lable at ScienceDirect

Biomaterials 106 (2016) 205e216

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Sequential and sustained release of SDF-1 and BMP-2 from silkfibroin-nanohydroxyapatite scaffold for the enhancement of boneregeneration

Xiaofeng Shen a, 1, Yanxia Zhang b, 1, Yong Gu a, Yun Xu a, Yong Liu a, Bin Li a, c,Liang Chen a, *

a Department of Orthopaedic Surgery, The First Affiliated Hospital of Soochow University, Suzhou, Jiangsu, 215006, PR Chinab Institute for Cardiovascular Science & Department of Cardiovascular Surgery of the First Affiliated Hospital, Soochow University, Suzhou, Jiangsu, 215007,PR Chinac Orthopedic Institute, Soochow University, Suzhou, Jiangsu, 215007, PR China

a r t i c l e i n f o

Article history:Received 5 June 2016Received in revised form14 August 2016Accepted 15 August 2016Available online 17 August 2016

Keywords:Bone regenerationControlled releaseStromal cell derived factor-1Bone morphogenetic protein-2Silk fibroinNano-hydroxyapatite

* Corresponding author.E-mail address: chenliang1972@sina.com (L. Chen

1 These two authors contributed equally to this wo

http://dx.doi.org/10.1016/j.biomaterials.2016.08.0230142-9612/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

In this study, a cell-free bone tissue engineering system based on a silk fibroin (SF)/nano-hydroxyapatite(nHAp) scaffold was developed, in which two bioactive molecules, stromal cell derived factor-1 (SDF-1)and bone morphogenetic protein-2 (BMP-2), were embedded and released in a sequential and controlledmanner to facilitate cell recruitment and bone formation, respectively. BMP-2 was initially loaded into SFmicrospheres, and these BMP-2 containing microspheres were subsequently encapsulated into the SF/nHAp scaffolds, which were successively functionalized with SDF-1 via physical adsorption. The resultsindicated rapid initial release of SDF-1 during the first few days, followed by slow and sustained releaseof BMP-2 for as long as three weeks. The composite scaffold significantly promoted the recruitment ofbone marrow mesenchymal stem cells (BMSCs) and osteogenic differentiation of them in vitro. Further,the in vivo studies using D-Luciferin-labeled BMSCs indicated that implantation of this composite scaf-fold markedly promoted the recruitment of BMSCs to the implanted sites. Enhanced bone regenerationwas identified at 12 weeks' post-implantation. Taken together, our findings suggested that the sequentialand sustained release of SDF-1 and BMP-2 from the SF/nHAp scaffolds resulted in a synergistic effect onbone regeneration. Such a composite system, therefore, shows promising potential for cell-free bonetissue engineering applications.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Large bone defects associated with trauma, tumor, and infectionfrequently require surgical intervention [1,2]. Currently, the trans-plantation of allogeneic and autologous bones remains the main-stay of treatment for bone defects. However, the risks of infection,rejection, and donor site morbidity and the limited supply preventits clinical application and remain serious problems that must beresolved [1,3,4]. Bone tissue engineering, which is promisingalternative strategy for inducing tissue regeneration, has thusattracted considerable attention [4,5]. Conventional tissue

).rk.

engineering involves the use of a combination of biomaterials,appropriate bioactive molecules, and cells, which may suffer fromthe problems related to long-term in vitro culture for the formationof new bones, such as the high cost of cell isolation, proliferationandmaturation, as well as the risk of pathogens [6]. Alternatively, insitu bone tissue engineering provides a better solution by takingadvantage of the in vivo microenvironment, in which the cell-free,bone-specific biomaterials alone or in combination with bioactivemolecules are directly implanted in the defect sites to recruit hostcirculating stem cells in situ, which thus aims to guide the osteo-genic differentiation of mesenchymal stem cells (MSCs) for boneregeneration [7,8].

Biomaterial comprises a key element in bone tissue engineeringand provides a three-dimensional (3D) scaffold to support cellgrowth and the formation of the extracellular matrix (ECM) [9]. Silkfibroin (SF), a biologically derived protein harvested from

X. Shen et al. / Biomaterials 106 (2016) 205e216206

domesticated silk worm (Bombyx mori) cocoons, has been recog-nized as a favorable scaffold material for bone tissue engineeringbecause of its unique mechanical properties, tunable biodegrada-tion rate, good preservation of bioactive molecule activity and theability to support the osteogenic differentiation of MSCs [10,11].Moreover, hydroxyapatite (HAp) is chemically similar to the naturalbone and exhibits excellent osteoconductivity [12,13]. It has beendemonstrated that the incorporation of nanoparticle HAp (nHAp)in SF scaffolds may promote bone regeneration through theosteoinductive effects of nHAp on bonemarrowmesenchymal stemcells (BMSCs) [12,13].

In addition to biomaterials, bioactive molecule comprisesanother key element in bone tissue engineering to endow thefunctionality of the biomaterial to stimulate the recruitment andosteogenic differentiation of BMSCs for bone regeneration [9].Stromal cell derived factor-1 (SDF-1 or CXCL12) is a member of theCXC chemokine family; the recognition of SDF-1 by its main re-ceptor, CXCR4, triggers the migration of CXCR4-positive stem cellsand progenitor cells to injury sites during the acute phase of bonerepair and participates in regeneration [14,15]. More importantly, ithas been reported that SDF-1 signaling contributed to the osteo-genic differentiation of mesenchymal C2C12 cells induced by bonemorphogenetic protein-2 (BMP-2) [14]. BMP-2 comprises animportant osteoinductive growth factor, which is approved by theUS Food and Drug Administration (FDA) for clinical use to inducebone formation via the enhancement of osteoblast progenitor cellrecruitment, angiogenesis, and the stimulation of the osteogenicdifferentiation of MSC [16e18]. Taking advantage of the uniqueproperties of SDF-1 and BMP-2, several groups have attempted toutilize their combination for bone regeneration [19,20]. However,the concomitant use of the BMP-2 and SDF-1 mixture had no ad-ditive effects on osteoblastic differentiation, cell migration orangiogenesis [20]. In Hwang's work, the sequential treatment ofcollagen scaffolds with BMP-2 following SDF-1 demonstrated thestrongest degree of new bone regeneration compared with thesimultaneous treatment with BMP-2 and SDF-1 [21]. This phe-nomenon may be explained by the enhancement of stem cellmigration in the early stage by SDF-1 and the activation of osteo-blastic differentiation in the later stages by BMP-2, considering thatSDF-1 is not required following the activation of osteogenic in-duction [19,21]. It should be noted that in Hwang's work, BMP-2was applied to the implanted scaffolds via periodic percutaneousinjections, which is not appropriate for some fracture sites in deeplocations of the body (e.g., lumbar vertebra). In addition, because ofthe uncontrolled and offsite release of BMP-2, high doses of BMP-2must be used to compensate for the short half-life of BMP-2 in vivoto achieve effective clinical bone regeneration [18], which maycause side effects, such as heterotopic bone formation, edema orinflammatory reactions [22e24]. It is suggested that this problemmay be solved by the encapsulation of BMP-2 into microspheresprior to incorporation into a polymeric scaffold [25]. For example,SF microspheres have been demonstrated to be efficient in deliv-ering BMP-2 with the controlled release kinetics for MSC chon-drogenesis [25e27].

In this work, we developed a novel scaffold for effective boneregeneration with the capability to sequentially release SDF-1, fol-lowed by the release of BMP-2 in a controlled manner. A compositeof SF and nHAp porous scaffold was used to incorporate SF micro-spheres encapsulated with BMP-2, and SDF-1 was then physicallyadsorbed onto the scaffold to construct the cell-free scaffold. The SFmicrospheres were prepared via laminar jet break-up technology toendow the high encapsulation efficiency of BMP-2 [28]. With thecombination of physically adsorbed SDF-1 and microspheresencapsulated with BMP-2 in the SF/nHAp scaffold, we expectedenhanced synergistic effects for bone regeneration, in which the

initial release of SDF-1 facilitates the recruitment of MSCs, and thecontrolled release of BMP-2 subsequently induces the osteogenicdifferentiation of MSCs. The effects of the sequential and sustainedrelease of SDF-1 and BMP-2 on the migratory and osteogenic ca-pacity of MSCs and the bone regeneration were assessed bydifferent and complementary assays.

2. Materials and methods

2.1. Preparation of scaffolds that contain both SDF-1 and BMP-2

2.1.1. Preparation of SF solutionsAqueous SF solutions from cocoons of B. mori (RudongXinsilu

Co., Ltd., Jiangsu, China) were prepared as previously described[13]. Briefly, the cocoons were boiled for 1 h in an aqueous solutionof Na2CO3 (0.5% w/w) and subsequently rinsed with water toremove the sericin. The extracted SF was solubilized in 9.3 M LiBr(Strem Chemicals Inc., MA, USA) solution at 60 �C for 4 h to producea 20% (w/v) solution. This solution was dialyzed against water for 4d with 8 times water changes in between. Finally, the purified SFsolutionwas dialyzed against polyethylene glycol 20,000 (Biosharp,Shanghai, China) powder to obtain a concentrated SF of 15.0%. Thesolutionwas filtered through a syringe filter and stored at 4 �C priorto use.

2.1.2. Preparation of SF microspheresThe SF microspheres were fabricated using the laminar jet

break-upmethod as previously describedwith a slightmodification[28]. A syringe connected to a nozzle with a diameter of 200 mmwas fixed on the micro-injection pump (LSP02-1B, Baoding LongerPrecision Pump Co., Ltd., China). The SF solution with concentra-tions at 2%, 4% and 6% (w/v) in the injector underwent jet break-upinto droplets under a high voltage of 10 kV. The produced pelletswere collected in a liquid nitrogen bath. The frozen spheres werefreeze dried (Alpha 1e2 LD plus, Christ, Germany) for 2 d. Thefreeze-dried microspheres were subsequently treated with meth-anol at room temperature for 30 min to induce SF structuraltransformation. Finally, the insoluble SF microsphere in water wasobtained after being lyophilized again. The SF concentration at 6%was used in the following experiment. To produce BMP-2 (orrhBMP-2, Shanghai Rebone Biomaterials Co. Ltd., China) loaded SFspheres, the growth factor was added to the aqueous SF solutionwith a mass ratio of BMP-2 to SF at 1:50 (w/w).

2.1.3. Fabrication of SF based composite scaffoldSF microspheres loaded with BMP-2 were incorporated into a

mixture of nHAp (50e100 nm; Aladdin, Shanghai, China) and SFsuspension to prepare the composite scaffolds via freeze drying;SDF-1 (or rhSDF-1a, Peprotech, USA) was then physically adsorbedonto the surface of the scaffold. First, nHAp was dispersed inphosphate buffer solution (PBS, 0.001 M, pH 6.8) by sonication for1 min to achieve a suspension concentration at 20% (w/v). ThenHAp and SF solution with a mass ratio of nHAp to SF at 1:20 (SI,supporting information, Fig. S1) was subsequently prepared viathe addition of the desired amount of nHAp/PBS into SF solution(6%, w/v) under sonication for 1 min and magnetic stirring for30 min. SF microspheres loaded with BMP-2 were suspended intothe mixture of SF/nHAp with the mass ratio of microspheres tonHAp at 2:3. The prepared composite suspension (50 mL) was thentransferred to a mold plate (5 mm in diameter and 3 mm in height)and frozen in liquid nitrogen for freeze drying. The freeze-driedscaffold was treated with methanol for 30 min to induce SF struc-tural transformation to develop a cylindrical insoluble SF scaffoldthat contained SF microspheres loaded with BMP-2 (BMP-2 (E)).The SF/nHAp scaffold was prepared and used as a control. The

X. Shen et al. / Biomaterials 106 (2016) 205e216 207

scaffold adsorbed BMP-2 directly without encapsulation into mi-crospheres (BMP-2 (P)) was prepared by dropping BMP-2 (2.0 mg in10 mL PBS) onto the control scaffold.

For the SDF-1 decorated scaffold, 400 ng SDF-1 in 10 mL PBS wasdropped onto the control scaffold, BMP-2 (P) scaffold and BMP-2 (E)scaffold to prepare the SDF-1 scaffold, SDF-1þBMP-2 (P) (S þ B (P))scaffold, and SDF-1þBMP-2 (E) (Sþ B (E)) scaffold, respectively. Thescaffold was subsequently stored at �20 �C and sterilized with ul-traviolet light (l ¼ 254 nm) for 4 h prior to use.

The abbreviations and the preparation of the scaffolds used inthis work were summarized in Table 1.

2.2. Characterization of microspheres and scaffolds

2.2.1. Scanning electron microscopyTo observe the morphology of the SF microspheres and scaf-

folds, the samples were examined using a scanning electron mi-croscope (SEM, S-4800; Hitachi, Kotyo, Japan) at an acceleratingvoltage of 10 kV. Prior to characterization, the samples were addeddirectly on top of conductive tapes mounted on the SEM samplestubs and sputter coated with gold for 60 s using gold sputtercoating equipment (SC7620, Quorum Technologies, UK).

2.2.2. Fourier transform infrared spectroscopyThe structure of the scaffold was analyzed via Fourier transform

infrared spectroscopy (FTIR, Nicolet 6700; Thermo scientific, USA).For each measurement, 128 scans were obtained with a resolutionof 4 cm�1, with wavelengths that ranged from 650 to 4000 cm�1.

2.3. Determination of SDF-1 and BMP-2 release from scaffold

The release of SDF-1 and BMP-2 from the scaffold in vitro wasquantified via ELISA. First, the scaffolds were incubated in 2 mL PBSwith vigorous shaking at 37 �C. At the desired time points (3 h, 12 h,24 h, 4 d, 7 d, 10 d, 13 d, 19 d, 25 d, 31 d and 34 d), PBS was collectedand stored at �20 �C prior to use, and the scaffold was incubated in2 mL fresh PBS again. Finally, the SDF-1 and BMP-2 concentrationsin the collected PBS were quantified with a SDF-1 and BMP-2 ELISAkit (R&D, USA) according to the manufacturer's instructions. Thecumulative release ratio was calculated as the ratio of the cumu-lativemass of SDF-1 or BMP-2 released at each time interval to theirinitial input amount in the scaffold, which is 2.0 mg and 400 ng forBMP-2 and SDF-1, respectively.

2.4. Migratory response of BMSCs to SDF-1 modified scaffold

To evaluate the recruitment capacity of SDF-1 on BMSCs, theTranswell chemotactic migration model was used. Briefly,1 � 104 cells were seeded in the upper chamber of a 24-wellTranswell plate (pore size: 8 mm, Corning, USA), and scaffoldswere placed in the lower chamber. After 24 h, the upper surface ofthe Transwell membrane was initially scraped with a cotton swabto remove the adherent cells and subsequently detached from theinserts. The cells that migrated to the lower side of the membrane

Table 1Sample abbreviations used in this work.

Name of sample Description of sample

Control Original scaffold of SF/nSDF-1 Scaffold with physicallyBMP-2 (P) Scaffold with physicallyBMP-2 (E) Scaffold with BMP-2 encS þ B (P) Scaffold with physicallyS þ B (E) Scaffold with physically

were fixed with 4% paraformaldehyde (Biosharp, Shanghai, China)for 30 min and stained with 0.1% crystal violet (Solarbio, Beijing,China) for 10 min. The number of cells on the lower surface of themembrane was counted in five random high power (200�)microscopic fields in each well.

2.5. Cell adhesion and proliferation on scaffold

Cylinder scaffolds were used for the cell adhesion studies. Priorto the cell culture work, the scaffolds were sterilized under UV,followed by immersion in OriCell™ SD Rat MSC Growth Medium(Cyagen, Guangzhou, China) overnight. BMSCs were seeded intoscaffolds in 96-well culture plate at a density of 2.0� 103 cells/well.The samples were placed in an incubator at 37 �C, 95% relativehumidity and 5% CO2 partial pressure. The culture medium wasreplaced every second day.

The morphology of the cells seeded on the cultured scaffoldsurface was examined via SEM. The cells were fixed in 4% para-formaldehyde at 4 �C for 1 h, followed by rinsing with deionizedwater three times. The samples were subsequently dehydratedthrough a gradient series of ethanol/water solution (10%, 20%, 35%,50%, 70%, 85%, and 100%). Finally, the sample was dried using theCO2 critical-point drying method and sputter coated with gold. Theelectron microscope was operated at 5 kV to image the samples.

Cell proliferation was measured using cell counting kit-8 re-agent (CCK-8, Dojindo, Kumamoto, Japan). At the desired timespoints (1 d, 3 d, 5 d, 7 d and 9 d), the BMSC-seeded scaffolds wereincubated in CCK-8 solution at 37 �C in an incubator with 5% CO2 for2 h. The solution obtained was measured at 450 nm using aspectrophotometer.

2.6. In vitro osteogenesis study

The alkaline phosphatase (ALP) activity, calcium content andexpression of osteocalcin (OCN) proteins and genes related toosteogenesis were measured to assess the osteogenic differentia-tion of BMSCs. Prior to the analyses, all scaffolds were sterilizedwith UV. Osteogenic induction medium (OM, Cyagen, Guangzhou,China) was used to evaluate the ALP activity, calcium content andgene expressions. The scaffolds were placed in the upper chamberof a 6-well Transwell plate (pore size: 8 mm, Corning, USA), and3 � 105 cells were cultured in the lower chamber.

2.6.1. Alkaline phosphatase activityThe activity of ALP, a widely exploited early biochemical marker

for osteogenic activity, was measured. After 7 d of culture, the cellswere fixed in 4% paraformaldehyde, followed by a reactionwith ALPstaining solution (Yeasen, Shanghai, China) for 20 min. The ALP-positive cells were stained blue and visualized with microscopicfields.

2.6.2. Calcium contentMineralization was evaluated by quantifying the formation of

calcium phosphate by cells using Alizarin Red S (ARS) staining

HAP without any biomoleculesadsorbed SDF-1adsorbed BMP-2apsulated into SF microspheresadsorbed SDF-1 and BMP-2adsorbed SDF-1 and BMP-2 encapsulated into SF microspheres

X. Shen et al. / Biomaterials 106 (2016) 205e216208

(Cyagen, Guangzhou, China). After 21 d of culture, the cells werewashed with PBS and fixed with 4% paraformaldehyde for 30 min,followed by staining with a 40 mM ARS solution for 15 min; thecells were subsequently washed with sufficient deionized water.The cells were visualized in microscopic fields.

2.6.3. OCN immunofluorescence stainingOCN proteins, which are commonly used as late markers of

osteogenic differentiation, were examined using immunofluores-cence staining. After 21 d, the cells were fixed with 4% para-formaldehyde, treated with 0.5% Triton X-100 (Sigma-aldrich, USA)and blocked with 2% BSA (Solarbio, Beijing, China). After washingwith PBS, the cells were stained against OCN primary antibody(Abcam, USA) at 10 mg/ml at 4 �C overnight. The cells were subse-quently incubated in the secondary antibody of goat anti-mouseIgG H&L (1:500, Cy3®; Abcam, USA) for 1 h at 37 �C, and thenuclei were stained with DAPI.

2.6.4. Expression of osteogenic genesThe expression of genes related to osteogenesis, including Alp,

Runt-related factor-2 (Runx2), Osteopontin (Opn), and Ocn wasanalyzed by quantitative reverse transcription polymerase chainreaction (RT-qPCR). The RT-qPCR assay was performed as previ-ously described to evaluate the expression levels of the genes ofBMSCs cultured with the scaffold at 3, 7 and 14 d [29,30]. Theprimer sequences used for PCR amplification are listed in Table 2.Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as aninternal control, and all primers were synthesized by Invitrogen.

2.7. In vivo bone formation study

Sprague-Dawley (SD) male rats were purchased from theExperimental Animal Center of Soochow University (Suzhou,China). The animal handling and surgical procedures were con-ducted in accordance with protocols approved by the Ethics Com-mittee at the First Affiliated Hospital of Soochow University.

2.7.1. Rat calvarial critical size defect model and scaffoldimplantation

The SD rats were randomly assigned to groups, including con-trol, scaffold of SDF-1, BMP-2 (P), BMP-2 (E), S þ B (P) and S þ B (E),to evaluate the osteogenic potential in a cranial defect on the im-plantation of scaffold with or without BMP-2 and/or SDF-1, as wellas the difference between the sequential release and the concom-itant release of BMP-2 and SDF-1 from the scaffold. The scaffoldswere implanted in defects (5 mm in diameter, one defect at thecenter of parietal bone for bioluminescence imaging (BLI) and twobilateral defects for micro computed tomography (mCT) imagingand histological analysis) in the calvarium of the SD, which wereprepared with a dental trephine drill.

Table 2Primers used for RT-qPCR.

Gene Primer/probe Sequence Ann. temp Tm (�C)

Alp Forward CGTCTCCATGGTGGATTATGCT 64.5Reverse CCCAGGCACAGTGGTCAAG

Runx2 Forward TCTTCCCAAAGCCAGAGCG 64.5Reverse TGCCATTCGAGGTGGTCG

Opn Forward GAGGAGAAGGCGCATTACAG 57Reverse AAACGTCTGCTTGTCTGCTG

Ocn Forward CATGAAGGCTTTGTCAGACT 57Reverse CTCTCTCTGCTCACTCTGCT

Gapdh Forward GGCAAGTTCAACGGCACAGT 57 and 64.5Reverse GCCAGTAGACTCCACGACAT

2.7.2. BMSC recruitment in vivoIn vivo BLI was performed using the IVIS lumina series III im-

aging system (Caliper Life Science, Hopkinton, MA). A total of 18 SDrats (average weight of 120 g) were divided into six groups andwere anesthetized using 2% isoflurane inhalation 3 days after thecreation of a calvarial defect model. Each rat was injected with1.0 � 106 BMSCs labeled with the firefly luciferase (Fluc) reportergene (Synchem, Shanghai, China) through the tail vein. After in-jection at 30 min, 1 d, 3 d, 7 d and 14 d whole-body images of eachrat were obtained. Imaging was conducted 20 min after the intra-peritoneal injection of the reporter probe of D-luciferin (Synchem,Shanghai, China). The BLI signals at the standardized region of in-terest (ROI) were investigated.

2.7.3. Micro-computed tomography analysisAfter 8 and 12 weeks post-surgery, a total of 72 SD rats (average

weight of 300 g) were sacrificed using CO2 suffocation, and thecalvarial specimen was harvested and fixed in 10% formalin for mCTimaging and histological analysis. The three dimensional (3D)structures of the regenerated bone tissue within the cranial defectarea were evaluated with mCT (SkyScan 1176, SkyScan, Aartselaar,Belgium) with the following settings: 65 kV, 385 mA, and 1 mm Alfilter. 3D reconstruction was performed with system software. Acylinder ROI of 4.8 mm in diameter was used for the bone volumefraction (bone volume (BV)/tissue volume (TV), means ± standarddeviations of 6 rats).

2.7.4. Histological procedureAt 12 weeks post-implantation, the harvest specimens were

fixed in 4% paraformaldehyde for 48 h and decalcified in 10% formicacid for 1 week at room temperature. The samples were dehydratedthrough an alcohol gradient and embedded in paraffin blocks. Five-micron-thick histological sections were cut at the center of theembedded specimens, followed by staining with hematoxylin andeosin (H&E) to assess the bone formation area. Images were ob-tained under a bright field microscope (Zeiss Axiovert 200, CarlZeiss Inc., Thornwood, NY, USA).

2.8. Statistical analysis

Experiments were performed in triplicate unless otherwiseindicated. The data were expressed as the means ± standard de-viations. Statistical analysis (GraphPad Software, Inc.; USA) wasevaluated using one-way analysis of variance (ANOVA) followed byTukey's multiple comparison test to evaluate the differences be-tween the groups. Differences at p < 0.01 and p < 0.05 wereconsidered statistically significant.

3. Results

3.1. Characterization of composite scaffolds

The SF/nHAp based scaffold impregnated with BMP-2 and/orSDF-1 was prepared as illustrated in Fig. 1A. First, SF microspheresloaded with BMP-2 were fabricated using laminar jet break-uptechnology. The diameter of the spheres was controlled byadjusting the SF concentration. The microspheres were subse-quently mixed with a suspension of SF and nHAp to prepare the SF/nHAp scaffolds loaded with BMP-2 (BMP-2 (E)); SF/nHAp scaffoldswere prepared as the control. Finally, the SDF-1 was physicallyadsorbed onto the control, BMP-2 (E) by dropping SDF-1 solutiononto the scaffold to obtain scaffolds of SDF-1 and S þ B (E),respectively. The difference between the sequential release andconcomitant release of SDF-1 and BMP-2 on the effect of boneregeneration was also investigated by preparing the scaffold of

Fig. 1. (A) Schematic illustration of the preparation of the S þ B (E) scaffold; the sequential release of SDF-1 and BMP-2 from scaffold facilitate BMSC homing, proliferation andosteogenic differentiation. SEM images of (B) untreated and (C) methanol-treated SF microspheres. Scale bar ¼ 100 mm; scale bar insert in B is 30 mm and C is 20 mm. (D) FTIR spectraof spheres and scaffold, and cross-section of SF scaffolds (E) without and (F) with SF microspheres loaded with BMP-2. Scale bar ¼ 500 mm; scale bar inset is 100 mm.

X. Shen et al. / Biomaterials 106 (2016) 205e216 209

S þ B (P), in which BMP-2 was directly adsorbed onto the controlscaffold instead of encapsulated into microspheres.

The influences of the SF concentration and methanol treatmenton the size and morphology of the microspheres were investigated.The diameter increased with increasing concentration. The sphereshad average diameters of 189 ± 32, 201 ± 34, and 227 ± 40 mm forthe SF concentrations at 2%, 4% and 6%, respectively. The SF con-centration at 6% was used in the following experiment. Methanoltreatment has beenwidely used to make SF microspheres insolublevia conformational transitions of SF from random coil to b-sheet[31,32]. As shown in Fig. 1B and C, the methanol treatment resultedin a decrease in the diameter of SF microspheres from 227 ± 40 to112 ± 28 mm, which is consistent with the results reported by othergroups [28]. Moreover, after treatment, the microspheres exhibiteda porous structure and clearly rougher surface compared with thenon-treated microspheres. A cross-section of these spheres indi-cated a porous structure, as shown in Fig. 1F. There are no obviouschanges in the morphology or size of the microspheres with theloading of BMP-2 (SI, Fig. S2). Microspheres loaded with BMP-2were incorporated into the suspension of SF and nHAp during thescaffold fabrication. The FTIR spectra of pure SF, SF/nHAp and SF/nHAp scaffold loaded with microspheres that contained BMP-2 areshown in Fig. 1D. The peaks at 1650 cm�1 (amide I) and 1516 cm�1

(amide II) correspond to the SF. The existence of nHAp in the SFscaffold is confirmed by the appearance of a peak at 1025 cm�1,which is ascribed to the PO4

3� characteristic peaks, and its intensitydecreased with the incorporation of the microsphere loaded withBMP-2. The SEM characterization of the scaffold (Fig. 1E and F)indicated that the microspheres are integrated and homogeneously

distributed within the scaffold (Fig. 1F). The average pore size of theSF scaffolds is 124 ± 42 mm, which decreased to 85 ± 25 mmfollowing microsphere incorporation.

3.2. In vitro release of SDF-1 and BMP-2

The release profiles of SDF-1 and BMP-2 from the SF/nHAPscaffolds loaded with SDF-1 and/or BMP-2 in vitro were investi-gated using the corresponding ELISA kits, and the results are pre-sented in Fig. 2. There was a burst release with 47% of the total SDF-1 released from the scaffolds in the first day; the release was sub-sequently slowed down with approximately 85% of the total SDF-1released after 7 d. In contrast, the overall release of BMP-2 from thescaffold was comparatively slow and sustained compared withSDF-1; a less initial burst release with 21% of the total BMP-2released was identified for the first day, and the BMP-2 wasreleased for as long as three weeks. At 10 d, nearly 90% of the SDF-1was released from the scaffold, whereas the remaining percent ofBMP-2 was more than 50%, indicating the sequential releasebehavior of the two bioactive molecules from S þ B (E) scaffold(Fig. 2A). For the scaffolds of SDF-1 and BMP-2 (E), irrespective ofthe combination, SDF-1 or BMP-2 exhibited nearly the same releasebehavior as that for the scaffold of S þ B (E) (SI, Fig. S3). While forthe scaffold of S þ B (P) inwhich BMP-2 was directly adsorbed ontoscaffold without using microspheres, the release profile of BMP-2was found to be similar to that of SDF-1. A burst release with 54%of BMP-2 released from the S þ B (P) scaffold was found in the firstday; at 10 d, 85% BMP-2 was already released from the S þ B (P)scaffold (Fig. 2B), indicating the concomitant release characteristic

Fig. 2. In vitro release profiles of SDF-1 and BMP-2 from scaffolds of (A) S þ B (E) and (B) S þ B (P).

X. Shen et al. / Biomaterials 106 (2016) 205e216210

from S þ B (P) scaffold in contrast to the sequential release char-acteristic from S þ B (E) scaffold.

3.3. In vitro cell migration

The properties of BMSCs were confirmed by expression of BMSCassociated surface markers by flow cytometry and immunofluo-rescence staining (SI, Figs. S4 and S5). The ability of SDF-1 from thescaffold to mobilize BMSCs in vitrowas evaluated using a Transwellsystem [15,33]. As shown in Fig. 3, cell migration across theTranswell membrane was identified as a result of the released SDF-1 from the scaffold. The numbers of the recruited BMSCs in thegroups that contained SDF-1 were significantly increasedcompared with the control or scaffold loaded with BMP-2 alone,and the recruited BMSCs of the scaffold loaded with SDF-1þBMP-2were slightly higher than the scaffold loaded with SDF-1 alone.

3.4. Cell adhesion and proliferation on the scaffold

The biocompatibility of the SF/nHAP scaffold loaded with SDF-1or/and BMP-2 was evaluated via a cell adhesion and growth assay.BMSCs were seeded and cultured on the scaffolds, and themorphology of the adhered cells was assessed via SEM after culturefor 1, 5 and 9 d. The SEM results in Fig. 4A demonstrated that BMSCsadhered to the surfaces of all scaffolds. As the culture timeincreased, there was an increase in the number of cells on thescaffolds, and the cell morphology changed from round to spindle-shaped, which indicates that the SF/nHAP based scaffolds arebiocompatible.

Fig. 3. Transwell migration assay of the effects of SDF-1 from the scaffolds on BMSC migcontained control or scaffold loaded with SDF-1 alone, BMP-2 alone or SDF-1þBMP-2 after 24row. (B) Quantitative comparison of migrated cells among the different groups, *p < 0.05 a

The proliferation of BMSCs on the scaffold loaded with BMP-2and/or SDF-1 was assessed using a CCK-8 assay from 1 to 9 d. Asshown in Fig. 4B, the cells exhibited a good proliferating ability onthe scaffolds loaded with BMP-2. After 3 d of culture, the CCK-8reading was slightly increased in the scaffolds with SDF-1þBMP-2compared with BMP-2 alone both in sequential and concomitantrelease systems (S þ B (E) vs BMP-2 (E) and S þ B (P) vs BMP-2 (P)).The results suggested that the composite scaffold with bothbioactive molecules has a good ability to support cell proliferation,and S þ B (E) scaffold can support cell proliferation in a relativelylong time. The cells exhibited the most apparent viability in thescaffold with SDF-1þBMP-2 compared with the other scaffolds (SI,Fig. S6), and the result is consistent with that of the CCK-8 assay.

3.5. In vitro osteogenesis study

The ability of scaffolds to promote osteogenesis was assessed.The ALP activity, calcium content and osteogenesis marker of OCNof the BMSC culture with scaffolds incorporated with SDF-1 and/orBMP-2 were investigated, and the results are presented in Fig. 5. Allthe scaffolds contained BMP-2 exhibited higher ALP activity (noremarkably difference among these scaffolds) compared with thescaffolds without BMP-2. The scaffolds of S þ B (E) and BMP-2 (E)exhibited an obvious increase in the calcium content and OCNcontent measurements in 21 d, when compared to the Sþ B (P) andBMP-2 (P); the scaffold of S þ B (E) exhibited the highest levelamong all groups.

To further investigate the effects of SDF-1 and BMP-2 on oste-ogenic differentiation, the expression of osteogenic genes,

ration. (A) Representative microscopic images (25�) of cells cultured in media thath of culture; the second row represents higher magnification images (400� of the first

nd **p < 0.01 compared with control.

Fig. 4. Cell adhesion and proliferation assays of scaffolds. (A) SEM images of cell adhesion on scaffold loaded with SDF-1 or/and BMP-2 after seeding for 1 d, 5 d and 9 d. (B) Cellproliferation of BMSCs cultured on different scaffolds after 1 d, 3 d, 5 d, 7 d and 9 d using a CCK-8 kit. Statistically significant differences are indicated with * p < 0.05 and **p < 0.01compared with control.

Fig. 5. Osteogenesis of BMSCs measured by (A) ALP staining at 7 d, (B) Alizarin red staining at 21 d and (C) Immunofluorescence assays for OCN expression at 21 d, with the nucleusstained in blue (DAPI) and OCN stained in red. Scale bar ¼ 100 mm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of thisarticle.)

X. Shen et al. / Biomaterials 106 (2016) 205e216 211

including Alp, Runx2, Ocn and Opn, was quantified by RT-qPCR at 3,7 and 14 d. Fig. 6A and B shows the expression levels of the Alp andRunx2 genes for the BMSCs culturedwith the scaffolds incorporatedwith SDF-1 and/or BMP-2. The expression patterns were similar for

these two genes. The scaffolds contained BMP-2 exhibited asignificantly up-regulated expression of both genes compared withcontrol scaffold, and the scaffold of S þ B (E) exhibited the highestlevel at 7 d and 14 d. Fig. 6C and D shows the Opn and Ocn gene

Fig. 6. Detection of mRNA from selected osteogenic markers in BMSCs after 3 d, 7 d, and 14 d of incubation. The mRNA levels of (A) Alp, (B) Runx2, (C) Opn, and (D) Ocn werequantified (*p < 0.05 and **p < 0.01 compared with control).

X. Shen et al. / Biomaterials 106 (2016) 205e216212

expression results, and these two gene expression patterns werealso similar. There was no obvious difference among the six scaf-folds at 3 d. However, at a prolonged period (7 d and 14 d), therewere significantly increased levels of Opn and Ocn expression in thecells cultured in the scaffold of S þ B (E) compared with the otherscaffolds. The results demonstrated the promotion effect of thescaffold S þ B (E) incorporated with SDF-1þBMP-2 in a sequentialrelease manner on the stimulation of osteogenic genes in BMSCs.

3.6. In vivo studies

The migration of BMSCs injected through the rat tail vein to thedefects was assessed in vivo via BLI (Fig. 7). Cell recruitment to thetissue around the site of the scaffolds incorporated with SDF-1 and/or BMP-2was identified. At 30min and 1 day after the injection, theD-Luciferin-labeled BMSCswere predominately retained in thelungs for all groups, with the exception of a slightly enhancedmigration to the defect site for the SDF-1þBMP-2 group at 1 d. At3 d, the luminescence intensity increased for all groups, particularlyfor the groups contained SDF-1, and the strongest signal wasidentified in the scaffold loaded with SDF-1þBMP-2. As the timeincreased to 7 d and 14 d, a significant decrease in the intensity ofthe bioluminescence signals was identified for all groups; however,the SDF-1 and SDF-1þBMP-2 groups continued to exhibit a higherintensity than the other three groups without SDF-1.

To evaluate the ability of the scaffolds loaded with SDF-1þBMP-2 as grafts to facilitate bone formation within a full-thickness bonedefect in vivo, we surgically created bilateral cranial bone defects inthe rats. The implant area and bone formation in the defects wereimaged with mCT (Fig. 8A and B). The images indicated that thebone content in the scaffolds of BMP-2 (E) and S þ B (E) wasincreased compared with the other four scaffolds at both 8 and 12weeks. New bone was formed from the edge toward the center ofthe defects. The bone volume factionwas the highest in the scaffold

of S þ B (E) where SDF-1 and BMP-2 were released in a sequentialmanner. The bone content at 12 weeks was increased comparedwith 8 weeks, which indicates that the bone content in this groupincreased over time. At 12 weeks, the bone completely bridged theinjury site for the scaffold of S þ B (E). Histological results in Fig. 8Cindicated the formation of new bone with the typical structure ofthe graft group at 12 weeks post-implantation. There was a thickerbone matrix in the scaffolds of S þ B (E) compared with the othergroups.

4. Discussion

In situ tissue engineering for bone regeneration comprises apowerful strategy to treat bone defects, which are regulated by theinteraction between bone progenitor cells and their environment,including biomaterials and bioactive molecules (e.g., chemokinesand growth factors). The objective of this study was to develop acomposite scaffold to initially enhance MSC homing through thedelivery of the chemokine SDF-1 and then stimulate the osteogenicdifferentiation of these MSCs via the subsequent sustained releaseof BMP-2 from the scaffold. To accomplish this design, SF/nHApscaffolds were used to incorporate SF microspheres encapsulatedwith BMP-2, and the prepared scaffolds physically absorbed SDF-1to achieve cell-free scaffolds (S þ B (E)) for the enhanced boneregeneration used in in situ tissue engineering. Here, BMP-2 wasloaded into SF microspheres to fulfill the subsequent controlledrelease purpose. For comparison, the scaffold of S þ B (P) in whichSDF-1 and BMP-2 were released in a concomitant way was pre-pared. In this S þ B (P) scaffold, BMP-2 was directly dropped ontoscaffold instead of encapsulated into SF microspheres. nHAp wasassembled into the SF scaffold to promote bone regenerationthrough its osteoinductive effects [13].

Many techniques have been used to manufacture SF micro-spheres, including ball milling, spray-drying, double emulsion, self-

Fig. 7. BMSC migration to the defect cranial site. (A) Rats are shown at 30 min, 1, 3, 7 and 14 d post-injection; data indicate the systemic cell distribution of reporter cells. Red circlesindicate the ROI. (B) Quantitative analysis of the luminescent signal on the ROI in (A). *p < 0.05 and **p < 0.01 compared with control. (For interpretation of the references to colourin this figure legend, the reader is referred to the web version of this article.)

X. Shen et al. / Biomaterials 106 (2016) 205e216 213

assembly, microfluidics and electro-spraying [34]. Of thesemethods, the laminar jet break-up technology has been demon-strated to be an efficient approach to produce SF microspheres witha high encapsulation efficiency of biomolecules. This technology isalso scalable, simple, and may be operated using only aqueous SFsolutions. Wenk et al. reported that SF microspheres prepared viathis approach not only preserved the bioactivity of insulin-likegrowth factor I (IGF-I) but also enabled a striking sustainment ofits release profile [28]. SF microspheres were prepared from a seriesof SF solutions with different concentrations. The SF microspheresize increased with an increasing SF concentration from 2% to 6%,which was attributed to the process in which the higher SF con-centration induced a higher SF viscosity, thereby reducing thelikelihood of the break-up of the solution into small droplets.However, further increase in the SF concentration to 9% is difficultto fabricate SF microspheres because of the nozzle blockage. Pre-vious studies have confirmed that the SF concentration comprises akey factor in drug release properties, and the initial burst releaseand the release rates of the drugs encapsulated within the SF

microspheres were reduced with the SF concentration increase[28]. These findings are likely a result of the bigger spheres pre-pared at a high SF concentration, which reduces the solventpenetration and limits its disruption of the interaction betweenBMP-2 and SF that control release [35]. In our study, the SF con-centration at 6% was used to encapsulate BMP-2 for the sustainedrelease purpose.

The laminar jet break-up technology typically leads to the highencapsulation efficiency of drugs in SF microspheres [28], which isbeneficial for minimizing the drug loss and maintaining sustaineddrug release. The encapsulation efficiency was considered to be100% for the untreated SF spheres, which is comparable with pre-vious results [36]. To induce b-sheet formation and thus waterinsolubility, SF microspheres were treated with methanol; thistreatment may lead to the partial loss of the BMP-2 amountthrough leaching [28,37], which causes a decrease in the encap-sulation efficiency to 74% (the encapsulation efficiency was calcu-lated from the ratio of the cumulative BMP-2 released when therelease profile reached plateau to the total BMP-2 encapsulated in

Fig. 8. (A) mCT results of the specimen of calvarial critical-sized defects and (B) the bone volume faction. The circle plates in (A) represent original calvarial defects. The white areaabove the circle plate and light area in the plate represent newly formed bone. (C) Sectional structure of calvarias repaired by scaffolds at 12 weeks post-implantation (25�). Scalebar ¼ 1 mm. Bone-like structures were detected in all graft groups after H&E staining. The second row represents higher magnification images (200�) of the corresponding squareboxes in the first row in (C). NB, HB and RSF represent new bone, host bone and remnant silk fibroin, respectively. Red arrow shows remnant of microspheres. Blue arrows indicatenewly formed blood vessels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

X. Shen et al. / Biomaterials 106 (2016) 205e216214

the scaffold). However, it is suggested that the bioactivity ofencapsulated BMP-2 could be well maintained evenwith methanoltreatment [26,38]. BMP-2 loaded SF microspheres were subse-quently incorporated into an SF/nHAp suspension to prepare thecomposite scaffold, followed by physically adsorbed SDF-1 on thesurfaces to achieve an SDF-1þBMP-2 loaded scaffold of S þ B (E).The release behaviors of SDF-1 and BMP-2 in S þ B (E) scaffold aredifferent; the SDF-1 exhibited a burst release in the first day, fol-lowed by a slower release for a period of 7 d, whereas the release ofBMP-2 exhibited a substantially lower extent of burst release in thefirst day and a substantially longer release for up to 3 weeks. Whilefor the S þ B (P) scaffold, both SDF-1 and BMP-2 showed a burstrelease profile. This difference may be explained by the differentincorporation methods [34]. Two mechanisms of both the diffusionof BMP-2 and degradation of SF may participate in the BMP-2release process in S þ B (E) scaffold. The former causes a burstrelease in the early stage, whereas the latter leads to the prolongedrelease of BMP-2 frommicrospheres. SDF-1 or BMP-2 from Sþ B (P)scaffold predominately attached at the scaffold surface via physicaladsorption, whereas its burst release was mainly attributed to thediffusion mechanism [39].

We hypothesized that compared to the concomitant releasefrom S þ B (P), the sequential release of these two bioactive mol-ecules from S þ B (E) are beneficial for the promotion of boneregeneration because the initial burst release of SDF-1 promotesMSC homing to the graft areas in the early time, whereas the sus-tained release of BMP-2 maintains an effective concentration ofBMP-2 to promote the transformation of MSCs into osteoblasts. Totest this hypothesis and its potential application in bone defects,the in vitro cell migration, proliferation, and osteogenetic differ-entiation of BMSCs, as well as the in vivo bone formation wereassessed.

It has been reported that SDF-1 enhances the migration ofMSCs; however, controversial results exist regarding the thresholdof SDF-1 used to induce MSC mobilization [15,33]. Ji et al. indicatedthat the membranes loaded with SDF-1 in the range of 50e400 ngsignificantly induce MSC chemotaxis without dose-dependent ef-fects [15]. In our experiment, 400 ng SDF-1 was used. Both theconcomitant and sequential release of SDF-1 and BMP-2 has a

synergistic effect on the chemotactic capability of BMSCs, which isconsistent with previous reports [14,15,40]. This synergistic effectmay be explained by the fact that in addition to the SDF-1/CXCR4pathway involved in the process of BMSC homing induced bySDF-1 [14,15], BMP-2 may also recruit the cells [40], and theplacental growth factor (PIGF) pathway plays an important role inthe recruitment of BMSCs triggered by BMP-2 [16,40].

To determine the effects of SDF-1 and BMP-2 on the prolifera-tion and osteogenic differentiation of BMSCs in vitro, a CCK8 assay,the ALP activity, the calcium content and the osteogenesis markerof OCN were measured. The scaffolds that contained BMP-2induced significantly increased cell proliferation and osteoblasticactivity comparedwith the scaffolds without BMP-2. The treatmentwith SDF-1 and BMP-2 together resulted in an even higher prolif-eration and osteoblastic activity and a greater potential to accel-erate new bone formation compared with the scaffold loaded withBMP-2 alone, in both concomitant and sequential release systems.However, the controlled release of BMP-2 from the sequentialrelease system induced higher osteoblastic activity compared withthe concomitant release system at a prolonged period. Consideringthe addition of SDF-1 alone only slightly influences BMSC prolif-eration and ALP activity, it is suggested that SDF-1 alone cannoteffectively enhance osteogenic differentiation in vitro [21]. Theability of osteogenic differentiation of BMSCs cultured in the scaf-folds incorporated with SDF-1 and/or BMP-2 was also investigatedby evaluating the expressions of a series of osteogenic genes,including Alp, Runx2, Ocn and Opn. Alp and Runx2 are typically usedto confirm the early differentiation of osteoblastic cells, whereasOcn and Opn are representative indexes for further bone matura-tion [18,40,41]. It is thus reasonable that the up-regulation of Alpand Runx2 is initiated as early as 3 d, whereas Ocn and Opn begin at7 d. The scaffold of S þ B (E) exhibited the highest gene expressionfor all markers after culture for 7 d, which supports the previouslydescribed results regarding the calcium content and ALP activitymeasurement.

The migration of BMSCs is of great importance for the boneregeneration [42,43], however, so far there are fewer reports on thein vivo investigation of migration of BMSCs compared with corre-sponding in vitro investigation. Tracking the injected cells via

X. Shen et al. / Biomaterials 106 (2016) 205e216 215

imaging offers the feasibility to determine their survival, migration,immunogenicity and their function in bone regenerating in livinganimals. Among the optical imaging techniques, BLI is a promisingtool for semi-quantitative measurements of biological processesbecause of a strong relationship between live cell number and thebioluminescence signal detected both in vitro and in vivo [44,45]. Inthis work, D-Luciferin-labeled BMSCs were injected through the rattail vein and were found to be mainly trapped in the lungs within1 d after injection (Fig. 7), in consistent with previous report[46,47]. At 3 d, BMSCs were gradually recruited to the defect site forall groups. Particularly, the groups containing SDF-1 showedstronger signal compared with the other groups, suggesting thatSDF-1 is critical to induce themigration and homing of stem cells toa targeted site within the body [48]. The scaffolds of S þ B (P) andS þ B (E) exhibited slightly higher signal than the scaffold loadedwith SDF-1 alone, indicating that the released BMP-2 did notinterfere the function of SDF-1 but might provide a synergistic ef-fect to improve the recruitment of BMSCs to the defect sites, whichis in consistent with our experimental data of in vitro migrationassay. Bioluminescence decreased over time, suggesting that mostof the injected cells were dead after 14 d following the trans-plantation. The lower survival rate for the injected cells was alsofound by other group [47]; this phenomenon may likely be attrib-uted to the effect of immune rejection [44].

Besides recruiting the BMSCs to the implanted area, the SF/nHAp scaffolds loaded with SDF-1þBMP-2 may substantiallyenhance bone regeneration in rat cranial defects; After 12 weeks ofimplantation of S þ B (E), the bone completely bridged the injurysite as supported by the BLI, mCT and histological results. The extentof bone regeneration was formed in the order of S þ B (E) > BMP-2(E) > S þ B (P) > BMP-2 (P) > SDF-1 > control. There was almost nobone regeneration at the sham-surgery control sites (SI, Fig. S7). Incomparison with the in vitro results, the scaffolds loaded with SDF-1 alone may also promote bone regeneration. These phenomenamay be explained by the fact that there are many other types ofcells, such as osteoblast progenitor cells (OPCs) and endothelialprogenitor cells (EPCs), in the circulating blood [49,50]. OPCs andEPCs have the potential to facilitate osteogenesis [51] and angio-genesis [52], respectively. SDF-1 has an inherent ability to enhancethe recruitment of these cells [53,54], and BMP-2 has a potential topromote the osteogenic differentiation of bone marrow-derivedstem/progenitor cells and recruit the cells [40,55]; thus, therelease of these two factors from the scaffold may promote therecruitment of hematopoietic and bone marrow-derived stem/progenitor cells. The former cells will generate new blood vessels,whereas the latter cells exhibit the ability to be osteoblasts by BMP-2 stimulation and result in enhanced bone regeneration [40]. Inaddition, SDF-1 provides synergistic effects that support BMP-2-induced, osteogenic differentiation of mesenchymal C2C12 cells[56] or BMSC-mediated bone formation [19]. Hosogane et al.demonstrated a regulatory role of SDF-1 in BMP-2-induced osteo-genic differentiation of MSCs, in which perturbation of the SDF-1signaling affected the differentiation of MSCs towards osteoblasticcells in response to BMP-2 stimulation [14]. It was furtherdemonstrated that the effect of SDF-1 on BMP-2 induced osteo-genesis was mediated via intracellular Smad and Erk activation[14,21] or Smad and MAPK activation [56].

Angiogenesis and the vascularization capability are critical forthe survival of engineered grafts following implantation [57]. His-tological results (Fig. 8C) show the presence of blood vessel innewly formed bones. Thus, further experiments are being con-ducted in our group, such as testing angiogenesis after the im-plantation of S þ B (E). Moreover, histological results (Fig. 8C)indicate that most SF scaffold was degraded following an implan-tation period of 12 weeks; thus, the influence of the scaffold

degradation behavior on bone formation will also be investigated.

5. Conclusions

In the present study, we developed a SF/nHAp based scaffoldincorporated with SDF-1 and BMP-2 loaded SF microspheres toinduce the sequential and controlled release of SDF-1 and BMP-2.The rapid initial release of SDF-1 promoted BMSC recruitment,and the relatively slow and sustained release of BMP-2 facilitatedosteogenic differentiation both in vitro and in vivo. Compared withthe concomitant release, the sequential release of SDF-1 and BMP-2from scaffolds increased bone regeneration in rat cranial defects,and the bone completely bridged the injury site after 12 weeks ofimplantation. The application of these scaffolds may comprise apowerful platform in bone tissue engineering to treat bone defects.

Acknowledgments

This work was supported by the National Natural ScienceFoundation of China (81371930, 21604059), Key Talented ManProject of Jiangsu Province (RC2011102) and Standardized Diag-nosis and Treatment Project of Key Diseases in Jiangsu Province(BE2015641). Y. Zhang also acknowledges the Natural ScienceFoundation of Jiangsu Province (BK20160321) and the NaturalScience Foundation of the Jiangsu Higher Education Institutions ofChina (16KJB430029).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2016.08.023.

References

[1] R. Dimitriou, E. Jones, D. McGonagle, P.V. Giannoudis, Bone regeneration:current concepts and future directions, BMC Med. 9 (2011) 66.

[2] J. Raphel, M. Holodniy, S.B. Goodman, S.C. Heilshorn, Multifunctional coatingsto simultaneously promote osseointegration and prevent infection of ortho-paedic implants, Biomaterials 84 (2016) 301e314.

[3] S.F. Badylak, R.M. Nerem, Progress in tissue engineering and regenerativemedicine, Proc. Natl. Acad. Sci. U. S. A. 107 (2010) 3285e3286.

[4] D. Tang, R.S. Tare, L.Y. Yang, D.F. Williams, K.L. Ou, R.O.C. Oreffo, Biofabricationof bone tissue: approaches, challenges and translation for bone regeneration,Biomaterials 83 (2016) 363e382.

[5] M. Farokhi, F. Mottaghitalab, M.A. Shokrgozar, K.L. Ou, C. Mao,H. Hosseinkhani, Importance of dual delivery systems for bone tissue engi-neering, J. Control. Release 225 (2016) 152e169.

[6] R. Langer, Perspectives and challenges in tissue engineering and regenerativemedicine, Adv. Mater. 21 (2009) 3235e3236.

[7] A.M. Martins, C.M. Alves, F.K. Kasper, A.G. Mikos, R.L. Reis, Responsive and insitu-forming chitosan scaffolds for bone tissue engineering applications: anoverview of the last decade, J. Mater. Chem. 20 (2010) 1638e1645.

[8] I.K. Ko, S.J. Lee, A. Atala, J.J. Yoo, In situ tissue regeneration through host stemcell recruitment, Exp. Mol. Med. 45 (2013) e57.

[9] Q. Li, L. Ma, C. Gao, Biomaterials for in situ tissue regeneration: developmentand perspectives, J. Mater. Chem. B 3 (2015) 8921e8938.

[10] F. Mottaghitalab, H. Hosseinkhani, M.A. Shokrgozar, C. Mao, M. Yang,M. Farokhi, Silk as a potential candidate for bone tissue engineering, J. Control.Release 215 (2015) 112e128.

[11] J. Melke, S. Midha, S. Ghosh, K. Ito, S. Hofmann, Silk fibroin as biomaterial forbone tissue engineering, Acta Biomater. 31 (2016) 1e16.

[12] H. Zhou, J. Lee, Nanoscale hydroxyapatite particles for bone tissue engineer-ing, Acta Biomater. 7 (2011) 2769e2781.

[13] H. Liu, G.W. Xu, Y.F. Wang, H.S. Zhao, S. Xiong, Y. Wu, et al., Compositescaffolds of nano-hydroxyapatite and silk fibroin enhance mesenchymal stemcell-based bone regeneration via the interleukin 1 alpha autocrine/paracrinesignaling loop, Biomaterials 49 (2015) 103e112.

[14] N. Hosogane, Z. Huang, B.A. Rawlins, X. Liu, O. Boachie-Adjei, A.L. Boskey, etal., Stromal derived factor-1 regulates bone morphogenetic protein 2-inducedosteogenic differentiation of primary mesenchymal stem cells, Int. J. Biochem.Cell. Biol. 42 (2010) 1132e1141.

[15] W. Ji, F. Yang, J. Ma, M.J. Bouma, O.C. Boerman, Z. Chen, et al., Incorporation ofstromal cell-derived factor-1 alpha in PCL/gelatin electrospun membranes forguided bone regeneration, Biomaterials 34 (2013) 735e745.

X. Shen et al. / Biomaterials 106 (2016) 205e216216

[16] Y. Kimura, N. Miyazaki, N. Hayashi, S. Otsuru, K. Tamai, Y. Kaneda, et al.,Controlled release of bone morphogenetic protein-2 enhances recruitment ofosteogenic progenitor cells for de novo generation of bone tissue, Tissue Eng.Part A 16 (2010) 1263e1270.

[17] R.C. de Guzman, J.M. Saul, M.D. Ellenburg, M.R. Merrill, H.B. Coan, T.L. Smith, etal., Bone regeneration with BMP-2 delivered from keratose scaffolds, Bio-materials 34 (2013) 1644e1656.

[18] S. Suliman, Z. Xing, X. Wu, Y. Xue, T.O. Pedersen, Y. Sun, et al., Release andbioactivity of bone morphogenetic protein-2 are affected by scaffold bindingtechniques in vitro and in vivo, J. Control. Release 197 (2015) 148e157.

[19] S. Herberg, C. Susin, M. Pelaez, R.N. Howie, R.M. de Freitas, J. Lee, et al., Low-dose bone morphogenetic protein-2/stromal cell-derived factor-1b cotherapyinduces bone regeneration in critical-size rat calvarial defects, Tissue Eng. PartA 20 (2014) 1444e1453.

[20] C.H. Lee, M.U. Jin, H.M. Jung, J.T. Lee, T.G. Kwon, Effect of dual treatment withSDF-1 and BMP-2 on ectopic and orthotopic bone formation, Plos ONE 10(2015) e0120051.

[21] H.D. Hwang, J.T. Lee, J.T. Koh, H.M. Jung, H.J. Lee, T.G. Kwon, Sequentialtreatment with SDF-1 and BMP-2 potentiates bone formation in calvarialdefects, Tissue Eng. Part A 21 (2015) 2125e2135.

[22] E.J. Carragee, E.L. Hurwitz, B.K. Weiner, A critical review of recombinant hu-man bone morphogenetic protein-2 trials in spinal surgery: emerging safetyconcerns and lessons learned, Spine J. 11 (2011) 471e491.

[23] J.N. Zara, R.K. Siu, X. Zhang, J. Shen, R. Ngo, M. Lee, et al., High doses of bonemorphogenetic protein 2 induce structurally abnormal bone and inflamma-tion in vivo, Tissue Eng. Part A 17 (2011) 1389e1399.

[24] C.A. Tannoury, H.S. An, Complications with the use of bone morphogeneticprotein 2 (BMP-2) in spine surgery, Spine J. 14 (2014) 552e559.

[25] X. Wang, E. Wenk, X. Zhang, L. Meinel, G. Vunjak-Novakovic, D.L. Kaplan,Growth factor gradients via microsphere delivery in biopolymer scaffolds forosteochondral tissue engineering, J. Control. Release 134 (2009) 81e90.

[26] X. Wang, E. Wenk, A. Matsumoto, L. Meinel, C. Li, D.L. Kaplan, Silk micro-spheres for encapsulation and controlled release, J. Control. Release 117(2007) 360e370.

[27] L. Chen, H.L. Liu, Y. Gu, Y. Feng, H.L. Yang, Lumbar interbody fusion withporous biphasic calcium phosphate enhanced by recombinant bonemorphogenetic protein-2/silk fibroin sustained-released microsphere: anexperimental study on sheep model, J. Mater. Sci. Mater. Med. 26 (2015) 126.

[28] E. Wenk, A.J. Wandrey, H.P. Merkle, L. Meinel, Silk fibroin spheres as a plat-form for controlled drug delivery, J. Control. Release 132 (2008) 26e34.

[29] R.A. Perez, J.H. Kim, J.O. Buitrago, I.B. Wall, H.W. Kim, Novel therapeutic core-shell hydrogel scaffolds with sequential delivery of cobalt and bonemorphogenetic protein-2 for synergistic bone regeneration, Acta Biomater. 23(2015) 295e308.

[30] Q. Guo, C. Liu, J. Li, C. Zhu, H. Yang, B. Li, Gene expression modulation in TGF-3-mediated rabbit bone marrow stem cells using electrospun scaffolds ofvarious stiffness, J. Cell. Mol. Med. 19 (2015) 1582e1592.

[31] A.N. Mitropoulos, G. Perotto, S. Kim, B. Marelli, D.L. Kaplan, F.G. Omenetto,Synthesis of silk fibroin micro-and submicron spheres using a co-flow capil-lary device, Adv. Mater. 26 (2014) 1105e1110.

[32] Z. Luo, L. Jiang, Y. Xu, H. Li, W. Xu, S. Wu, et al., Mechano growth factor (MGF)and transforming growth factor (TGF)-beta 3 functionalized silk scaffoldsenhance articular hyaline cartilage regeneration in rabbit model, Biomaterials52 (2015) 463e475.

[33] Y.S. Liu, M.E. Ou, H. Liu, M. Gu, L.W. Lv, C. Fan, et al., The effect of simvastatinon chemotactic capability of SDF-1 alpha and the promotion of bone regen-eration, Biomaterials 35 (2014) 4489e4498.

[34] F. Mottaghitalab, M. Farokhi, M.A. Shokrgozar, F. Atyabi, H. Hosseinkhani, Silkfibroin nanoparticle as a novel drug delivery system, J. Control. Release 206(2015) 161e176.

[35] T. Yucel, M.L. Lovett, D.L. Keplan, Silk-based biomaterials for sustained drugdelivery, J. Control. Release 190 (2014) 381e397.

[36] P.C. Bessa, E.R. Balmayor, H.S. Azevedo, S. Nuernberger, M. Casal, M. vanGriensven, et al., Silk fibroin microparticles as carriers for delivery of humanrecombinant BMPs. Physical characterization and drug release, J. Tissue Eng.

Regener. Med. 4 (2010) 349e355.[37] A.B. Li, J.A. Kluge, N.A. Guziewicz, F.G. Omenetto, D.L. Kaplan, Silk-based sta-

bilization of biomacromolecules, J. Control. Release 219 (2015) 416e430.[38] S. Hofmann, C. Foo, F. Rossetti, M. Textor, G. Vunjak-Novakovic, D.L. Kaplan, et

al., Silk fibroin as an organic polymer for controlled drug delivery, J. Control.Release 111 (2006) 219e227.

[39] X. Huang, C.S. Brazel, On the importance and mechanisms of burst release inmatrix-controlled drug delivery systems, J. Control. Release 73 (2001)121e136.

[40] J. Ratanavaraporn, H. Furuya, H. Kohara, Y. Tabata, Synergistic effects of thedual release of stromal cell-derived factor-1 and bone morphogenetic protein-2 from hydrogels on bone regeneration, Biomaterials 32 (2011) 2797e2811.

[41] M. Bruderer, R.G. Richards, M. Alini, M.J. Stoddart, Role and regulation ofRunx2 in osteogenesis, Eur. Cells Mater. 28 (2014) 269e286.

[42] S. Li, Q. Tu, J. Zhang, G. Stein, J. Lian, P.S. Yang, et al., Systemically transplantedbone marrow stromal cells contributing to bone tissue regeneration, J. Cell.Physiol. 215 (2008) 204e209.

[43] L.M. Yu, Q.S. Tu, Q.Q. Han, L. Zhang, L. Sui, L.L. Zheng, et al., Adiponectinregulates bone marrow mesenchymal stem cell niche through a unique signaltransduction pathway: an approach for treating bone disease in diabetes,Stem Cells 33 (2015) 240e252.

[44] P.E. de Almeida, J.R.M. van Rappard, J.C. Wu, In vivo bioluminescence fortracking cell fate and function, Am. J. Physiol. Heart Circ. Physiol. 301 (2011)H663eH671.

[45] R.J. Hajjar, D.P. Cormode, Tracking cell therapy bioluminescence lighting theway, JACC-Cardiovasc. Imag. 5 (2012) 56e58.

[46] F. Granero-Molto, J.A. Weis, M.I. Miga, B. Landis, T.J. Myers, L. O'Rear, et al.,Regenerative effects of transplanted mesenchymal stem cells in fracturehealing, Stem Cells 27 (2009) 1887e1898.

[47] S. Huang, L.L. Xu, Y.X. Sun, Y.F. Zhang, G. Li, The fate of systemically admin-istrated allogeneic mesenchymal stem cells in mouse femoral fracture healing,Stem Cell Res. Ther. 6 (2015) 206.

[48] V. Sordi, M.L. Malosio, F. Marchesi, A. Mercalli, R. Melzi, T. Giordano, et al.,Bone marrow mesenchymal stem cells express a restricted set of functionallyactive chemokine receptors capable of promoting migration to pancreaticislets, Blood 106 (2005) 419e427.

[49] G.Z. Eghbali-Fatourechi, J. Lamsam, D. Fraser, D. Nagel, B.L. Riggs, S. Khosla,Circulating osteoblast-lineage cells in humans, N. Engl. J. Med. 352 (2005)1959e1966.

[50] Y. Fan, J. Ye, F. Shen, Y. Zhu, Y. Yeghiazarians, W. Zhu, et al., Interleukin-6stimulates circulating blood-derived endothelial progenitor cell angiogenesisin vitro, J. Cereb. Blood Flow. Metab. 28 (2008) 90e98.

[51] S. Otsuru, K. Tamai, T. Yamazaki, H. Yo shikawa, Y. Kaneda, Circulating bonemarrow-derived osteoblast progenitor cells are recruited to the bone-formingsite by the CXCR4/stromal cell-derived factor-1 pathway, Stem Cells 26 (2008)223e234.

[52] T. Asahara, T. Murohara, A. Sullivan, M. Silver, R. vanderZee, T. Li, et al.,Isolation of putative progenitor endothelial cells for angiogenesis, Science 275(1997) 964e967.

[53] A. Aiuti, I.J. Webb, C. Bleul, T. Springer, J.C. GutierrezRamos, The chemokineSDF-1 is a chemoattractant for human CD34(þ) hematopoietic progenitorcells and provides a new mechanism to explain the mobilization of CD34(þ)progenitors to peripheral blood, J. Exp. Med. 185 (1997) 111e120.

[54] D.K. Jin, K. Shido, H.G. Kopp, I. Petit, S.V. Shmelkov, L.M. Young, et al., Cyto-kine-mediated deployment of SDF-1 induces revascularization throughrecruitment of CXCR4(þ) hemangiocytes, Nat. Med. 12 (2006) 557e567.

[55] C.M. Li, C. Vepari, H.J. Jin, H.J. Kim, D.L. Kaplan, Electrospun silk-BMP-2 scaf-folds for bone tissue engineering, Biomaterials 27 (2006) 3115e3124.

[56] W. Zhu, O. Boachie-Adjei, B.A. Rawlins, B. Frenkel, A.L. Boskey, L.B. Ivashkiv, etal., A novel regulatory role for stromal-derived factor-1 signaling in bonemorphogenic protein-2 osteogenic differentiation of mesenchymal C2C12cells, J. Biol. Chem. 282 (2007) 18676e18685.

[57] J.M. Kanczler, R.O.C. Oreffo, Osteogenesis and angiogenesis: the potential forengineering bone, Eur. Cells Mater. 15 (2008) 100e114.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具