Poly (l-lactide-co-caprolactone) scaffolds enhanced with poly (β-hydroxybutyrate-co-β-

hydroxyvalerate) microspheres for cartilage regeneration

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2013 Biomed. Mater. 8 025005

(http://iopscience.iop.org/1748-605X/8/2/025005)

Download details:

IP Address: 162.105.25.84

The article was downloaded on 02/03/2013 at 05:04

Please note that terms and conditions apply.

View the table of contents for this issue, or go to the journal homepage for more

Home Search Collections Journals About Contact us My IOPscience

IOP PUBLISHING BIOMEDICAL MATERIALS

Biomed. Mater. 8 (2013) 025005 (13pp) doi:10.1088/1748-6041/8/2/025005

Poly (L-lactide-co-caprolactone) scaffoldsenhanced with poly (β-hydroxybutyrate-co-β-hydroxyvalerate) microspheres forcartilage regenerationChao Li1, Jingjing Zhang1, Yijiang Li1, Shamus Moran2, Gilson Khang3

and Zigang Ge1,4,5

1 Department of Biomedical Engineering, College of Engineering, Peking University, Beijing 100871,People’s Republic of China2 The Wallace H Coulter Department of Biomedical Engineering, College of Engineering,Georgia Institute of Technology, Atlanta, GA 30332, USA3 Department of Polymer Nano Science and Technology, Chonbuk National University, Jeonju 561752,Korea4 Center for Biomedical materials and Tissue Engineering, Academy for Advanced InterdisciplinaryStudies, Peking University, Beijing 100871, People’s Republic of China

E-mail: gez@pku.edu.cn

Received 7 September 2012Accepted for publication 7 January 2013Published 5 February 2013Online at stacks.iop.org/BMM/8/025005

AbstractBiodegradable polymers, either as porous scaffolds or microspheres, have been investigatedbroadly for cartilage tissue engineering. A combination of these two forms of materials couldpotentially maximize their benefits. In this study, porous poly (L-lactide-co-ε-caprolactone)(PLCL) scaffolds were integrated with poly (β-hydroxybutyrate-co-β-hydroxyvalerate)(PHBV) microspheres to enhance the mechanical properties of the scaffolds as well as topotentially regulate cell behavior through altering surface topography. PHBV microspheresfabricated with an emulsion solvent evaporation method were incorporated into PLCLscaffolds (0%, 20%, 40% and 50% W/W). Compressive modulus, surface topography andporosity of the composite scaffolds were evaluated, and in vitro and in vivo chondrogenesiswithin the chondrocyte-laden scaffolds was investigated by examining proliferation ofchondrocytes and the deposition of glycosaminoglycan (GAG) and type II collagen. Theresults showed significant enhancement of the compressive modulus of the scaffoldsincorporated with PHBV microspheres, while Young’s modulus of the scaffolds with 50%PHBV incorporation was 3.3 times higher than PLCL scaffolds alone. The porosity of thecomposite scaffolds was kept constant for all levels of PHBV incorporation. Though the PLCLscaffolds incorporated with microspheres showed no significant effects on adhesion at 6 h aswell as in vitro cartilage formation and proliferation of the chondrocytes at both 2 weeks and4 weeks, total contents of GAG and type II collagen excreted increased significantly with time.The chondrocyte-laden scaffolds formed cartilage-like tissues at 4 and 8 weeks afterimplantation in nude mice, with increased staining density of type II collagen and GAG overtime. In conclusion, incorporation of PHBV microspheres not only enhanced the compressive

5 Author to whom any correspondence should be addressed.

1748-6041/13/025005+13$33.00 1 © 2013 IOP Publishing Ltd Printed in the UK & the USA

Biomed. Mater. 8 (2013) 025005 C Li et al

modulus of PLCL scaffolds, but could also serve as scaffolding structures for cartilaginoustissue formation.

S Online supplementary data available from stacks.iop.org/BMM/8/025005/mmedia

(Some figures may appear in colour only in the online journal)

1. Introduction

Once damaged, articular cartilage has a poor abilityto regenerate, mainly due to lack of blood and nervesupply, constrained cell motion, limited ability of existingchondrocytes to proliferate, and a complex physiological andmechanical environment [1]. Cartilage tissue engineering isone of the most effective strategies to regenerate cartilagedefects and has been widely investigated and adoptedclinically; however, there are still many challenges such asin vitro chondrocyte dedifferentiation, poor integration withhost cartilage and lack of scaffolds with mechanical propertiescompatible with cartilage [2]. Scaffolds used for cartilagetissue engineering should not only withstand mechanicalstimulus to maintain structural integrity, but also transfermechanical signals to individual chondrocytes to regulate theirfunctionality. At best, scaffolds could support and regulatethe organization and subsequent proliferation of seeded cellsthrough controlled release of biochemical factors and differentphysical/chemical properties [3].

Mechanical properties of biomaterials are of greatimportance for cartilage regeneration, especially the aggregatecompressive Young’s modulus and viscoelasticity. To avoiddeterioration under repetitive physiological loads, biomaterialsshould have a compressive Young’s modulus comparableto native cartilage, which varies between 0.4 and 0.8 MPa[4]. Proteoglycans and collagen, as well as their interaction,provide articular cartilage with the unique physical propertiesto withstand mechanical stress [1]. An anisotropic 3Dwoven poly-glycolic acid structure has been fabricated withmechanical properties similar to native cartilage [5]. However,further study of this structure has yet to be reported in cartilagetissue engineering, which could partially be attributed toincompliance of its unnatural structures with natural cartilage.In order to attain the required strength, multiple scaffoldsmade of different materials (chitosan, collagen, Poly (lactic-co-glycolic acid), Poly (lactic acid), Poly (β-hydroxybutyrate-co-β-hydroxyvalerate) and different structures (sponges,hydrogels and woven structures) have been developed[6–9]. However, additional strategies to chemically crosslinkthe material are required to maintain sustainable strength orfor optimization of internal structures.

Viscoelastic properties of biomaterial scaffolds similarto natural cartilage are also critical for scaffolds to maintaintheir structures under repetitive physiological loads as well asto transfer proper mechanical signals to chondrocytes withinthe scaffolds. Cyclic mechanical stimuli could regulate thephenotype of chondrocytes [10], while proper viscoelasticproperties of the biomaterials could facilitate transmission ofthe mechanical stimuli to both cartilage and chondrocytes.

Three kinds of biodegradable elastomers including poly (1,8-octanediol citrate) (POC), poly (glycerol-sebacate) (PGS)and poly-ε-caprolactone (PCL) have been shown to enhancethe quality of engineered cartilage under dynamic compressionin vitro [8, 11]. Both POC and PGS have a three-dimensionalcrosslinked network of random coils analogous to vulcanizedrubber, while PCL is made of linear molecular chains withoutcrosslinks. Their chemical properties, such as hydrophilicityas well as rate and pattern of degradation, can be tailoredby grafting hydrophobic moieties to the hydroxyl groups.However, the aforementioned elastomers have a notably lowYoung’s modulus and high yield strain, which are incompatiblefor cartilage regeneration. Poly (L-lactide-co-ε-caprolactone,PLCL), made of poly lactic acid (PLA) and PCL, with amolar ratio of 7:3 is a highly elastic and linear biodegradablepolymer with a relatively mild degradation, which avoidsan abrupt drop in pH value during degradation, similar toPLA and PGA scaffolds [12, 13]. However, the relativelylow compressive modulus should be enhanced when usedin cartilage regeneration. Three methods could be used toimprove the strength of the scaffold materials: choosingmaterials which are of great strength, improving strengthwith chemical crosslinking and incorporating stiffer materialswithin the elastic scaffolds. Until recently, methods to furtherimprove the compressive strength of pure PLCL scaffolds wereinadequate due to the fact that PLCL is a linear elastomerwhich lacks the capability of being enhanced throughcrosslinking [2, 3].

Integration of materials with a high Young’s modulusis a promising strategy for improving the mechanicalproperties of viscoelastic scaffolds. Elastic PLCL was blendedwith stiff PLGA to improve the compressive Young’smodulus to 12.9 MPa, which could maintain its architectureduring the in vitro experiment [14]. Biocompatible poly(β-hydroxybutyrate-co-β-hydroxyvalerate) (PHBV) with arelatively high elastic modulus (1.2 GPa) is a promisingcandidate for improving the mechanical properties of scaffolds[15]. PHBV copolymers with 8% hydroxyvalerate (HV)have been used for cartilage regeneration, which facilitatedchondrogenic differentiation in vitro and in vivo [16]. PHBVcould not only help to maintain chondrogenic phenotypeof pre-differentiated human adipose-derived stem cells, butcell/PHBV constructs could also produce neo-cartilage in aheterotopic site [17]. Furthermore, incorporation of PHBVmicrospheres may regulate the functionality of chondrocytesthrough altering the surface topography and provide thepotential to include bioactive factors [18].

Porous PLCL scaffolds have previously been fabricatedfor cartilage tissue engineering; however, their lowcompressive Young’s modulus hindered their utility [3, 19, 20].

2

Biomed. Mater. 8 (2013) 025005 C Li et al

Figure 1. The schematic fabrication of PHBV microspheres-PLCL composite scaffolds.

In this study, PHBV microspheres were integrated intothe PLCL scaffolds with an aim to improve the Young’smodulus and increase surface topography for the attachmentof chondrocytes. The composite scaffolds were fabricatedwith a combination of porogen-leaching and lyophilizationmethods. Morphology, compressive modulus and porosityof the composite scaffolds were characterized. Proliferationand differentiation of the chondrocytes on the scaffolds wereevaluated in vitro, while extracellular matrix (ECM) depositionwas investigated in vitro and in vivo.

2. Experimental section

2.1. Preparation of PHBV microspheres

PHBV microspheres were fabricated using an emulsionsolvent evaporation technique [21]. A 2% PHBV solution wasmade by dissolving 0.5 g of PHBV (8% hydroxyvalerate (HV)content, Goodfellow Cambridge Limited, UK) in 25 mL ofmethylene chloride. The solution was then quickly pouredinto 500 mL of deionized water containing 0.5% (W/V)polyvinyl alcohol (PVA124, Guoyao Chemical ReagentsLimited, Beijing, China) under agitation at a speed of 1000 rpmfor 12 h, until the organic solvent was completely evaporated.The microspheres were collected after centrifugation at4000 rpm for 3 min, washed three times with deionized waterto remove PVA, and then freeze-dried for 24 h.

2.2. PHBV microsphere size analysis

Size distribution of the PHBV microspheres was determinedusing a light-scattering particle size analyzer (Matersizer 2000,Malvern Instrument Ltd, UK). The desiccated microsphereswere suspended in 1000 mL of distilled water and analyzedafter continuous stirring.

2.3. Integration of PHBV microspheres with PLCL scaffolds

Three-dimensional PLCL scaffolds were fabricated with aporogen-leaching method, as described previously [22]. 0.5 gof PLCL (7:3; Mw 23 000 Da, Daigang Biomaterials Inc.Jinan, China) was dissolved in acetone. Following fabricationof PLCL scaffolds, 0, 0.1, 0.2 or 0.25 g PHBV microsphereswith 8.0 g sodium chloride (200–300 μm) were suspended inthe PLCL solution to fabricate PLCL, 20% PHBV/PLCL, 40%PHBV/PLCL and 50% PHBV/PLCL scaffolds, respectively.After agitation, the composite mixtures were cast in glassmolds before exposure to air for 48 h to fully evaporatethe solvent. The mixtures were then rinsed to remove NaClcrystals. Composite scaffolds were frozen at −20 ◦C beforebeing lyophilized for 24 h. After lyophilization, the acquiredscaffolds were stored in a desiccator until usage (figure 1).

2.4. Scanning electron microscopy

The scaffolds were sectioned using a scalpel, before beingfixed and sputter coated with gold for 80 s at 18 mA. Internalstructures of the scaffolds and PHBV microspheres wereobserved by scanning electron microscopy (SEM; Quanta200FEG, FEI, USA) at an accelerating voltage of 5 kV.

2.5. Porosity of the scaffolds

Porosity of the scaffolds was estimated using a previousmethodology [23]. Sizes of the scaffolds were measured witha vernier caliper, and the weight of the scaffolds was recorded.The scaffolds were then immersed in absolute alcohol for 2 hand weight was assessed once more. Porosity was calculatedas (Ws-Wd)/ρ/V, where ρ represented the density of alcoholand V represented volume of the scaffolds, respectively.

3

Biomed. Mater. 8 (2013) 025005 C Li et al

2.6. Compression test

An unconfined compression test was carried out with anInstron 5843 mechanical testing system (Instron Corporation,America). Compressive loads were applied to individualspecimens in a PBS bath using a stainless steel indenter. Allsamples were pre-loaded three times to a 10% strain beforea constant loading was applied at a displacement of 0.5 mmmin−1 until 45% of strain was reached. The measured thicknesswas converted to the strain of the sample (ε = 1 − L/L0, whereL0 and L represent the thickness before and after compression,respectively). Young’s modulus (E = σ/ε, where σ and ε

denote the stress and strain of the sample, respectively) wasdetermined automatically. This process was repeated on fourspecimens for each sample type.

2.7. Isolation and culture of primary chondrocytes

Primary chondrocytes were isolated from articular cartilageof an adult pig (Yorkshire, 10–12 months). Cartilage slicescollected from the femoral condyle were first digested with0.25% TrypLE (Invitrogen) for 30 min and then with 0.25%(w/v) type II collagenase solution (17101-015, Gibco, USA)in Dulbecco’s modified Eagle’s medium (DMEM, 31600-034, Invitrogen, USA) with 1 g tissue/5 mL collagenasesolution. The mixture was incubated for 12–16 h at 37 ◦Cwith intermittent agitation before the cells were collectedthrough a strainer mesh (200 mesh, 74 μm, Solarbio, Fr)and counted with a hemocytometer. The chondrocytes wereadded at a density of 4000 cells cm−2, cultured with DMEM,and supplemented with 10% fetal bovine serum (FBS,SV30087.02, Gibco) with 100 U mL−1 PS (Penicillin, 0741;Streptomycin, 0832, Amresco, USA). The cell cultures wereincubated at 37 ◦C in a humidified environment with 5% CO2.Culture media was changed every 72 h, and cells were passagedwhen 80% confluence was reached. Passage 1 chondrocyteswere used in the experiment.

2.7.1. Preparation of the scaffolds and cell seeding. Thescaffolds were cut into disks of length × width × height =4 × 4 × 2 mm3 and sterilized with 75% (V/V) ethanol solutionfor 2 h followed by 30 min of ultraviolet radiation. All scaffoldswere pre-soaked with DMEM for 24 h before cell seeding.30 μL of cell suspension (1 × 107 cells mL−1) were loadeddrop-wise onto the top of the scaffolds and incubated at 37 ◦Cfor 60 min before 100 μL of culture medium was added toimmerse the scaffolds. The immersed scaffolds were incubatedat 37 ◦C in a humidified incubator of 5% CO2.

2.7.2. Attachment and proliferation of the chondrocytes. Cellattachment (6 h) and cell proliferation (2 and 4 weeks) wereevaluated with Hoechst 33258. The cells were digested with1 mL of 3 mg mL−1 proteinase K (H10091, Merck, Germany)overnight at 57 ◦C at pre-scheduled time points. One hundredmicroliters of 0.1 mg mL−1 of H33258 was used to dye 100 μLdigested solution in 96-well plates (Berthold). Fluorescenceintensity was measured on a microplate reader (Synergy,BioTek, USA) at excitation and emission wavelengths of 360

and 465 nm, respectively. A calibration curve was obtainedfrom DNA standard solutions with known concentrations ofcells. Four replicates were used for each time point.

2.7.3. Live/dead characterization. After the chondrocyteswere cultured in the scaffolds (length × width × height =4 × 4 × 2 mm3) for prescheduled periods (6 h, 2 w), thescaffolds were sequentially incubated with 100 μL of 2 μLmL−1 fluorescein diacetate solution (FDA, F7378, Sigma) for15 min and followed by 100 μL of 5 μg mL−1 of propidiumiodide solution (PI, P4170, Sigma) for another 5 min at37 ◦C. The constructs were evaluated with confocal laserscanning microscopy (CLSM, LSM510, Zeiss, Germany) atexcitation wavelength of 488 nm and emission wavelengths of550–670 nm. Morphology and spatial distribution of the cellswere evaluated.

2.7.4. Glycosaminoglycan quantification. Sulfated GAGconcentration of chondrocytes/scaffolds constructs wasdetermined with dimethylmethylene blue (DMMB, 341088,Sigma) after in vitro culture for 2 or 4 weeks [24]. Theconstructs were digested in 3 mg mL−1 proteinase K(H10091, Merck, Germany) overnight at 57 ◦C. 1.0 mL ofworking DMMB solution (16 μg mL−1 DMMB 2.5% ethanol,1 M guanidine hydrochloride (GuHCl), 0.2 g L−1 sodiumformate and 2% formic acid) was mixed with 100 μLof digestive solution, vortexed for 30 min and centrifugedat 9600g for 10 min. After discarding the supernatant,1.0 mL of decomplexation solution (50 mM sodium acetatesolution buffer (pH 6.8), 10% propan-1-ol, 4 M GuHCl)was added and incubated for 30 min. Absorbance wasmeasured at 656 nm. The standard curve was determinedas follows: 0.01 g chondroitin sulfate (CS) (Sigma; CAS:39455-18-0) was dissolved into 10 mL ionized water,and the solution was diluted to 40/20/10/4/2 times toform 0.025/0.05/0.1/0.25/0.5 mg mL−1 CS solutions. Inaddition, 2/3/4/5 mg ml−1 CS solutions were also prepared.Absorbance was measured at 656 nm. Contents of sGAGwere extrapolated from the standard curve. The process wascompleted three times.

2.7.5. Quantitative PCR. The samples were lysed in 1.0 mLTrizol (15596-026, Invitrogen) for 10 min after in vitroculture for 2 or 4 weeks. The total RNA was extracted fromthe cell-scaffold constructs following the manufacturer’sinstructions. cDNA was reverse transcribed using iScriptTM

cDNA synthesis kit (Bio-Rad, Hercules, CA) followingmanufacturer’s instructions. Real-time PCR was performedusing the Power SYBR Green PCR Master Mix (AppliedBiosystem, Foster City, CA) on Applied Biosystems 7500Real-Time PCR System (Applied Biosystem) at 95 ◦C for15 min followed by 40 cycles of 15 s denaturation at 94 ◦C,30 s annealing at 55 ◦C and 30 s elongation at 72 ◦C. Eachsample was repeated three times for each gene of interest.Genes of interest were normalized to the reference geneglyceraldehydes-3-phosphate dehydrogenase (GAPDH).Gene expression was calculated as 2−��Ct. The following

4

Biomed. Mater. 8 (2013) 025005 C Li et al

Figure 2. Size distribution of PHBV microspheres.

forward and reverse primers were used for amplification [25]:for GAPDH, forward 5′-ATGGTGAAGGTCGGAGTGAA-3′; reverse 5′-AATGAAGGGGTCATTGATGG-3′; foraggrecan, forward 5′-CATCACCGAGGGTGAAGC-3′;reverse 5′-CCAGGGGCAAATGTAAAGG-3′; for type IIcollagen, forward 5′-TGAGAGGTCTTCCTGGCAAA-3′;reverse 5′-GAAGTCCCTGGAAGCCAGAT-3′. The processwas completed three times.

2.8. In vivo studies

The animal study was approved by the Institutional AnimalCare and Usage Committee (IACUC) of Peking University.Six 7-week-old male athymic mice (SPF, Vital River, China)were used for this study. An intramuscular injection of2.5 mg mL−1 of ketamine (50 mg kg−1) was administeredbefore three mid-sagittal skin incisions of 5 mm over the backof the mice were made. Each construct, with porcine cellscultured in vitro for 10 days, was inserted into subcutaneouscavities separately, and a total of six constructs were implantedin each mouse. Subcutaneous tissue and skin were suturedlayer by layer before covering incisions with penicillin powderand gauze. Mice were killed at 4 or 8 weeks, and the implantswere harvested. Two replicates were assessed.

2.8.1. Histological and immunohistological assessment. Thesamples were fixed in 10% (v/v) buffered formalin, dehydratedand embedded for histological analysis. The specimenswere sectioned into 5 μm thick sections and stained withhematoxylin and eosin (H&E). For alcian blue staining, thetissue sections were incubated with 0.5% alcian blue (Sigma-Aldrich) in 0.1M HCl for 30 min and counterstained withnuclear fast red (Sigma-Aldrich). For immunohistochemistrystaining, tissue sections were incubated with hydrogenperoxide and pepsin for 20 min separately. The sections wereincubated for an hour with a 1:500 dilution of monoclonalantibodies of type II collagen (Clone 6B3; Chemicon Inc.,Temecuela, CA) and then incubated with biotinylated goatanti-mouse (Lab Vision Corporation, Fremont, CA) for30 min. The sections were incubated with streptavidin,peroxidase and 3, 3′-diaminobenzidine before counterstainingwith hematoxylin.

2.9. Statistical analysis

SPSS V 18.0 (SPSS Inc., IL, USA) was utilized to analyze thedata, and a one-way analysis of variance (ANOVA) statisticaltest (P < 0.05, Tukey) was used to compare the differencebetween each group. All data were expressed as mean ±standard deviation.

3. Results

3.1. Characterization of PHBV microspheres

Fabricated PHBV microspheres had a relatively narrowparticle size distribution with a median particle diameter of37.2 μm (90% of the microspheres distributed from 22.7to 61.4 μm) (figure 2). PHBV microspheres had sphericalshapes and rough surfaces. Diameters of most of the PHBVmicrospheres were less than 50 μm, which was in accordancewith the findings of the size distribution analysis (figure 3).

3.2. Characterization of different scaffolds

More microspheres were seen in the composite scaffolds withhigher percentages of PHBV microsphere integration, thoughpore sizes of the composite scaffolds remained constant. Themicrospheres were uniformly distributed across and integratedwithin the PLCL matrix (figure 4). PHBV microspheres wereprominent in the internal surface of the matrix which changedthe surface topography of composite scaffolds on the microscale. The porosities of the scaffolds remained relativelyunchanged with integration of the PHBV microspheres,varying from 83.0 ± 3.6% (pure PLCL scaffold), 83.5 ±7.1% (20% PHBV) to 88.5 ± 3.5% (40% PHBV) (P > 0.05).Even when the concentration of PHBV microspheres was 50%(84.9 ± 2.3%), the porosity of the composite scaffolds showedno significant variation compared with pure PLCL scaffolds(figure 5). The water contact angle of the PLCL film was87.3 ± 2.8◦, which was not significantly different from thecontact angle of the PHBV film (88.4 ± 5.1◦) (figure S1available at http://stacks.iop.org/BMM/8/025005/mmedia).

5

Biomed. Mater. 8 (2013) 025005 C Li et al

(A)(B)

Figure 3. SEM photograph of PHBV microspheres: (A) magnification of 200 × ; (B) magnification of 1000 × .

(A) (B)

(C) (D)

Figure 4. SEM photograph of different composite scaffolds: (A) pure PLCL; (B) 20% PHBV microspheres/PLCL; (C) 40% PHBVmicrospheres/PLCL; (D) 50% PHBV microspheres/PLCL.

6

Biomed. Mater. 8 (2013) 025005 C Li et al

Figure 5. Porosity of PLCL scaffold, 20% PHBV/PLCL, 40% PHBV/PLCL and 50% PHBV/PLCL composite scaffolds (n = 4).

Figure 6. Compressive Young’s modulus of PLCL scaffold, 20% PHBV/PLCL, 40% PHBV/PLCL and 50% PHBV/PLCL compositescaffolds, a statistical difference between all groups except for 40% PHBV/PLCL scaffolds and 50% PHBV/PLCL scaffolds (n = 6,∗P < 0.05 compared with PLCL scaffold).

3.3. Compressive modulus of the scaffolds

Compressive strength of composite scaffolds increasedsubstantially from 37 ± 5 kPa (PLCL scaffolds) to77 ± 17 kPa (20% PHBV/PLCL) and 113 ± 15 kPa(40% PHBV/PLCL) to 123 ± 29 kPa (50% PHBV/PLCL)(figure 6) with the integration of PHBV microspheres. Therewas statistical difference between all groups except for 40%PHBV/PLCL scaffolds and 50% PHBV/PLCL scaffolds. Asmall amount of PHBV microspheres became detached fromthe PLCL matrix under the pre-loading stress, while thecomposite PHBV/PLCL scaffolds kept intact.

3.4. Morphology and distribution of chondrocytes in thescaffolds

The number of adhered cells on each group of compositescaffolds increased gradually with increasing culture time(P < 0.05). However, there was no significant variationbetween groups at 6 h, 2 weeks and 4 weeks (figure 7).Live–dead staining indicated that more than 95% of cells weredistributed homogeneously and viable in all the scaffolds withfewer than 5% of cells dead at 6 h and 2 weeks (figure S4available at http://stacks.iop.org/BMM/8/025005/mmedia).The chondrocytes appeared round in shape at 6 h in allthe scaffolds, while spreading of chondrocytes was observed

7

Biomed. Mater. 8 (2013) 025005 C Li et al

Figure 7. DNA assay applied to investigate chondrocytes attachment at 6 h and proliferation at 2 and 4 weeks after cell seeding onto PLCLscaffold, 20% PHBV/PLCL and 40% PHBV/PLCL composite scaffolds. There is no significant difference among different scaffolds types,but the number of adhered cells increased gradually with increasing culture time. All the numbers of chondrocytes are shown as means ±SD 32 (n = 4).

at 2 weeks. However, there was no difference betweenthe morphology of chondrocytes grown in the pure PLCLscaffolds and PHBV microspheres/PLCL composite scaffolds(figure 8).

3.5. In vitro cartilage formation of constructs

Though the GAG production increased significantly with timein all the scaffolds (P < 0.05), there was no significantdifference among different scaffolds at both 2 and 4 weeks(figure 9).

Gene expression of aggrecan in all scaffolds increasedwith time, but no significant difference between differentgroups was observed, which was inconsistent with resultsof GAG quantification. Total type II collagen expressionincreased with culture time in all scaffolds (P < 0.05), whilechondrocytes expressed no difference in type II collagen after2 weeks and 4 weeks of culturing among different scaffolds(P > 0.05) (figure 10).

3.6. Histology of in vivo implantation

All mice were kept healthy without infection until killed, andall 24 implants were available for histology. H&E stainingshowed that the cells distributed evenly in all scaffolds(figures 11A, D and G). More cells and ECM were seen in allscaffolds at 8 weeks (figure 12 A, D and G). The structure of thescaffolds was maintained up to 8 weeks, with the microspheresembedded within the secreted ECM.

Uneven and similar Alcian Blue and type II collagenstaining was seen in all groups at 4 weeks (figures 11B, Eand H). The staining became denser and more homogeneousfor all groups at 8 weeks when compared to the 4 weektime point (figure 12B, E and H). The staining was strongeraround the PHBV microspheres (within 5 μm) than the

peripheral area (more than 5 μm) at 4 weeks (P < 0.05, figureS5 available at http://stacks.iop.org/BMM/8/025005/mmedia).However, this localized concentration became inconspicuouswhen the staining became stronger and homogeneous at 8weeks (P > 0.05) (figure S6 available at stacks.iop.org/BMM).

4. Discussion

Mechanical properties of scaffolds such as compressiveYoung’s modulus and elasticity are critical for cartilageregeneration [26]. Though elastic synthetic polymers havehighly elastic properties, their compressive moduli are stillmuch lower than native cartilage [11, 20]. There are threeways to enhance the strength of scaffolds: choosing materialswhich are of great strength, improving strength with chemicalcrosslinking and incorporating stiffer materials within theelastic scaffolds. Compared with choosing stiffer materials orincreasing cross-links, incorporating stiffer materials, such asmicrospheres, has some advantages over the other strategiesincluding maintaining porosity and increasing roughness ofthe surfaces. Another strategy to improve compressive Young’smodulus is to decrease the porosity of the composite scaffolds.However, this strategy is not optimal since high porosityscaffolds offer advantages for tissue engineering.

PHBV-enhanced PLCL composite scaffolds could havean advantage in transferring external mechanical stimuli, bothdynamic and static, to individual cells. It has been reported thatproper cyclic mechanical stimuli could promote functionalityof chondrocytes within scaffolds by selecting scaffolds withproper mechanical properties [20]. Composite scaffolds couldmeet these requirements through combining properties ofdifferent materials. Though PLCL scaffolds had relativelygood recovery ratios and similar viscoelastic propertiesto native cartilage, their compressive modulus was onemagnitude less than native cartilage [3]. Integration of PHBV

8

Biomed. Mater. 8 (2013) 025005 C Li et al

(A) (B)

(C) (D)

(E) (F)

Figure 8. CLSM image showing chondrocytes on PLCL scaffold, 20% PHBV/PLCL and 40% PHBV/PLCL composite scaffold at 6 h and2 weeks. (A) and (B): chondrocytes on PLCL at 6 h and 2 weeks; (C) and (D): chondrocytes on 20% PHBV microspheres/PLCL scaffold at6 h and 2 weeks; (E) and (F): chondrocytes on 40% PHBV microspheres/PLCL scaffold at 6 h and 2 weeks. The live cells were dyed byFDA (green) and dead cells were stained by PI (red). Cell seeding density: 200 000/scaffold.

microspheres with a relatively high compressive Young’smodulus effectively enhanced compressive Young’s modulusof the composite scaffolds [15]. The limitation to furtherenhance Young’s modulus lies in the PHBV microspheresjust mixing into the scaffolds and only supplying withphysical support. When the percentage of PHBV microspheresincreased to 40%, the microspheres were almost completely

distributed throughout the matrix of the scaffolds. Once thePHBV microsphere proportion increased beyond 40%, therewas no increase in compressive Young’s modulus of thescaffolds. As the purpose of adding PHBV microspheres wasto enhance the compressive Young’s modulus of the scaffolds,the 50% PHBV/PLCL was omitted from the study sinceit did not increase Young’s modulus further. Most PHBV

9

Biomed. Mater. 8 (2013) 025005 C Li et al

Figure 9. Quantitative analysis of GAG levels associated with thecell-scaffolds constructs engineered in vitro at 2 and 4 weeks.Data shown are means ± SD (n = 3, ∗P < 0.05 means the samescaffolds compared with different time point).

microsphere diameters used in this study were below 50 μm,and the median particle diameter of the microspheres was37.2 μm. The particle-leaching method has been broadlyemployed to fabricate porous scaffolds, in which porosity,pore size distribution and inter-connectivity can be preciselyregulated by tailoring the fabrication parameters such as thesize of sodium chloride particles and amount of porogen. In ourstudy, sodium chloride particles with diameters (200–300 μm)were employed to create pores in the PLCL scaffold matrix.Therefore, the addition of relatively small PHBV microspheres(37.2 μm) had little effect on pore size of the PLCL scaffolds.Without any chemical crosslinking, the PHBV microsphereswere only incorporated into the PLCL matrix and thus no effecton scaffold degradation was observed (figure S7 available athttp://stacks.iop.org/BMM/8/025005/mmedia). It took morethan 1 year for PLCL scaffolds to degrade completely in vitro,while 81.3% mass of the PLCL scaffolds remained for morethan 15 weeks in vivo [5], and it took more than 1 year forPHBV scaffolds to degrade completely in vitro [27]. Our invitro degradation experiments of PLCL and composite 40%PHBV/PLCL scaffolds also confirmed this finding, and only10% of total mass of PLCL and 40% PHBV/PLCL scaffoldshad been degraded after 2 months (figure S7 available athttp://stacks.iop.org/BMM/8/025005/mmedia). Although thedegradation times of the two different polymers varied slightly,in vivo cartilage tissue formation was found to occur by3 months, well before scaffold degradation. During the pre-loading stress, a small amount of PHBV microspheres becamedetached from the PLCL matrix which probably did not impairthe structure or the ultimate Young’s modulus of compositePHBV/PLCL scaffolds, as the detached PHBV microsphereswere those which loosely integrated with the PLCL matrix.Therefore, this new system could provide an intact structurefor cartilage tissue regeneration.

Micro-scale surface topography has a great influence oncell attachment and proliferation [28–30]. Previous studiesdemonstrated that topography affects different cell types invarying ways and that different cell types prefer specific

Figure 10. Real-time quantification of Aggr and Col2 expression inchondrocytes cultured in PLCL scaffold, 20% PHBV/PLCL and40% PHBV/PLCL composite scaffolds. Expression was normalizedto GAPDH and presented as fold changes relative to level inchondrocytes. Data shown are means ± SD (n = 3, ∗P < 0.05means the same scaffolds compared with different time point).

surface topographies. Mesenchymal stem cells have beenshown to maintain their phenotype and high proliferationrate on island-patterned surfaces with 100 μm compared to60 μm island-patterned surface and flat topography [31].Cell adhesion and proliferation analysis showed that PHBVmicrospheres integrated into scaffolds did not significantlyimprove chondrocyte adhesion and proliferation, which mayindicate that changing surface topography using this size ofmicrospheres has little effect on chondrocyte adhesion andproliferation [32]. In order to distinguish the effects of the twomaterials, PHBV and PLCL, on the chondrocytes, the watercontact angles, cell activit, attachment and proliferation onthe PHBV and PLCL films were evaluated separately. All theresults suggested that there were no significant differences incell behavior on the PHBV and PLCL films (P > 0.05) (seesupporting information, figures S1, S2 and S3, available athttp://stacks.iop.org/BMM/8/025005/mmedia). In theory, themaximum area of human cell spreading is approximately4000 μm2 [33]. The area of islands formed by PHBVmicrospheres was about 2500 μm2, indicating that one PHBVmicrosphere could only attach one cell, but if the distancebetween PHBV microspheres was less than 20 μm, cellswould be able to spread across multiple islands. With anincreasing proportion of PHBV microspheres, the surfacetopography of composite scaffolds changed greatly. Once

10

Biomed. Mater. 8 (2013) 025005 C Li et al

(A) (B) (C)

(D) (E) (F)

(G) (H) (I)

Figure 11. Histology of explants at 4 weeks: HE staining for PLCL scaffolds/chondrocytes, 20% PHBV microspheres/PLCLscaffolds/chondrocytes, 40% PHBV microspheres/PLCL scaffolds/chondrocytes (A, D and G); Alcin Blue staining for PLCLscaffolds/chondrocytes, 20% PHBV microspheres/PLCL scaffolds/chondrocytes, 40% PHBV microspheres/PLCL scaffolds/chondrocytes (B, E and H); Immunological staining of type II Collagen for PLCL scaffolds/chondrocytes, 20% PHBV microspheres/PLCLscaffolds/chondrocytes, 40% PHBV 33 microspheres/PLCL scaffolds/chondrocytes (C, F and I)

(A) (B) (C)

(D) (E) (F)

(G) (H) (I )

Figure 12. Histology of explants at 8 weeks: HE staining for PLCL scaffolds/chondrocytes, 20% PHBV microspheres/PLCLscaffolds/chondrocytes, 40% PHBV microspheres/PLCL scaffolds/chondrocytes (A, D and G); Alcin Blue staining for PLCLscaffolds/chondrocytes, 20% PHBV microspheres/PLCL scaffolds/chondrocytes, 40% PHBV microspheres/PLCL scaffolds/chondrocytes (B, E and H); immunological staining of type II Collagen for PLCL scaffolds/chondrocytes, 20% PHBV microspheres/PLCLscaffolds/chondrocytes, 40% PHBV microspheres/PLCL scaffolds/chondrocytes(C, F and I)

11

Biomed. Mater. 8 (2013) 025005 C Li et al

the amount of PHBV microspheres reached 40%, a largenumber of PHBV microspheres emerged on the surface ofthe PLCL matrix and joined closely with each other. Whenchondrocytes were loaded onto the composite scaffolds, theytended to attach onto the PHBV microspheres at the surfaceof composite scaffolds, which may induce more cartilageextracellular matrix formation around the PHBV microspheresat 4 weeks. Higher ECM deposition was localized around thePHBV microspheres in the composite constructs suggestingthat chondrocytes adhering to the PHBV microspheres mighthave higher functionality. However, there was no significantvariation on overall GAG and type II collagen content or geneexpression among different scaffolds at 4 weeks. A previousstudy also demonstrated that although the incorporation ofPHBV microspheres into the PLGA matrix significantlyaltered the roughness of the matrix and increased thecompressive modulus of the composite scaffolds, the cellswere opt to adhere onto or among the PHBV microspheres inthe matrix [18].

Although there is not an advantageous cellular response toPHBV microsphere incorporation into PLCL scaffolds, apartfrom acquiring a higher compressive modulus scaffold, thisnew PHBV microsphere/PLCL system can also be potentiallyused to deliver bioactive factors or for controlled releaseof drugs while also functioning as a scaffold for cartilageregeneration. Differentiation of chondrocytes and secretion ofECM could further be regulated by immobilization of bioactivemolecules onto PHBV microspheres in newly fabricatedscaffolds in vitro.

5. Conclusions

Current PHBV microsphere/PLCL composite scaffolds havea relatively higher compressive strength than traditional PLCLscaffolds, good pore interconnectivity. The in vitro and invivo evaluation of chondrocytes in the composite scaffoldsdemonstrated that the new systems were biocompatible andcould serve as scaffolding structures for cartilaginous tissueformation. Future modifications to the scaffolds with bioactivemolecules incorporated with the PHBV microspheres couldalso potentially serve as a new system in delivery of bioactivefactors or drugs for functional cartilage tissue engineering.

Acknowledgments

This work is supported by National Basic Research Program ofChina (973 Program) (2011CB707500 and 2012CB619100)and National Natural Science Foundation of China grant(81271722) and the Wallace H. Coulter GT/Emory-PKU BMECollaborative Research Seed Grant Program. The authors alsothank Dr Zheng Yang for constructive advice.

References

[1] Setton L A, Elliott D M and Mow V C 1999 Alteredmechanics of cartilage with osteoarthritis: humanosteoarthritis and an experimental model of jointdegeneration Osteoarthritis Cartilage 7 2–14

[2] Li C et al 2012 A viscoelastic chitosan-modifiedthree-dimensional porous poly(L-lactide-co-ε-caprolactone)scaffold for cartilage tissue engineering J. Biomater. Sci.Polym. Ed. 23 405–24

[3] Ge Z et al 2012 Functional biomaterials for cartilageregeneration J. Biomed. Mater. Res. A100 2526–36

[4] Athanasiou K A et al 1991 Interspecies comparisons of in situintrinsic mechanical-properties of distal femoral cartilageJ. Orthop. Res. 9 330–40

[5] Jeong S I et al 2004 Morphology of elasticpoly(L-lactide-co-ε-caprolactone) copolymers and in vitroand in vivo degradation behavior of their scaffoldsBiomacromolecules 5 1303–9

[6] Kuo Y C and Ku I N 2008 Cartilage regeneration by novelpolyethylene oxide/chitin/chitosan scaffoldsBiomacromolecules 9 2662–9

[7] Jung Y et al 2010 The effect of hybridization of hydrogels andpoly(L-lactide-co-epsilon-caprolactone) scaffolds oncartilage tissue engineering J. Biomater. Sci., Polym. Ed.21 581–92

[8] Gong Z et al 2010 Use of synovium-derived stromal cells andchitosan/collagen type I scaffolds for cartilage tissueengineering Biomed. Mater. 5 055005

[9] Lu H et al 2011 Culture of bovine articular chondrocytes infunnel-like collagen-PLGA hybrid sponges Biomed. Mater.6 045011

[10] Jung Y et al 2008 Cartilaginous tissue formation using amechano-active scaffold and dynamic compressivestimulation J. Biomater. Sci., Polym. Ed. 19 61–74

[11] Jeong C G and Hollister S J 2010 A comparison of theinfluence of material on in vitro cartilage tissue engineeringwith PCL, PGS, and POC 3D scaffold architecture seededwith chondrocytes Biomaterials 31 4304–12

[12] Kesireddy V et al 2006 Development of growth factor/insulinrelease for articular cartilage regeneration in PLGAscaffolds Tissue Eng. 12 1073

[13] Yang Z et al 2012 Improved mesenchymal stem cellsattachment and in vitro cartilage tissue formation onchitosan-modified poly(L-lactide-co-epsilon-caprolactone)scaffold Tissue Eng. A 18 242–51

[14] Kim J Y and Cho D-W 2009 Blended PCL/PLGA scaffoldfabrication using multi-head deposition systemMicroelectron. Eng. 86 1447–50

[15] Savenkova L et al 2000 Mechanical properties andbiodegradation characteristics of PHB-based films ProcessBiochem. 35 573–9

[16] Kose G T et al 2005 Tissue engineered cartilage on collagenand PHBV matrices Biomaterials 26 5187–97

[17] Ye C et al 2009 PHB/PHBHHx scaffolds and humanadipose-derived stem cells for cartilage tissue engineeringBiomaterials 30 4401–6

[18] Huang W et al 2010 PHBV microspheres—PLGA matrixcomposite scaffold for bone tissue engineering Biomaterials31 4278–85

[19] Jung Y et al 2008 Application of an elastic biodegradablepoly(L-lactide-co-epsilon-caprolactone) scaffold forcartilage tissue regeneration J. Biomater. Sci., Polym. Ed.19 1073–85

[20] Jung Y et al 2008 Cartilage regeneration with highly-elasticthree-dimensional scaffolds prepared from biodegradablepoly(L-lactide-co-epsilon-caprolactone) Biomaterials29 4630–6

[21] Wang Y J et al 2007 Fabrication, characterization andlong-term in vitro release of hydrophilic drug usingPHBV/HA composite microspheres Mater. Lett.61 1071–6

12

Biomed. Mater. 8 (2013) 025005 C Li et al

[22] Ho M H et al 2004 Preparation of porous scaffolds by usingfreeze-extraction and freeze-gelation methods Biomaterials25 129–38

[23] Liu X and Ma P X 2009 Phase separation, pore structure, andproperties of nanofibrous gelatin scaffolds Biomaterials30 4094–103

[24] Yin J et al 2011 Characterization of human primarychondrocytes of osteoarthritic cartilage at varying severityChin. Med. J. 124 4245–53

[25] Wang C-C et al 2011 A highly organized three-dimensionalalginate scaffold for cartilage tissue engineering preparedby microfluidic technology Biomaterials 32 7118–26

[26] Temenoff J S and Mikos A G 2000 Review: tissue engineeringfor regeneration of articular cartilage Biomaterials21 431–40

[27] Sultana N and Khan T H 2012 In vitro degradation of PHBVscaffolds and nHA/PHBV composite scaffolds containinghydroxyapatite nanoparticles for bone tissue engineeringJ. Nanomater. 2012 190950

[28] Parker J A T C et al 2002 The effect of bone anchoring andmicro-grooves on the soft tissue reaction to implantsBiomaterials 23 3887–96

[29] Parker J A T C et al 2002 Soft tissue response tomicrotextured silicone and poly-L-lactic acid implants:fibronectin pre-coating versus radio-frequency glowdischarge treatment Biomaterials 23 3545–53

[30] Yap F L and Zhang Y 2007 Protein and cell micropatterningand its integration with micro/nanoparticles assemblyBiosens. Bioelectron. 22 775–88

[31] Lee I C et al 2009 The behavior of mesenchymal stem cells onmicropatterned PLLA membranes J. Biomed. Mater. Res. A91A 929–38

[32] Darling E M, Hu J C Y and Athanasiou K A 2004 Zonal andtopographical differences in articular cartilage geneexpression J. Orthop. Res. 22 1182–7

[33] Chen C S et al 1998 Micropatterned surfaces for control ofcell shape, position, and function Biotechnol. Progr.14 356–63

13