obtained aseptically from three donors (male, 40e 65 years old) with informed consent. Bone marrow specimen was collected from the disposed aspirates using a 10 ml syringe. The aspirates were immediately mixed with sodium-heparin (10000U/ml), and diluted in five volumes of phosphatebuffered saline. The cell suspension was then fractionated on a Percoll gradient (1.077 g/cm3 of density, Pharmacia) and centrifuged at 800  g for 30 min. The BMSC-enriched interface fraction was collected and plated onto a 10 cm dish containing 10 ml Dulbecco’s Modified Eagles Medium with 1 g/ml glucose (DMEM/LG), 10% FBS, 1 P/S/A (penicillin/streptomycin/fungizone). The medium was changed every four days. When the cells reached 80% confluence, they were trypsinized and passaged into new 10-cm dishes at a cell density of 5  105 cells/dish. The cells were sub-cultured till passage 3 (P3). P3 cells were then seeded at a cell density of 6.5  103 cells/cm2 and subjected to various studies. The remaining cells were collected, resuspended in 10% DMSO in FBS to a concentration of 2  106 cells/ml, and then stored in liquid nitrogen for later use. 2.4. Differentiation assay The multipotency characteristics of BMSCs toward osteogenic, chondrogenic and adipogenic differentiations were assessed. For osteogenic differentiation of human BMSCs, cells were cultured in DMEM/LG medium supplemented with 10% FBS, 50 mg/ml L-ascorbate-2-phosphate, 107 M dexamethasone and 10 mM bglyceralphosphate for 21 days. The chondrogenic differentiation of BMSCs was achieved by high-density micromass culture in the chondrogenic medium (DMEM/ F12, 5% FBS, 107 M dexamethasone, 50 mg/ml L-ascorbate-2-phosphate, and 10 ng/ ml TGF-b1) for 21 days [17,18]. Briefly, BMSCs were suspended in DMEM/F12 medium at the density of 1  107 cells/ml. Droplets of 10 ml cell suspension were loaded into the culture dishes to form cell aggregates on the substratum. The droplets of high density cells were allowed to stand at 37  C for 2 h, and then the chondrogenic medium was carefully loaded into the culture dishes. For adipogenic differentiation, BMSCs were induced in DMEM/HG medium in the presence of 10% FBS, 106 M dexamethasone, 0.5 mM methyl-isobutyl-methyl-xanthine, 0.2 mM indomethacin, and 10 mg/ml insulin for 21 days. 2.5. Flow cytometry analysis BMSCs were fixed with ethanol overnight at 20  C. Aliquots of 5  105 cells were incubated separately with each of the fluorochrome-conjugated antibodies against a panel of cell surface markers, including STRO-1, CD105, CD73, CD29, CD44, CD90w, CD34, CD14 and CD45 for 45 min at 4  C. Cells were resuspended in Con’s tube (BD) containing 200 ml of phosphate-buffered saline/1% bovine serum albumin and analyzed by Flow Cytometry using the FACS Calibur system (BD Biosciences). 2.6. Immunofluorescence staining To evaluate the effect of type II collagen-coated surface on BMSCs, cells were cultured on either type I collagen-coated, type II collagen-coated or non-coated plates for designated intervals before subjected to immunofluorescence staining. After fixation, cells were blocked with blocking buffer (10% BSA, 0.3% Triton X-100) for 30 min, followed by incubation with each specific primary antibody for 2 h and with fluorescence conjugated-secondary antibody for 1 h at room temperature (RT). The cells were washed 3 times (10 min each) with PBS after the incubation. Cells were further stained with 0.1 mg/ml of DAPI (blue) at RT for 30 min to visualize the nuclei. For F-actin filament staining, BMSCs with or without osteogenic induction were allowed to attach on variously coated plates for 120 min, followed by formaldehyde fixation and immunostaining with Alexa FluorÒ 594 conjugated-phalloidin (red) for 20 min at RT. Cells were further stained with 0.1 mg/ml of DAPI (blue) at RT for 30 min to visualize the nuclei. 2.7. Calcium deposition assay To detect calcium deposition, the differentiated BMSCs were fixed with 4% formaldehyde at RT for 30 min, and rinsed rapidly with distilled water. Then, 1 ml of pH 4.2 Alizarin Red S solution was added to cover cell surface for 5 min, followed by washing thoroughly with distilled water. The calcium deposits exhibited as orange red sediments on the cell surface and were recorded microscopically. 2.8. Total RNA isolation and RT-PCR Total RNA of cells after various treatments was extracted with TRIzolÒ reagent and stored at 80  C for later use. RNA (500 ng) was then reverse-transcribed in a

2682

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

20 ml reaction mixture. Reverse-transcription was performed according to the protocol described by the manufacturer (Superscript III Kit, Invitrogen). Aliquots of cDNA specimens in each group were further amplified by conventional PCR (PlatinumÒ Taq system, Invitrogen) or realtime PCR (Roche Lightcycler System) using specific primer sets as follow: RUNX2: forward, 50 -TACCAGACCGAGACCAACAGAG-30 , and reverse, 50 -CACCACCGGGCTCACGTCGC-30 ; SOX9: forward, 50 -ATCTGAAGAAGGAGAGCGAG-30 , and reverse, 50 -TCAGAAGTCTCCAGAGCTTG-30 ; AGN: forward, 50 -TGCATTCCACGAAGCTAACCT-30 , and reverse, 50 -CGCCTCGCCTTCTTGAAAT-30 ; COL2A1: forward, 50 -TACCCCAATCCAGCAAACGT-30 , and reverse, 50 TGTTTCGTGCAGCCATCCT-30 ; osteocalcin: forward, 50 -CCACCGAGACACCATGA-30 , and reverse, 50 -CAGCCATTGATACAGGTAGC-30 ; osteonectin: forward, 50 -AGGTATCTGTGGGAGCTAATC-30 , and reverse, 50 -ATTGCTGCACACCTTCTC-30 ; ITGA2: forward, 50 -GTGCCTTTGGACAAGTGGTT-30 , and reverse, 50 -GGGCAACTCTGTGCTTGAT30 ; ITGB1: forward, 50 -ATGAATGAAATGAGGAGGATTACTTCG-30 , and reverse, 50 AAAACACCAGCAGCCGTGTAAC-30 ; fatty acid binding protein: forward, 50 GGAAAGTCAAGAGCACCATA-30 , and reverse, 50 -CATGACGCATTCCACCAC-30 ; b-Actin: forward, 50 -CGAGGACTTTGATTGCAC-30 , and reverse, 50 -TATCACCTCCCCTGTGTG30 .The expression intensities of each gene were normalized by using b-Actin gene product as internal control. 2.9. Western blotting Whole cell lysates were prepared by protein extracting buffer containing 50 mM TriseHCl, pH 7.4, 150 mM NaCl, 0.1% SDS and protease inhibitor cocktail (Sigmae Aldrich). Total proteins were size-fractionated on SDS-PAGE and transferred to PVDF membrane. The membrane was blocked with 5% non-fat dry milk prior to incubating with each primary antibody (1:1000 dilution), followed by respective HRPconjugated secondary antibody (1:2000 dilution). The membrane was washed with Tris-buffered saline/0.1% TWEEN-20 after each incubation, and visualized with ECL reagent (Thermo Scientific, IL, USA). The activation of FAK, ERK and JNK were evaluated by antibodies specific to the activated, tyrosine-phosphorylated and threonine-phosphorylated FAK, ERK and JNK, respectively. 2.10. Immunoprecipitation Whole cell lysates were prepared by protein extracting buffer containing 50 mM TriseHCl, pH 7.4, 150 mM NaCl, 0.1% SDS and protease inhibitors. Aliquots of cell lysates were incubated with 2 mg of anti-phosphoserine antibody for 2 h at 4  C, followed by precipitation with 20 ml of protein A-agarose beads for 1 h at 4  C. The immunoprecipitated proteins were subsequently blotted with RUNX2 antibody. 2.11. Collagen coating Type I or type II collagen solutions were diluted with 5 mM acetic acid to the designated concentration for surface coating. Tissue culture plates were coated with type I collagen or type II collagen at a concentration of 100 mg/ml for 2 h at RT (room temperature). The remaining collagen solution was removed, and the dishes were washed with PBS twice. After washings, the coated dishes were then dried and UVsterilized in the cell culture hood. The coated dishes were stored at 4  C before use. HA/TCP scaffolds were coated with type I collagen, type II collagen, or type I þ type II collagen at a final concentration of 100 mg/ml or with 5 mM acetic acid (as noncoating control) for 2 h at RT. After incubation, the remaining solution was removed, and the collagen/acetic acid treated HA/TCP scaffolds were washed with PBS. The ECM-coated scaffolds were air-dried in the cell culture hood with UV-light on for 1 h and stored at 4  C until use. 2.12. Cell attachment analysis Aliquots of the BMSCs were separately pre-treated with various integrin blocking antibodies for 30 min, followed by plated on the collagen-coated or noncoated plates without wash. After cultured for 12 h in the presence of the respective blocking antibody, the cells were fixed with 4% formaldehyde and examined under microscopic observation. Morphologies of the BMSCs from various experimental groups were recorded with a IX81 inverted Microscope (Olympus, Tokyo, Japan) with a RT3 CCD camera (SPOT Imaging Solutions, MI, USA). The relative cell spreading area was quantified from 4 random images of selected fields of the culture plates from each repeated experiment by Image J software (NIH Image, MD, USA). 2.13. Animal surgery A rat bone defect model was established to evaluate the bone healing potency of type II collagen-HA/TCP substitute. A total of 16 male mature SD rats (8wk-old) with the average weight of 300 g were divided into 4 groups of 4 animals for non-treated (control), agarose-HA/TCP, type I collagen-HA/TCP, and type II collagen-HA/TCP treated groups. For the preparation of bone substitutes, 600 mg of HA/TCP powder (granule size: 63e250 um) was mixed with 600 ul of either type I collagen, type II collagen or agarose solution (1 mg/ml) before the operation. Surgery was performed with anesthesia of IM injection of 50 mg/kg ketamine and 10 mg/kg xylazine mixture. After shaving each animal’s right hind limb with iodine-alcohol disinfection, a 5 mm segmental defect at the middle of the femur shaft was created using a

0.5 mm drill. Under sterile condition, the proximal and distal segments of femur were fixated with a 1.2 mm intramedullary rod. Aliquots of 60 ul collagen-HA/TCP or agarose-HA/TCP bone substitutes were applied at the segmental defect site and covered with a patch of 10 mm2 gel foam. The soft tissue was closed in two layers with absorbable sutures and the wound region was disinfected again after surgery. The animals were allowed to move freely without any restrictions. Animals were followed by x-ray, micro-computed tomography and foot print analysis at 28 and 35 days post-surgery. At the end of the treatment, the rats were euthanized and the femur were harvested and stored in 10% buffered formalin for histological analysis. 2.14. Footprint analysis Foot print analysis of the experimental animals was performed at 7, 14, 21, 28, 35 days after the operation as described elsewhere [19,20]. The hind feet of the experimental animals were dipped in carbon ink and placed on a paper-lined walkway (10 cm  50 cm) that led into a darkened cage box. Footprints left on the paper as the rat walked down the track were measured for comparison between the experimental and control sides. For each animal, 3 footprints out of the 5 steps were measured. The footprints were analyzed by following parameters: (1) print length (PL), defined as the distance from the heel to the third toe; (2) toe spread (TS), defined as the distance from the first to the fifth toe, and (3) intermediate toe spread (ITS), defined as the distance from the second to the fourth toe. The ratio factor of each experimental group is defined as: (1) print length factor (PLF) ¼ (EPL  NPL)/NPL, (2) toe spread factor (TSF) ¼ (ETS  NTS)/NTS, and (3) intermediary toe spread factor (ITSF) ¼ (EITS  NITS)/ NITS, where E is for the experimental side and N is for the normal control. The ratio factors were calculated by Baine-Mackinnone-Hunter sciatic function index (SFI) with the formula: SFI ¼ 38.3  PLF þ 109.5  TSF þ 13.3  ITSF  8.8. 2.15. Computed tomography analysis CT analysis of the experimental animals was performed at 35 days post-surgery. SD Rats were maintained under general anesthesia during the scanning procedure. Each rat is placed in a sample holder in the cranial-caudal direction and scanned using a high-resolution micro-CT system (Triumph X-O CT System) at a spatial resolution of 80 mm (voxel dimension) and 1024  1024 pixel matrices. The X-ray image and multiple CT section of the lesion site were captured. A volume rendering model is created using AMIDE software to reveal the morphology of the repaired tissue of the surgery site. 2.16. Statistical analysis For the quantitative assays, each datum point represents mean  SD of three independent experiments or an experiment of triplicate assay. The statistical analysis was performed with one-way ANOVA and Duncan’s Multiple Range Test. The statistical significances among the experimental groups were indicated with superscripts of consecutive alphabets. Groups labeled with consecutive superscript letters, indicate that the statistic difference between the two groups is at p value less than 0.05, and thus, are considered significantly different.

3. Results 3.1. Surface marker identification and differentiating assay of BMSCs The BMSCs were isolated as described, expanded, and cell morphology was closely monitored. Fig. 1A show the colony forming units in P0 (passage 0) primary culture. Fibroblast-like cell phenotype was observed even after 4 passages (Fig. 1B). The surface markers of BMSCs were characterized by flow cytometry analysis as shown in Fig. 1C. A phenotype of STRO-1þ CD105þ CD73þ CD29þ CD44þ CD90wþ CD34 CD14 CD45 cell surface markers was identified in the P4 (passage 4) BMSCs. To examine the multipotency of the cells used in this study, the P4 BMSCs were further treated with various induction medium for 21 days to test their differentiating ability toward osteoblasts, chondrocytes, and adipocytes. After 21 days of osteogenic induction, the culture dish showed significant reddish precipitation after alizarin red S staining, representing the mineralization of the culture (Fig. 1D). Similarly, alcian blue staining exhibited abundant GAG deposition in the high-density micromass culture after 21 days of chondrogenic induction (Fig. 1E). In the case of adipogenic differentiation, the induced cells exhibited red (in the web version) oil droplets after staining with oil red O stain after 21 days of induction (Fig. 1F). The marker mRNA expressions were also evaluated after each induction to further confirm the differentiation

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

ability of BMSCs. After 21 days of osteogenic induction, cells exhibited increasing levels of osteonectin and osteocalcin mRNA expression. Likewise, after 21 days of chondrogenic induction, the up-regulation of either SOX9 or COL2A1 mRNA expression level was detected. As to the adipogenic differentiation potential, the expression of significant fatty acid binding protein mRNA was also found after 21 days of induction (Fig. 1G). These data demonstrate that the primarily isolated BMSCs from surgical patients possess multi-lineage differentiation potential into osteogenic, chondrogenic, and adipogenic lineage cells. 3.2. The effect of type II collagen on BMSC osteogenesis To evaluate the modulatory effect of type II collagen on BMSC osteogenic differentiation, the cells were cultured on type I

2683

collagen-coated, type II collagen-coated, or non-coated control culture plates. After attachment, BMSCs were treated with osteogenic medium for 4, 8, 12 and 16 days to detect the initiation of calcium deposition of BMSCs on variously coated plates. The induced BMSCs were subjected to formaldehyde fixation and alizarin red S staining [21]. BMSCs cultured on type I collagen-coated and type II collagen-coated plates exhibited significant calcium deposition by day 12; much earlier than that did on the non-coated control plates that showed calcification by day 16. At day 16, BMSCs on either type I collagen-coated or type II collagen-coated plates showed a much greater level of calcium deposition than cells on the non-coated plates (Fig. 2A). As shown in Fig. 2B, each bar presents the mean level of alizarin red S-stained calcium deposition in each group of triplicate after quantification. To further characterize the osteogenic induction effect of type II collagen, RUNX2 expression of

Fig. 1. Cell morphology, differentiation potential and surface marker expressions of human bone marrow-derived BMSCs: (A) A colony forming unit is shown under bright field observation in the passage 0 BMSC culture. (B) A fibroblast-like phenotype of cells is observed after 4 passages of BMSC culture. The scale bars indicate 100 mm. (C) An aliquot of 5  105 BMSCs were separately incubated with each surface marker antibody at 4  C for 30 min, and subjected to FACS analysis. The BMSCs were found to exhibit a phenotype of STRO-1þ CD105þ CD73þ CD29þ CD44þ CD90wþ CD34 CD14 CD45. Histogram overlays show the FL1 (FITC) or FL2 (PE) intensity corresponding to the positive staining of each cell surface marker (solid line) and the intensity for the isotype control (dotted line). (D) Alizarin red staining, (E) Alcian blue staining, and (F) Oil red O staining of BMSCs after osteogenic, chondrogenic and adipogenic induction for 21 days, respectively. Upper left insert in (D) shows the gross view of the whole cell culture dish of the osteogenic mediuminduced BMSCs after calcium deposition staining. Upper left insert in (F) shows the adipogenesis-induced BMSCs with positively-stained oil droplets at 200 magnification. The scale bars indicate 100 mm. (G) For each differentiation assay, the corresponding marker mRNA expressions were evaluated to examine the differentiation status of BMSCs to osteogenic, chondrogenic, and adipogenic lineage cells. Non-induced BMSCs (cultured in basal medium) were used as control to validate the results.

2684

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

BMSCs cultured on type I collagen-coated, type II collagen-coated, or non-coated control culture plates were also analyzed by immuno-staining. After 16 days of culture in the osteogenic medium, BMSCs of each group were fixed and stained with RUNX2 antibody conjugated with FITC. Under fluorescent microscope, BMSCs cultured on the type II collagen-coated plates exhibited a greater level of RUNX2 expression than did cells either on the type I collagen-coated plates or the non-coated control plates (Fig. 2C). To examine further the regulatory effects of type II collagen on osteogenic marker gene expression in BMSCs, the mRNA expression patterns of RUNX2, osteocalcin, SOX9, and AGN in the BMSCs cultured with osteogenic induction on either type II collagen-coated or non-coated plates were assessed up to 10 days (Fig. 2D). In the CII-coated group, the RUNX2 mRNA level began to elevate since day 6 and significantly up-regulated at day 8 and day 10. Similarly, the osteocalcin mRNA level increased after day 6, and gradually elevated by day 8 and sharply elevated by day 10. However, in the non-coated group, the mRNA expression levels of both RUNX2 and osteocalcin slightly elevated without statistical significance during the same induction interval. As to the chondrogenic marker, SOX9 mRNA expression levels in both groups were gradually decreased during the 10 day treatment. By day 10, the SOX9 expression levels in both the non-coated and type II collagen-coated groups reduced to a 0.4-fold and 0.25-fold of that at day 0, respectively. Likewise, the AGN expression levels of non-coated group and type II collagen-coated group also decreased to about 0.15-fold and 0.4-fold of that at day 0, respectively. The data demonstrate that type II collagen-coated surface significantly enhanced the level of osteogenic marker mRNA expression and calcium deposition of BMSCs during the osteogenic induction. 3.3. The expression of potential type II collagen receptor in osteogenic medium-induced BMSCs Integrin a2b1 complex are known to serve as a primary receptor for native type II collagen through a RGD-independent pathway [22]. To explore the possible direct association of a2b1 integrin with the regulatory role of type II collagen in BMSC osteogenic differentiation, the level of activated integrin a2b1 complex (VLA-2) and the mRNA expression levels of ITGA2 and ITGB1 integrin subunits were monitored during the osteogenic induction for 10 days. By FACS analysis, the VLA-2 expression on BMSC surface increased by 1.8-fold over control level on day 4 post-osteogenic induction, and then gradually reduced from day 6 to a lower level than that of the control by day 10 (Fig. 3A). Subsequently, the expression levels of both integrin subunits were also found to be up-regulated after the exposure of BMSCs to osteogenic medium. The mRNA expression of ITGA2 increased 2.5-fold and reached a peak level about 3-fold at day 4 after the induction. In the case of ITGB1, the mRNA expression level was elevated at day 2, and declined after day 4 (Fig. 3B). The results imply that BMSCs exhibited an elevation of both integrin a2b1 mRNA level and activated VLA-2 complex expression during the early stage of osteogenic differentiation. 3.4. The interaction of osteogenic medium-induced BMSCs with type II collagen-coated surface To further evaluate the interaction of BMSCs with type II collagen during the early stage of osteogenic differentiation, BMSCs with or without a 4-day pre-induction with osteogenic medium were seeded on either type I collagen-coated, type II collagencoated or non-coated plates for 2 h to examine the actin filament (F-actin) distribution and cell spreading morphology (Fig. 4A). Both type I collagen-coated or type II collagen-coated plates facilitated

cell spreading and F-actin extension in BMSCs comparing to that on the non-coated control plates, either with or without osteogenic medium induction. Interestingly, after 4-day induction with the osteogenic medium, BMSCs exhibited a more expanded morphology with positive F-actin staining than the non-induced group (w/o osteogenic induction) on the type II collagen-coated plates. The observation implies that osteogenic induction alters the cell spreading morphology and F-actin distribution of BMSCs on type II collagen-coated surface. 3.5. The role of integrin subtypes in BMSC-type II collagen interaction Integrins a1b1, a2b1, a5b1 and avb3 have been reported to bind type I and/or type II collagens [22,23]. To examine the role of these integrins in the interaction of BMSCs with type II collagen during early stage of osteogenic differentiation, the osteogenic mediuminduced BMSCs were treated with the functional blocking antibodies of these integrin subtypes. After induced with osteogenic medium for 4 days, BMSCs were treated with functional blocking antibodies against a1b1, a2b1, a5b1, or avb3 integrins for 30 min, and then seeded on the type II collagen-coated plates. After 12 h of cultivation, cells were fixed and examined for their morphology under microscope. BMSCs attached on the type II collagen-coated plates showed a more extended cell morphology and greater cell spreading area (Fig. 4D) than that on the non-coated control plates (Fig. 4C). However, after the pre-treatment of a2b1 blocking antibody, the spreading of the osteogenic medium-induced BMSCs on type II collagen-coated surface was significantly reduced (Fig. 4F); when treated with a1b1 blocking antibody, the spreading of the osteogenic medium-induced BMSCs was moderately reduced (Fig. 4E). Furthermore, the pre-treatment with a5b1 (Fig. 4G) or avb3 (Fig. 4H) blocking antibodies only slightly alter the spreading of the osteogenic medium-induced BMSCs on the type II collagencoated plates. Fig. 4B represents the quantitative data of average cell spreading area for each group of repeated experiments. These results indicate that the interaction of BMSCs with type II collagencoated surface is mainly mediated through integrin a2b1. 3.6. Type II collagen-induced signaling in osteogenic mediuminduced BMSCs To study the molecular mechanism of type II collagen for modulating BMSC osteogenesis, the activation of integrin-related FAK signaling pathway by exogenous type II collagen was examined. After 4 days of osteogenic induction, BMSCs cultured on noncoated control culture plates were treated with 200 mg/ml of type II collagen and collected at 15, 30, 60, 90, 120 min after treatment. The cell lysates were analyzed by Western Blot analysis. FAK was found to be sequentially phosphorylated at Tyr-397 (peaked at 15 min), Tyr-576/577 (peaked at 30 min), and Tyr-925 (peaked at 60 min). The activation of downstream MAP kinases was also followed. The phosphorylation of JNK was observed at 15 min post type II collagen exposure, while that of ERK1/2 showed no significant change (Fig. 5A). Subsequently, the short term effect of type II collagen on RUNX2 activation was also examined. The phosphorylation level of RUNX2 was elevated and reaching peak state at 90 min, and then down regulated after 120 min of exposure to type II collagen (Fig. 5B). The result implied that type II collagen can directly activate RUNX2 through FAK-JNK signaling but not through FAK-ERK1/ 2 cascade. To further determine the role of integrin a2b1 in type II collagen activated signaling in osteogenic medium-induced BMSCs, an antibody of integrin a2b1 was used to block the binding of exogenous type II collagen to the cells. After 4 days of osteogenic

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691 2685

Fig. 2. The calcium deposition level, RUNX2 fluorescent staining, and mRNA expression patterns of BMSCs cultured on variously coated plates. (A) Control: non-coated plates; CI-Coated: BMSCs cultured on 100 mg/ml type I collagencoated plates; CII-Coated: BMSCs cultured on 100 mg/ml type II collagen-coated plates. (B) Data represent the quantified mean levels of alizarin red S staining of various groups of triplicate assays. (C) For RUNX2 staining, BMSCs of each group were cultured in osteogenic medium for 16 days, and stained with FITC-conjugated RUNX2 antibody (green). The cell nuclei were stained with DAPI (blue). The scale bars indicate 50 mm. (D) Messenger RNA expression patterns of osteogenic and chondrogenic marker genes in BMSCs with osteogenic medium treatment. The expression intensities of each gene were normalized by b-Actin as internal control. The statistical significance among the experimental groups was indicated with superscripts of consecutive alphabets. Groups labeled with different superscript letters are significantly different at p < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

2686

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

Fig. 3. The expression of activated integrin a2b1 complex (VLA-2) and their mRNA expression levels. (A) FACS analysis for VLA-2 complex expression on BMSCs during osteogenic induction. The levels of VLA-2 complex on BMSCs cultured in osteogenic medium were examined at days 0, 2, 4, 6, 8, and 10 by using FACS analysis. Histogram overlays show the FL1 (FITC) intensity corresponding to VLA-2 positive staining (solid line) and the intensity for the isotype control (dash line). The numbers in the graph indicate the relative intensity (RI) of the fluorescence. (B) Messenger RNA expression patterns of ITGA2 and ITGB1 integrin subunits in BMSCs during osteogenic induction. BMSCs were cultured on 6 well culture plates in the osteogenic medium for 12 days. Total RNA was extracted at day 0, 2, 4, 8, and 12 and the relative mRNA expression levels of ITGA2 and ITGB1 were analyzed by realtime PCR. The expression intensities of each gene were normalized by using b-actin as internal control. The statistical significance among the experimental groups was indicated with superscripts of consecutive alphabets. Groups labeled with different superscript letters are significantly different at p < 0.05.

induction, BMSCs were pre-treated with a2b1 blocking antibody for 30 min, followed by exposure to 200 mg/ml of type II collagen for another 30 min, then subjected to western analysis. The pretreatment of the BMSCs with a2b1 antibody abolished the type II collagen-activated FAK and JNK phosphorylation (Fig. 5C). However, the phosphorylation of ERK1/2 remained unchanged. The result revealed that type II collagen-induced FAK-JNK phosphorylation was integrin a2b1-dependant. All together, the data indicate that integrin a2b1 is the predominant receptor for type II collagen, and directly activates FAK-JNK signaling cascade in the osteogenic medium-induced BMSCs. 3.7. Mineral deposition of BMSCs on type II collagen-coated HA/TCP scaffolds To evaluate the potential of type II collagen in orthopedic application, the BMSCs were seeded into non-coated control HA/ TCP scaffold, or HA/TCP scaffolds coated with either type I collagen (100 mg/ml), type II collagen (100 mg/ml), or 1:1 mixture of type I þ type II collagen (at equivalent concentration of 100 mg/ml). After 21 days of in vitro culture in the osteogenic medium, the

bone scaffolds were fixed and subjected to SEM analysis. As shown in Fig. 6A, the blank HA/TCP scaffold showed a smooth surface without any deposits. The non-coated control HA/TCP scaffold showed a relatively smooth surface with cells attached (Fig. 6B). On type I collagen-coated HA/TCP scaffold, a minor extent of mineral deposition was observed at the region where the cells attached to the scaffold (Fig. 6C). An apparently eroded HA/TCP surface and obvious mineral deposition was found on the type I þ type II collagen-coated surface of the scaffold (Fig. 6D). Interestingly, a great amount of mineral deposition and crystal-forming structure were recognized on type II collagen-coated surface of the scaffold, where the attached cells were partially calcified (Fig. 6E). In short, type II collagen-coated HA/TCP scaffolds showed a better mineralization potential for the osteogenic differentiation of BMSCs than did the control or the type I collagen-coated scaffolds. Furthermore, mineralization extent on the scaffold coated with a mixture of type I collagen and type II collagen at equal ratio is not better than with type II collagen alone. These data demonstrate the beneficial potential of type II collagen for the immediate improvement of commercial available, ready-made bone scaffold to enhance BMSC calcification through surface modification.

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

2687

Fig. 4. The F-actin filament staining and cell attachment assay of BMSCs to variously coated plates. (A) For F-actin filament staining, BMSCs with or without osteogenic induction were allowed to attach on variously coated plates for 120 min, followed by staining with Alexa FluorÒ 594 conjugated-phalloidin (red) and DAPI (blue). The scale bars represent 20 mm in length. Morphologies of osteogenic medium-induced BMSCs attached on the (C) control (non-coated) plates; (D) type II collagen-coated plates; and those of BMSCs, attached on the type II collagen-coated plates, either pre-treated with (E) anti-a1b1 blocking antibody; (F) anti-a2b1 blocking antibody; (G) anti-a5b1 blocking antibody; and (H) anti-avb3 blocking antibody were presented. Inserted frames represent the zooming-in images of the attached cells in each group. The scale bars represent 100 mm in length. (B) Relative cell spreading area from each group was quantified by Image J software. The statistical significance among the experimental groups was indicated with superscripts of consecutive alphabets. Groups labeled with different superscript letters are significantly different at p < 0.05. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.8. The effect of type II collagen-coated HA/TCP bone substitute on segmental defect repair To evaluate the effect of type II collagen in combination with HA/ TCP substitute on bone repair, a segmental bone defect model was established in rats to study the bone regeneration potency of type II collagen-HA/TCP. Briefly, a 5-mm segmental defect was created at the middle shaft of the right hind femur in each group of SD rats. The agarose-HA/TCP, type I collagen-HA/TCP or type II collagenHATCP substitutes were applied at the created defect site

followed by coverage with a gel foam patch around the repaired region. A non-treated control group with no substitute filled in the defect site was performed as well. The post-surgery X-ray and CT analyses revealed the morphology of the regenerated bone tissue at the defect site of each group (Fig. 7A). The white arrows indicate the surgery sites, and the corresponding regions are revealed in the CT section (white frames) and volume rendering models. As shown in the X-ray and CT image, the fully reunion of the femur shaft was only observed in the type I collagen-HA/TCP and type II collagenHA/TCP groups. Interestingly, among the treated groups, the

2688

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

Fig. 5. Type II collagen-activated signaling cascade in the osteogenic medium-induced BMSCs. (A) The phosphorylation patterns of JNK, ERK and FAK in the osteogenic mediuminduced BMSCs in response to 200 mg/ml of type II collagen were accessed by western blot analyses. (B) The activation of RUNX2 in the osteogenic medium-induced BMSCs upon to the addition of 200 mg/ml of type II collagen was also evaluated at designated intervals. Cell lysates of various experimental groups were immunoprecipitated (IP) with antiphosphoserine antibody, followed by SDS-PAGE analysis and immunoblotting (IB) using anti-RUNX2 antibody. Total RUNX2 and b-actin in the whole cell lysate (INPUT) were immunoblotted with anti-RUNX2 and anti-b-Actin antibodies. (C) The effects of integrin a2b1 blocking antibody on type II collagen-induced FAK, JNK, and ERK1/2 activation in the BMSCs. CII: BMSCs treated with 200 mg/ml of type II collagen only; CII þ anti-a2b1: BMSCs pre-treated with a2b1 blocking antibody followed by type II collagen exposure.

defected femur treated with type II collagen-HA/TCP showed a greater bone mass and callus formation than those with agaroseHA/TCP group and type I collagen-HA/TCP group. Fig. 8 represents the hematoxylin-eosin stained sections of the reunion site of each group. The data indicate that in the non-treated control group and agarose-HA/TCP group, there were mainly cartilage and fibrous tissues presented in the defect sites, which was lacking of regenerated mature bone tissue. However, in the type I collagen-HA/TCP and type II collagen-HA/TCP group, a mature bone tissue was observed at the repairing site. The type I collagen-HA/TCP group exhibited a woven bone morphology (new bone formation) with a region of cartilage-like tissue remained at the reunion gap. However, the type II collagen-HA/TCP group showed dense woven bone tissue and marrow formation at the reunion site with a cartilaginous external callus, which revealed a more matured bone tissue at the reunion site than the type I collagen-HA/TCP group. The functional evaluation of the hind limb motor function of the rats from each group was assayed by footprint analysis. The SFI score of the experimental animals from each group at 7, 14, 21, 28, 35 days post-surgery were presented (Fig. 7B). The data indicate that at day 7, both type I collagen-HA/TCP and type II collagen-HA/ TCP groups showed a significant better improvement on walking function comparing to the non-treated control group and the agarose-HA/TCP group. Among them, type II collagen-HA/TCP group showed a higher average score then did type I collagenHA/TCP during day 7 to day 21. After 28 days of treatment, both group showed equivalent outcomes on SFI scoring and higher scoring than the non-treated control group and the agarose-HA/ TCP group, though not significant. 4. Discussion The importance of type II collagen in the early stage of bone development has been implicated in previous studies [4e6]. Type II

collagen expression in at least a subset of bone-forming cells is sustained over a longer period of time in normal bone development without histological evidence of cartilage formation [9]. It is also suggested that type II collagen might serve as substrates for calcification and present in the region destined to be mineralized [24,25]. These notions offer a clue for the working hypothesis that type II collagen matrix may provide an osteogenesis-inductive microenvironment for BMSCs during the early stage of bone development. The data obtained in this study support this hypothesis. Type II collagen-coated surface significantly accelerated the initiation and enhanced the amount of calcium depositions and RUNX2 expression of BMSCs culture in osteogenic medium (Fig. 2). The underlying mechanism for the osteogenic modulating effect of type II collagen was also investigated. The BMSCs after osteogenic medium induction showed an enhanced cell spreading morphology and Factin distribution on type II collagen-coated surface (Fig. 4). Furthermore, cell attachment assay indicated that the functional blocking antibody of integrin a2b1 inhibits the attachment of the osteogenic medium-induced BMSCs to type II collagen-coated surface (Fig. 4B). These results are in correspondence with the data that BMSCs exhibited an elevated level of integrin a2b1 complex after 4 days of osteogenic induction (Fig. 3A). It is thus concluded that a short-term treatment with the osteogenic medium triggers the BMSCs to partially differentiate and express a relatively high level of integrin a2b1. The up-regulated integrin a2b1 complex facilitates the responsiveness of the partially differentiated BMSCs to type II collagen. The predominant type of integrin complex that associated with the binding of the partially differentiated BMSCs with type II collagen was also evaluated in this study. The data from cell attachment assay indicated that the binding of type II collagen to the osteogenic medium-induced BMSCs was mainly through integrin a2b1 (Fig. 4B). It further implies that integrin a2b1 might predominantly responsible for the

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

2689

Fig. 6. SEM observation of HA/TCP scaffolds coated with/without type I or type II collagen and cultured with/without BMSCs in the osteogenic medium for 21 days. BMSCs were seeded into (B) control (non-coated), (C) type I collagen-coated, (D) mixture of type I and type II collagen (1:1) coated, and (E) type II collagen-coated HA/TCP scaffolds. (A) represents the blank HA/TCP scaffold without BMSCs seeded in. After 21 days of in vitro cultivation in the osteogenic medium, the bone scaffolds were fixed and subjected to SEM analysis. The upper-right inserts show the high-magnification views of the HA/TCP surfaces at the designated frame from each group.

interaction of type II collagen with the partially differentiated BMSCs during the early stage of osteogenic differentiation, and caused the differential responses of BMSCs to type II collagencoated surface before and after the osteogenic induction (Fig. 4A). Type I collagen was reported to trigger the activation of FAK-RafERK signaling, and resulted in the phosphorylation and translocation of Runx2 [26e28]. However, in this study, Type II collagen was shown to activate the FAK-JNK signaling cascade followed by RUNX2 activation in the osteogenic medium-induced BMSCs (Fig. 5). JNK inhibitor was reported not to significantly affect the expression level of RUNX2 in osteoblasts [29]. Thus, the type II collagen induced-JNK activation observed in this study might alter the RUNX2 activity at the post-translational level. This was further confirmed by the observation that RUNX2 phosphorylation was up-

regulated after the serial activation of FAK Y576/Y577 and JNK. Furthermore, the inhibition of integrin a2b1 by the functional blocking antibody eliminated the observed FAK-JNK signaling (Fig. 5C). This observation indicated that the FAK-JNK signaling cascade triggered by the type II collagen is integrin a2b1dependant. It is consequently speculated that integrin a2b1 conduct the upstream activation of the observed FAK-JNK signaling cascade. Through the binding with integrin a2b1, type II collagen triggers the activation of RUNX2 via FAK-JNK signaling, and promotes the subsequent osteogenic differentiation of the BMSCs. A scheme of type II collagen-modulated differentiation of BMSC was proposed in this study. After osteogenic medium induction, the expression of integrin a2b1 in BMSC is up-regulated, which results in the increased response of the cell to type II collagen. This cell

2690

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

Fig. 7. Morphological and functional analysis of the repaired femur of each experimental group. (A) X-ray and CT analysis of the defect sites of rat femurs from each group. The white arrows indicate the defect sites in the femur shaft in the x-ray images. The corresponding regions are revealed in the CT section (white frames) and the volume rendering models. (B) The SFI scoring of animals in each repairing group at 7, 14, 21, 28, 35 day after the surgery. Reference SFI score at “0” represents normal function, while at “100” means complete loss of function. Control: rat femur with defect surgery; Agarose-HA/TCP: rat femur with surgical defect plus implantation of agarose-HA/TCP substitute; CI-HA/TCP: rat femur with surgical defect plus implantation of type I collagen-HA/TCP substitute; CII-HA/TCP: rat femur with surgical defect plus implantation of type II collagen-HA/TCP substitute. The statistical significance among the experimental groups was indicated with superscripts of consecutive alphabets. Groups labeled with different superscript letters are significantly different at p < 0.05.

model represents a transient stage of BMSCs during the early phase of endochondral ossification process. At this stage, the partially differentiated BMSCs express a relatively higher level of integrin a2b1, and increase their responsiveness to the pre-existed cartilaginous matrix. As a result, the interaction of the partially differentiated BMSCs with the cartilaginous matrix triggers the FAK-JNK signaling and RUNX2 activation, thus, promotes the subsequent osteogenic differentiation and terminal calcification of the tissue. In

accordance with the observation of this study, Castano-Izquierdo et al reported that after pre-cultured in the osteogenic medium for 4 days, the induced BMSCs showed a greater ability to regenerate rat cranium with higher extent of bone formation and better union of the defect site, [30] although the mechanism was not defined. The current study further presents a detailed fundamental basis to interpret the results from the previous study. Furthermore, the presence of cartilaginous template during bone development,

Fig. 8. The histological analysis of the repairing defect sites of the rat femur bones. (A) non-treated control group, (B) agarose-HA/TCP group, (C) type I collagen-HA/TCP group, and (D) type II collagen-HA/TCP group. The triangles indicate the fibrous tissue; the asterisks point to the new bone formation regions at the reunion site; and the arrows point to the cartilage-like tissues. (E) The morphology of the dense woven bone with marrow formation at the reunion site of the type II collagen-HA/TCP group; (F) The cartilaginous external callus at the reunion site of the type II collagen-HA/TCP group.

L.-H. Chiu et al. / Biomaterials 35 (2014) 2680e2691

which involves an endochondral ossification-like process, also exists during fracture healing [31]. The present study provides a molecular mechanism for the preformation of the cartilaginous tissue during the fracture healing process. As demonstrated in this study, the implanted type II collagen-HA/TCP stimulates more significant callus formation at the defect site comparing to that does the other groups. Interestingly, the histological data showed that the implantation with type II collagen-HA/TCP substitute, rather than the type I collagen-HA/TCP substitute, resulted in a more dense woven bone tissue with marrow formation at the reunion site, and significant cartilaginous external callus. This implies that the new bone is forming by an endochondral ossificationlike process. 5. Conclusion A distinct role of type II collagen in modulating BMSC osteogenic differentiation via integrin a2b1-FAK-JNK signaling was demonstrated. Type II collagen facilitates osteogenesis of BMSCs better than did type I collagen. The results provide an insight into the role of and molecular mechanism for BMSC-cartilaginous matrix interaction in the early stage bone development. Furthermore, a type II collagen-coated HA/TCP bone scaffold was also demonstrated to have the potential of expediting bone repair in a segmental defect rat model. Based on this established conception, new therapeutic approaches may be developed by applying type II collagen for faster and better bone regeneration in future clinical practice. Acknowledgments This study was supported by the National Science Council of Taiwan via grant funding: NSC98-3112-B-038-002 (YHT),NSC993112-B-038-001 (YHT), NSC 100-2314-B-038-013 (YHT) and NSC 100-2120-M-038-002 (WFL, YHT). This study was also supported in part by Core Facility Center, Office of Research and Development at Taipei Medical University and the Molecular and Genetic Imaging Core/National Research Program for Genomic Medicine at National Yang-Ming University. References [1] Ornitz DM, Marie PJ. FGF signaling pathways in endochondral and intramembranous bone development and human genetic disease. Genes Dev 2002;16:1446e65. [2] Prall WC, Haasters F, Heggebo J, Polzer H, Schwarz C, Gassner C, et al. Mesenchymal stem cells from osteoporotic patients feature impaired signal transduction but sustained osteoinduction in response to BMP-2 stimulation. Biochem Biophys Res Commun 2013;440:617e22. [3] Kronenberg HM. Developmental regulation of the growth plate. Nature 2003;423:332e6. [4] Hiltunen A, Metsaranta M, Virolainen P, Aro HT, Vuorio E. Retarded chondrogenesis in transgenic mice with a type II collagen defect results in fracture healing abnormalities. Dev Dyn 1994;200:340e9. [5] Vandenberg P, Khillan JS, Prockop DJ, Helminen H, Kontusaari S, Ala-Kokko L. Expression of a partially deleted gene of human type II procollagen (COL2A1) in transgenic mice produces a chondrodysplasia. Proc Natl Acad Sci U S A 1991;88:7640e4. [6] Li SW, Prockop DJ, Helminen H, Fassler R, Lapvetelainen T, Kiraly K, et al. Transgenic mice with targeted inactivation of the Col2 alpha 1 gene for collagen II develop a skeleton with membranous and periosteal bone but no endochondral bone. Genes Dev 1995;9:2821e30. [7] Farrell E, van der Jagt OP, Koevoet W, Kops N, van Manen CJ, Hellingman CA, et al. Chondrogenic priming of human bone marrow stromal cells: a better route to bone repair? Tissue Eng Part C Methods 2009;15:285e95.

2691

[8] Janicki P, Kasten P, Kleinschmidt K, Luginbuehl R, Richter W. Chondrogenic pre-induction of human mesenchymal stem cells on beta-TCP: enhanced bone quality by endochondral heterotopic bone formation. Acta Biomater 2010;6: 3292e301. [9] Szabova L, Yamada SS, Wimer H, Chrysovergis K, Ingvarsen S, Behrendt N, et al. MT1-MMP and type II collagen specify skeletal stem cells and their bone and cartilage progeny. J Bone Miner Res 2009;24:1905e16. [10] Chiu LH, Yeh TS, Huang HM, Leu SJ, Yang CB, Tsai YH. Diverse effects of type II collagen on osteogenic and adipogenic differentiation of mesenchymal stem cells. J Cell Physiol 2012;227:2412e20. [11] Warstat K, Meckbach D, Weis-Klemm M, Hack A, Klein G, de Zwart P, et al. TGF-beta enhances the integrin alpha2beta1-mediated attachment of mesenchymal stem cells to type I collagen. Stem Cells Dev 2010;19:645e56. [12] Mizuno M, Kuboki Y. Osteoblast-related gene expression of bone marrow cells during the osteoblastic differentiation induced by type I collagen. J Biochem 2001;129:133e8. [13] Salasznyk RM, Klees RF, Hughlock MK, Plopper GE. ERK signaling pathways regulate the osteogenic differentiation of human mesenchymal stem cells on collagen I and vitronectin. Cell Commun Adhes 2004;11:137e53. [14] Huang Z, Tian J, Yu B, Xu Y, Feng Q. A bone-like nano-hydroxyapatite/collagen loaded injectable scaffold. Biomed Mater 2009;4:055005. [15] Maehara H, Sotome S, Yoshii T, Torigoe I, Kawasaki Y, Sugata Y, et al. Repair of large osteochondral defects in rabbits using porous hydroxyapatite/ collagen (HAp/Col) and fibroblast growth factor-2 (FGF-2). J Orthop Res 2010;28:677e86. [16] Masuda T, Kawai T, Anada T, Kamakura S, Suzuki O. Quality of regenerated bone enhanced by implantation of octacalcium phosphate-collagen composite. Tissue Eng Part C Methods 2010;16:471e8. [17] Carlberg AL, Pucci B, Rallapalli R, Tuan RS, Hall DJ. Efficient chondrogenic differentiation of mesenchymal cells in micromass culture by retroviral gene transfer of BMP-2. Differentiation 2001;67:128e38. [18] Denker AE, Haas AR, Nicoll SB, Tuan RS. Chondrogenic differentiation of murine C3H10T1/2 multipotential mesenchymal cells: I. Stimulation by bone morphogenetic protein-2 in high-density micromass cultures. Differentiation 1999;64:67e76. [19] Ao Q, Fung CK, Tsui AY, Cai S, Zuo HC, Chan YS, et al. The regeneration of transected sciatic nerves of adult rats using chitosan nerve conduits seeded with bone marrow stromal cell-derived Schwann cells. Biomaterials 2011;32: 787e96. [20] Sarikcioglu L, Demirel BM, Utuk A. Walking track analysis: an assessment method for functional recovery after sciatic nerve injury in the rat. Folia Morphol (Warsz) 2009;68:1e7. [21] Stanford CM, Jacobson PA, Eanes ED, Lembke LA, Midura RJ. Rapidly forming apatitic mineral in an osteoblastic cell line (UMR 106-01 BSP). J Biol Chem 1995;270:9420e8. [22] Tuckwell DS, Ayad S, Grant ME, Takigawa M, Humphries MJ. Conformation dependence of integrin-type II collagen binding. Inability of collagen peptides to support alpha 2 beta 1 binding, and mediation of adhesion to denatured collagen by a novel alpha 5 beta 1-fibronectin bridge. J Cell Sci 1994;107(Pt 4): 993e1005. [23] Schneider GB, Zaharias R, Stanford C. Osteoblast integrin adhesion and signaling regulate mineralization. J Dent Res 2001;80:1540e4. [24] Kirsch T, von der Mark K. Isolation of human type X collagen and immunolocalization in fetal human cartilage. Eur J Biochem 1991;196:575e80. [25] von der Mark K, von der Mark H. The role of three genetically distinct collagen types in endochondral ossification and calcification of cartilage. J Bone Joint Surg Br 1977;59-B:458e64. [26] Takeuchi Y, Suzawa M, Kikuchi T, Nishida E, Fujita T, Matsumoto T. Differentiation and transforming growth factor-beta receptor down-regulation by collagen-alpha2beta1 integrin interaction is mediated by focal adhesion kinase and its downstream signals in murine osteoblastic cells. J Biol Chem 1997;272:29309e16. [27] Schlaepfer DD, Hunter T. Evidence for in vivo phosphorylation of the Grb2 SH2-domain binding site on focal adhesion kinase by Src-family proteintyrosine kinases. Mol Cell Biol 1996;16:5623e33. [28] Egan SE, Giddings BW, Brooks MW, Buday L, Sizeland AM, Weinberg RA. Association of Sos Ras exchange protein with Grb2 is implicated in tyrosine kinase signal transduction and transformation. Nature 1993;363:45e51. [29] Matsuguchi T, Chiba N, Bandow K, Kakimoto K, Masuda A, Ohnishi T. JNK activity is essential for Atf4 expression and late-stage osteoblast differentiation. J Bone Miner Res 2009;24:398e410. [30] Castano-Izquierdo H, Alvarez-Barreto J, van den Dolder J, Jansen JA, Mikos AG, Sikavitsas VI. Pre-culture period of mesenchymal stem cells in osteogenic media influences their in vivo bone forming potential. J Biomed Mater Res A 2007;82:129e38. [31] Einhorn TA. The cell and molecular biology of fracture healing. Clin Orthop Relat Res 1998:S7e21.