Acta Biomaterialia 6 (2010) 455–465

Contents lists available at ScienceDirect

Acta Biomaterialia

journal homepage: www.elsevier .com/locate /actabiomat

An injectable scaffold: rhBMP-2-loaded poly(lactide-co-glycolide)/hydroxyapatitecomposite microspheres

Hong Shen, Xixue Hu, Fei Yang, Jianzhong Bei, Shenguo Wang *

Institute of Chemistry, Chinese Academy of Sciences, Center for Molecular Sciences, Zhongguan Cun, Haidian Qu, Beijing 100080, China

a r t i c l e i n f o a b s t r a c t

Article history:Received 12 February 2009Received in revised form 6 July 2009Accepted 9 July 2009Available online 15 July 2009

Keywords:PLGA/HA(50/50)rhBMP-2Composite microspheresCell affinityInjectable scaffold

1742-7061/$ - see front matter � 2009 Acta Materialdoi:10.1016/j.actbio.2009.07.016

* Corresponding author. Tel./fax: +86 10 62581241E-mail addresses: wangsg@hotmail.com, wangsg@

Poly(lactide-co-glycolide)/hydroxyapatite(50/50) (PLGA/HA(50/50)) composite microspheres were fabri-cated and treated with a mixture of 0.25 M NaOH aqueous solution and ethanol (v/v = 1/1) at 37 �C. Theproperties of untreated and treated PLGA/HA(50/50) composite microspheres were determined and com-pared. The results showed that the surface roughness, HA content and hydrophilicity of the treated PLGA/HA(50/50) composite microspheres increased with treatment time. However, the treatment time shouldbe kept within 2 h in order to maintain the shape of the PLGA/HA(50/50) microspheres. At the same time,a degradation study showed that both the untreated and treated microspheres degraded gradually withtime, with the treated microspheres degrading faster in the first 4 weeks. The rhBMP-2-loaded PLGA/HA(50/50) composite microspheres were prepared by solution dipping treated PLGA/HA(50/50) compos-ite microspheres. Mouse OCT-1 osteoblast-like cells were cultured on the untreated, treated and rhBMP-2-loaded PLGA/HA(50/50) composite microspheres and the cell affinity of the various microspheres wasassessed and compared. It was found that the surface-treated PLGA/HA(50/50) composite microspheresclearly promoted osteoblast attachment, proliferation and alkaline phosphatase activity. It was consid-ered that the hydrophilicity, osteoconductivity and surface roughness were increased by the increasein the HA component, which facilitated cell growth. Moreover, the rhBMP-2 loaded on the treatedPLGA/HA(50/50) composite microspheres could be slowly released and further enhanced osteoblast dif-ferentiation. The good cell affinity and enhanced osteogenic potential of the rhBMP-2-loaded PLGA/HAcomposite microspheres indicate that they could be used as an injectable scaffold.

� 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Bone grafts, including autografts, allografts and xenografts, havebeen widely used for repairing bone defects originated fromtrauma, tumor resection, bone fractures, infections and skeletalabnormities [1–3]. Although some of them are effective for boneregeneration, many defects, such as limited availability, immunereaction, transfer of pathogens and non-biodegradability, have lim-ited the development of bone grafts [4,5]. Bone tissue engineeringusing osteogenic cells, osteoinductive growth factors and scaffoldsalone or in combination appears to be the most promising alterna-tive to existing therapies for bone repair and regeneration [6]. Oneof the most significant challenges of this technique is to design andfabricate suitable biodegradable scaffolds that can support celladhesion, growth, proliferation and differentiation, and guide theprocess of tissue formation.

Recently, the development of scaffolds for tissue engineeringhas focused on the design and fabrication of biomimetic scaffoldmaterials that can interact with surrounding tissues by biomolec-

ia Inc. Published by Elsevier Ltd. A

.iccas.ac.cn (S. Wang).

ular recognition. Biomimetic material is designed to elicit specificcellular responses and directly form new tissue through some spe-cific interaction. Hydroxyapatite (HA) (Ca10(PO4)6(OH)2), as abioceramic, is an effective component for biomimetic materialssince its chemical and structural characteristics are similar to themineral phase of native bone [7]. Because of its good biocompati-bility and osteoconductivity, HA has been widely used as a bone-filling material in dental and orthopedic surgery [8,9]. However,drawbacks such as difficulty of shaping, poor mechanical strength,brittleness and slow degradation rate still limit the direct applica-tion of HA as a scaffold for bone tissue engineering [10–12]. It ishoped that composite scaffolds of HA and other biomaterials canovercome the defects mentioned above [13–15].

Much attention has also been paid recently to composites of HAand synthetic polylactone-type biodegradable polymers, such aspoly(L-lactide), polyglycolide and their copolymer poly(lactide-co-glycolide) (PLGA), since the polymers possess good mechanicalproperties, low immunogenicity and toxicity, and an adjustabledegradation rate. The combination of HA and polylactone-typepolymer can be expected to combine the best features of bothmaterials to obtain the optimum scaffold for bone tissue engineer-ing, since it possesses fundamental characteristics such as bioactiv-

ll rights reserved.

456 H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465

ity, biomechanical similarity, processability and biodegradability.It can also reduce the acidity of the degradation products of thepolylactone-type polymers [16].

Most polylactone/HA composite scaffolds have been fabricatedinto films, rods, plates, blocks and foam by the solvent evaporationmethod [17], the electrospinning method[18,19], the solvent cast-ing and particulate leaching (SC/PL) method [20,21], the compres-sion molding method [22,23], the phase separation method[24,25], the gas foaming and particulate leaching (GF/PL) method[26,27], the indirect solid free form (SFF) method [28], etc.Although the geometries and configurations of pre-shaped polylac-tone/HA composite scaffolds can be continually improved withadvancements in the fabrication method, it is still difficult to pro-duce the complicated shapes of scaffolds required to meet clinicalneeds. Considering that microspheres exhibit flowability and canbe injected as a three-dimensional scaffold into various shapedbone defects for bone repair, polylactone/HA composite micro-sphere-type scaffolds could provide more versatile applicationsthan pre-shaped scaffolds. On the other hand, microsphere-typescaffolds can also perform tissue repair and gene therapy whenloaded with special growth factors or drugs. Microsphere-typescaffolds also possess other advantages, such as needing only aminor incision and a more convenient operation for the scaffoldtransplantation. The polylactone/HA composite microspheres havegenerally been fabricated by an emulsion–solvent evaporationmethod and studied as drug delivery vehicles [29,30]. However, afew studies have reported on the application of polylactone/HAcomposite microspheres as injectable scaffolds in tissueengineering.

On the other hand, although it has been reported that HA canenhance osteoblast growth and differentiation, whether HA can in-duce bone formation by itself it still a matter of contention. Onoet al. [31] thought HA could not induce bone formation by itself,particularly in sites where bone does not form normally. However,Habibovic et al. [32] indicated the osteoinductive properties of HAin ectopic sites. Therefore, it is necessary to combine the polylac-tone/HA composite microspheres with another bone inductor topromote and expedite bone formation.

Bone morphogenetic proteins (BMPs) are potent bone induc-tors, which control osteogenesis. They play a crucial role in cellgrowth and differentiation in a variety of cell types, including oste-oblasts [33]. To improve the efficiency of administering BMPs, thecombination of BMPs with biomaterials has received considerableattention [34–40]. BMPs are combined on biomaterials mainly byphysisorption [35,37], electrostatic interaction [39] or covalentbinding [40]. Of the BMPs, BMP-2 is most extensively researched,and has been used in clinical applications since it is readily avail-able and has a very strong osteoinductive activity [41,42]. It hasbeen reported that BMP-2 can promote the maturation of commit-ted cells to become more differentiated osteoblasts and inducebone formation in ectopic and orthotopic sites in vivo [33,43,44].

The aim of this study was to develop a BMP-2-loaded polylac-tone/HA composite microsphere-type scaffold as an injectablebone tissue engineering scaffold. First, the PLGA/HA(50/50) com-posite microspheres were treated with a mixture of NaOH aqueoussolution and ethanol. The morphology and properties of the treatedPLGA/HA(50/50) composite microspheres were investigated byscanning electron microscopy (SEM), energy-dispersive spectros-copy (EDS), gel permeation chromatography (GPC) and water up-take. At the same time, the degradation of untreated and treatedmicrospheres was studied and compared. The treated PLGA/HA(50/50) composite microspheres were also combined withrecombinant human bone morphogenetic protein-2 (rhBMP-2).Finally, adhesion, proliferation and differentiation of cells on therhBMP-2-loaded PLGA/HA(50/50) composite microspheres wereevaluated and compared with that on untreated PLGA/HA(50/50)

and treated PLGA/HA(50/50) composite microspheres using themouse OCT-1 osteoblast-like cell as a model cell in vitro.

2. Materials and methods

2.1. Materials

Glycolide and lactide were purchased from Acros Chemica, N.V.and purified twice by recrystallization from ethyl acetate. Stannousoctoate (Sigma, A.R.) was used without further purification. Ethylacetate was dried by P2O5 overnight and then distilled. Hydroxyap-atite (4.86 lm of average particle size) was purchased from Sinop-harm Chemical Reagent Co., Ltd., China. rhBMP-2 was produced bythe Fourth Military Medical University, China. Poly(vinyl alcohol)(PVA, average Mn = 77,000, 87–89% hydrolyzed) was purchasedfrom Tianjin Zongheng Chemical Company, China. Dulbecco’s mod-ified Eagle’s medium (DMEM) and fetal bovine serum (FBS) werepurchased from Invitrogen Corporation, America. Trypsin and eth-ylenediaminetetraacetic acid (EDTA) were obtained from Sigma.The protein assay kit and alkaline phosphatase (ALP) kit were fromNanjing Jianchen Bioengineering Institute, China.

2.2. Synthesis of PLGA

PLGA (molar ratio of lactide/glycolide = 70/30, Mw = 112,000)was synthesized from L-lactide and glycolide under high vacuumin the presence of stannous octoate as a catalyst (0.05 wt.%) at160 �C for 20 h according to the literature [45].

2.3. Fabrication of PLGA/HA(50/50) composite microspheres

PLGA/HA(50/50) composite microspheres were fabricated by anemulsion–solvent evaporation method. Briefly, 200 mg of PLGAwas completely dissolved in 8 ml of dichloromethane and 200 mgof HA was added to the PLGA solution and stirred thoroughly to forma well-dispersed mixture. The mixture was then emulsified in a PTFEtube by ultrasonication under 300 W of output for 30 s. The formedsolid-in-oil emulsion was subsequently dropped into 200 ml ofexternal PVA aqueous solution (1% (w/v) concentration) and furtherstirred at 500–600 rpm at room temperature for 2–3 h to evaporatethe organic solvent. Finally, the formed microspheres were collectedby centrifugation, washed five times with distilled water and freeze-dried using a lyophilizer (ALPHA 1–2 LD) to sublimate the remainedwater and obtain free-flowing PLGA/HA(50/50) microspheres. Theaverage diameter of the PLGA/HA(50/50) composite microsphereswas in a range of 50–120 lm. The dried microspheres were storedat 4 �C before used.

2.4. Surface treatment

In accordance with the previously reported method [46], thePLGA/HA(50/50) composite microspheres were immersed in a mix-ture of 0.25 M NaOH aqueous solution and ethanol (v/v = 1/1) for apredetermined period (0.5, 1, 2 and 3 h) of incubation at 37 �C,then rinsed with deionized water (3 � 10 min) and freeze-driedusing a lyophilizer.

The untreated PLGA/HA(50/50) composite microspheres, andthe PLGA/HA(50/50) composite microspheres treated with themixture of 0.25 M NaOH aqueous solution and ethanol (v/v = 1/1)for 1 and 2 h at 37 �C were abbreviated as UT-PLGA/HA, T1-PLGA/HA and T2-PLGA/HA, respectively.

2.5. Characterization of surface property

After various PLGA/HA(50/50) composite microspheres weredried and sputter-coated with gold, their surface morphology them

H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465 457

was observed by SEM (Hitachi S-530, Japan) under an acceleratingvoltage of 5.0 kV.

The surface element content of the various PLGA/HA(50/50)composite microspheres was examined by EDS (Phoenix, USA).

2.6. Hydrophilicity measurement

After various PLGA/HA(50/50) composite microspheres wereimmersed in distilled water for 72 h at room temperature, the ex-cess water was removed by filter paper. The water uptake of thevarious microspheres was calculated according to the followingformula:

water uptakeð%Þ ¼ ½ðW2 �W1Þ=W1� � 100

where W1 and W2 are the mass of the microspheres before and afterbeing immersed in water, respectively.

2.7. Molecular weight measurement

The weight-average molecular weight (Mw) and number-aver-age molecular weight (Mn) of PLGA in the various PLGA/HA(50/50) composite microspheres were determined by gel permeationchromatography (GPC). GPC measurements were carried out on aWaters 510 apparatus equipped with ShodexGPC KF-800 columnsat 35 �C. Chloroform was used as the eluent at a flow rate of1.0 ml min�1. The molecular weight of the copolymers was cali-brated with polystyrene standards.

2.8. Degradation study

For the degradation study, 40.0 mg of untreated or treatedPLGA/HA(50/50) composite microspheres was added to centrifugetubes with 8 ml of pH 7.4 phosphate-buffered saline (PBS), whichwas renewed every week, and the centrifuge tube was placed ina thermostated shaking water bath with constant shaking at37 ± 1 �C. At a predetermined interval the specimen was takenout and rinsed three times with deionized water. Finally, the spec-imen was dried to constant mass by freeze-drying and weighed.The mass loss of the specimen was calculated as:

mass loss ð%Þ ¼ ðW0 �WdÞ=W0 � 100

where W0 and Wd represent the initial and dried mass of the de-graded specimen, respectively.

The molecular weight of PLGA in the degraded microsphereswas measured by GPC.

2.9. The rhBMP-2 loading procedure of the treated PLGA/HA(50/50)composite microspheres

A 5 ll aliquot of distilled water containing 500 ng of rhBMP-2was dropped into 10 mg of T2-PLGA/HA and completely absorbedby the microspheres because the volume of rhBMP-2 solutionwas less than that of water taken up by the microspheres. Themicrospheres were left overnight at 4 �C, and then dried byfreeze-drying to obtain the rhBMP-2-loaded PLGA/HA(50/50) com-posite microspheres.

The rhBMP-2-loaded PLGA/HA(50/50) composite microsphereswere abbreviated as T2-PLGA/HA/BMP.

2.10. rhBMP-2 release measurement of the T2-PLGA/HA/BMP

A 10.0 mg aliquot of T2-PLGA/HA/BMP was immersed in 300 llof release medium, which comprised PBS (pH 7.4) supplementedwith 0.1% bovine serum albumin (Sigma), to perform the releasestudy. The microspheres were incubated at 37 �C for up to 3 weeksunder static conditions. At preset time intervals, the incubation

solution was collected and replaced with fresh release medium.The amount of released rhBMP-2 in the collected medium wasdetermined by an ELISA Kit (Human BMP-2 Quantikine ELISA Kit,R&D Systems, Minneapolis, MN).

2.11. Cell culture

2.11.1. Preparation of cellsOCT-1 osteoblast-like cells were derived from osteocalcin pro-

moter-driven SV-40 T-antigen transgenic mouse calvarias [47]and supplied by the Institute of Molecular Biology Study of NankaiUniversity (China). Briefly, cells were incubated at 37 �C in DMEMsupplemented with 10% FBS and 100 U cm�3 each of penicillin andstreptomycin in a 5% CO2 incubator. When the cells had grown toconfluence, they were detached by trypsin/EDTA (0.05% (w/v) tryp-sin/0.02% (w/v) EDTA) and resuspended in fresh culture medium tothe correct concentration for seeding onto various PLGA/HA(50/50)composite microspheres.

2.11.2. Cell culture on microspheresFor cell culture, 10.0 mg of various PLGA/HA(50/50) composite

microspheres was sterilized by 70% ethanol for 1 h. The ethanolwas then exchanged with an excess amount of distilled waterand subsequently washed with PBS (pH 7.4) three times. TheOCT-1 cell suspension, containing 1.0 � 105 cells, was seeded ontothe microspheres and the suspension including the microspheresand cells was then incubated in a siliconized spinner flask, whichwas placed in a 5% CO2 incubator at 37 �C and rotated at 5 rpm.

2.11.3. Observation of cell morphology and distributionAfter OCT-1 cells had been cultured for 7 and 14 days on the

various PLGA/HA(50/50) composite microspheres, the micro-spheres were washed with PBS and fixed with 2.5% glutaraldehydefor 24 h at 4 �C. The microspheres were then dehydrated through aseries of graded alcohols, free-dried and sputter-coated with gold.Finally, the morphology of cells was observed by SEM.

The distribution of OCT-1 cells on the various PLGA/HA(50/50) composite microspheres was observed after cell culture for7 days. To visualize the viable cells on the microspheres undera confocal laser scanning microscope (CLSM; Zeiss LSM 510, Ger-many), the microspheres were transferred to fresh medium con-taining 25 ng ml�1 fluorescein diacetate (FDA) and incubated for30 min. The microspheres were then washed with PBS and ob-served under a CLSM. In the process, FDA penetrated throughthe cell membranes and was hydrolyzed into fluorescein by via-ble cells, which were then excited at 488 nm under the CLSM[48].

2.11.4. Determination of cell attachment efficiency and proliferationAfter the OCT-1 cells had been seeded on various PLGA/HA(50/

50) composite microspheres and incubated for 6 h, the micro-spheres were collected and rinsed with PBS three times. The cellsattached to the microspheres were then digested by trypsin/EDTAand counted. Finally, the cell attachment efficiency was calculatedaccording to the following formula:

cell attachment efficiency ð%Þ ¼ ðN1=N0Þ � 100

where N1 and N0 are the number of attached cells and seeded cells,respectively.

Cell proliferation was similarly determined by counting thenumber of cells on various PLGA/HA(50/50) composite micro-spheres after culture for 1, 4 and 7 days.

2.11.5. MTT assayAfter OCT-1 cells had been cultured on various PLGA/HA(50/

50) composite microspheres for 1, 4 and 7 days, the viability

458 H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465

and proliferation of OCT-1 cells were determined by MTT assay.At each predetermined interval, 300 ll of fresh culture mediumwas added to each flask after the original culture medium wasremoved. Then 8 ll of MTT solution (5 mg ml�1 in PBS) was

Fig. 1. Surface morphology of PLGA/HA(50/50) composite microspheres after different tr0.5 h; (e and f) 1 h; (g and h) 2 h; (i and j) 3 h. The arrows indicate the HA.

added to each flask, followed by incubation at 37 �C with 5%CO2 for 4 h. The upper medium was removed carefully and theintracellular formazan was dissolved in 300 ll of 0.04 mol l�1

HCl/isopropanol, which was added to each flask. The absorbance

eatment times (the left was � 800 and the right was � 2000). (a and b) 0 h; (c and d)

Table 1Surface element composition of PLGA/HA(50/50) composite microspheres treatedwith 0.25 M NaOH/C2H5OH (v/v = 1:1) for different times at 37 �C.

Treatment time (h) Element content (wt.%)

C O P Ca

0 53.09 ± 4.06 42.15 ± 3.06 1.43 ± 1.02 3.33 ± 1.190.5 44.82 ± 3.01 40.05 ± 3.19 4.60 ± 1.8 8 10.53 ± 2.031 34.27 ± 2.59 33.87 ± 2.79 8.81 ± 1.94 23.04 ± 3.402 18.61 ± 2.70 26.32 ± 2.11 14.91 ± 2.38 40.16 ± 4.893 4.21 ± 1.09 30.19 ± 2.69 16.29 ± 2.21 49.31 ± 4.19

H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465 459

of produced formazan was measured at 570 nm with micro-platereader (ZS-2, Beijing).

2.11.6. Alkaline phosphatase assayPLGA/HA(50/50) composite microspheres with OCT-1 cells were

collected from the flask at various culture time intervals (3, 7, 10and 14 days). The microspheres were washed with PBS and thensuspended in 0.1% Triton X-100. After the OCT-1 cells on the micro-spheres had been frozen and thawed repeatedly, the suspensionsolution was centrifuged and the supernatants collected to mea-sure the ALP activity.

ALP activity was determined by using disodium phenyl phos-phate as the substrate at pH 10. Each reaction was initiated withdisodium phenyl phosphate and allowed to proceed for 15 min at37 �C, before the developer potassium ferricyanide was quicklyadded. The absorbance at 520 nm was measured with an ultravio-let spectrophotometer (752-type ultraviolet grating spectropho-tometer, Shanghai). The ALP activity was normalized by totalintracellular protein synthesis and thus expressed as units pergram of protein. Total protein content was determined at 595 nmusing a protein assay kit (Nanjing Jianchen Bioengineering Insti-tute, China) according to manufacturer’s instructions.

2.12. Statistical analysis

The data were expressed as means ± standard deviations (SD)(n = 3 or 4). Statistical comparisons were performed using Stu-dent’s t-test or a two-way analysis of variance using Turkey’s Mul-tiple Comparison test. p-values < 0.05 were considered statisticallysignificant.

3. Results and discussion

3.1. Morphology of PLGA/HA(50/50) composite microspheres

The surface topography structure of the PLGA/HA (50/50) com-posite microspheres before and after the alkali treatment was ob-served by SEM and are shown in Fig. 1. It can be noted that,compared with the relatively smooth surface of UT-PLGA/HA(Fig. 1a and b), there were more broken points and cracks inwhich HA could be clearly observed on surfaces of the alkali-trea-ted microspheres (Fig. 1d, f, h and j). Moreover, the surface rough-ness of the alkali-treated PLGA/HA(50/50) microspheres hadincreased with increasing alkali treatment time. The surface etch-ing of the composite microspheres should be attribute to thehydrolysis of PLGA on the surface of the microspheres in the0.25 M NaOH aqueous solution and ethanol (v/v = 1/1) at 37 �C.

Table 2Hydrophilicity of PLGA/HA(50/50) composite microspheres treated with 0.25 M NaOH/C2H

Treatment time (h) 0 0.5Water uptake (%) 10.10 ± 3.09 29.72 ± 4.12

It could be further observed that, in a low concentration alkalisolution (0.25 M) and less than 2 h of treatment time, only theouter layer of the composite microspheres had been treated andthe bulk of microspheres had not been affected (Fig. 1c, e andg), so the PLGA/HA(50/50) composite microspheres still kept theiroriginal shape. However, when the treatment time was longerthan 2 h, the microspheres were broken down and the PLGA/HA(50/50) composite microspheres changed to an irregular shape(Fig. 1i). Therefore, to keep the regular shape of the PLGA/HA(50/50) composite microspheres, the alkali treatment time should belimited to within 2 h.

3.2. Surface composition of PLGA/HA(50/50) composite microspheres

The surface element content of the PLGA/HA(50/50) compositemicrospheres before and after treatment was examined by EDS, asshown in Table 1. The results showed that the surface calcium andphosphorus contents of the UT-PLGA/HA were only 3.33 and 1.43%,respectively, but they had drastically increased after alkali treat-ment. After treating for 0.5 h, the content of calcium and phospho-rus had increased to 10.53 and 4.60%, respectively. With furthertreatment time up to 2 h, the surface element content of PLGA/HA(50/50) composite microspheres reached 40.16% for calciumand 14.91% for phosphorus, respectively. It could be seen that theCa/P ratio calculated from the EDS measurement was not consis-tent with the hydroxyapatite content. It was considered that theCa/P ratio of the microspheres’ surface had been influenced bymany factors, including the smoothness of the microspheres. Fromthe above results, it could be concluded that the HA content of themicrospheres’ surface would be increased after treatment. The HAhad been exposed to the surface of the microspheres by theremoval of PLGA during the alkali etch.

3.3. Hydrophilicity of PLGA/HA(50/50) composite microspheres

The hydrophilicity of various PLGA/HA(50/50) composite micro-spheres was determined by water uptake, as shown in Table 2. Itcan be seen that the water uptake of UT-PLGA/HA was only10.10%, whereas that of the alkali-treated PLGA/HA(50/50) com-posite microspheres had been enhanced remarkably with increas-ing treatment time. After 2 h of treatment time, the water uptakeof PLGA/HA(50/50) composite microspheres had reached 56.63%,which was over five times of that of the UT-PLGA/HA. This resultcould be attributed to exposure of the hydrophilic component –HA – and the hydrophilic carboxyl and hydroxyl groups, whichwere produced by cleavage of the ester bonds of the PLGA, as wellas the formation of a porous structure in the compositemicrospheres.

3.4. Determination of alkaline treatment time of PLGA/HA(50/50)composite microspheres

The effect of alkaline treatment time on the molecular weight ofthe PLGA component of PLGA/HA(50/50) composite microsphereswas examined by GPC, as shown in Table 3. The results indicatethat Mw, Mn and molecular weight distribution (Mw/Mn) did notapparently change within 2 h of alkaline treatment. Therefore, itcan be inferred that only the surface layer of the microsphereshad been modified while the bulk property of the microspheres

5OH (v/v = 1:1) for different times at 37 �C.

1 2 340.51 ± 3.22 56.63 ± 5.78 68.81 ± 4.90

40

50

60

70

80

90*#*#

*

men

t effi

cien

cy (%

)

Table 3Molecular weight of the PLGA component in PLGA/HA(50/50) composite micro-spheres treated with 0.25 M NaOH/C2H5OH (v/v = 1:1) for different times at 37 �C.

Treatment time (h) Mw (�104) Mn (�104) Mw/Mn

0 11.20 6.30 1.780.5 11.73 6.77 1.731 11.38 6.69 1.702 11.05 6.28 1.763 9.90 5.50 1.80

Table 4Change in molecular weight of the PLGA component and mass loss of UT-PLGA/HAwith degradation time.

Degradation time (weeks) Mw (�104) Mn (�104) Mw/Mn Mass loss (%)

0 11.20 6.30 1.78 –1 10.86 6.24 1.74 0.13 ± 0.062 10.58 6.19 1.71 3.29 ± 0.743 9.89 6.03 1.64 3.57 ± 0.574 8.82 5.80 1.52 4.24 ± 0.80

0 5 10 15 20 250

10

20

30

40

50

60

70

80

90

Cum

ulat

ive

fract

ion

(%)

Time (days)

Fig. 2. Cumulative release of rhBMP-2 from the T2-PLGA/HA/BMP, where 500 ng ofrhBMP-2 was loaded into 10 mg of microspheres.

460 H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465

had not been altered within 2 h of treatment. Moreover, it was no-ticed that the molecular weight of the PLGA component of PLGA/HA(50/50) composite microspheres treated for 0.5 or 1 h was a lit-tle higher than that of untreated microspheres. It might be that thelow molecular weight PLGA on the surface of composite micro-spheres was first hydrolyzed and dissolved in the treating solution.However, after 3 h of treatment, the molecular weight of PLGAcomponent of PLGA/HA(50/50) composite microspheres had obvi-ously decreased. This result revealed that the bulk of compositemicrospheres had been modified by alkali treatment. Thus, treat-ment for 2 h in a mixture of 0.25 M NaOH aqueous solution andethanol (v/v = 1/1) at 37 �C is the optimal condition for improvingthe surface properties of the composite microspheres withoutaltering their bulk properties.

3.5. Change in the mass of microspheres and the molecular weight ofthe PLGA component during degradation

Mass loss of UT-PLGA/HA and T2-PLGA/HA, and of the molecu-lar weight of the PLGA component in both of them, during degra-dation in vitro were measured and are summarized in Tables 4and 5, respectively. It can be seen that the loss of mass of boththe microspheres increased with increasing degradation time, butthere was no obvious difference within 4 weeks. The reasons forthe loss of mass of the microspheres were that the degraded prod-ucts of the PLGA component and a part of the HA component weredissolved in the degradation medium At the same time, the Mw, Mn

and Mw/Mn of the PLGA component in both the microspheres grad-ually decreased with increasing degradation time. This phenome-non could be induced by the degradation of the PLGA and theproduction of lower molecular weight PLGA. However, the molec-ular weight of the PLGA component of T2-PLGA/HA decreased fas-ter than that of UT-PLGA/HA. For UT-PLGA/HA, the Mw decreasedfrom 11.20 � 104 to 8.82 � 104 within 4 weeks, but for T2-PLGA/

Table 5Change in molecular weight of the PLGA component and mass loss of T2-PLGA/HAwith degradation time.

Degradation time (weeks) Mw (�104) Mn (�104) Mw/Mn Mass loss (%)

0 11.05 6.28 1.76 –1 10.54 6.16 1.71 0.11 ± 0.072 10.08 6.08 1.66 3.31 ± 0.633 9.27 5.84 1.59 3.63 ± 0.604 7.98 5.42 1.47 4.57 ± 0.58

HA it decreased from 11.05 � 104 to 7.98 � 104 in the same time.These results might be attributed to the porous surface structureand the better hydrophilicity of the T2-PLGA/HA.

3.6. rhBMP-2 release behavior of T2-PLGA/HA/BMP

T2-PLGA/HA possesses several advantages for loading rhBMP-2.First, the improved hydrophilicity of the treated PLGA/HA compos-ite microspheres can permit more rhBMP-2 water solution to becompletely absorbed and diffused into the microspheres, whichwould ensure that rhBMP-2 is evenly dispersed on the micro-spheres. Secondly, the richer HA on the surface of the treatedmicrospheres has an inherent capacity to bind rhBM-2, becauseHA has a high binding affinity for proteins [36,49–51]. Thirdly,the polar hydroxyl and carboxylic acid terminal groups on the trea-ted microspheres, which were generated by the breaking of the es-ter bond of PLGA during alkali treatment, can provide many sitesfor the capture of rhBMP-2 by polar interaction and hydrogenbonding [52]. The loaded rhBMP-2 on the T2-PLGA/HA exhibitedsustained release, as shown in Fig. 2. A moderate burst releasecan be seen during the first day, with 21.99% of the rhBMP-2 beingreleased. After the moderate burst release, rhBMP-2 was continu-ously released from T2-PLGA/HA/BMP such that the cumulative ex-

0

10

20

30

T1-PLGA/HAUT-PLGA/HA T2-PLGA/HA T2-PLGA/HA/BMP

Atta

ch

Fig. 3. Attachment efficiency of OCT-1 cells on various PLGA/HA(50/50) compositemicrospheres cultured for 6 h. *p < 0.05: significant against the attachmentefficiency of UT-PLGA/HA. #p < 0.05: significant against the attachment efficiencyof T1-PLGA/HA.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

741

*#*

*#*#*#

**#*#

*

Abso

rban

ce

Culture time (days)

UT-PLGA/HA T1-PLGA/HA T2-PLGA/HA T2-PLGA/HA/BMP

Fig. 5. MTT–tetrazolium assay of OCT-1 cells cultured on various PLGA/HA(50/50)composite microspheres for different time periods. *p < 0.05: significant against theproliferation and viability of OCT-1 cells on UT-PLGA/HA at the corresponding day.#p < 0.05: significant against proliferation and viability of OCT-1 cells on T1-PLGA/HA at the corresponding day.

0 1 2 3 4 5 6 7 8

456789

101112131415161718192021

Cel

l num

bers

( *10

4 cells

/10m

g)

UT-PLGA/HA T1-PLGA/HA T2-PLGA/HA T2-PLGA/HA/BMP

Time (days)

*#

**

*

*

*#

*#

*#

Fig. 4. OCT-1 cell number on various PLGA/HA(50/50) composite microspherescultured for different time. *p < 0.05: significant against the cell number of UT-PLGA/HA. #p < 0.05: significant against the cell number of T1-PLGA/HA.

H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465 461

tent of rhBMP-2 released after 21 days was 79.31%. This sustainedrelease from the PLGA/HA composite microspheres could be attrib-uted to the high binding affinity between HA and rhBMP-2, as wellas the polar interaction and hydrogen bonding between PLGA andrhBMP-2. Hence, alkali treatment appeared to have a profound ef-fect on the loading of rhBMP-2 onto the PLGA/HA compositemicrospheres.

Fig. 6. CLSM images of OCT-1 cells cultured on various PLGA/HA(50/50) composite micrPLGA/HA/BMP.

3.7. Cell affinity of PLGA/HA(50/50) composite microspheres

OCT-1 osteoblast-like cell was used as a model cell to becultured on the UT-PLGA/HA, T1-PLGA/HA, T2-PLGA/HA andT2-PLGA/HA/BMP in vitro. The cell affinity of the four kinds ofmicrospheres was identified and compared by measuring cell

ospheres for 7 days. (a) UT-PLGA/HA; (b) T1-PLGA/HA; (c) T2-PLGA/HA; and (d) T2-

462 H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465

attachment, proliferation, morphology, distribution and expressionof differentiated osteoblastic function (ALP) on the microspheres.

3.7.1. Attachment and proliferation of OCT-1 cells on PLGA/HA(50/50)composite microspheres

The attachment efficiency of OCT-1 cells on various PLGA/HA(50/50) composite microspheres after 6 h cell culture was sum-marized in Fig. 3. It could be seen that all the microspheres allowedadhesion of the OCT-1 cells on them; however, the cell attachmentefficiencies on the various PLGA/HA(50/50) composite micro-spheres were different. The OCT-1 cell attachment efficiency onUT-PLGA/HA was 46.2%, which was obviously less than that onthe other microspheres. The OCT-1 cell attachment efficiency onthe T2-PLGA/HA and T2-PLGA/HA/BMP was especially high, reach-

Fig. 7. Morphology of OCT-1 cells cultured on various PLGA/HA(50/50) composite microHA; (c and d) T1-PLGA/HA; (e and f) T2-PLGA/HA; (g and h) T2-PLGA/HA/BMP. The arro

ing 77.3 and 78.4% respectively, which was much higher than thaton the T1-PLGA/HA (65.7%).

The OCT-1 cell numbers on various PLGA/HA(50/50) compositemicrospheres after cultured for different time are compared inFig. 4. It can be seen that all the types of composite microspheresallowed the proliferation of OCT-1 cells on their surface, but withinthe culture period in vitro the numbers of cells on the various PLGA/HA(50/50) composite microspheres were different. The prolifera-tion rate on the UT-PLGA/HA was the slowest, the cell number beingthe smallest after culture for 7 days, at only 8.6 � 104 cells/10 mg.By contrast, the proliferation rate of OCT-1 cells on the T1-PLGA/HA was faster than on the UT-PLGA/HA, increasing from 6.6 � 104

to 13.7 � 104 cells/10 mg in 7 days. The proliferation rates of OCT-1 cells on the T2-PLGA/HA and T2-PLGA/HA/BMP microspheres

spheres for 7 days (the left was � 400, and the right was � 800). (a and b) UT-PLGA/ws indicate the cells.

0

4000

8000

12000

16000

20000

24000

28000

*

*#$

*

*

*#

*#$*#$

*#

**#

**#

*#

141073

UT-PLGA/HA T1-PLGA/HA T2-PLGA/HA T2-PLGA/HA/BMP

ALP

activ

ity (U

/g p

rote

in)

Culture time (days)

Fig. 8. ALP activity assay of OCT-1 cells cultured on various PLGA/HA(50/50)composite microspheres for different time periods. ALP activity was determined asenzyme activity units (U) per gram of protein.*p < 0.05: significant against ALPactivity of OCT-1 cells on UT-PLGA/HA at the corresponding day. #p < 0.05:significant against ALP activity of OCT-1 cells on T1-PLGA/HA at the correspondingday. $p < 0.05: significant against ALP activity of OCT-1 cells on T2-PLGA/HA at thecorresponding day.

H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465 463

were the fastest, with the numbers of cells after 7 days of incuba-tion being twice those after the first day of incubation.

3.7.2. Viability and distribution of OCT-1 cells on PLGA/HA(50/50)composite microspheres

The viability of OCT-1 cells on the various PLGA/HA(50/50)composite microspheres was determined by MTT assay, whichmeasured the activity of mitochondria dehydrogenase of livingcells after culturing for 1, 4 and 7 days, as shown in Fig. 5. It canbe seen that the number of viable OCT-1 cells cultured on the trea-ted PLGA/HA(50/50) composite microspheres was higher than thaton the UT-PLGA/HA. Moreover, the number of viable OCT-1 cells onthe T2-PLGA/HA and T2-PLGA/HA/BMP was higher than that on theT1-PLGA/HA.

The CLSM observation further confirmed the existence of viableOCT-1 cells on the surface of PLGA/HA(50/50) composite micro-spheres after incubating for 7 days. The round black regions inFig. 6 are microspheres, while the green bright regions representviable OCT-1 cells stained by FDA. It can be seen that the OCT-1cells could grow on all the microspheres regardless of their surfacecontent of HA, but the cell numbers and distribution on the fourkinds of PLGA/HA(50/50) composite microspheres were different.The OCT-1 cells tended to attach and grow in the gaps betweenthe UT-PLGA/HA microspheres (Fig. 6a). However, the cells weredistributed more densely and more evenly on the treated PLGA/HA(50/50) composite microspheres (Fig. 6b–d). Moreover, the cellsgrew and proliferated along the whole surface of treated PLGA/HA(50/50) composite microspheres.

3.7.3. Morphology of OCT-1 cells on PLGA/HA(50/50) compositemicrospheres

The SEM photographs of OCT-1 cells incubated on the four kindsof PLGA/HA(50/50) composite microspheres for 7 days are shownand compared in Fig. 7. It can be seen that the OCT-1 cells culturedon all PLGA/HA(50/50) composite microspheres bridged betweenthe microspheres with elongated shapes (Fig. 7a, c, e and g). Manycells on all the PLGA/HA(50/50) composite microspheres hadspread and maintained physical contact with each other throughfilopodia or lamellopodia within 7 days (Fig. 7b, d, f and h). How-ever, compared with UT-PLGA/HA (Fig. 7a and b) and T1-PLGA/HA (Fig. 7c and d), more polygonal and spindle-shaped cells at-tached and spread well on the T2-PLGA/HA (Fig. 7e and f) andT2-PLGA/HA/BMP (Fig. 7g and h).

Connecting the results of the cell distribution and morphologi-cal observations with the results of the cell attachment, prolifera-tion and viability revealed that the treated PLGA/HA(50/50)composite microspheres were more favorable for OCT-1 cellattachment and growth than the UT-PLGA/HA. Moreover, cellattachment and growth on T2-PLGA/HA and T2-PLGA/HA/BMPwere better than T1-PLGA/HA. It is considered that the improvedcell attachment and growth by alkali treatment could be explainedfrom two aspects – the surface topography and surface chemistryof the PLGA/HA(50/50) composite microspheres. The surfacetopography may be an important factor affecting cell attachmentand growth, with a rough surface being more favorable for cellattachment and growth [53–55]. As shown in Fig. 1, the surfaceof PLGA/HA(50/50) composite microspheres became rougher withincreasing treatment time. More importantly, the surface chemis-try of the PLGA/HA(50/50) composite microspheres changed dueto the alkali treatment. In this study, the EDS data (Table 1) showthat increasing the alkali treatment time enhanced the HA contenton the surface of PLGA/HA(50/50) composite microspheres, whichnot only led directly to an increase in hydrophilicity, but also intro-duced a component suitable for cell attachment and growth.Hydrophilicity as a main characteristic of biomaterials plays animportant role in determining the cell affinity of the materials.

Many studies have shown that a moderate increase in hydrophilic-ity is favorable for cell attachment and growth [56–58]. However,there was no obvious difference in the cell attachment, morphol-ogy, viability and proliferation between T2-PLGA/HA and T2-PLGA/HA/BMP after the cells were cultured on them for 7 days.This means that the 50 ng/mg rhBMP-2-loaded onto the T2-PLGA/HA had no obvious effect on these parameters within 7 days.Whether a dose-dependent effect of loaded rhBMP-2 on cellattachment, morphology, viability and proliferation exists requiresfurther investigation.

3.7.4. Differentiation of OCT-1 cells on PLGA/HA(50/50) compositemicrospheres

The differentiated function of OCT-1 cells was evaluated by mon-itoring the ALP activity of the cells. ALP is a marker enzyme of OCT-1cells and probably involves a mineralization process. As shown inFig. 8, the activity of ALP increased significantly within 14 days onall of the PLGA/HA(50/50) composite microspheres. However, thecells on the treated PLGA/HA(50/50) composite microspheresshowed higher activity than those on the UT-PLGA/HA. Furthermore,the ALP activity of cells on the T2-PLGA/HA was higher than that onthe T1-PLGA/HA. The ALP activity of OCT-1 cells on the T2-PLGA/HA/BMP was the highest of all the PLGA/HA(50/50) composite micro-spheres. The results showed that the ALP activity of OCT-1 cells onthe treated PLGA/HA(50/50) composite microspheres was enhancedand increased further with increasing treatment time. This was be-cause osteoconductive HA on the surface of PLGA/HA(50/50) com-posite microspheres gradually increased with increasingtreatment time. On the other hand, the highest ALP activity ofOCT-1 cells on the T2-PLGA/HA/BMP should attribute to therhBMP-2 being evenly loaded on the T2-PLGA/HA and then undergo-ing sustained release. Therefore, although the loading of 50 ng/mgrhBMP-2 on the T2-PLGA/HA had not obvious effect on the OCT-1cell attachment, morphology, viability and proliferation, the osteo-inductive activity of the loaded rhBMP-2 had an obvious promotingeffect on osteoblast differentiation.

High-magnification SEM micrographs of OCT-1 cells culturedfor 2 weeks on the four kinds of microspheres revealed consistentresults of ALP activity, as shown in Fig. 9. It can be seen that glob-ular accretions, which are indicative of calcification [59], were

Fig. 9. Scanning electron micrographs of OCT-1 cells cultured for 2 weeks on various PLGA/HA(50/50) composite microspheres (�10,000). (a) UT-PLGA/HA; (b) T1-PLGA/HA;(c) T2-PLGA/HA; and (d) T2-PLGA/HA/BMP. The arrows indicate the globular accretions.

464 H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465

fewest and smallest on the UT-PLGA/HA (Fig. 9a). There were morecalcified globular accretions on the T2-PLGA/HA (Fig. 9b) than onthe T1-PLGA/HA (Fig. 9c). Among all of the microspheres, the calci-fied globular accretions on the T2-PLGA/HA/BMP (Fig. 9d) were themost numerous and largest. The results indicate that the osteoin-ductivity of rhBMP-2 combined with the osteoconductivity of HAaccelerated the calcification of osteoblasts on the PLGA/HA(50/50) composite microspheres.

4. Conclusions

Alkali treatment is an effective way to modify the surface ofPLGA/HA(50/50) composite microspheres. With increasing treat-ment time, the HA content of the surface, the roughness and thehydrophilicity of the PLGA/HA(50/50) microspheres all graduallyincreased. For the PLGA/HA(50/50) composite microspheres tomaintain their regular shape, the alkali treatment time should belimited to within 2 h. The decrease in molecular weight of PLGAcomponent of T2-PLGA/HA was faster than that of UT-PLGA/HAduring the first 4 weeks of degradation. Since cell adhesion andgrowth were influenced by hydrophilicity, roughness and the HAcomponent, which changed with the duration of the alkalitreatment, a 2 h treatment time was chosen as the optimal timefor enhancing cell affinity of the PLGA/HA(50/50) compositemicrospheres.

rhBMP-2 could be loaded on the T2-PLGA/HA by a simple solu-tion dipping method. The loaded rhBMP-2 could be slowly releasedfrom the T2-PLGA/HA/BMP in vitro and enhanced the differentia-tion of osteoblasts on the composite microspheres. rhBMP-2-loaded PLGA/HA composite microspheres, possessing good cellaffinity and osteogenic potential, would make promising injectablescaffolds.

Acknowledgement

This research was supported by a grant from Major State BasicScience Research and Development Program of China (973,No.2005CB5227074).

Appendix A. Figures with essential colour discrimination

Certain figures in this article, particularly Figures 4 and 6, are dif-ficult to interpret in black and white. The full colour images can befound in the on-line version, at doi:10.1016/j.actbio.2009.07.016).

References

[1] Parrish FF. Homografts of bone. Clin Orthop Relat Res 1972;87:36–42.[2] Brown KLB, Cruess RL. Bone and cartilage transplantation in orthopaedic

surgery: a review. J Bone Joint Surg Am 1982;64:270–9.[3] DeBoer HH. The history of bone grafts. Clin Orthop Relat Res 1988;226:292–8.[4] Bonfiglio M, Jeter WS. Immunological response to bone. Clin Orthop

1972;87:19–27.[5] Griffon DJ, Junlop DG, Howie CR, Pratt JN, Gilchrist TJ, Simth N. An ovine model

to evaluate the biological properties of impacted morselized bone graftsubstitutes. J Biomed Mater Res 2001;56:444–51.

[6] Laurencin CT, Attawia MA, Lu LQ, Borden MD, Lu HH, Gorum WJ, et al.Poly(lactide-co-glycolide)/hydroxyapatite delivery of BMP-2-produccing cells:a regional gene therapy approach to bone regeneration. Biomaterials2001;22:1271–7.

[7] Rizzi SC, Heath DJ, Coombes AG, Bock N, Textor M, Downes S. Biodegradablepolymer/hydroxylapatite composites: surface analysis and initial attachmentof human osteoblasts. J Biomed Mater Res 2001;55:475–86.

[8] Tanizawa Y, Sawamura K, Suzuki T. Reaction characteristics of dental andsynthetic apatites with AL-3+ and LA-3+ ions in acidic solutions. J Chem SocFaraday Trans 1990;86:4025–9.

[9] Laurencin CT, Attawia M, Borden MD. Advancements in tissue engineered bonesubstitutes. Curr Opin Orthop 1999;10:445–51.

[10] Van LP, Li F, Keustermans JP, Streydio JM, Delannay F, Munting E. The influence ofhigh sintering temperatures on the mechanical properties of hydroxyapatite. JMater Sci Mater Med 1995;6:8–13.

[11] Pohunkova H, Adam M. Reactivity and the fate of some composite bioimplantsbased on collagen in connective tissue. Biomaterials 1995;16:67–71.

[12] Wang M. Developing bioactive composite materials for tissue replacement.Biomaterials 2003;24:2133–51.

[13] Boccaccini AR, Blaker JJ. Bioactive composite materials for tissue engineeringscaffolds. Expert Rev Med Devices 2005;2:303–17.

[14] Zhao F, Yin YJ, William W, Lu J, Leong CY, Zhang WY, et al. Preparation andhistological evaluation of biomimetic three-dimensional hydroxyap-atite/chitosan–gelatin network composite scaffolds. Biomaterials 2002;23:3227–34.

[15] Yunoki S, Ikoma T, Monkawa A, Ohta K, Kikuchi M, Sotome S, et al. Control ofpore structure and mechanical property in hydroxyapatite/collagen compositeusing unidirectional ice growth. Mater Lett 2006;60:999–1002.

[16] Yang F, Cui WJ, Xiong Z, Liu L, Bei JZ, Wang SG. Poly(L-lactide-co-glycolide)/tricalcium phosphate composite scaffold and its various changes duringdegradation in vitro. Polym Degrad Stabil 2006;91:3065–73.

H. Shen et al. / Acta Biomaterialia 6 (2010) 455–465 465

[17] Zhou ZH, Ruan JM, Zhou ZC, Zou JP. Synthesis and properties of compositebiomaterials based on hydroxyapatite and poly(L-lactide). Polym-Plast Technol2008;47:496–501.

[18] Jose MV, Thomas V, Johnson KT, Dean DR, Nyairo E. Aligned PLGA/HAnanofibrous nanocomposite scaffolds for bone tissue engineering. ActaBiomater 2009;5:305–15.

[19] Sui G, Yang XP, Mei F, Hu XY, Chen GQ, Deng XL, et al. Poly-L-lactic acid/hydroxyapatite hybrid membrane for bone tissue regeneration. J BiomedMater Res 2007;82A:445–54.

[20] Marra KG, Szem JW, Kumta PN, DiMilla PA, Weiss LE. In vitro analysis ofbiodegradable polymer blend/hydroxyapatite composites for bone tissueengineering. J Biomed Mater Res 1999;47:324–35.

[21] Kothapalli CR, Shaw MT, Wei M. Biodegradable HA-PLA 3-D porous scaffolds:effect of nano-sized filler content on scaffold properties. Acta Biomater2005;1:653–62.

[22] Hasegawa S, Ishii S, Tamura J, Furukawa T, Neo M, Matsusue Y, et al. A 5–7year in vivo study of high-strength hydroxyapatite/poly(L-lactide)composite rods for the internal fixation of bone fractures. Biomaterials2006;27:1327–32.

[23] Ishii S, Tamura J, Furukawa T, Nakamura T, Matsusue Y, Shikinami Y, et al.Long-term study of high-strength hydroxyapatite/poly(L-lactide) compositerods for the internal fixation of bone ractures: a 2–4-year follow-up study inrabbits. J Biomed Mater Res 2003;66B:539–47.

[24] Wei G, Ma PX. Structure and properties of nano-hydroxyapatite/polymercomposite scaffolds for bone tissue engineering. Biomaterials 2004;25:4749–57.

[25] Zhang R, Ma PX. Poly(a-hydroxyl acids)/hydroxyapatite porous composites forbone–tissue engineering. I. Preparation and morphology. J Biomed Mater Res1999;44:446–55.

[26] Kim SS, Ahn KM, Park MS, Lee JH, Choi CY, Kim BS. A poly(lactide-co-glycolide)/hydroxyapatite composite scaffold with enhanced osteoconductivity. J BiomedMater Res 2007;80A:206–15.

[27] Kim SS, Park MS, Jeon O, Chio CY, Kim BS. Poly(lactide-co-glycolide)/hydroxyapatite composite scaffolds for bone tissue engineering. Biomaterials2006;27:1399–409.

[28] Taboas JM, Maddox RD, Krebsbach PH, Hollister SJ. Indirect solid free formfabrication of local and global porous, biomimetic and composite 3D polymer–ceramic scaffolds. Biomaterials 2003;24:181–94.

[29] Qiu X, Han Y, Zhuang X, Chen X, Li Y, Jing X. Preparation of nano-hydroxyapatite/poly(L-lactide) biocomposite microspheres. J Nanopart Res2007;9:901–8.

[30] Shi X, Wang Y, Ren L, Gong Y, Wang DA. Enhancing alendronate release from anovel PLGA/hydroxyapatite microspheric system for bone repairingapplications. Pharm Res 2009;26:422–30.

[31] Ono I, Yamashita T, Jin HY, Ito Y, Hamada H, Akasaka Y, et al. Combination ofporous hydroxyapatite and cationic liposomes as a vector for BMP-2 genetherapy. Biomaterials 2004;25:4709–18.

[32] Habibovic P, Yuan H, van der Valk CM, Meijer G, van Blitterswijk CA, de GrootK. 3D microenvironment as essential element for osteoinduction bybiomaterials. Biomaterials 2005;26:3565–75.

[33] van den Dolder J, de Ruijter AJE, Spauwen PHM, Jansen JA. Observations on theeffect of BMP-2 on rat bone marrow cells cultured on titanium substrates ofdifferent roughness. Biomaterials 2003;24:1853–60.

[34] Baldwin SP, Saltzman WM. Materials for protein delivery in tissue engineering.Adv Drug Deliv Rev 1998;33:71–86.

[35] Xiong Z, Yan YN, Wang SG, Zhang RJ, Zhang C. Fabrication of porous scaffoldsfor bone tissue engineering via low-temperature deposition. Scripta Mater2002;46:771–6.

[36] Rai B, Teoh SH, Hutmacher DW, Cao T, Ho KH. Novel PCL-based honeycombscaffolds as drug delivery systems for rhBMP-2. Biomaterials 2005;26:3739–48.

[37] Kloss FR, Gassner R, Preiner J, Ebner A, Larsson K, Hächl O, et al. The role ofoxygen termination of nanocrystalline diamond on immobilisation of BMP-2and subsequent bone formation. Biomaterials 2008;29:2433–42.

[38] Chung YI, Ahn KM, Jeon SH, Lee SY, Lee JH, Tae G. Enhanced bone regenerationwith BMP-2 loaded functional nanoparticle–hydrogel complex. J ControlRelease 2007;121:91–9.

[39] Jeon O, Song SJ, Kang SW, Putnam AJ, Kim BS. Enhancement of ectopic boneformation by bone morphogenetic protein-2 released from aheparin-conjugated poly(L-lactic-co-glycolic acid) scaffold. Biomaterials2007;28:2763–71.

[40] Puleo DA, Kissling RA, Sheu MS. A technique to immobilize bioactive proteins,including bone morphogenetic protein-4 (BMP-4) on titanium alloy.Biomaterials 2002;23:2079–87.

[41] Bessho K, Carnes DL, Cavin R, Ong JL. Experimental studies on bone inductionusing low-molecular-weight poly (DL-lactide-co-glycolide) as a carrier forrecombinant human bone morphogenetic protein-2. J Biomed Mater Res2002;61:61–5.

[42] Kato M, Toyoda H, Namikawa T, Hoshino M, Terai H, Miyamoto S, et al.Optimized use of a biodegradable polymer as a carrier material for the localdelivery of recombinant human bone morphogenetic protein-2 (rhBMP-2).Biomaterials 2006;27:2035–41.

[43] Jeon O, Song SJ, Yang HS, Bhang SH, Kang SW, Sung MA, et al. Long-termdelivery enhances in vivo osteogenic efficacy of bone morphogenetic protein-2compared to short-term delivery. Biochem Biophys Res Commun2008;369:774–80.

[44] Xiao YT, Xiang LX, Shao JZ. Bone morphogenetic protein. Biochem Biophys ResCommun 2007;362:550–3.

[45] Gilding DK, Reed AM. Biodegradable polymers for use in surgery-polyglycolic/polyactic acid homo and copolymers. Polymer 1979;20:1459–64.

[46] Yang J, Wan YQ, Tu CF, Cai Q, Bei JZ, Wang SG. Enhancing the cell affinity ofmacroporous poly(L-lactide) cell scaffold by a surface modification methodconvenient. Polym Int 2003;52:1892–9.

[47] Chen D, Chen H, Feng JQ, Windle JJ, Koop BA, Harris HA, et al. Osteoblastic celllines derived from a transgenic mouse containing the osteocalcin promoterdriving the SV-40 T-antigen. Mol Cell Diff 1995;3:193–212.

[48] Bancel S, Hu WS. Confocal scanning microscopy examination of celldistribution in macro porous microcarriers. Biotechnol Prog 1996;12:398–402.

[49] Boccacini AR, Roether JA, Hench LL, Maquet V, Jerome R. A compositesapproach to tissue engineering. Ceram Eng Sci Proc 2002;23:805–16.

[50] Matsumoto T, Okazaki M, Inoue M, Yamaguchi S, Kusunose T, Toyonaga T, et al.Hydroxypatite particles as a controlled release carrier of protein. Biomaterials2004;25:3807–12.

[51] Ziegler J, Wohlfart UM, Kessler S, Breitig D, Gunther KP. Adsorption and releaseproperties of growth factors from biodegradable polymer scaffolds. BiochemBiophys Res Commun 2002;292:144–52.

[52] Shen H, Hu XX, Bei JZ, Wang SG. The immobilization of basic fibroblast growthfactor on plasma-treated poly(lactide-co-glycolide). Biomaterials 2008;29:2388–99.

[53] Fewster SD, Coombs RRH, Kitson J, Zhou S. Precise ultrafine surface texturing ofimplant materials to improve cellular adhesion and biocompatibility.Nanobiology 1994;3:201–10.

[54] Clapper JD, Pearce ME, Guymon CA, Salem AK. Biotinylated biodegradablenanotemplated hydrogel networks for cell interactive applications.Biomacromolecules 2008;9:1188–94.

[55] Thian ES, Ahmad Z, Huang J, Edirisinghe MJ, Jayasinghe SN, Ireland DC, et al.The role of electrosprayed apatite nanocrystals in guiding osteoblastbehaviour. Biomaterials 2008;29:1833–43.

[56] Tamada Y, Ikada Y. Fibroblast growth on polymer surfaces and biosynthesis ofcollagen. J Biomed Mater Res 1994;28:783–9.

[57] Saltzman WM, Oarsons-wringerter P, Leong KW, Lin S. Fibroblast andhepatocyte behavior on synthetic polymer surface. J Biomed Mater Res1991;25:741–59.

[58] Webb K, Hlady V, Tresco PA. Relative importance of surface wettability andcharged functional groups on NIH 3T3 fibroblast attachment, spreading, andcytoskeletal organization. J Biomed Mater Res 1998;41:422–30.

[59] Li C, Vepari C, Jin HJ, Kim HJ, Kaplan DL. Electrospun silk-BMP-2 scaffolds forbone tissue engineering. Biomaterials 2006;27:3115–24.