cl

ko

ces,

Keywords:Biodegradable materialsPLA/PLGA nanoparticlesmPEG-PLGA microparticlesControlled drug delivery

a dresent study aimed at developing biodegradable nanocomposite microparticles

Journal of Controlled Release 156 (2011) 1120

Contents lists available at ScienceDirect

Journal of Contro

evengineering therapies to restore damaged orworn-out parts of the body[1]. The current strategies have utilized the autologous stem cells incombination with supportive bioresorbable matrices and bioactivemolecules [2]. However, one of the biggest challenges in this approach isto create a three-dimensional structure, namely a functional cell scaffoldthat can promote the proliferation and differentiation of cells in order toregenerate the new tissue [3]. In the previous studies the injectablescaffolds have provided great opportunities for cell transplantation anddelivery of growth enhancing components with minimal invasivesurgery [4,5], thus the development of injectable scaffolds has gainedgreat attention in theelds of autologous cell therapy and drugdelivery.

microparticles have been reported previously in numerous differentapplications for the delivery of biomolecules, which provided a greatpotential for controlled drug release [14,15]. Properties of these systemscan easily be tuned using different PLA/PLGA polymers with varyingcompositions or molecular weights for obtaining the desired releaserate, as the drug release is predominantly controlled either by diffusionor by erosion depending on the polymer compositions and molecularweight [16,17]. Furthermore, some modied copolymer types, e.g.PEGylated PLA/PLGA, offer a more hydrophilic and exible surface,which could provide better cell afnity and biocompatibility [18].Therefore, bioactivity of scaffolds can be achieved by loading peptides,Various biomaterials, either natural (chitobrin [9]) or synthetic (polycaprolactone [10][11]), have been widely utilized to develop

Corresponding author at: Department of PharmaceuFaculty of Pharmaceutical Sciences, University of CopenhDenmark. Tel.: +45 3533 6346.

E-mail address: hmn@farma.ku.dk (H.M. Nielsen).

0168-3659/$ see front matter 2011 Elsevier B.V. Aldoi:10.1016/j.jconrel.2011.07.013nce of cell-based tissue

Polylactide/poly(lactide-co-glycolide) (PLA/PLGA) polymers are a classof materials accepted by FDA for injection in this area because of theirbiocompatibility and biodegradability [12,13]. PLA/PLGA-based nano-/The past decades have seen the emergeComposite scaffoldsTissue engineering

1. Introductionprepare nanoparticles by the double emulsion method. Uniform and spherical nanoparticles were obtained atan average size of 270300 nm. The thrombin receptor activator peptide-6 (TRAP-6) was successfully loadedin PLDL70 and PLGA85 nanoparticles. During the 30 days' release, PLDL70 nanoparticles showed sustainablerelease with only 30% TRAP-6 released within the rst 15 days, while almost 80% TRAP-6 was released fromPLGA85 nanoparticles during the same time interval. The release mechanism was found to depend on thecrystallinity and composition of the nanoparticles. Subsequently, mPEG-PLGA nanocomposite microparticlescontaining PLDL70 nanoparticles were produced by the ultrasonic atomization method and evaluated tosuccessfully preserve the intrinsic particulate properties and the sustainable release prole, which wasidentical to that of the nanoparticles. Good cell adhesion of the human broblasts onto the nanocompositemicroparticles was observed, indicating the desired cell biocompatibility. The presented results thusdemonstrate the development of nanocomposite microparticles tailored for sustainable drug release forapplication as injectable cell scaffolds.

2011 Elsevier B.V. All rights reserved.san [6], alginate [7,8],, polypropylene fumaratebiodegradable scaffolds.

proteins or bioacbased nano-/mic

Physiologicalregulatory factomatrix moleculeThrombin acts assites, thus it is berepair [19]. A nu

tics and Analytical Chemistry,agen, DK-2100 Copenhagen ,

l rights reserved.nd poly(L-lactide-co-glycolide) (PLGA85) were used toAccepted 9 July 2011Available online 20 July 2011with sustained drug delivery properties thus potentially being suitable for autologous stem cell therapy. Semi-crystalline poly(L-lactide/DL-lactide) (PLDL70) aReceived 20 April 2011 bioactive molecules. The pBiodegradable nanocomposite microparticell scaffolds

Yanhong Wen a, Monica Ramos Gallego b, Lene FeldsEva Horn Mller a, Hanne Mrck Nielsen a,a Department of Pharmaceutics and Analytical Chemistry, Faculty of Pharmaceutical Scienb Coloplast A/S, Global R&D, Biomaterials, Holtedam 1, DK-3050 Humlebk, Denmark

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

Article history: Injectable cell scaffolds play

j ourna l homepage: www.e lses as drug delivering injectable

v Nielsen b, Lene Jorgensen a, Hanne Everland b,

University of Copenhagen, Universitetsparken 2, DK-2100, Copenhagen, Denmark

ual role in tissue engineering by supporting cellular functions and delivering

lled Release

i e r.com/ locate / jconre ltive molecules into and releasing from the PLA/PLGA-roparticles to guide cell growth and tissue formation.ly, tissue repair is guided by the stimuli of localrs, such as different growth factors or extracellulars, to promote specic responses to the target cells [5].a serine proteinase to mediate the clotting at woundlieved to participate in regulating all stages of tissuember of growth and chemotactic factors are released

2. Materials and methods

2.1. Materials

Poly(L-lactide:DL-lactide molar ratio 70:30) (PLDL70) and poly(L-lactide:glycolide molar ratio 85:15) (PLGA85) were provided by PuracBiochem (Gorinchem, The Netherlands) at a purity of 95%. Methyl-poly(ethylene glycol)-poly(DL-lactide:glycolide molar ratio 50:50)(mPEG-PLGA), randomly arranged with 6.25% (w/w) mPEG (Mw2 kDa), was provided by Coloplast A/S (Humlebk, Denmark) at apurity of 99%. Thrombin receptor activator peptide-6 (TRAP-6) with a

12 Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120from the platelets and endothelial cells following the thrombintreatment [20]. Some studies reported that the encapsulation ofthrombin-related peptides in cell scaffolds, such as the TP508 peptide,which represents the non-proteolytic receptor-binding domain ofthrombin, can be benecial for bone regeneration [21,22]. It isbelieved that thrombin-induced cellular activity is mediated bybinding to a G-protein-coupled cell surface receptor, i.e. theprotease-activated receptor PAR-1 [23]. The present study includesthe use of another thrombin-related peptide, i.e. thrombin receptoractivator peptide-6 (TRAP-6), as TRAP-6 was previously determinedas the most potent sequence for activation of PAR-1 and mimics someof the thrombin activities [24].

It is challenging for the manufacturing process to fulll all therequirements of scaffolds with a desirable three-dimensional mor-phology and mechanical structure, which could sufciently supportthe cells and provide an ideal environment for cell growth. Surfaceand interior morphology, e.g. porosity and interconnectivity, areessential to cell adhesion and cell proliferation, but also important tothe diffusion of nutrients and waste compounds [25,26]. Manytechniques have been used to prepare microparticles with appropri-ate morphology and structure. The water-in-oil-in-water (w/o/w)double emulsion-solvent evaporation method is often used to obtainPLGA microparticles with different porosities and morphologies[27,28]. This method can be modied in order to achieve a poroussurface or interior structure of microparticles by using a gas foamingmethod or by the addition of hydrogen peroxide [27,29]. However, inan industrial perspective, the double emulsion method may to someextent have limitations with regard to up-scaling. Ultrasonic atom-ization is an alternative in being a gentle, versatile and reliabletechnique for microparticle preparation, which has been proven toencapsulate peptide or protein drugs without damaging the drugintegrity and efcacy [30,31]. The major advantage of this method isthat different sizes of uniform microspheres can be produced byadjusting the vibration frequency, the polymer concentration and thesurface tension, so that the desirable morphology and drug encapsu-lation of the microspheres can be obtained [32]. As ultrasonicatomization is an easily up-scalable and industrially reliable method,it has been used to fabricate the microparticles with differentencapsulation properties in this study.

Among the wide variety of different scaffolds, the commonapproach is to bring cells into the defect area and to support thecellular functions during the early phases of the regeneration process[33]. A simple delivery system used for cartilage repair will be themicrocarriers or microparticles combined with the chondrocytes,which will be injected into the cartilage defects through anarthroscopic procedure and localized by a xative precursor such asbrin glue [34]. Ideally, the chondrocytes may be cultured in thepresence of the microparticles for a period of time prior to use andoptionally in addition with the adhesives such as the bronectin,which could facilitate cell adhesion and in-growth as reportedpreviously [35]. Thus the chondrocytes are attached onto the surfaceor within the structure to form the particlecell constructs withsustained delivery of bioactive molecules capable of facilitating thecartilage formation.

The objective of this study was to develop a novel type ofbiodegradable nanocomposite microparticles as drug deliveringinjectable cell scaffolds. Two semi-crystalline polymers, namely poly(L-lactide/DL-lactide)(PLDL70) and poly(L-lactide-co-glycolide)(PLGA85), were used to prepare TRAP-6-loaded nanoparticles. Thecharacteristics and the release mechanism of the two types ofnanoparticles with different composition and crystallinity wereinvestigated. The nanoparticles showing superior sustained drugrelease were further incorporated into the mPEG-PLGAmicroparticlesto produce nanocompositemicroparticles, whichwere examinedwithregard to physical characteristics, drug release property and cell

adhesion.peptide sequence SFLLRN (Mw 748.9 Da, 95% purity) was purchasedfrom CASLO Laboratory Aps. (Lyngby, Denmark). Polyvinyl alcohol(PVA) 403 was from Kuraray Chemical Co. Ltd. (Osaka, Japan).Coumarin-6 (N99.0% purity) was purchased from Sigma-Aldrich. Allthe other chemicals were obtained commercially at analytical grade.

2.2. RP-HPLC analysis of TRAP-6 peptide

The TRAP-6 concentration was analyzed by RP-HPLC (LaChrom,Hitachi Ltd., Tokyo, Japan) using an auto injector (Hitachi-Merck,LaChrom L7200). The separation of TRAP-6 was done on a C18 column(50 mm4.6 mm, 2.6 m, Kinetex, Phenomenex, Allerd, Denmark)at a constant ow of 0.8 ml/min with uorescence detection of thephenylalanine in the TRAP-6 (Ex 258 nm/Em 282 nm). A mobilephase composed of solvent A (95% H2O/5% acetonitrile/0.1%TFA) andsolvent B (5% H2O/95%acetonitrile/0.1%TFA) at an A:B ratio of 85:15was used. The limit of detection (LOD) and the limit of quantication(LOQ) were obtained to be 0.3 g/ml (n=3) and 1 g/ml (n=3) byinjection of 30 l samples. All samples were injected three times.

2.3. Preparation of nanoparticles

The properties of the polymers used for the preparation ofnanoparticles are listed in Table 1.

PLDL70 and PLGA85 were selected due to their solubilityproperties in dichloromethane (DCM) and acetone.

TRAP-6 loaded nanoparticles were prepared by the water-in-oil-in-water (w/o/w) double emulsion-solvent evaporation method.Briey, 250 l of TRAP-6 solution (0.8 mg/ml in phosphate-bufferedsaline (PBS), pH 7.4) was added to 500 l of 20 mg/ml PLDL70 orPLGA85 solution in DCM. The mixture was emulsied by sonicationfor 90 seconds using a UP100H ultrasonic processor (HielscherUltrasonics GmbH, Teltow, Germany) to form the primary emulsion.1 ml of 2% (v/v) PVA in MilliQ water was immediately added to theprimary emulsion and whirlmixed well for 60 seconds. The mixturewas sonicated for another 60 seconds to produce the second emulsionbefore transfer and consequent dilution to 5 ml of 2% (v/v) aqueousPVA solution. The resulting double emulsion was stirred magneticallyfor 3 hours to evaporate all the organic solvent. As controls, non-loaded nanoparticles were prepared in the same way, except that250 l PBS buffer was added to the polymer solution instead of theTRAP-6 solution. For imaging, uorescent nanoparticles were pre-pared by adding 0.25% (w/w) coumarin-6 to the DCM phase.

Table 1Properties of polymers used for nanoparticles.

Polymers L:Gratioa

Mwb

(kDa)Meltingrange(C)

Inherentviscosity(dl/g)

Solubilityc

DCM Acetone

PLDL70 100:0 850 113.1120.3 3.85 S NSPLGA85 85:15 560 126.9144.4 3.11 S NS

a Lactide:glycolide ratio.b Average molecular weight.c Solubility reported on the data sheet from the material provider, S: soluble, NS:insoluble.

13Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120Nanoparticles were subsequently collected by ultracentrifugation(Eppendorf Refrigerated Microcentrifuge, Hamburg, Germany) inEppendorf tubes at 18,000g and 4 C for 12 min. All nanoparticleswere washed twice with MilliQ water and dried in a freeze dryer(Epsilon 24 LSC freeze dryer, Osterode am Harz, Germany) aspreviously reported [36] for further usage. At least 3 batches of eachtype of nanoparticle were prepared.

2.4. Characterization of nanoparticles

2.4.1. Particle size and zeta potentialParticle size and size distribution of all nanoparticles were

determined by dynamic light scattering, applying photon correlationspectroscopy (PCS). Freeze-dried nanoparticles were easily re-dispersed in MilliQ water applying an ultrasonic water bath andwhirlmixer. The mean particle diameter (Z-average) and polydisper-sity index (PDI) of the nanoparticles were measured by intensity at25 C using a Malvern NanoZS (Malvern Instruments Ltd., Worcester-shire, UK) equipped with a 633-nm laser and 173 detection optics.

The zeta potential of the nanoparticles was measured by the laserDoppler electrophoresis technique using the Malvern NanoZS. Ananosphere size standard (602.7 nm, Duke Scientic Co., USA) anda zeta potential transfer standard (686.8 mV Malvern Instru-ments Ltd., Worcestershire, UK) were used to verify the performanceof the instrument. All the samples were measured in triplicate.Malvern DTS v.5.10 software (Malvern Instruments Ltd., Worcester-shire, UK) was used for data acquisition and analysis. For viscosity andrefractive index, values of pure water were used.

2.4.2. MorphologyThe surface morphology and size range of the nanoparticles were

observed by atomic force microscopy (AFM Nanoscope V, VeecoInstruments Inc., NY, USA). Nanoparticles were re-suspended inMilliQ water and the suspensions were drop-cast onto the Si/SiO2substrates. Subsequently, nanoparticles were scanned by AFM in thetapping mode after the evaporation of the water. Representativeimages were selected from imaging 3 different samples.

2.4.3. Encapsulation efciencyThe encapsulation efciency (EE) of TRAP-6 loaded into nanopar-

ticles was determined by extracting the TRAP-6 from the freeze-driednanoparticles. First, 1 mg of nanoparticles was dissolved in 200 l ofchloroform and then 0.5 ml of PBS buffer was added. The mixture waswhirlmixed and kept rotating end-over-end for 2 hours to facilitatethe TRAP-6 partitioning into the aqueous phase. The organic andaqueous phases were separated by ultracentrifugation at 18,000g and4 C for 20 min. The supernatants in the water phase were collected,and the TRAP-6 content was quantied by RP-HPLC as described inSection 2.2. Three separate samples were extracted, and at least threedifferent batches were measured. The encapsulation efciency (%)was calculated as the ratio of the measured drug loading to thetheoretical drug loading (Eq. (1)).

Encapsulation efficiency EE% = Measured drug loadingTheoretical drug loading

100%

1

2.5. Preparation of nanocomposite microparticles

The nanocomposite microparticles were prepared by an ultrasonicatomization method developed at Coloplast A/S as described [34].Briey, 40 mg of freeze-dried PLDL70 nanoparticles loaded withTRAP-6 or coumarin-6 were re-dispersed in 20 ml of a 5% (w/w)

mPEG-PLGA solution in acetone. The suspension was delivered via asyringe pump into a 25 KHz ultrasonic atomizer (at nozzle, Sono-tekCo., NY, USA) at ow rate 1 ml/min. The atomized microdrops wereformed through the vibration of the ultrasonic nozzle and weresprayed out when the energy on the atomizing surface reached theamplitude needed. After evaporation of the solvent, the microdropswere hardened to form the microparticles. Conventional microparti-cles (i.e. without nanoparticles)were produced in the sameway, exceptthat 0.4 mg TRAP-6 powder was dispersed in the mPEG-PLGA solutionbefore being pumped into the atomizer. Both the nanocompositemicroparticles and the conventional microparticles were collected inethanol kept at low temperature. The particles were then collected anddried under vacuum for 48 hours for further analysis.

2.6. Nanocomposite microparticle characterization

2.6.1. Particle sizeParticle size parameters (D50%, D90%) were determined by laser

diffraction using a Malvern Mastersizer 2000 (Malvern Instrument Ltd.,Worcestershire, UK). Approximately 10 mg of nanocomposite micropar-ticles or conventional microparticles were dispersed in MilliQ water andadded into a small volume sample dispersion unit of theMastersizer. Theaverage diameter was measured by volume and plotted as particle sizedistributions. The span value describes the size distribution as it is ameasure of the width of the volume distribution relative to the mediandiameter, as described in Eq. (2), whereN% (N=10, 50, 90) is the volumepercentage of microspheres with diameters of DN%. Each of the samples(n=3) was measured in triplicate.

Span = D90%D10% =D50%; 2

2.6.2. MorphologyThe microparticle surface and interior morphology were deter-

mined by a scanning electron microscope (SEM, JSM-5200 JEOL Ltd.,Tokyo, Japan). Microparticles were sliced to an approx. thickness of2 mm by cryostat sectioning (CM1100, Leica Microsystems, Nussloch,Germany). Dried microparticles or sliced particle sections werespread out on a metal specimen stub and gold coated by sputtercoating prior to the imaging. Representative images were selectedfrom imaging of 2 different samples.

2.6.3. Nanoparticle distributionVisualization of the nanoparticle distribution in the composite

particles was achieved by incorporating coumarin-6 (Ex 464 nm/Em518 nm) into the PLDL70 nanoparticles, as described in Section 2.3.Imaging was done by confocal laser scanning microscopy (CLSM)using a LSM 510 microscope equipped with an argon laser (458 and488 nm) and a HeNe laser (543 nm) (Carl Zeiss GmBH, Jena,Germany). Data analysis was done using LSM 510 software. Imagesare representative of at least ve sample examinations, each from twodifferent batches.

2.7. In vitro release studies

2.7.1. In vitro TRAP-6 release from nano- and microparticlesIn vitro release studies were performed with PLDL70 and PLGA85

nanoparticles, as well as with the conventional microparticles andnanocomposite microparticles. Non-loaded particles served as con-trols. For nanoparticles, 10 mg of non-loaded or TRAP-6-loadednanoparticles (PLDL70 and PLGA85, respectively) were suspendedin 1 ml PBS (pH 7.4) using protein low-bind eppendorf tubes; formicroparticles, 100 mg non-loaded or TRAP-6-loaded conventionalmicroparticles and nanocomposite microparticles were suspended in1 ml PBS (pH 7.4). Both the nanoparticles and microparticles werepre-wetted in PBS using an ultrasonic water bath for 10 min. All the

suspensions were incubated at 370.1 C in a linear shaking water

14 Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120bath at 0.18 min1 (Grant Instruments Ltd., Cambridgeshire, UK). Atpredetermined time internals, all samples were ultracentrifuged at18,000g for 10 min before aliquots of supernatants (125 l) werecollected. After replenishing with an equivalent volume of PBS, allrelease samples were re-dispersed using an ultrasonic water bath. Thecollected supernatants were stored at 20 C before analysis by RP-HPLC. Each release study was done in triplicate.

In order to evaluate the physical state of the polymers andnanoparticles in the release study, original PLDL70 and PLGA85polymers, freshly-made blank PLDL70 and PLGA85 nanoparticles, aswell as nanoparticles at selected time points during the release study,were collected, freeze dried and analyzed by differential scanningcalorimetry (DSC) (TAC7/DX, Perkin Elmer, Waltham, USA). Thermo-grams were obtained on 25 mg samples heated in sealed aluminumpans at a scanning range rate of 10 C/min in a nitrogen atmosphere inthe range 20180 C(n=3). Pyris-Instrument managing software(Perkin Elmer, MA, USA) was used for data analysis.

2.8. In vitro cell adhesion

2.8.1. Fibroblast cultureHuman broblasts isolated from mamma reductions by explana-

tion were cultured in Dulbecco's Modied Eagle Medium (DMEM)containing 10% (v/v) fetal calf serum (FCS) (D10), antibiotics(penicillin (100 U/ml) and streptomycin (100 g/ml)). The broblastswere trypsinated, counted and cultured for 3 days together with thenanocomposite as well as conventional microparticles in T75 in Poly-HEMA-coated asks. The initial culturing density of the cells was1104 cells/ml, and the particle concentration was 1.5 mg/ml.

2.8.2. Critical point dryingAt day 3 of seeding, the broblasts together with the micropar-

ticles, samples were washed and prepared for imaging by an E3100critical point dryer (Quorum Technologies Ltd., East Grinstead, UK). Inbrief, particles seeded with cells were washed 2 times with PBSfollowed by xation for 1 and 4 hours in 1% and 3% (v/v), respectively,glutaraldehyde in MilliQ. The cell-seeded particles were subsequentlydehydrated by ethanol. For this, samples were placed in a gradient ofethanol, i.e. 30%, 50%, 70%, 80%, 90%, 95% and 100% (v/v), each for15 min. After dehydration, further substitution with CO2-miscibleliquid-amyl acetate was carried out through a series of graded bathsbetween ethanol and amyl acetate, i.e. 75:75, 50:50, 25:75 (15 mineach), and eventually the samples were immersed in 100% (v/v) amylacetate and placed in the chamber of the critical point dryer. Thechamber was lled with liquid CO2, and subsequently amyl acetatewas removed completely. Samples were immersed in liquid CO2 for 1hour before increasing the temperature to the critical point of CO2(31.5 C, 1100 psi) at which liquid CO2 vaporized without surfacetension effects, which maintained the intact morphology of cellsamples. The dried samples were sputter coated and imaged by SEMas described above.

3. Results

3.1. Characterization of nanoparticles

PLDL70 and PLGA85 nanoparticles loaded with TRAP-6 werereproducibly obtained by the double emulsion-solvent evaporationmethod. AFM images showed both spherical PLDL70 and PLGA85nanoparticles with a relatively smooth surface (Fig. 1). The PLDL70and PLGA85 nanoparticles had a mean size range of 275300 nmwitha narrow size distribution and zeta potential ranging from 35.1 to40.5 mV. Compared to the non-loaded nanoparticles, TRAP-6loading did not signicantly affect the particle size and zeta potential.Coumarin-6 loaded PLDL70 nanoparticles for the confocal imaging

showed identical size distribution (data not shown). The encapsula-tion efciency was improved by process optimization from initial 5%to 15.4% and 10.3% for PLDL70 and PLGA85 nanoparticles, respective-ly. The characteristics of nanoparticles are summarized in Table 2.

3.2. In vitro TRAP-6 release from nanoparticles

Accumulative release of TRAP-6 from PLDL70 and PLGA85 nano-particleswasmonitored over 30 days. PLDL70 nanoparticles displayed asustained release prole (Fig. 2a), which can be divided into threephases: (i) initial burst release of 10% TRAP-6 during the rst day; (ii)slow sustained release of additional 20% peptide over 15 days; (iii)faster release of an additional 50% of the encapsulated TRAP-6 until the30th day. After the 30 days, over 20% TRAP-6 was recovered from theremaining PLDL70 nanoparticles, providing a mass balance of 955%.In contrast, PLGA85 nanoparticles showed a rapid release with almost50% initial burst release in the rst day and 80% released over 15 days.Hardly any TRAP-6 was recovered from the PLGA85 nanoparticles after30 days, which shows the complete release of TRAP-6 from PLGA85nanoparticles. As a consequence of the release proles, PLDL70nanoparticles with higher sustained release were selected for thefurther formulation of nanocomposite microparticle.

In order to explain the different release proles of these two types ofnanoparticles, DSC thermograms of the pure polymer and nanoparticleson day 0, day 1, day 10 and day 18 of the release experiment wereanalyzed. The samples for the DSC corresponded to different phases ofthe release prole as indicated in Fig. 2. Overall, the DSC datademonstrated differences in the phase transition observed betweenthe polymer and freshly-made nanoparticles, as well as among thenanoparticles from the release experiment (Fig. 3). The glass transitiontemperature (Tg) of PLDL 70 nanoparticles was observed at approxi-mately 40 C and increased slightly during the release experiment, yet,the Tg of PLGA 85 nanoparticles did not seem to increase until the 18thday. The PLDL70 polymer showed twomelting peaks (Fig. 3a), in whichthe higher peak at 113123 C was identical to the melting point(Table 1), while the lower peak at 6065 C probably indicated themelting point of the existing L-lactide oligomers in the PLDL70 polymer[37], which may be the residuals from the polymer production or adegradation product formed during the storage. However, there was nomelting peak in the formulated PLDL70 nanoparticles (Fig. 3b), whichindicates that the PLDL70 in the nanoparticles was in the amorphousstate caused by the preparation process. But on day 1 and day 10 afterinitiation of the release study, the peak at 6065 C reappeared,probably indicating certain crystalline growth (Figs. 3c and d). At day18, the peak at 6065 C disappeared (Fig. 3e), which most likelyindicates an amorphous state of PLDL70 nanoparticles due to theabsenceof anymeltingpeak. Similar to the PLDL70polymer, thePLGA85polymer also demonstrated a semi-crystalline structure with twomelting peaks, which were the melting points of PLGA85 at 126144 C (Table 1) and L-lactide oligomers at 6065 C, respectively(Fig. 3f). The two peaks were still present in the thermograms of thefreshly-made PLGA85 nanoparticles, i.e. at day 0 (Fig. 3g), demonstrat-ing that, unlike the PLDL70 nanoparticles, the crystalline structure wasnot highly affected by the preparation process. The PLGA85 nanopar-ticles also showed an increase of the peak at 6065 C during the releasebetween day 1 and day 10 (Fig. 3h and i), but it was not observed at day18 which is similar to that of PLDL70 nanoparticles. In all samples ofPLGA85, including the polymer and nanoparticles, the higher meltingpeak was observed (Fig. 3fj), indicating that the PLGA nanoparticleswere inpartially crystalline structure at all timesduring the release up to18 days.

3.3. Characterization of nanocomposite microparticles

According to the release prole, TRAP-6-loaded PLDL70nanoparticleswere selected and incorporated into mPEG-PLGA microparticles to form

nanocompositemicroparticles. As shown in Table 3, PLDL70/mPEG-PLGA

not shown); however, the interior structure demonstrated relativelymore hollow structure compared to the nanocomposite (Fig. 4c andd).

Fig. 5 shows the distribution of PLDL70 nanoparticles labeled withcoumarin-6 in the composite microparticles. Horizontal scanning ofthe nanocomposite microparticles in 4 m steps demonstrated thatthe PLDL70 nanoparticles (in green) were observed primarily on theouter layers of the microparticles, most likely due to the much denserstructures of the outer layers which contained more nanoparticles.Leakage of coumarin-6 from the nanoparticles during the processingof the nanocomposite particles was checked for and observed not tohave occurred.

3.4. Sustained release from nanocomposite microparticles

In order to investigate the release properties of PLDL70/mPEG-PLGAnanocompositemicroparticles, TRAP-6 release from the nanocompositemicroparticles was studied over 30 days in vitro (Fig. 6). Compared tothe conventional microparticles, which resulted in more than 70% burstrelease within the rst day, the nanocomposite nanoparticles showed aprolonged release prole similar to that of PLDL70 nanoparticles,indicating that the properties of the nanoparticles were not affected bythe process of ultrasonic atomization. The results show that the

15Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120nanocomposite microparticles showed a mean particle size of76.1 m. Compared to the conventional microparticles with a meanparticle size of 82.6 m, the inclusion of nanoparticles did notsignicantly change the particle size of composites (P=0.3). The sizerange of the nanocomposite microparticles prepared by ultrasonicatomization was in an acceptable dimension (80100 m) for use asthe injectable cell scaffolds [38].

SEM images showed the spherical nanocomposite microparticleswith uniform size and porous surface (Fig. 4a and b). The conventionalmicroparticles displayed similar size and surface morphology (data

nanocompositemicroparticles have the ability to preserve the sustainedrelease prole of the PLDL70 nanoparticles [39,40].

Fig. 1. Atomic force microscopy (AFM) images of (a) PLDL70 nanoparticles, scalebar=500 nm and (b) PLGA85 nanoparticles, scale bar=250 nm.

Table 2Characteristics of the prepared nanoparticles.

Polymers Loadeddrug

Diameter(nm)

PDIa Zeta potential(mV)

EEb(%)

PLDL70 276.86.5 0.1290.013 35.11.6 TRAP-6 286.43.5 0.1860.011 37.22.0 15.44.5

PLGA85 276.65.2 0.1630.009 38.31.5 TRAP-6 292.29.6 0.1310.012 40.51.2 10.33.6

Samples represent the mean of three determinationsstandard deviation.a Polydispersity index.b Encapsulation efciency.Fig. 2. In vitro TRAP-6 release from (a) PLDL70 nanoparticles and (b) PLGA85nanoparticles over 30 days (MeanSD; n=3). Black arrows indicated the selected

time points when the release samples were analyzed by DSC.

16 Y. Wen et al. / Journal of Controlled Release 156 (2011) 11203.5. Cell attachment on the nanocomposite microparticles

After culturing the broblasts for 3 days, it was observed that thecells attached well on the surface of PLDL70/mPEG-PLGA nanocom-posite microparticles (Fig. 7). The cells with a size of approximately10 m preferably adhered in-between the nanocomposite micropar-ticles (as indicated by the red circles) forming broblast-particleaggregates during the cultivation, which was indicative of good cellbiocompatibility. Particlecell constructs were evidently observed,which would allow three-dimensional cell growth and eventuallyformation of new tissue along with the degradation of the micropar-ticles. A cell proliferation study showed that the broblasts culturedon the nanocomposite microparticles had a 4-fold higher proliferationthan cells in suspension (data not shown).

4. Discussion

Cell scaffolds are the key components in the tissue regenerationparadigm, as they not only function as the structural support to allowcell adhesion, proliferation and growth, but also if possible as adelivery depot for the controlled release of bioactive molecules. In

Fig. 3. DSC curves of polymers and nanoparticles at different time intervals during therelease: (a) PLDL70 polymers; (b, c, d, e) PLDL70 nanoparticles from the release on day0, 1, 10 and 18, respectively; (f) PLGA85 polymers; (g, h, i, j) PLGA85 nanoparticles fromthe release on day 0, 1, 10, 18, respectively.

Table 3The characteristics of the conventional and nanocomposite microparticles.

Microparticles Loaded nanoparticles (w/w) D50% (m) Span value

Conventional 82.68.5 1.70.5Nanocomposite 4% PLDL70 76.19.4 1.50.2

Samples represent the mean of three determinationsstandard deviation.many studies, the development of composite scaffolds has been infocus for demonstrating superior properties with regard to cell afnityand controlled drug release, e.g. for bone or cartilage regeneration[41,42]. With the aim of developing injectable cell scaffolds withsustained drug delivery function, nanocomposite microparticles weretailor-designed by incorporating TRAP-6-loaded PLDL70 nanoparti-cles into mPEG-PLGA microparticles. This study displayed thecharacterization and careful selection of the appropriate nanoparticlesas the drug delivery system, and subsequently the incorporation intocomposite microparticles as cell scaffolds with controlled releaseproperties.

Two biodegradable polymers with crystalline structure, PLDL70and PLGA85 were chosen for preparation of the nanoparticles basedon their insolubility in acetone, which is essential to the following stepfor successful preparation of the microparticulate cell scaffold.Nanoparticles were prepared by the double emulsion-solvent evap-oration method, which is well known to be applicable for reproduc-ibly obtaining fairly homogenous nanospheres with an average size of200250 nm [36,43,44]. PLDL70 and PLGA85 nanoparticles showedgood colloidal stability without aggregation and easy re-dispersionafter freeze drying. However, the challenge when using the doubleemulsion method is the resulting low encapsulation efciency ofhydrophilic drugs to the PLGA polymer matrices, which is largely dueto the low afnity of such drugs to the hydrophobic polymers. Somestudies demonstrated that the loading of hydrophilic drugs can beimproved by increasing the partitioning of the drug to the organicphase, for instance by tuning the pH of aqueous phase [45],complexation with phospholipids [46] or ion-pairing agents [47,48].Other studies also demonstrated that optimizing the processingparameters via a factorial design can help to achieve higher drugloading [36]. As described [49], the encapsulation efciency of water-soluble drug was mostly around 5%20%, and in the present study theoptimizations increased the TRAP-6 encapsulation from around 5%15%. Further enhancement of the encapsulation efciency may beachievable, but it was not the primary aim of the present study.

Even though the PLDL70 and PLGA85 nanoparticles displayed asimilar surface morphology and encapsulation efciency, the releasestudy showed different release proles, which is probably due to thedifferences in crystallinity and composition between PLDL70 andPLGA85. DSC thermal analysis data demonstrated changes in thephase transition before and after the production of nanoparticles, aswell as during the release period. Considering the originally semi-crystalline structure of both PLDL70 and PLGA85 polymers, freshly-made PLDL70 nanoparticles (at day 0) seemed to be in the amorphousstructure, while PLGA85 nanoparticles were still partially in thecrystalline state. During the rst 10 days of the release study, thecrystallization process of L-lactide oligomers occurred in both PLDL70and PLGA85 nanoparticles, probably due to the exposure to moistureand the temperature of the incubation condition [50]. The DSCanalysis supported the release proles, indicating that that the releasefrom PLDL70 nanoparticles was both diffusion and erosion controlled,especially as seen from the triphasic release of TRAP-6 from PLDL70nanoparticles [51]. In the rst 15 days' lag time, the PLDL70nanoparticles displayed a much lower initial burst release thanPLGA85 nanoparticles. This was probably because of the betterdispersion of drug in the amorphous polymer matrix with gooddrugpolymer interactions, which consequently decreased the diffu-sion rate from PLDL70 nanoparticles, while the zero-order releasestarting from day 15 was most likely dominated by the initiation ofsurface erosion [52], which was conrmed by the DSC curve showingthe crystalline structure changed to amorphous on day 18. Themechanism of the release from the PLGA85 nanoparticles is alsoexpected to be a result of diffusion. The existing crystalline structurein PLGA85 nanoparticles hampered the dispersion of drug in thematrices, which consequently resulted in rapid initial diffusion from

the polymer matrices. Previous studies also reported similar ndings

17Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120that crystalline particles resulted in a much higher initial burst releasethan the amorphous ones [50,53]. In addition, the slight increases inthe Tg of both nanoparticles were due to the physical aging of the

Fig. 4. Scanning electron microscopic (SEM) images of (a) nanocomposite micropar-ticles, scale bar=100 m; (b) Surface porosity of nanocomposite microparticles, scalebar=1 m; (c) Interior structure of conventional microparticles, scale bar=5 m; (d)Interior structure of nanocomposite microparticles, scale bar=10 m.polymer during the storage time [54], during which the polymerchains approached the equilibrium state and the cohesional entan-glements were formed by chain relaxation; therefore more energy isneeded to disengage the tangled chain structures, which resulted inthe Tg shifting to a higher temperature [55].

Besides the crystallinity, the composition of the polymers is alsolikely to be related to the drug release proles. Many studies haveshown that the L:G ratio of PLGA polymers strongly inuences thedrug release [56,57]. The rather fast initial burst release from PLGA85nanoparticles could also be caused by the more hydrophilic PLGAcompared to PLA in the PLDL70. The more hydrophilic polymer has ahigher tendency to take up the water and hydrate the polymermatrices, which would relax the polymer chain and increase thediffusion of drug. Furthermore, the inuence would be moresignicant on the release of a drug with more hydrophilic property,like TRAP-6 has compared to a hydrophobic drug [58], as thehydrophilic drug has a higher afnity to the water than to thepolymer matrix.

The sustained drug release was considered an imperative criterionfor selecting PLDL70 nanoparticles to be incorporated into mPEG-PLGA microparticles. For the purpose of obtaining successful com-posite scaffolds, many reports have been published on blends ofvarious multifunctional materials to enhance the properties of thescaffolds [41,59], and on imbedding nano-structures into the scaffoldsto obtain more controllable structures and properties [60,61]. The

Fig. 5. Distribution of PLDL70 nanoparticles (in green) in nanocomposite microparti-cles. Images from left to right represent horizontal scans in 4 m steps obtained byconfocal microscopy.

TRAP-6 release from nanocomposite microparticles conrmed thesustained release ability of scaffolds, which in turn indicated that thePLDL70 nanoparticles were not inuenced by the process and that theintrinsic properties were preserved in the composites. This can mostlikely be explained by the mild ultrasonic vibration during theatomization, as the energy applied during the ultrasonic atomizationwas merely a few watts, which would not affect the structure oractivity of encapsulated drug [30,31]. A similar study showed thatthrombin-related peptide-TP 508-loaded PLGA/PEG microparticleswere fabricated with poly (propylene fumarate)(PPF) to produce

the same time as providing the morphology suitable for broblastadhesion.

release of a bioactive drug over 30 days have been developed forpotential used as injectable cell scaffolds. The drug releasemechanism

Fig. 6. In vitro TRAP-6 release from conventional mPEG-PLGA microparticles (opencircle) and PLDL70/mPEG-PLGA nanocomposite microparticles (closed circle).

18 Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120PLGA/PEG/PPF injectable composite scaffolds and a variety of releaseproles were obtained by changing different processing parameters[42]. Therefore, loading and releasing of peptides, protein or otherFig. 7. SEM images of broblasts cultivated on nanocomposite microparticles for 3 days;the circles show the adhesion of broblasts on the composites (scale bar: Upper100 m; lower left 50 m; lower right 10 m).from two types of nanoparticles based on PLDL70 and PLGA85 wasshown to be related to the polymer crystallinity and composition. ThePLDL70 nanoparticles showed a more prolonged release and werecomposited with mPEG-PLGA microparticles to form nanocompositemicroparticles. The composites thus produced demonstrated asimilarly sustainable release as the nanoparticles. In addition, goodcell adhesion onto the nanocomposite microparticles further contrib-uted to the potential for use as injectable cell scaffolds. These resultsprovide a novel method for obtaining injectable composites withtailored drug delivery properties. Future work will be able toapply this method for the delivery of other bioactive substancesto achieve better in vitro and in vivo effects for autologous stem celltherapy.

Acknowledgements

The authors are grateful for nancial support from The DanishNational Advanced Technology Foundation (HTF), Drug ResearchAcademy (DRA), Hrslev Foundation and the Danish Agency forScience, Technology and Innovation. We acknowledge the expertassistance of Dorthe Kyed rbk, Niels Peter Aae Christensen (Facultyof Pharmaceutical Sciences, University of Copenhagen), ZhongmingWei (Faculty of Natural Sciences, University of Copenhagen), TinaRomme and Mette Trnes (Coloplast A/S) with the SEM, DSC, AFM,slicing of microparticles and culturing broblasts, respectively. Theauthors would also like to acknowledge Niels Coley for languagecorrections, Marco van derWeert (Faculty of Pharmaceutical Sciences,University of Copenhagen) and Lisbeth Grndahl (University ofThe cell adhesion study indicated good cell biocompatibility of thenanocomposite microparticles, as the broblasts readily attached tothe composite microparticles and built up cellparticle constructs in3 days. The release of TRAP-6 from the composite scaffolds would bebenecial to the inammatory and proliferative phases of woundhealing by regulating the cell functions [19]. It has been demonstratedin vivo that the stimulation by TRAP-6 released from PLGAmicroparticles promoted the cell proliferation and led to accelerationof tissue repair [19,63]. Therefore the drug delivering compositescaffolds could be potentially applied for the sustained release ofdifferent growth factors such as broblast growth factor (FGF) orinsulin-like growth factor (IGF) for the stimulation of bone or cartilageregeneration [5].

5. Conclusion

Novel nanocomposite microparticles demonstrating sustainedbiomolecules are achievable for different applications using thedesign of composite scaffolds.

Aiming at developing injectable cell scaffolds, the nanocompositemicroparticles of a size b100 m were considered to be applicablewithout the risk of blocking the syringe needle. Besides the size,surface and interior morphology are critical to the cell adhesion andgrowth. For instance, the porous scaffolds are more favorable to cellattachment and migration by providing a sufcient surface area and athree dimensional network [25,62], but unfortunately the porousstructure normally exhibited a quite rapid initial burst release due tothe diffusion through the large pores [27]. The challenge is thus toretain the relatively porous structure and simultaneously maintain asustained release of drugs. In this study, the nanocompositemicroparticles demonstrated a good ability to control the release atQueensland) for valuable suggestions.

19Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120References

[1] P. Bianco, P.G. Robey, Stem cells in tissue engineering, Nature 414 (6859) (2001)118121.

[2] C. Zandonella, Tissue engineering: the beat goes on, Nature 421 (6926) (2003)884886.

[3] M. Sittinger, D.W. Hutmacher, M.V. Risbud, Current strategies for cell delivery incartilage and bone regeneration, Current Opinion in Biotechnology 15 (5) (2004)411418.

[4] J.D. Kretlow, L. Klouda, A.G.Mikos, Injectablematrices and scaffolds for drug deliveryin tissue engineering, Advanced Drug Delivery Reviews 59 (45) (2007) 263273.

[5] S.H. Lee, H. Shin, Matrices and scaffolds for delivery of bioactive molecules in boneand cartilage tissue engineering, Advanced Drug Delivery Reviews 59 (45)(2007) 339359.

[6] S. Ladet, L. David, A. Domard, Multi-membrane hydrogels, Nature 452 (7183)(2008) 7679.

[7] C.S. Tan, A. Jejurikar, B. Rai, T. Bostrom, G. Lawrie, L. Grndahl, Encapsulation of aglycosaminoglycan in hydroxyapatite/alginate capsules, Journal of BiomedicalMaterials Research. Part A 91A (3) (2009) 866877.

[8] G. Coppi, V. Iannuccelli, E. Leo, M.T. Bernabei, R. Cameroni, Protein immobilizationin crosslinked alginate microparticles, Vol. 19, 2002, pp. 3744.

[9] J.C. Becker, M. Beckbauer, W. Domschke, H. Herbst, T. Pohle, Fibrin glue, healing ofgastric mucosal injury, and expression of growth factors: results from a human invivo study, Gastrointestinal Endoscopy 61 (4) (2005) 560567.

[10] E.Y. Teo, S.-Y. Ong, M.S. Khoon Chong, Z. Zhang, J. Lu, S. Moochhala, B. Ho, S.-H.Teoh, Polycaprolactone-based fused deposition modeled mesh for delivery ofantibacterial agents to infected wounds, Biomaterials 32 (1) (2011) 279287.

[11] J.W. Lee, K.S. Kang, S.H. Lee, J.-Y. Kim, B.-K. Lee, D.-W. Cho, Bone regeneration using amicrostereolithography-produced customized poly(propylene fumarate)/diethylfumarate photopolymer 3D scaffold incorporating BMP-2 loaded PLGA micro-spheres, Biomaterials 32 (3) (2011) 744752.

[12] J.-M. L, X. Wang, C. Marin-Muller, H. Wang, P.H. Lin, Q. Yao, C. Chen, Currentadvances in research and clinical applications of PLGA-based nanotechnology,Expert Review of Molecular Diagnostics 9 (4) (2009) 325341.

[13] L. Zhang, S.P. Schwendeman, in: S. Valle, E. Escher, W.D. Lubell (Eds.), Peptides forYouth, Vol. 611, Springer, New York, 2009, pp. 611613.

[14] S. Acharya, S.K. Sahoo, PLGA nanoparticles containing various anticancer agentsand tumour delivery by EPR effect, Advanced Drug Delivery Reviews 63 (3)(2011) 170183.

[15] J.S. Golub, Y.-T. Kim, C.L. Duvall, R.V. Bellamkonda, D. Gupta, A.S. Lin, D. Weiss, W.Robert Taylor, R.E. Guldberg, Sustained VEGF delivery via PLGA nanoparticlespromotes vascular growth, American Journal of Physiology. Heart and CirculatoryPhysiology 298 (6) (2010) H1959H1965.

[16] D. Klose, F. Siepmann, K. Elkharraz, J. Siepmann, PLGA-based drug deliverysystems: importance of the type of drug and device geometry, InternationalJournal of Pharmaceutics 354 (12) (2008) 95103.

[17] R. Bodmeier, K.H. Oh, H. Chen, The effect of the addition of low molecular weightpoly(DL-lactide) on drug release from biodegradable poly(DL-lactide) drugdelivery systems, International Journal of Pharmaceutics 51 (1) (1989) 18.

[18] J. Vange, K. Schwach-Abdellaoui, H. Everland, P.S. Nielsen, B. Nielsen, L.K.Jespersen, L.F. Nielsen, Degradable hydrophilic block copolymers with improvedbiocompatibility for soft tissue regeneration, US Patent WO/2007/101443 (2007)

[19] T.N. Dugina, E.V. Kiseleva, M.A. Lange, T.V. Vasil'eva, C. Grandls, E.A.Markvicheva, Z.D. Bespalova, M.E. Pal'keeva, S.M. Strukova, Effect of syntheticpeptide thrombin receptor agonist encapsulated in microparticles based on lacticand glycolic acid copolymer on healing of experimental skin wounds in mice,Bulletin of Experimental Biology and Medicine 138 (5) (2004) 463466.

[20] G. Grandaliano, A.J. Valente, H.E. Abboud, A novel biologic activity of thrombin:stimulation of monocyte chemotactic protein production, The Journal ofExperimental Medicine 179 (5) (1994) 17371741.

[21] L.R. Amir, G. Li, T. Schoenmaker, V. Everts, A.L. Bronckers, Effect of thrombinpeptide 508 (TP508) on bone healing during distraction osteogenesis in rabbittibia, Cell and Tissue Research 330 (1) (2007) 3544.

[22] Y. Wang, C. Wan, G. Szke, J.T. Ryaby, G. Li, Local injection of thrombin-relatedpeptide (TP508) in PPF/PLGA microparticlesenhanced bone formation duringdistraction osteogenesis, Journal of Orthopaedic Research 26 (4) (2008) 539546.

[23] S.M. Strukova, T.N. Dugina, I.V. Chistov, M. Lange, E.A. Markvicheva, S. Kuptsova, V.P.Zubov, E. Glusa, Immobilized Thrombin Receptor Agonist Peptide AcceleratesWound Healing in Mice, Vol. 7, 2001, pp. 325329.

[24] K.Y. Hui, J.A. Jakubowski, V.L. Wyss, E.L. Angleton, Minimal sequence requirementof thrombin receptor agonist peptide, Biochemical and Biophysical ResearchCommunications 184 (2) (1992) 790796.

[25] H.J. Chung, I.K. Kim, T.G. Kim, T.G. Park, Highly open porous biodegradablemicrocarriers: in vitro cultivation of chondrocytes for injectable delivery, TissueEngineering 14 (5) (2008) 607615.

[26] Q. Hou, P.A. De Bank, K.M. Shakesheff, Injectable scaffolds for tissue regeneration,Journal of Materials Chemistry 14 (13) (2004) 19151923.

[27] S.E. Bae, J.S. Son, K. Park, D.K. Han, Fabrication of covered porous PLGAmicrospheres using hydrogen peroxide for controlled drug delivery andregenerative medicine, Journal of Controlled Release 133 (1) (2009) 3743.

[28] J. Wang, B.M. Wang, S.P. Schwendeman, Characterization of the initial burstrelease of a model peptide from poly(D, L-lactide-co-glycolide) microspheres,Journal of Controlled Release 82 (23) (2002) 289307.

[29] H.J. Chung, I.K. Kim, T.G. Kim, T.G. Park, Highly Open Porous BiodegradableMicrocarriers, Vitro Cultivation of Chondrocytes for Injectable Delivery, Vol. 14,2008, pp. 607615.[30] B. Bittner, T. Kissel, Ultrasonic atomization for spray drying: a versatile techniquefor the preparation of protein loaded biodegradable microspheres, Journal ofMicroencapsulation 16 (3) (1999) 325341.

[31] Y. Yeo, K. Park, A new microencapsulation method using an ultrasonic atomizerbased on interfacial solvent exchange, Journal of Controlled Release 100 (3)(2004) 379388.

[32] B. Albertini, N. Passerini, L. Rodriguez, Evaluation of ultrasonic atomization as anew approach to prepare ionically cross-linked chitosan microparticles, Journal ofPharmacy and Pharmacology 57 (7) (2005) 821829.

[33] R. Stoop, Smart biomaterials for tissue engineering of cartilage, Injury 39 (2008)7787 (1, Supplement 1).

[34] H. Everland, P. Samuelsen, J. Vange, C. Clausen, M.R. Gallego, Compositions andmethods for augmentation and regeneration of living tissue in a subject, US PatentWO/2009/106641 (2009)

[35] C. Bouf, O. Thomas, C. Bony, A. Giteau, M.-C. Venier-Julienne, C. Jorgensen, C.Montero-Menei, D. Nol, The role of pharmacologically active microcarriersreleasing TGF-[beta]3 in cartilage formation in vivo by mesenchymal stem cells,Biomaterials 31 (25) (2010) 64856493.

[36] D. Cun, C. Foged, M. Yang, S. Frkjr, H.M. Nielsen, Preparation and character-ization of poly(dl-lactide-co-glycolide) nanoparticles for siRNA delivery, Int JPharm 390 (1) (2010) 7075.

[37] S.J. de Jong, S.C. De Smedt, J. Demeester, C.F. van Nostrum, J.J. Kettenes-van denBosch, W.E. Hennink, Biodegradable hydrogels based on stereocomplexformation between lactic acid oligomers grafted to dextran, Journal ofControlled Release 72 (13) (2001) 4756.

[38] K.W. Chun, H.S. Yoo, J.J. Yoon, T.G. Park, Biodegradable PLGA microcarriers forinjectable delivery of chondrocytes: effect of surface modication on cellattachment and function, Biotechnology Progress 20 (6) (2004) 17971801.

[39] F. Yang, X. Qu, W. Cui, J. Bei, F. Yu, S. Lu, S. Wang, Manufacturing and morphologystructure of polylactide-type microtubules orientation-structured scaffolds, Bio-materials 27 (28) (2006) 49234933.

[40] G. Chen, D. Liu, N. Maruyama, H. Ohgushi, J. Tanaka, T. Tateishi, Cell adhesion ofbone marrow cells, chondrocytes, ligament cells and synovial cells on a PLGA-collagen hybrid mesh, Materials Science and Engineering C 24 (68) (2004)867873.

[41] M. Prabaharan, M.A. Rodriguez-Perez, J.A. de Saja, J.F. Mano, Preparation andcharacterization of poly(L-lactic acid)-chitosan hybrid scaffolds with drug releasecapability, Journal of Biomedical Materials Research. Part B, Applied Biomaterials81B (2) (2007) 427434.

[42] E.L. Hedberg, A. Tang, R.S. Crowther, D.H. Carney, A.G. Mikos, Controlled release ofan osteogenic peptide from injectable biodegradable polymeric composites,Journal of Controlled Release 84 (3) (2002) 137150.

[43] K. Avgoustakis, A. Beletsi, Z. Panagi, P. Klepetsanis, A.G. Karydas, D.S. Ithakissios,PLGA-mPEG nanoparticles of cisplatin: in vitro nanoparticle degradation, in vitrodrug release and in vivo drug residence in blood properties, Journal of ControlledRelease 79 (13) (2002) 123135.

[44] T. Grner, R. Gref, D. Michenot, F. Sommer, M.N. Tran, E. Dellacherie, Lidocaine-loaded biodegradable nanospheres. I. Optimization of the drug incorporation intothe polymer matrix, Journal of Controlled Release 57 (3) (1999) 259268.

[45] F. Tewes, E. Munnier, B. Antoon, L. Ngaboni Okassa, S. Cohen-Jonathan, H.Marchais, L. Douziech-Eyrolles, M. Souc, P. Dubois, I. Chourpa, Comparative studyof doxorubicin-loaded poly(lactide-co-glycolide) nanoparticles prepared bysingle and double emulsion methods, European Journal of Pharmaceutics andBiopharmaceutics 66 (3) (2007) 488492.

[46] F. Cui, K. Shi, L. Zhang, A. Tao, Y. Kawashima, Biodegradable nanoparticles loadedwith insulin-phospholipid complex for oral delivery: preparation, in vitrocharacterization and in vivo evaluation, Journal of Controlled Release 114 (2)(2006) 242250.

[47] Y. Chen, F. Wang, H.A.E. Benson, Effect of formulation factors on incorporation ofthe hydrophilic peptide dalargin into PLGA and mPEG-PLGA nanoparticles,Peptide Science 90 (5) (2008) 644650.

[48] D. Ranney, P. Antich, E. Dadey, R. Mason, P. Kulkarni, O. Singh, H. Chen, A.Constantanescu, R. Parkey, Dermatan carriers for neovascular transport target-ing, deep tumor penetration and improved therapy, Journal of ControlledRelease 109 (13) (2005) 222235.

[49] F. Ito, H. Fujimori, K. Makino, Factors affecting the loading efciency of water-soluble drugs in PLGA microspheres, Colloids and Surfaces. B, Biointerfaces 61 (1)(2008) 2529.

[50] Y. Aso, S. Yoshioka, T. Terao, Effects of storage on the physicochemical propertiesand release characteristics of progesterone-loaded poly(L-lactide) microspheres,International Journal of Pharmaceutics 93 (13) (1993) 153159.

[51] B.S. Zolnik, P.E. Leary, D.J. Burgess, Elevated temperature accelerated releasetesting of PLGA microspheres, Journal of Controlled Release 112 (3) (2006)293300.

[52] S.C.J. Loo, Z.Y.S. Tan, Y.J. Chow, S.L.I. Lin, Drug release from irradiated PLGA andPLLA multi-layered lms, Journal of Pharmaceutical Sciences 99 (7) (2010)30603071.

[53] S. Izumikawa, S. Yoshioka, Y. Aso, Y. Takeda, Preparation of poly(L-lactide)microspheres of different crystalline morphology and effect of crystallinemorphology on drug release rate, Journal of Controlled Release 15 (2) (1991)133140.

[54] A. Rawat, D.J. Burgess, Effect of physical ageing on the performance ofdexamethasone loaded PLGA microspheres, International Journal of Pharmaceu-tics 415 (12) (2011) 164168.

[55] K. Liao, D. Quan, Z. Lu, Effects of physical aging on glass transition behavior of poly(L-lactide), European Polymer Journal 38 (1) (2002) 157162.

[56] A. Budhian, S.J. Siegel, K.I. Winey, Controlling the in vitro release proles for asystem of haloperidol-loaded PLGA nanoparticles, International Journal ofPharmaceutics 346 (12) (2008) 151159.

[57] L. Mu, S.S. Feng, A novel controlled release formulation for the anticancer drugpaclitaxel (Taxol): PLGA nanoparticles containing vitamin E TPGS, Journal ofControlled Release 86 (1) (2003) 3348.

[58] C. Wischke, S.P. Schwendeman, Principles of encapsulating hydrophobic drugs inPLA/PLGA microparticles, International Journal of Pharmaceutics 364 (2) (2008)298327.

[59] T. Guo, J. Zhao, J. Chang, Z. Ding, H. Hong, J. Chen, J. Zhang, Porous chitosan-gelatinscaffold containing plasmid DNA encoding transforming growth factor-[beta]1 forchondrocytes proliferation, Biomaterials 27 (7) (2006) 10951103.

[60] K.S. Jack, S. Velayudhan, P. Luckman, M. Trau, L. Grndahl, J. Cooper-White, Thefabrication and characterization of biodegradable HA/PHBV nanoparticlepolymercomposite scaffolds, Acta Biomaterialia 5 (7) (2009) 26572667.

[61] X. Shi, B. Sitharaman, Q.P. Pham, F. Liang, K. Wu,W. Edward Billups, L.J. Wilson, A.G.Mikos, Fabrication of porous ultra-short single-walled carbon nanotubenanocomposite scaffolds for bone tissue engineering, Biomaterials 28 (28)(2007) 40784090.

[62] M.H. Sheridan, L.D. Shea, M.C. Peters, D.J. Mooney, Bioabsorbable polymerscaffolds for tissue engineering capable of sustained growth factor delivery,Journal of Controlled Release 64 (13) (2000) 91102.

[63] A. Rusanova, A. Makarova, S. Strukova, E. Markvicheva, L. Gorbachyova, K.Stashevskaya, T. Vasil'eva, E. Sidorova, Z. Bespalova, C. Grandls, Thrombinreceptor agonist peptide immobilized in microspheres stimulates reparativeprocesses in rats with gastric ulcer, Bulletin of Experimental Biology andMedicine142 (1) (2006) 3538.

20 Y. Wen et al. / Journal of Controlled Release 156 (2011) 1120

Biodegradable nanocomposite microparticles as drug delivering injectable cell scaffolds1. Introduction2. Materials and methods2.1. Materials2.2. RP-HPLC analysis of TRAP-6 peptide2.3. Preparation of nanoparticles2.4. Characterization of nanoparticles2.4.1. Particle size and zeta potential2.4.2. Morphology2.4.3. Encapsulation efficiency

2.5. Preparation of nanocomposite microparticles2.6. Nanocomposite microparticle characterization2.6.1. Particle size2.6.2. Morphology2.6.3. Nanoparticle distribution

2.7. In vitro release studies2.7.1. In vitro TRAP-6 release from nano- and microparticles

2.8. In vitro cell adhesion2.8.1. Fibroblast culture2.8.2. Critical point drying

3. Results3.1. Characterization of nanoparticles3.2. In vitro TRAP-6 release from nanoparticles3.3. Characterization of nanocomposite microparticles3.4. Sustained release from nanocomposite microparticles3.5. Cell attachment on the nanocomposite microparticles

4. Discussion5. ConclusionAcknowledgementsReferences