lable at ScienceDirect

Biomaterials 32 (2011) 3404e3412

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

Cell affinity for bFGF immobilized heparin-containing poly(lactide-co-glycolide)scaffolds

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

BNLMS, State Key Laboratory of Polymer Physics & Chemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China

a r t i c l e i n f o

Article history:Received 29 December 2010Accepted 13 January 2011Available online 5 February 2011

Keywords:bFGFImmobilizationSlow releaseHeparinPLGAScaffold

* Corresponding author. Tel./fax: þ86 10 62581241E-mail address: wangsg@iccas.ac.cn (S. Wang).

0142-9612/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.biomaterials.2011.01.037

a b s t r a c t

In order to effectively and uniformly immobilize basic fibroblast growth factor (bFGF) to thick PLGAscaffold, the heparin-conjugated PLGA (H-PLGA) was synthesized at the first by reaction betweenheparin and a low molecular weight PLGA. Then heparin-containing PLGA (H-PLGA/PLGA) scaffold wasfabricated by blending the H-PLGAwith a high molecular weight PLGA. Finally, bFGF was immobilized onthe H-PLGA/PLGA scaffold mainly by static electricity action between them. The effect of H-PLGA contenton bFGF binding efficiency of the H-PLGA/PLGA scaffolds was investigated. It was found that bFGFbinding efficiency increased with increasing H-PLGA content. The bound bFGF can release in vitro slowlyfrom the H-PLGA/PLGA scaffolds and last over two weeks. The released bFGF has still preserved itsbioactivity. The attachment and growth of mouse 3T3 fibroblasts on the H-PLGA/PLGA scaffolds werebetter than that on the PLGA scaffold, however bFGF immobilized H-PLGA/PLGA scaffolds showed muchbetter cell affinity. Therefore, the method to use the H-PLGA/PLGA scaffold for immobilizing bFGF is notonly effective for slow delivering bFGF with bioactivity, but also can be used for fabricating thick scaffoldwhere bFGF could be combined and uniformly distributed.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction by aqueous solution state. In result, it could be only retained at the

Recently, the research of combination of growth factors intopolylactone-type biodegradable polymeric scaffolds had beenextensively noticed by biomaterial and tissue engineeringresearchers [1e4]. Polylactone-type biodegradable polymers, suchas poly(L-lactide) (PLLA), polyglycolide (PGA) and their copolymerpoly(lactide-co-glycolide) (PLGA) etc possess non-toxicity, lowimmunogenicity, good mechanical property, and adjustabledegradation rate [5,6], however, lack of cell recognition sites, poorhydrophilicity and lower surface energy of the polymers will affectcell to attach and grow on the polymeric scaffolds [7,8]. On theother hand, it is considered that growth factors are polypeptidesthat can transmit signals to modulate cellular activities [9].Administration of exogenous growth factors also showed potentialtherapeutic results for tissue regeneration, bone healing, woundhealing, as well as angiogenesis in vivo [10e14], and had positiveeffect on adhesion, proliferation, differentiation, migration, andgene expression of a variety of cell types in vitro [13,15e18].However, since growth factor is easy to denature in presence ofwater and under higher temperature, bioactivity of the growthfactor cannot keep for a long termwhen it is injected into the body

.

All rights reserved.

wound sites for very short duration and half-life of the growthfactor became very short caused by its susceptibility to enzymaticand thermal degradation in vivo [19e21]. Therefore, an effectivegrowth factor delivery system is required to overcome shortcomingof the growth factor for clinical therapy.

Considering most of polymers are hydrophobic, a hopefulstrategy is to combine growth factor into a hydrophobic polymerfor avoiding the growth factor to contact water by the hydrophobicpolymeric barrier. After combination of growth factor into poly-lactone-type polymeric scaffolds, the scaffolds can possess not onlyadequate mechanical strength, required biodegradation rate andmorphological structure, but also effective delivery behavior of thegrowth factor for actively guiding and accelerating cell attachment,migration, proliferation and differentiation in the scaffolds.However, the challenge of the technique is how to incorporate thewater-soluble growth factors into the hydrophobic polymericscaffolds evenly.

It was reported that growth factors can be incorporated directlyinto the polymer scaffolds during [2,22e24] or after the scaffoldfabrication [25e27]. However, because fabrication of polymerscaffolds must use organic solvent, bioactivity of the growth factorwill be damaged and reduced. On the other hand, the poorhydrophilicity and lack of functional group of the polymers oftenresult in lower growth factor loading efficiency and the growth

H. Shen et al. / Biomaterials 32 (2011) 3404e3412 3405

factor could not be tightly bound to the polymers by solutiondipping method after the scaffold fabrication.

Previously authors ever reported a plasma treatment combiningwith growth factor anchorage method, which can effectivelyimmobilize growth factor on PLGA scaffolds [28,29]. This method israpid, clean and without organic solvent pollution. However, usingthe plasma treatment method the depth of immobilized growthfactor in the scaffold is limited because of straight irradiation andweaker trans-permeability of the plasma ray. In result the innerpore surface of the thicker scaffold will be difficult to be modified,on the other hand, the outer surface of the scaffold will be possibleto deform and degrade seriously if high power and long treatmenttime are administrated [30]. Therefore, although the plasmatreatment combining with growth factor anchorage method is aneffective method for immobilizing growth factors on the PLGAscaffolds, the defect of the method is the treated thickness of thescaffolds is limited. So, the urgent requirement is to develop a moreeffective method for uniformly immobilizing growth factor intothick scaffolds.

It was reported that introducing heparin into polylactone-typescaffolds can be easy to immobilize growth factor into the scaffoldsbyabilityofheparinbindinggrowth factor [31,32].Heparin is ahighlysulfated macromolecular polysaccharide which can associate withthe cell surface and it is one component of extracellular matrix[31,33]. It is well accepted that the specific electrostatic interactionscan occur between the negatively charged sulfate groups of heparinand positively charged amino acid residues of proteins [31,33,34],and the electrostatic interaction can enhance binding affinity of theheparin to a number of growth factors such as bFGF, vascular endo-thelial growth factor (VEGF), transforming growth factor-b (TGF-b),platelet-derived growth factor (PDGF), nerve growth factor (NGF),bonemorphogenetic proteins (BMPs) and enable the growth factorsto diffuse out in a sustained manner [31,35,36].

Heparin can be introduced into the polylactone-type polymericscaffolds by physical sorption, ion reaction and covalent binding,but the combined heparin by simple physical sorption method isunstable and uneven compared with other methods. Sometimesthe stability of heparin bound on the polymeric scaffolds by ionreaction also could not meet the application demand. Althoughamong three of them, the covalent bound heparin was the moststable, it is difficult to directly conjugate large number of heparininto the polymeric scaffolds by chemical method since there werevery few functional groups (only two end groups) in backbone ofthe polymers, especially in the case of using high molecular weightpolylactone-type polymers.

In this research a modified method for immobilizing growthfactor was developed. At the first, a lower molecular weight PLGAwas used to synthesize heparin-conjugated PLGA (H-PLGA). Thenthe H-PLGA was used to blend with a high molecular weight PLGAto obtain heparin-containing PLGA (H-PLGA/PLGA). By means ofadjusting component ratio of the H-PLGA and the high molecularweight PLGA, a series of H-PLGA/PLGA scaffolds which containingdifferent content of heparinwere obtained. Finally, bFGFwas boundinto the H-PLGA/PLGA scaffolds to fabricate bFGF immobilized H-PLGA/PLGA scaffolds. The bFGF release profile of the bFGF immo-bilized H-PLGA/PLGA scaffolds as well as bioactivity of the releasedbFGF were determined, and then adhesion and growth of mouse3T3 fibroblasts in vitro in the bFGF immobilized H-PLGA/PLGAscaffolds were determined, compared and discussed.

2. Materials and methods

2.1. Materials

L-Lactide and glycolide were purchased from PURAC (the Netherlands) andpurified by recrystallization twice in dried ethyl acetate. Highmolecularweight PLGA

(Mw¼ 127,000,molar ratio of lactyl/glycotyl¼ 70/30)was prepared by ring-openingpolymerization of L-lactide and glycolide under high vacuum at 160 �C for 20 h in thepresence of stannous octoate (SIGMA, German) as catalyst (0.05 wt% of L-lactide andglycolide) [37]. Lower molecular weight PLGA (Mw ¼ 33,000, molar ratio of lactyl/glycotyl¼ 70/30) was prepared also by ring-opening polymerization of L-lactide andglycolide under 160 �C but for 30 min and concentration of stannous octoate was0.1 wt% of L-lactide and glycolide. Heparin sodium salt (Beijing Biodee BiotechnologyCo., Ltd, China, produced by porcine intestinal mucosa, 150 unit/mg), dicyclohex-ylcarbodiimide (DCC, Aldrich), 4-(dimethyl amino) pyridine (DMAP, Aldrich) and allother chemicals were used as received. The 1,4-dioxane, chloroform and otherreagents were of analytical quality and directly used without further treatment.

2.2. Synthesis of H-PLGA

H-PLGA was prepared by direct coupling reaction of DCC/DMAP chemistryaccording to the reference [38]. Briefly, 0.6 g heparin was firstly dissolved ina mixture of formamide (30 mL) and N,N-dimethylformamide (30 mL), and then0.6 g lowmolecular weight PLGAwas dissolved in a mixture of formamide (100 mL)andN,N-dimethylformamide (100mL). DCC (0.01 g) and DMAP (0.006 g) were addedto above heparin solution and stirred for 10 min. Then, the above PLGA solution wasdropped into the reaction solution system and stirred at 50 �C for 12 h undernitrogen atmosphere. After the coupling reaction, the reaction system wasconcentrated and then precipitated using excess ethanol. After the precipitate waswashed by distilled water, the precipitate was dissolved again by chloroform and theproduced solution was re-precipitated using excess ethanol. Finally, the precipitatewas filtered out off the system and dried at 35 �C for 24 h in vacuum for eliminatingthe residual solvent to obtain the H-PLGA.

2.3. Preparation of H-PLGA/PLGA film and scaffold

H-PLGA/PLGA filmwas fabricated by a solution casting technique. Firstly, a blendof H-PLGA and high molecular weight PLGA with certain weight ratio was dissolvedin chloroform to form 5 wt% composite solution. Then the blend solution was castinto a poly (tetrafluoroethylene) (PTFE) mould. After solvent had evaporated in air atroom temperature, the formed film was removed from the mould and performedde-solvent thoroughly under vacuum at room temperature for 48 h.

H-PLGA/PLGA scaffold was manufactured by an improved solideliquid phaseseparation method. Firstly H-PLGA and high molecular weight PLGA with certainweight ratio were dissolved in dioxane to form 8 wt% H-PLGA/PLGA blend solution.Then the blend solution was pushed into a column container which was full ofa certain weight of sieved NaCl granules (200e280 mm) to form a composite andthen the composite was maintained at 0 �C over 24 h to perform solideliquid phaseseparation completely. After solvent was removed by freeze-drying for 3 days, theformed matrix column was put into distilled water to leach the NaCl out. Thedistilled water was renewed every 3 h until no chloric ion could be detected bydropping of AgNO3 aqueous solution. The fabricated H-PLGA/PLGA scaffold wasdried and kept in a desiccator for usage.

2.4. Determination of contact angle and porosity

Contact angle of various films to deionized water was measured on air surface ofthe films using a FACE CA-D-type Contact Angle Meter (Kyowa Kaimenka-gaku Co.,Ltd). Ten independent determinations at different sites of a film were averaged.

Porosity of various PLGA scaffolds was determined according to a methodreported previously [39]. At first, the scaffold was cut into short column witha certain diameter (D) and height (H), and the volume (Vw) of the scaffold includingthe pore volume (Vp) and the skeleton volume of the scaffold (Vs) were calculated. Adensity bottle was filled with ethanol (density re) at 30 �C and weighed (W1). Thescaffold sample (weight Ws) was put into the above density bottle and kept at 30 �Cfor 1 h under vacuum to remove any trapped air in the pores/tubules. Then thedensity bottle was filled with ethanol and kept at 30 �C again for 30 min and all theoverflowed ethanol was cleaned away carefully. Finally the density bottle wasweighed again (W2). Parameters of the scaffold including the volume of the wholescaffold (Vw), the volume of pores (Vp), the volume of the scaffold skeleton (Vs), andthe porosity (3) of the scaffold were calculated as follows:

Vw ¼ Vp þ Vs ¼ p� ðD=2Þ2�H

Vs ¼ ðW1 þWs �W2Þ=re3 ¼ 1� Vs=Vw

2.5. Characteristics of heparin on the surface of H-PLGA/PLGA scaffold

Heparin content on the surface of H-PLGA/PLGA scaffolds was determined by thetoluidine blue colorimetric method [38,40]. The column type H-PLGA/PLGA scaffoldwith 8mm of diameter and 4 mm of height was placed in 1mL of 0.2% NaCl solution,then1mLof toluidineblue solution (0.05 gof toluidine bluewasdissolved in 1000mL

4000 3500 3000 2500 2000 1500 1000 500Wavenumber(cm-1)

PLGA

H-PLGA

Heparin

Fig. 1. FTIR spectra of H-PLGA, PLGA and heparin. The arrows indicated the maindifference in absorb peaks of H-PLGA and PLGA.

H. Shen et al. / Biomaterials 32 (2011) 3404e34123406

of 0.01 N HCl aqueous solution containing 0.2% NaCl) was added and removed innerair by decompression. After 10 min of vibration, 2.0 mL of hexane was added andmixed by vortex to allowed phase separation. Then absorbance of the aqueous layersat 631 nm was determined by a UV spectrophotometer (752 Ultraviolet GratingSpectrophotometer, Shanghai). According to a standard curve of absorbance at631 nm with different concentration of heparin prepared by the same method, theheparin content on the surface of H-PLGA/PLGA scaffolds can be quantified.

To observe and compare distribution of heparin on different H-PLGA/PLGAscaffold, a series of column type PLGA scaffolds with 8 mm of diameter and heightwere treated by the same method above. After the scaffolds were immersed intotoluidine blue for 10 min, they were washed by distilled water for removing freetoluidine blue. Finally, the scaffolds were respectively split into two pieces by thinand sharp razor blade for observing and recording inner color of the scaffolds bytaking photos.

2.6. Determination of bound bFGF

The column type PLGA scaffold and H-PLGA/PLGA scaffold with 8 mm ofdiameter and 4 mm of height were placed in a 48-well plate. 1 mL bFGF solution(500 ng/mL of bFGF (human recombinant, 154 amino acid, Mw ¼ 17.2 kDa, Pepro-TechAsia) in pH 7.4 phosphate buffer saline (PBS)) was added into each well. Afterremoving inner air by decompression, the PLGA scaffolds were incubated in thebFGF solution for 1 h at room temperature on a shaker. The supernatants werecollected respectively. Then bFGF bound scaffolds were washed with PBS for twotimes. All the washing liquid were also collected and mixed with previous super-natants respectively. The amount of bFGF in the collected mixed solution wasassayed using a Quankine Immunoassay kit according to the manufacture’sinstruction (Human bFGF Quankine ELISA kit, R & D Systems, Minneapolis, MN,USA). Binding efficiency of bFGF to the PLGA scaffolds was evaluated according to thefollowing formula:

Binding efficiency ð%Þ ¼ ½ðWa �WbÞ=Wa� � 100

where Wa and Wb are weight of bFGF respectively in PBS solution before and afterincubation of the PLGA scaffolds.

2.7. Release determination of bound bFGF

Above (Section 2.6) mentioned bFGF bound PLGA (PLGA/bFGF) and bFGF boundH-PLGA/PLGA(70/30) (H-PLGA/PLGA(70/30)/bFGF) scaffolds were respectivelyimmersed in 1 mL of pH 7.4 PBS including 1% bovine serum albumin (Sigma) at 37 �Cunder static condition to perform bFGF release test for 3 weeks. At preset timeintervals, the incubation solution was collected and renewed with fresh releasemedium. The amount of bFGF in the collected release medium was determined bythe same Quankine Immunoassay kit according to the manufacture’s instruction.

2.8. Preparation of cells

Mouse 3T3 fibroblasts were supplied by the Chinese Academy of MilitaryMedical Sciences. The cells were incubated at 37 �C in DMEM (Gibco) supplementedwith 10% fetal bovine serum (Gibco) and 100 U/cm3 each of penicillin and strepto-mycin in a 5% CO2 incubator. When the cells had grown to confluence, they weredetached by trypsin/EDTA (0.05%/0.02%) (Sigma) and seeded onto various samples.

2.9. Bioactivity assay of released bFGF

Mouse 3T3 fibroblasts at a density of 1.5 � 104 cells/well were evenly seededonto 24-well culture plate with addition of 1 mL of DMEM basal medium including5% fetal bovine serum. The column type PLGA, H-PLGA/PLGA(70/30), PLGA/bFGF andH-PLGA/PLGA(70/30)/bFGF scaffolds with 8 mm of diameter and 4 mm of heightwere respectively immersed in the DMEM basal medium of the well, and cultured at37 �C under 5% CO2 atmosphere. As the positive control, mouse 3T3 fibroblasts werecultured in 1 mL DMEM basal medium including 5% fetal bovine serum and freshbFGF added daily at the same concentration as that of the bFGF released fromH-PLGA/PLGA(70/30)/bFGF scaffolds. The culture medium was renewed daily. Atpredetermined interval, the cells on the wells were digested by trypsin/EDTA andthe number of cells in the wells was counted.

2.10. Determination of cell attachment efficiency

Various column type PLGA scaffolds with 8 mm of diameter and 4 mm of heightwere placed in a 24-well culture plate. 80 mL ofmouse 3T3 fibroblast suspensionwitha density of 8 � 105 cells/mL was seeded on them. After 3 h of cell culture in thescaffolds, the scaffolds were removed out of thewell. Cells remained in thewell weredigested by trypsin/EDTA and number of cells in the wells was counted. Finally, cellattachment efficiency of scaffolds was calculated according to the following formula:

Cell attachment efficiency ð%Þ ¼ 100ðN1 � N2Þ=N1

where N1 and N2 are the number of seeded cells and remained cells in the well,respectively.

2.11. MTT assay

Various PLGA scaffolds were placed into a 24-well culture plate and 80 mL ofmouse fibroblast suspension with a cell density of 5 � 105 cells/mL was seeded intothe scaffolds, respectively. The cell-seeded scaffolds were maintained at 37 �C under5% CO2 atmosphere for 3 h, and then 1.5 mL of culture medium was added to eachwell. At predetermined interval,15 mL ofMTTsolution (5mg/mL in PBS) was added toeach well, followed by incubation at 37 �C for 4 h to MTT formazan formation. Thenupper mediumwas carefully removed and the intracellular formazan was dissolvedby adding 800 mL of 0.04 mol/L HCl/iso-propanol to each well. The absorbance ofproduced formazanwas measured at 570 nmwith microplate reader (ZS-2, Beijing).

2.12. Observation of cell morphology

Various PLGA scaffolds were located in a 24-well culture plate. 80 mL of mousefibroblast suspension (8 � 105 cells/mL) was seeded on the scaffolds and thencultured at 37 �C under 5% CO2 atmosphere for 3 h prior to addition of 1.5 mL culturemedium into culture plate. After cultured for 2 weeks, the scaffolds were taken outfrom culture plate and washed with PBS, then fixed with 2.5% glutaraldehyde for24 h at 4 �C. After the scaffolds were dehydrated with a series of graded alcohols,dried and sputter-coated with gold, cell morphology on the scaffolds was observedby SEM (Hitachi S-4300, Japan).

2.13. Statistical analysis

The data were expressed as means � standard deviations (SD) (n ¼ 3 or 4).Statistical comparisons were performed using the Student’s t-test. p-values <0.05were considered statistically significant.

3. Results and discussion

3.1. Preparation and determination of H-PLGA

H-PLGA was prepared by direct coupling reaction of DCC/DMAPchemistry. Fourier-transformed infrared spectrometer (FTIR;Bruker EQUINOX 55, Germany) was used to confirm the reactionsduring the synthesis. FTIR spectra of PLGA, heparin and H-PLGAwere shown in Fig. 1. It could be seen that FTIR spectrum of H-PLGAcontained main characteristic peaks of PLGA and heparin, whichincluded a broader peak of NeH bond and/oreOH group stretchingvibration at 3300e3500 cm�1, CeH bond adsorption peaks at 2997and 2946 cm�1, stretching at 1756 and 1699 cm�1 that respectivelyidentified with eCOOe groups of PLGA and heparin, and a charac-teristic adsorption peak of sulfonated group (eSO3) at 1303 cm�1.

H. Shen et al. / Biomaterials 32 (2011) 3404e3412 3407

Compare to pure heparin, peak positions of heparin in the H-PLGAappeared to shift, which could derive from the influence ofPLGA coexisting in a backbone chain. However, compare to purePLGA, the peak positions of PLGA in the H-PLGA have no obviouschange. These results indicated the formation of H-PLGA polymer.The sulfur content of H-PLGA was determined to be 0.49% byelemental analyzer (Vario EL Ⅲ, Germany).

3.2. Properties of H-PLGA/PLGA film and scaffold

According to the preparation method described in section 2.3,three kinds of H-PLGA/PLGA films and scaffolds with 30/70, 50/50and 70/30 weight ratio of H-PLGA to high molecular weight PLGAwere obtained.

It was observed that all the H-PLGA/PLGA films possessed densestructure and smooth surface which were similar with that of PLGAand H-PLGA films (data not shown). The result indicated that bothcomponents of H-PLGA and high molecular weight PLGA had goodmiscibility and could mix into uniform phase in the H-PLGA/PLGA.Hydrophilicity of the H-PLGA/PLGA films was identified by deter-mination of contact angle to water of the films, as shown in Table 1.It could be seen that the contact angle to water of the H-PLGAsurface (52.4�) was smaller than that of PLGA surface, and thecontact angle of H-PLGA/PLGA surface to water increased withincreasing PLGA content.

Microscopic morphology and macroscopic appearance of thePLGA and H-PLGA/PLGA scaffolds were shown in Fig. 2 and Fig. 3a,respectively. It could be seen from the Fig. 3a that shape size ofH-PLGA/PLGA (70/30), H-PLGA/PLGA (50/50) and H-PLGA/PLGA(70/30) scaffolds was similar with that of PLGA scaffold, but theshape size of H-PLGA scaffold was smaller than that of PLGA.Moreover, the morphology structure of PLGA and H-PLGA/PLGAscaffolds has not obvious difference (Fig. 2). All of the scaffoldspossessed about 97.9% of the porosity, and uniform porous struc-ture, as well as mainly 200e280 mm of pore size which dependedon the granule size of porogen NaCl.

3.3. Distribution and quantification of heparin on the surface ofH-PLGA/PLGA scaffolds

The distribution of heparin on H-PLGA/PLGA scaffolds wasobserved by toluidine blue staining method, as shown in Fig. 3. Itcould be seen that color of PLGA scaffold was still white afterimmersed in the toluidine blue solution for 10 min, but the color ofthe H-PLGA/PLGA scaffolds was purple, and the color had graduallybecome deep with content of the H-PLGA increasing. Additionally,it could be also seen that the purple complex of toluidine blue andheparin evenly distributed in whole the scaffolds. It indicatedheparin was evenly introduced on the H-PLGA/PLGA scaffolds byblending of H-PLGA and PLGA, and the toluidine blue aqueoussolution could penetrate into inner of the H-PLGA/PLGA scaffolds tocouple with heparin to produce the purple complex, becausehydrophilicity of H-PLGA/PLGA scaffolds had been improved byintroducing the hydrophilic heparin. The result meant that water-soluble matter could enter into heparin modified PLGA scaffoldsand react with heparin, which was hardly influenced by thicknessof scaffolds.

Table 1Water contact angles of PLGA and H-PLGA/PLGA surfaces.

Polymer PLGA Mw ¼ 127,000 PLGA Mw ¼ 33,0

Contact angle to water (deg.) 78.3 � 0.6 78.0 � 0.9

The amount of heparin on the surface of H-PLGA/PLGA scaffoldswas further quantitatively determined by the toluidine blue color-imetric method. The amount of heparin on H-PLGA/PLGA scaffoldsincreased with increasing the H-PLGA content in the H-PLGA/PLGAscaffolds. The amount of heparin on the surface of H-PLGA/PLGA(30/70), H-PLGA/PLGA(50/50) and H-PLGA/PLGA(70/30) scaffolds with8 mm of diameter and 4 mm of height was 1.106 � 0.415 mg,2.149 � 0.361 mg and 3.085 � 0.380 mg, respectively.

3.4. Bind of bFGF on H-PLGA/PLGA scaffolds

The binding efficiency of bFGF to various H-PLGA/PLGA scaffoldswas determined and comparedwith that of PLGA scaffold, as shownin Fig. 4. It could be seen that the bFGF binding efficiency of H-PLGA/PLGA scaffolds was higher than that of the PLGA scaffold (22.6%),and increased from 35.3 to 71.3% with increasing the content of H-PLGA from 30 to 70%. The bFGF binding efficiency of H-PLGA/PLGA(70/30) scaffold was over four times higher than that of thePLGA scaffold. It demonstrated that introducing heparin on thePLGA scaffold could increase bFGF binding efficiency of the scaffold.

Since bFGF is a positive charge protein at physiological pH(pI ¼ 10.0), there may be strong electrostatic interactions betweenbFGF and the negative charge heparin. In the case of the H-PLGA/PLGA scaffold with higher H-PLGA content, the higher content ofthe heparin can provide more binding sites for bFGF by electrostaticinteraction, in result the binding efficiency of bFGF was enhanced.However, since each PLGAmolecule has only one terminal carboxylgroup and one terminal hydroxyl group, there were only a fewterminal carboxyl and hydroxyl groups on surface of the PLGAscaffold. As a result, the PLGA scaffold can only provided a few sitesfor binding bFGF and bFGF binding efficiency was the lowest.

On the other hand, improvement of hydrophilicity of theH-PLGA/PLGA scaffolds can also improve bFGF solution permeatinginto the H-PLGA/PLGA scaffolds for contacting the heparin, whichwould be favorable for more bFGF to be bound on the scaffolds.

3.5. bFGF release behavior of the H-PLGA/PLGA/bFGF scaffolds

The release profiles of bFGF from PLGA/bFGF and H-PLGA/PLGA(70/30)/bFGF scaffoldswere identified byELISA, as shown in Fig. 5. Itcould be seen that the PLGA/bFGF scaffold showed a rapid release ofbFGF at the first day, then followed by a very slow release. However,for the H-PLGA/PLGA(70/30)/bFGF scaffold, a continuous bFGFrelease pattern had remained over twoweeks after amoderate burstrelease. The different bFGF release profile of the two scaffolds couldbe attributed to the strong electrostatic interaction between bFGFand heparin. Since the polar O-containing groups on the PLGAscaffold were quite low, most of the bFGF would be adsorbed on thePLGA scaffold by a weak interaction, which resulted in the easilydesorption of the bFGF mainly by a quick diffusion process uponcontacting with aqueous medium. However, since many bFGFehe-parin complexes had formed by strong electrostatic interaction inthe H-PLGA/PLGA(70/30)/bFGF scaffolds, the bFGF had beenreleased slowly from H-PLGA/PLGA(70/30)/bFGF scaffold andinvolved two distinct release mechanisms. One was an initial burstrelease of the bFGF from H-PLGA/PLGA(70/30)/bFGF scaffold whichwas caused mainly by a quickly dissolving of the surface adsorbed

00 H-PLGA/PLGA H-PLGA

(30/70) (50/50) (70/30)

74.1 � 0.7 68.8 � 0.7 60.0 � 1.8 52.4 � 1.5

Fig. 2. Morphology structure of PLGA and H-PLGA/PLGA scaffolds. (a) PLGA; (b) H-PLGA/PLGA(30/70); (c) H-PLGA/PLGA(50/50); (d) H-PLGA/PLGA(70/30).

H. Shen et al. / Biomaterials 32 (2011) 3404e34123408

bFGF, and the other was a sustained bFGF release controlled bya thermodynamic equilibrium between the bFGFeheparincomplexes and free bFGF in release medium. Hence, using H-PLGA/PLGA scaffolds to bind bFGF was an effective method for controllingand sustaining bFGF release.

3.6. Bioactivity assay of the released bFGF

The biological activity of the bFGF released from various PLGAscaffolds was evaluated by measuring ability of the bFGF to stim-ulate proliferation of 3T3 fibroblasts. As shown in Fig. 6, 3T3fibroblasts in the basal medium containing PLGA or H-PLGA/PLGA

Fig. 3. Photograph of PLGA and H-PLGA/PLGA scaffolds before and after toluidine bluestaining. (a) View of scaffolds before toluidine blue staining. (b) Surface appearance ofscaffolds after toluidine blue staining; (c) Inner appearance of scaffolds after toluidineblue staining. The photograph of H-PLGA scaffold after toluidine blue staining was notshown because the scaffold had disorganized in the toluidine blue solution.

(70/30) scaffold showed low proliferation. Although 3T3 fibroblastsin the basal medium containing PLGA/bFGF scaffold exhibitedmodest proliferation in the first two days, the proliferation wasobviously lower than that in the basal medium containing H-PLGA/PLGA(70/30)/bFGF scaffold thereafter. It was likely due to a rapidrelease of bFGF from the PLGA scaffold at the first day and then veryfew release of the bFGF in the following days. However, it could beseen that the proliferation of 3T3 fibroblasts in the basal mediumcontaining the H-PLGA/PLGA(70/30)/bFGF scaffold was rapid. After8 days of culture, the cell growth in the basal medium containingthe H-PLGA/PLGA(70/30)/bFGF scaffold was not any statisticallydifferent from that in the basal medium containing free bFGF dailyadded. The results meant bFGF sustainedly released from theH-PLGA/PLGA/bFGF scaffold had preserved its bioactivity whichwas almost same as that of fresh bFGF daily added to the basalmedium in a free form.

3.7. Cell attachment and growth on bFGF immobilized H-PLGA/PLGA scaffolds

Attachment efficiency of 3T3 fibroblasts on various PLGA scaf-folds after 3 h cell culturewas summarized in Fig. 7. It could be seenthat the cell attachment efficiency of various H-PLGA/PLGA scaf-folds increased with increasing H-PLGA content. Cell attachmentefficiency on the PLGA scaffold was only 48.7%, but that on the H-PLGA/PLGA(70/30) scaffold reached to 71.2%, which was signifi-cantly greater than that on the PLGA scaffold. It was considered thatbecause hydrophilicity of the H-PLGA/PLGA scaffolds was graduallyimproved with increasing H-PLGA content, and the moderatehydrophilicity was favorable for cell attachment. Additionally, afterbinding bFGF to PLGA or H-PLGA/PLGA (30/70) scaffolds, althoughcell attachment efficiency of the scaffolds had no statistical differ-ence compare with that of corresponding scaffolds, the cell

0

10

20

30

40

50

60

70

80 *#$

*#

*

H-PLGA/PLGA(50/50)

H-PLGA/PLGA(70/30)

H-PLGA/PLGA(30/70)

PLGA

Bind

ing

effic

ienc

y (%

)

Fig. 4. Binding efficiency of bFGF to PLGA and H-PLGA/PLGA scaffolds. *p < 0.05; significant against the binding efficiency of bFGF to PLGA scaffold. #p < 0.05; significant against thebinding efficiency of bFGF to H-PLGA/PLGA(30/70) scaffold. $p < 0.05; significant against the binding efficiency of bFGF to H-PLGA/PLGA(50/50) scaffold.

H. Shen et al. / Biomaterials 32 (2011) 3404e3412 3409

attachment efficiency of bFGF immobilized H-PLGA/PLGA(50/50)and H-PLGA/PLGA(70/30) scaffolds was higher than that of theH-PLGA/PLGA(50/50) and H-PLGA/PLGA(70/30) scaffolds. Theresult could be attributed to the amount of bound bFGF on the H-PLGA/PLGA(50/50) and H-PLGA/PLGA(70/30) scaffolds was enoughfor increasing cell attachment, but the amount of bound bFGF onthe PLGA or H-PLGA/PLGA (30/70) scaffolds was insufficient toinfluence cell attachment on the scaffolds.

Proliferation and viability of 3T3 fibroblasts cultured on variousPLGA scaffolds were determined by MTT assay after cultured for 10days in vitro, as shown in Fig. 8. It could be seen that the prolifer-ation and viability of cells on the H-PLGA/PLGA scaffolds werehigher than those on the PLGA and PLGA/bFGF scaffolds, and theproliferation and viability of cells increased with increasing theH-PLGA content. Moreover, the proliferation and viability of cells on

0 5 10 15 20 250

20

40

60

80

100

120

PLGA/bFGFH-PLGA/PLGA(70/30)/bFGF

Cum

ulat

ive

rele

ase

of b

FGF

(ng/

scaf

fold

)

Time(days)

Fig. 5. Cumulative release of bFGF from PLGA/bFGF and H-PLGA/PLGA(70/30)/bFGFscaffolds.

the bFGF immobilized H-PLGA/PLGA scaffolds were significantlyhigher than those on the corresponding H-PLGA/PLGA scaffolds.The result clearly showed that the proliferation and viability of cellson the scaffolds had enhanced after bFGF immobilized on theH-PLGA/PLGA scaffolds.

Fig. 9 showed morphology of 3T3 fibroblasts on cross section ofvarious PLGA scaffolds after cultured for two weeks. It could be seenthat the cells on PLGA and PLGA/bFGF scaffolds (Fig. 9a and b) werespherical and tended to get together. Although the cells on theH-PLGA/PLGA (30/70) scaffold still were spherical, they attachedtightly on the porewall of the scaffold (Fig. 9c).With H-PLGA contentfurther increasing in theH-PLGA/PLGAscaffold, cells gradually spreadand grew along porewall of the scaffold (Fig. 9e and g). Furthermore,after the H-PLGA/PLGA scaffolds bound bFGF, cells spread better andmore cells were observed in the pore of the scaffolds (Fig. 9d,f and h).

0 2 4 6 80

5

10

15

20

25

30

*

*

*

*

PLGAPLGA/bFGFH-PLGA/PLGA(70/30)H-PLGA/PLGA(70/30)/bFGFPostive control

Cel

l num

ber (

*104

)

Time(day)

Fig. 6. Proliferation of 3T3 fibroblasts cultured with various PLGA scaffolds. *p > 0.05;insignificant against positive control.

0

10

20

30

40

50

60

70

80

90*#$&

*#$

*

*#$

*#*

H-PLGA/PLGA(70/30)

H-PLGA/PLGA(50/50)

H-PLGA/PLGA(30/70)

PLGA

Atta

chm

ent e

ffici

ency

Before binding bFGFAfter binding bFGF

Fig. 7. Attachment efficiency of 3T3 fibroblasts on PLGA and H-PLGA/PLGA scaffolds before and after binding bFGF. *p < 0.05; significant against the attachment efficiency of 3T3fibroblasts on PLGA scaffold before and after binding bFGF. #p < 0.05; significant against the attachment efficiency of 3T3 fibroblasts on H-PLGA/PLGA(30/70) scaffold before andafter binding bFGF. $p < 0.05; significant against the attachment efficiency of 3T3 fibroblasts on the H-PLGA/PLGA(50/50) scaffold before binding bFGF. &p < 0.05; significant againstthe attachment efficiency of 3T3 fibroblasts on the H-PLGA/PLGA(70/30) scaffold before binding bFGF.

H. Shen et al. / Biomaterials 32 (2011) 3404e34123410

Especially on the bFGF immobilized H-PLGA/PLGA (50/50) andH-PLGA/PLGA (70/30) scaffolds, a great many of cells spread wellalong pore wall of the scaffolds (Fig. 9f and h).

The cell culture results above revealed that the attachment andgrowth of the 3T3 fibroblasts on PLGA scaffold were improved by

0.0

0.1

0.2

0.3

0.4

0.5

0.6

*$

H-PLGA/PLGA(30/70)

PLGA

Before binding bFGFAfter binding bFGF

abso

rban

ce *

Fig. 8. MTT-tetrazolium assay of 3T3 fibroblasts cultured on PLGA and H-PLGA/PLGA scaproliferation and viability of 3T3 fibroblasts on PLGA scaffold before and after binding bFGFPLGA(30/70) scaffold before binding bFGF. #p < 0.05; significant against proliferation an&p < 0.05; significant against proliferation and viability of 3T3 fibroblasts on H-PLGA/PLGA

introducing H-PLGA and increasedwith increasing H-PLGA content.BecauseH-PLGAmight improve hydrophilicity of the scaffold, it wasfavorable for cell attachment and growth to migrate into inside ofthe scaffold. On the other hand, itwas obvious that after bFGF boundto H-PLGA/PLGA scaffold, the attachment and growth of 3T3

*$#&*$#*$#

*$

H-PLGA/PLGA(70/30)

H-PLGA/PLGA(50/50)

ffolds before and after binding bFGF for 10 days. *p < 0.05; significant against the. $p < 0.05; significant against proliferation and viability of 3T3 fibroblasts on H-PLGA/d viability of 3T3 fibroblasts on H-PLGA/PLGA(50/50) scaffold before binding bFGF.(70/30) scaffold before binding bFGF.

Fig. 9. Morphology observation of 3T3 fibroblasts on cross section of PLGA and H-PLGA/PLGA scaffolds before and after binding bFGF cultured for 2 weeks (500�). (a) PLGA beforebinding bFGF; (b) PLGA after binding bFGF; (c) H-PLGA(30/70) before binding bFGF; (d) H-PLGA(30/70) after binding bFGF. (e) H-PLGA(50/50) before binding bFGF; (f) H-PLGA(50/50)after binding bFGF. (g) H-PLGA(70/30) before binding bFGF; (h) H-PLGA(70/30)after binding bFGF. The arrows indicated the representative cells.

H. Shen et al. / Biomaterials 32 (2011) 3404e3412 3411

H. Shen et al. / Biomaterials 32 (2011) 3404e34123412

fibroblasts on the H-PLGA/PLGA scaffold had further enhanced.Based on high affinity of the heparin to the bFGF, the bFGF could beevenly immobilized on the PLGA scaffold and the immobilized bFGFcould sustain release and keep bioactivity to accelerated cellattachment and growth. Thus, the bFGF immobilized H-PLGA/PLGAscaffold is a potential and promising scaffold to be used in tissueengineering.

It should be noted that the heparin introduced into the PLGAscaffold can inhibit certain cells growth [41,42]. So an optimalmethod should be selected to immobilize growth factor on poly-lactone-type scaffolds according to tissue engineering demand.

4. Conclusions

In this study, heparin was introduced into PLGA scaffold byblending H-PLGA and a high molecular weight PLGA. The intro-duction of heparin on the PLGA scaffolds can increase the bindingsites of the PLGA scaffold to bFGF, which can effectively enhancebFGF binding on surface of the PLGA scaffold, which is uninfluencedby thickness of the scaffold. The binding bFGF efficiency of H-PLGA/PLGA scaffolds increased with increasing the content of H-PLGA inthe scaffolds. The bFGF bound on the H-PLGA/PLGA scaffolds can beslowly released over twoweeks in vitro. The released bFGF from theH-PLGA/PLGA/bFGF scaffolds can still maintain its bioactivity.Furthermore, the bFGF immobilized H-PLGA/PLGA scaffold hasgood cell affinity. Thus, the H-PLGA/PLGA scaffold combining withbFGF could be usable in delivering bFGF, and the bFGF immobilizedH-PLGA/PLGA scaffold is a hopeful scaffold for tissue engineering.

Acknowledgments

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

Appendix

Figures with essential color discrimination. Certain figures inthis article, particularly Figs. 1,3,5 and 6, are difficult to interpret inblack and white. The full color images can be found in the onlineversion, at doi:10.1016/j.biomaterials.2011.01.037.

References

[1] Biondi M, Ungaro F, Quaglia F, Netti PA. Controlled drug delivery in tissueengineering. Adv Drug Delivery Rev 2008;60:229e42.

[2] Yan YN, Xiong Z, Hu YY, Wang SG, Zhang RJ, Zhang C. Layered manufacturingof tissue engineering scaffolds via multi-nozzle deposition. Mater Lett2003;57:2623e8.

[3] Ma ZW, Gao CY, Gong YH, Shen JC. Cartilage tissue engineering PlLA scaffoldwith surface immobilized collagen and basic fibroblast growth factor.Biomaterials 2005;26:1253e9.

[4] Allen KD, Athanasiou KA. Scaffold and growth factor selection in temporo-mandibular joint disc engineering. J Dent Res 2008;87:180e5.

[5] Hu XX, Shen H, Yang F, Bei JZ, Wang SG. Preparation and cell affinity ofmicrotubular orientation-structured PLGA(70/30) blood vessel scaffold.Biomaterials 2008;29:3128e36.

[6] Liu XH, Won YJ, Ma PX. Surface modification of interconnected porous scaf-folds. J Biomed Mater Res 2005;74A:84e91.

[7] Shen H, Hu XX, Yang F, Bei JZ, Wang SG. Combining oxygen plasma treatmentwith anchorage of cationized gelatin for enhancing cell affinity of poly(lactide-co-glycolide). Biomaterials 2007;28:4219e30.

[8] Shen H, Hu XX, Yang F, Bei JZ, Wang SG. An injectable scaffold: rhBMP-2-loaded poly(lactide-co-glycolide)/hydroxyapatite composite microspheres.Acta Biomater 2010;6:455e65.

[9] Babensee JE, Mcintire LV, Mikos AG. Growth factor delivery for tissue engi-neering. Pharm Res 2000;17:497e504.

[10] Fu X, Shen Z, Chen Y, Xie J, Guo Z, Zhang M, et al. Randomized placebo-controlled trial of use of topical recombinant bovine basic fibroblast growthfactor for second-degree burns. Lancet 1998;352:1661e4.

[11] RadomskyML, ThompsonAY, Poser JW. Potential role offibroblast growth factorin enhancement of fracture healing. Clin Orthop Relat Res 1998;355S:283e93.

[12] Kirker-head CA. Potential applications and delivery strategies for bonemorphogenetic proteins. Adv Drug Delivery Rev 2000;43:65e92.

[13] Allori AC, Sailon AM, Warren SM. Biological basis of bone formation, remod-eling, and repair e part I: biochemical signaling molecules. Tissue Eng Part B-Rev 2008;14:259e73.

[14] Dereka XE, Markopoulou CE, Vrotsos IA. Role of growth factors on periodontalrepair. Growth Factors 2006;24:260e7.

[15] Yoon DM, Fisher JP. Effects of exogenous IGF-1 delivery on the early expres-sion of IGF-1 signaling molecules by alginate embedded chondrocytes. TissueEng Part A 2008;14:1263e73.

[16] Li C, Vepari C, Jin HJ, Kim HJ, Kaplan DL. Electrospun silk-BMP-2 scaffolds forbone tissue engineering. Biomaterials 2006;27:3115e24.

[17] LiuAC,GotliebAI. Transforming growth factor-beta regulates in vitroheart valverepair by activated valve interstitial cells. Am J Pathol 2008;173:1275e85.

[18] Farre J, Roura S, Prat-Vidal C, Soler-Botija C, Llach A, Molina CE, et al. FGF-4increases in vitro expansion rate of human adult bone marrow-derivedmesenchymal stem cells. Growth Factors 2007;25:71e6.

[19] Saltzman WM. Growth factor delivery in tissue engineering. MRS Bull1996;21:62e5.

[20] Tabata Y, Takebayashi Y, Ueda T, Ikada Y. A formulation method using d, l-lactic acid oligomer for protein release with reduced initial burst. J ControlRelease 1993;23:55e64.

[21] Nimni ME. Polypeptide growth factors: targeted delivery systems. Biomate-rials 1997;18:1201e25.

[22] Whang K, Tsai DC, Nam EK, Aitken M, Sprague SM, Patel PK, et al. Ectopic boneformation via rhBMP-2 delivery from porous bioabsorbable polymer scaffolds.J Biomed Mater Res 1998;42:491e9.

[23] Thula TT, Schultz G, Tran-Son-Tay R, Batich C. Effects of EGF and bFGF onirradiated parotid glands. Ann Biomed Eng 2005;33:685e95.

[24] Wang SG, Cai Q, Hou JW, Bei JZ, Zhang T, Yang J, et al. Acceleration effect ofbasic fibroblast growth factor on the regeneration of peripheral nerve througha 15 mm gap. J Biomed Mater Res 2003;66A:522e31.

[25] Xiong Z, Yan YN, Wang SG, Zhang RJ, Zhang C. Fabrication of porous scaffoldsfor bone tissue engineering via low-temperature deposition. Scr Mater2002;46:771e6.

[26] Jeon O, Rhie JW, Kwon IK, Kim JH, Kim BS, Lee SH. In vivo bone formationfollowing transplantation of human adiposeederived stromal cells that arenot differentiated osteogenically. Tissue Eng: Part A 2008;14:1285e94.

[27] Cowan CM, Aghaloo T, Chou YF, Walder B, Zhang XL, Soo C, et al. MicroCTevaluation of three-dimensional mineralization in response to BMP-2 doses invitro and in critical sized rat calvarial defects. Tissue Eng 2007;13:501e12.

[28] Shen H, Hu XX, Bei JZ, Wang SG. The immobilization of basic fibroblast growthfactor on plasma-treated poly(lactide-co-glycolide). Biomaterials 2008;29:2388e99.

[29] Shen H, Hu XX, Yang F, Bei JZ, Wang SG. The bioactivity of rhBMP-2 immo-bilized poly(lactide-co-glycolide) scaffolds. Biomaterials 2009;30:3150e7.

[30] Wan YQ, Tu CF, Yang J, Bei JZ, Wang SG. Influences of ammonia plasmatreatment on modifying depth and degradation of poly(l-lactide) scaffolds.Biomaterials 2006;27:2699e704.

[31] Jeon O, Song SJ, Kang SW, Putnam AJ, Kim BS. Enhancement of ectopic boneformation by bone morphogenetic protein-2 released from a heparin-conju-gated poly(l-lactic-co-glycolic acid) scaffold. Biomaterials 2007;28:2763e71.

[32] Yoon JJ, Chung HJ, Lee HJ, Park TG. Heparin-immobilized biodegradablescaffolds for local and sustained release of angiogenic growth factor. J BiomedMater Res Part A 2006;79A:934e42.

[33] Zhang L, Furst EM, Kiick KL. Manipulation of hydrogel assembly and growthfactor delivery via the use of peptideepolysaccharide interactions. J ControlRelease 2006;114:130e42.

[34] Spillmann D, Lindahl U. Glycosaminoglycan protein interactions e a questionof specificity. Curr Opin Struct Biol 1994;4:677e82.

[35] Lin H, Zhao YN, Sun WJ, Chen B, Zhang J, Zhao WX, et al. The effect ofcrosslinking heparin to demineralized bone matrix on mechanical strengthand specific binding to human bone morphogenetic protein-2. Biomaterials2008;29:1189e97.

[36] Lee AC, Yu VM, Lowe JB, Brenner MJ, Hunter DA, Mackinnon SE, et al.Controlled release of nerve growth factor enhances sciatic nerve regeneration.Exp Neurol 2003;184:295e303.

[37] Gilding DK, Reed AM. Biodegradable polymers for use in surgery-polyglycolic/polylactic acid homo and copolymers. Polymer 1979;20:1459e64.

[38] Jee KS, Park HD, Park KD, Kim YH, Shin JW. Heparin conjugated polylactide asa blood compatible material. Biomacromolecules 2004;5:1877e81.

[39] Shi GX, Cai Q, Wang CY, Lu N, Wang SG, Bei JZ. Fabrication of cell scaffold ofpoly(l-lactic acid) and poly(l-lactic-co-glycolic acid) and biocompatibility.Polym Adv Technol 2002;13:227e32.

[40] Smith PK, Mallia AK, Hermanson GT. Colorimetric method for the assay ofheparin content in immobilized heparin preparations. Anal Biochem1980;109:466e73.

[41] Clowes AW, Karnovsky MJ. Suppression by heparin of smooth muscle cellproliferation in injured arteries. Nature 1977;265:625e6.

[42] Nikitovic D, Zafiropoulos A, Tzanakakis GN, Karamanos NK, Tsatsakis AM.Effects of glycosaminoglycans on cell proliferation of normal osteoblasts andhuman osteosarcoma cells depend on their type and fine chemical composi-tions. Anticancer Res 2005;25:2851e6.