Development of an injectable PHBV microparticles-GG hydrogel hybrid system for regenerative medicine

Development

of

an

injectable

PHBV

microparticles-GG

hydrogel

hybrid

system

for

regenerative

medicine

Daniela

P.

Pacheco

a,b,c,

_*,

_Maria

_H.

_Amaral

c

_,

_Rui

_L.

_Reis

a,b

_,

_Alexandra

_P.

_Marques

a,b

_,

Vítor

M.

Correlo

a,b,

*

a_3B’s

ResearchGroup–Biomaterials,BiodegradablesandBiomimetics,UniversityofMinho,HeadquartersoftheEuropeanInstituteofExcellenceonTissue EngineeringandRegenerativeMedicine,AvePark,4806-909Taipas,Guimarães,Portugal

b_{ICVS/3B’s–}

PTGovernmentAssociateLaboratory,Braga/Guimarães,Portugal

c _{LaboratoryofPharmaceuticalTechnology/CentreofResearchinPharmaceuticalSciences,FacultyofPharmacy,UniversityofPorto,4050-313Porto,Portugal}

Received22August2014

Receivedinrevisedform13November2014 Accepted14November2014

Availableonline18November2014

1. Introduction Anumberofdifferentpolymers,eithersyntheticornatural, have been exploited forthe production of biodegradableDDS. These systems are cleaved through chemical or enzyme-catalyzed hydrolysisthusallowingthereleaseoftheactiveagentina controlled mannerand ata sustainedconcentration, within its therapeutic window.Thereleaseprofilesoftheactiveagentsfrom biodegradable polymers, in addition to a dependence with the degradation mechanismandkineticsofthepolymers(Ciçeketal., 1995;Mietal., 2002;ZhangandChu,2002),canbecontrolledbya numberofother factors, such as physicochemical properties of the polymersand active agents (Calandrelli et al., 2002; Abraham et al., 2003), thermodynamic compatibility between the polymers and active agents (Liu et al., 2004), as well as the shape of the devices (Chen and Lee, 2001; Tunón et al., 2003; Fulzele et al., 2004).

Among the biodegradable polymersthat have been explored for the production ofDDS,poly(lactide-co-glycolide) (Wenet al., 2013), poly(L-lactic acid)(Correiaet al.,2013),chitosan(Champa

andBhat, 2010), alginate (Iwanaga et al., 2013), starch-poly-

e

-caprolactone (Balmayor et al., 2009) and polyhydroxyalkanoates (PHAs) are the most used. From the PHA family, polyhydroxybutyrate (PHB) and its co-polymer with hydroxyvalerate (PHBV) inparticular, in addition

Thedesignandproductionofefficientdrugdeliverysystems (DDS) areofvitalimportanceintissueengineering(TE)and regenerative medicine(RM)(Biondietal.,2008).Ideally,DDS intendtosustaina particular active agents_’ concentration in patients’ blood and/or tissues for a defined and extended time, avoiding repeated administration(Amassetal.,1998;Kostand Langer,2001).Injectable DDSthathavebeenproposedsofarrely onliposomes(Monteiroetal., 2013),micelles(Xuetal.,2013), polymericmicro-andnanoparticles (Kozielskietal.,2013;Liang etal.,2013),cyclodextrins(Dhuleetal., 2012),amongothers (Oliveiraetal.,2011).

* Correspondingauthorat:3B’sResearchGroup–Biomaterials,Biodegradables and Biomimetics, Department of Polymer Engineering, University of Minho, HeadquartersoftheEuropeanInstituteofExcellenceonTissueEngineeringand RegenerativeMedicine,AvePark,Taipas,4806-909Guimarães,Portugal;ICVS/3B’s, AssociateLaboratory, PT Government AssociateLaboratory, Guimarães, Braga, Portugal.Tel.:+351918236302.

toTEapplications(Zonarietal.,2014),havebeenalsoproposedas delivery systems for various active agents (Shishatskaya et al.,

2008;Chen andTong, 2012;Kabilan et al.,2012).PHAs

micro-particles(MPs)havebeendevelopedbydifferent microencapsula-tion approaches that rely on solvent-evaporation methods, emulsification–diffusion,amongothers.Thefirstapproachoffers advantagesover othermethods, since it is highlyreproducible, presents the ability to serve as carriers of hydrophilic and hydrophobic active agents with high incorporation efficiency, which release profile can also be controlled by theprocessing factors(Ogawaetal.,1988;Yeoetal.,2001).Ontheotherhand,the emulsi_{fication–diffusion} is a very versatiletechnique, in which different types of emulsions can be used, though oil/water emulsionsareofinterestbecausetheyusewaterasthenonsolvent, simplerandmoreeconomic,sinceitfacilitatesthewashingstep andminimizesagglomeration(PintoReisetal.,2006).However, thismethodcanonlybeappliedtohydrophobicactiveagents,and limitations are imposed by the scale-up of the high energy requirementsinhomogenization(Soppimathetal.,2001).Relying onthesemicroencapsulationmethods,differentapplicationshave beenproposed. Instudiesofantibioticdelivery,gentamicinwas incorporatedintoPHAsMPstopreventinfectionsassociatedwith boneimplantsandscaffolds(Francisetal.,2011),whilst Rodríguez-ContrerasproducedPHAsMPstodeliverydoxycycline( Rodríguez-Contrerasetal.,2013).Moreover,Polettoandco-workersadopted theemulsification–diffusiontechniqueusingabinarymixtureof differentproportionsofchloroformandethanolaimingtodecrease thesizeoftheparticles(Polettoetal.,2008).

Althoughthemicroparticulatesystemspresent many advan-tages,uponinjectiontheyarecontinuouslyexposed toexternal mechanicaldeformation,leadingtouncontrollabledisplacements (Fitzgeraldetal.,1987;Griffith,2000;ChanandSo, 2005).This greatlyaffects the concentration of active agents at thetarget tissues,whichcanleadtoimpairedoutcomes,suchas,forexample, limited tissue regeneration (Lemperle et al., 2004). Moreover, whenin freemovementtheparticulate systemcanbeexpelled beforethereleaseoftheactiveagent.Therefore,extensiveefforts are being made to tackle this issue while harnessing the advantages of particulate systems, including their minimally invasiveadministrationandthecontrolleddeliveryoftheactive agent. In this sense, the incorporation of MPs in injectable hydrogels was hypothesized in order to develop a minimally-invasiveDDSthatwouldprovideMPslocalizeddeliveryavoiding uncontrolledmigration through thehuman body,along witha sustainedreleaseofanactiveagent.

Fromtheclinicalpointofview,injectablehydrogelsformedbyin situ polymerization are highly desirable due to: (1) enabling minimallyinvasiveapplication;(2)highcapacitytoadapttothe tissuedefect;and(3)highcompatibilitywithincorporatedactive agentsandcellsifdesired(Jin,2012).Inthisregard,gellangum(GG) hydrogelsthathavebeenshowntobeasuitableplatformtosupport and deliver cells in different TE applications as extensively characterizedinourGroup,wereselected(Coutinhoetal.,2010; Oliveiraetal.,2010;daSilvaetal.,2014).Thedoubleemulsi_fication– solventevaporationmethoddescribedbyOgawaetal.(1988)was adoptedtodevelop PHBVMPs, carrying bovineserum albumin (BSA) or dexamethasone (Dex) as hydrophilicand hydrophobic active agents’ models, respectively. This method is highly reproducible and has beenextensively used tocreate particles withasphericalandinnercore-shellstructurethatcanserveas carrierofwater-solubleandinsolubleactiveagents.The character-isticsoftheproducedMPsarelargelydependentonthetypeof polymer(hydrophobic/hydrophilic)and activeagentused.Thus, this methodwas modified,togetherwith theproperties of the hydrogeltotailorthedeliveryprofileoftheactiveagents.Inthis way,anewinsitugellinginjectablesystem,incorporatingPHBV

MPs loaded with both relevant hydrophilic and hydrophobic molecules,whichallowstheirlocalizedandcontrolledreleaseis hereinproposed.

2. Materialsandmethods 2.1.PreparationofPHBVMPs

The PHBV (molecular weight (Mw)=425,692gmol
1) (PHB IndustrialS.A.,Brazil)MPswerepreparedbyusingamodification of thedoubleemulsification–solventevaporationmethod previ-ouslydescribedbyOgawa(Ogawaetal.,1988).Brie_fly,3mLofa 0.5% (w/v) poly(vinyl alcohol) (PVA) (Sigma–Aldrich, Germany) aqueoussolution(W1)wasmixedwith15mLofa3%(w/v)PHBV solution in chloroform–ethanol (O). Two different ratios of chloroform–ethanolwereused inorder toverifyitseffect over thereleaseprofileoftheactiveagents,namely100:0chloroform: ethanol (absence of ethanol) and 90:10 chloroform:ethanol (presenceofethanol).Themixturewasemulsifiedfor3minusing anUltra-TurraxT18(IKA-Werke,Germany)inordertoobtainthe firstemulsion(W1/O).Theprimaryemulsionwas thendropped into150mLof1%(w/v)PVAsolution(W2)andhomogenized(RW 16basicIKA-Werke,Germany)during5minformingthesecondary emulsion (W1/O/W2).The finalW1/O/W2 emulsionwas leftto evaporate under magnetic stirring and the obtainedMPs were collected by centrifugation, washed at least three times with deionized water and freeze-dried. BSA was dissolved at a concentrationof0.1%(w/v)withinthefirstaqueousphase(W1), whilst Dex was dissolved in the organic phase at a final concentration of 3.310
3% (w/v). Unloaded-MPs were also prepared,ascontrol,usingthesameprotocolpreviouslydescribed, butintheabsenceoftheactiveagents.

2.2.IncorporationofPHBVMPsinGGhydrogels

ForthepreparationoftheinjectableGGhydrogels,apreviously describedprocedurewasfollowed(daSilvaetal.,2014).GelzanCM (Sigma–Aldrich,Germany)powderwasmixedatroom tempera-turewithdeionizedwaterattwodifferentconcentrations,0.75% (w/v)and1.25%(w/v),underconstantstirring.Thesolutionwas heated at 90C and kept at this temperature for 30min. Afterwards, 30mg of active agent-loaded PHBV MPs were dispersed with 20mL of GG solution. Subsequently, calcium chloride(VWR,USA)wasaddedtothissolutionataconcentration of0.18%(w/v)toactascrosslinker.Finally,thesolutionwasleftat roomtemperatureduringapproximatelyonehour, allowingthe formationofthegel,whichwasthenmaintainedina phosphate-bufferedsaline(PBS)solution.

2.3.QuantificationoftheMPs’productionyield

The yield of the particles prepared under the different conditionsdefinedinSection2.1wascalculatedusingEq.(1). yieldð%Þ¼wx

wy 100

(1)

TheyieldwascalculatedbasedontheweightofMPs(wx)and the weight of the compounds used to prepare the MPs (wy), polymerorpolymerplusactiveagent.

2.4.QuantificationoftheincorporationefficiencyofDexandBSAinto thePHBVMPs

AimingatdeterminingtheBSAincorporationefficiency(IE)into the MPs, 10mg of BSA- or Dex-loaded MPs were completely

dissolved in chloroform with vigorous shaking and at room temperaturefor24h.ThisallowsthedissolutionofPHBVandat thesametimepermitsthereleaseoftheentrappedactiveagent. Then,10mLofPBSwasaddedinordertodissolvetheBSA/Dex. Aftercentrifugation,the supernatant was collectedand filtered througha0.2mmmembrane.Theconcentrationofthe incorpo-ratedBSAwasmeasuredusingtheMicroBCATM_{ProteinAssayKit} (Pierce,IL)followingmanufacturer’sinstructions.Opticaldensity oftheBSA-andDex-containingsolutionswasmeasuredat562and 242nm,respectively,inaSynergyHTmulti-detectionmicroplate reader (Bio-Tek’s Gen5TM_{, USA). The amounts of BSA and Dex} withintheMPswereobtainedfromthecalibrationcurves.TheIEof theactiveagentswithinPHBVMPswascalculatedusingEq.(2)as follows:

IEð%Þ¼½BSAor DexwithinMPs

Initial½BSAor Dex 100 (2)

2.5.Invitroreleasestudies

InvitroreleaseofBSAandDexfromPHBVMPswasevaluatedup to21days.Forthat,30mgofMPswereincubatedin10mLofPBS (0.01M,pH7.4),andmaintainedinaprecisionwaterbath(Grant, UK)at60rpmand 37.00.5C. Atdefinedtime points,1mLof supernatantwascollectedandanequivalentamountoffreshPBS at37.00.5C was addedto maintainthe total volumeof the sample.TheamountofreleasedBSAandDexwasdeterminedas describedabove.Asimilarprocedurewasfollowedtodetermine thereleaseprofileofBSAandDexfromthePHBVMPsembedded withintheGGhydrogels.

2.6.Analysisofthekineticsandmechanismofrelease

ThereleasemechanismofBSAandDexfrombothPHBVMPs and PHBV MPs embedded in GG hydrogels was analyzed by applyingtheresultsindifferentkineticmodels(CostaandSousa Lobo,2001).Thezero-orderthatexplainstheactiveagentrelease rate as an independent phenomenon of the active agent concentration (Donbrow and Samuelov, 1980). The first-order usuallyapplied toactiveagent dissolution studies(Gibaldiand

Feldman,1967;Wagner,1969).TheHiguchithatisconcernedon

the release of water soluble and low soluble active agents incorporatedin semi-solid and/orsolid matrices, and describes theactiveagentreleaseas adiffusionprocessbasedonFickian diffusion (Higuchi, 1961, 1963). And the Korsmeyer–Peppas, generallyusedtoanalyzethereleaseofpharmaceuticalpolymeric dosageforms,whenthereleasemechanismisnotwellknownor whenmorethanonetypeofreleasephenomenacouldbeinvolved

(KorsmeyerandPeppas,1981;Korsmeyeretal.,1983;Costaand

SousaLobo,2001),weretheappliedmodels.Thedetermination

coefficient(R2_{)wasusedasanindicatorofthebestfittingofthe} data for each model. R2 _{and model parameters of the release} profiles were calculated using SigmaPlot 12.5 software (Systat SoftwareInc.,USA).

2.7.Laserdiffractionspectrometry

Thesize distributionof thepreparedMPs was measuredby laserdiffractionspectrometry(CoulterLS230,CoulterElectronics, USA).Thedriedsamplesweresuspendedina0.2%(w/v)Tween1 80 (Sigma–Aldrich, Germany) solution and sonicated for 5min withan ultra-sound probe(Sonoswiss SW 6H, Sonoswiss1 AG,

Fig.1.Unloaded-PHBVMPsphysicochemicalproperties.(A)YieldoftheprocessandsizeofthePHBVMPsproducedunderdifferentconditions.*p<0.05,**p<0.01 comparingthePHBVMPsproducedintheabsenceandinthepresenceofethanol.(B)SEMmicrographsofPHBVMPsobtainedinthe(i–iii)absenceand(iv–vi)presenceof ethanolintheorganicphase.Thesequentialmicrographsrepresent(iandiv)anoverviewoftheobtainedMPs,(iiandv)thedetailoftheirsurfaceand(iiiandvi)respective internalmorphology.

Switzerland) before measurement. The obtained homogeneous suspensionwasusedfortheparticlesizedistributionanalysis. 2.8.Fouriertransforminfraredspectroscopy

Fouriertransforminfraredspectroscopy(FTIR)analyses,under Transmissionmode,wereperformedinordertoanalyzePHBVMPs (unloaded,loadedandafterrelease)chemicalcomposition.With thispurpose,1mgofMPsweremixedwith40mgofpotassium bromide(KBr)andthenprocessedintoadiscina manualpress (161–1100handpress,Piketechnologies,Madison,WI).FTIR-KBr spectra(IRPrestige21,Shimadzu,Japan)wererecordedat40scans witharesolutionof4cm
1,from1400to4000cm
1.

2.9.Scanningelectronmicroscopy

Themorphologyof thePHBVMPsandPHBVMPsembedded withinGGhydrogelswasanalyzed beforeand aftertheinvitro release studies,byscanningelectronmicroscopy (SEM)using a S360microscope(LeicaCambridge,UK).Beforebeinganalyzed,the differentstructuresweregoldsputtercoated(Fisonsinstruments, UK)for2minat15mA.Micrographswererecordedat5.0kVand differentmagnifications.

2.10.Rheologicalanalysis

Cone–platerheometrywasconductedforGGhydrogelswith and without embedded PHBV MPs to assess their rheological behaviorasfunctionoftemperatureandtime.Forthispurpose,in eachmeasurement,avolumeof2mLofthesamplewasplacedin thebottomplateof therheometer(MalvernKinexusRotational Rheometer,UK)and allowedtostabilizefor1minataconstant temperatureof 5C beforestarting theanalysis. Measurements wereperformedbyincreasingthetemperaturefrom5 to50C(at

aheatingrateof1C/min)andapplyingaconstantshearstressof 0.1Pa.

2.11.Statisticalanalysis

The results of at least three independent experiments are presentedasmeanstandarddeviation(SD).Theyield,particle size and IE data were analyzed using the one-way analysis of variance(ANOVA)(

a

=0.05)andTukey’smultiplecomparisontest using Graphpad Prism version 5 (Graphpad Software, USA). Significancedifferencesweresetforp<0.05.

3.Results

3.1.EffectofprocessingmethodoverPHBVMPsproperties

The PHBV MPs wereproduced following a double emulsi fi-cation–solventevaporationmethodusingamixtureofchloroform: ethanolintwodifferentproportionsinordertoassesstheeffectof theprocessingconditionsovertheMPspropertiesandsubsequent incorporationcapacityandreleaseprofileofBSAandDex.Theyield in the preparation of PHBV MPs (Fig. 1A) revealed to be significantly(p<0.05)affectedbythesolventcomposition. The resultsshowedthatthepresenceofethanolintheorganicphase ledtoadecreasedof6%intheyield.Likewise,thesizedistribution of the MPs was also significantly (p<0.01) affected by the processing conditions. The meanparticlesize of unloaded-MPs obtainedinthepresenceofethanol(around58

m

m)wasfoundto belargerthaninitsabsence(around42

m

m),withadistributionof rangefrom15.6to76.4

m

mand5.6to91.1

m

m,respectively.

Concerning the morphology of the produced MPs (Fig.1B), perfectsphericalshapeMPswereobtainedindependentlyofthe processing conditions (Fig. 1B; i and iv). However, a detailed analysisofMPssurfacetopographyrevealedroughsurfaces,both

Fig.2. BSAandDexloaded-PHBVMPsphysicochemicalproperties.(A)YieldoftheprocessandsizeofthePHBVMPsproducedunderdifferentconditions.*p<0.05,**p<0.01, ***p<0.001comparingtheloadedandunloaded-PHBVMPsproducedunderthesameconditionsandincorporatingthesameactiveagent.(B)SEMmicrographsofBSA loaded-PHBVMPsobtainedinthe(iandii)absence;and(iiiandiv)presenceofethanolintheorganicphase.Thesequentialmicrographsrepresentan(iandiii)overviewof theobtainedMPsand(iiandiv)thedetailoftheirsurface.

withthe presenceofsomemicro-size pores,but withdifferent topography(Fig.1B;iiandv).Moreover,bothinthepresenceand absenceofethanol,MPspresentedaporouscore(Fig. 1B;iiiandvi). Nonetheless,theadditionofethanoltotheorganicphaseledtothe formationoflargerporeintheinnerstructure(Fig.1B;vi). 3.2.Loaded-MPsproperties

3.2.1.BSAandDexentrapment

TheimpactoftheincorporationofactiveagentswithinPHBV MPs over the process yield and particle size was determined (Fig.2A).Whilenosigni_ficantdifferenceswereobservedinterms ofyieldwhenproducingunloaded-PHBVMPs(Fig.1A)andactive agent-loadedMPs(Fig.2A)(p>0.05),thesizeoftheloaded-PHBV MPs was affected by the incorporation of the active agents (Fig.2A).Intheabsenceofethanol,themeansizeofthe loaded-PHBV MPs, independently of the incorporated agent, was significantlyhigherthantheunloaded-MPs.Thisdifference was notobservedinthepresenceofethanol.Furthermore,themean MPssizesignificantlyincreasedwiththemolecularweight(Mw)of theactiveagentthatwasincorporated.Thus,theBSA-loadedPHBV MPsexhibitedahighermeanparticlesizewhencomparedwiththe BSA-loadedPHBVMPs(Fig.2A).

TheSEManalysisoftheloaded-PHBVMPsconfirmednotonlythe particlesizemeasurements,butalsothe maintenanceof thespherical morphologyafterincorporationoftheactiveagents(Fig.2B;iand iii),whichwasindependentoftheincorporatedmolecule.

TheIEofBSAandDexwithinthePHBVMPsobtainedunder differentexperimentalconditionsrevealedtobeindependentof the processing conditions (p>0.05). An IE around 66% was achieved for BSA both in presence and absence of ethanol (Fig. 3A). Although not significantly different (p>0.05), the presence of ethanol in the organic phase resulted in higher (around 92% vs 82%) IE of Dex. Moreover, the Dex IE was significantlyhigherthanBSA.

TheentrapmentofBSAandDexwithinPHBVMPswasfurther confirmedbyFTIRanalysis(Fig.3BandC).TheIRpeaksofpureBSA at 1652 and 1531cm
1 are assigned to stretching vibration of amideI (mainly C¼O stretching vibrations), and amide II (the couplingofbendingvibrateofN–HandstretchingvibrateofC–N)

bands, respectively. The FTIR spectra of BSA-loaded PHBVMPs revealedthesecharacteristicbandsofboththepolymerandthe protein,indicatingthatthehydrophilicactiveagentwas success-fullyincorporated.Similarly,theFTIRanalysisalsoconfirmedthe successfulentrapmentofDexinsidethePHBVMPs(Fig.3C),since itscharacteristicpeaksarevisibleat1700and1660cm
1,which corresponds to the carbonyl stretching vibration of C-17 dihydroxiacetone side chain, and a peak at 1660cm
1 due to carbonyl stretching vibration of C3 A-ring associated with the1630cm
1O–Hstretchingbandofinterlamellarwater. 3.2.2.BSAreleasefromPHBVMPs

ThereleaseprofilefromBSA-loadedPHBVMPsproducedinthe absenceof ethanol indicated that 76%of the BSAwas released withinthefirst8hreachingareleaseofapproximately78%ofthe incorporated BSA that was sustained up to day 21 (Fig. 4A). Similarly,theBSAloaded-MPsproducedinthepresenceofethanol releasedabout78%of theincorporatedBSAwithinthefirst8h. After21daysofincubation,approximately80%ofBSAwasreleased fromBSA-loadedPHBVMPs.TheseresultsfittheHiguchimodel. The BSA release was confirmed byFTIR analysis (Fig. 4B). The spectraofBSA-loadedPHBVMPsproducedbothintheabsenceand presenceofethanol,revealedareductionintheintensityofthe BSAcharacteristicbands(Fig.4B)upto21days.Atthistimepoint, the presence these were still verified corresponding to the remainingproteinentrappedintheMPs.

TheMPsincorporatingBSA,monitoredbySEM,showedminor morphological and topographical alterations after the in vitro releasestudies(Fig.5C).Moreover,MPsstructureandintegritywas maintained.

3.2.3.DexreleasefromPHBVMPs

ThereleaseprofileofDexfromDex-loadedPHBVMPsproduced intheabsenceofethanolindicatedthat55%ofDexwasreleased withinthefirst8handamaximumofapproximately69%ofthe incorporatedDexatreleaseduptoday21.Inopposition,89%ofthe incorporatedDexintheMPsproducedinthepresenceofethanol wasreleasedwithinthefirst8handnoDexwaspresentintheMPs after5days.ThereleaseofDexfromtheparticlesobtainedinthe absence of ethanol followed the Higuchi model, whereas the

Fig.3. PHBVMPscharacterizationafterincorporationofBSAandDex(A)IEofBSAandDexwithinthePHBVMPs.(B)FTIRspectraofBSAbeforeandafterincorporationwithin PHBVMPs.(C)FTIRspectraofDexbeforeandafterincorporationofDexwithinPHBVMPs.ThecharacteristicbandsofBSAandDexarehighlightedwiththegrayregion.

releasekineticsofDexfromtheMPsproducedinthepresenceof ethanol fitted the Korsmeyer–Peppas model. Therefore, in this case,sincenissmallerthan0.43,aclassicalFickiandiffusionis predicted.

TheDexrelease fromthePHBVMPs wasconfirmedby FTIR analysis(Fig.5B).WhiletheFTIRspectraofDex-loadedPHBVMPs, producedintheabsenceofethanol,revealedareduction inthe

intensityoftheDexcharacteristicbandsduetothereleasebuta still present at day21,those bands wereabsentin thespectra corresponding to the Dex-loaded PHBV MPs produced in the presenceofethanolaftertherelease(Fig.5B).

As observedfor theMPs incorporating BSA, theDex-loaded PHBVMPsmorphologyand shapewerenotaffectedbytheDex release(datanotshown).

Fig.4.BSAreleasefromloaded-PHBVMPs.(A)InvitroreleaseprofileofBSAfromBSA-loadedPHBVMPsproducedunderdifferentconditions.(B)FTIRspectraofBSA-loaded PHBVMPsbeforeandafter21daysofinvitrorelease.ThecharacteristicbandsofBSAarehighlightedwiththegrayregion.(C)SEMmicrographsofBSA-loadedPHBVMPs obtainedinthe(i)absenceand(ii)presenceofethanol,after21daysofinvitroreleasestudies.

Fig.5.Dexreleasefromloaded-PHBVMPs.(A)InvitroreleaseprofileofDexfromDex-loadedPHBVMPsproducedunderdifferentconditions.(B)FTIRspectraofDex-loaded PHBVMPsbeforeandafter21daysofinvitrorelease.ThecharacteristicbandsofDexarehighlightedwiththegrayregion.

Fig.6. BSA-loadedPHBVMPsembeddedintheGGhydrogelsvisualizedby(i)SEMand(ii)fluorescencemicroscopyafterFITClabelling.In(i)theuppercornermicrograph representsacross-sectionofthebiphasicstructure.

3.3.InjectableGG/PHBVMPs 3.3.1.Injectablesystemproperties

The embeddingof the PHBV MPs produced under different conditions in the 0.75% and 1.25% (w/v) GG solutions was confirmed.Aftergelation,theMPswerehomogeneouslydispersed withinthehydrogel, maintainingtheirspherical shape (Fig.6). Moreover,PHBVMPswerecompletelyembeddedwithintheGG matrices,asitispossibletoobserveinFig.6;ii.

The rheologicalanalysisoftheGGhydrogelwithorwithout PHBVMPs,showedthat,independentlyoftheconcentrationofGG, thesol_–geltransitiontemperatureofthesystemswasaround39_C

(Fig.7)andgelswereformedwithin20s.Thistemperaturewas barely affected by the presence of the PHBV MPs within the hydrogel, however, as expected the viscosity of the 0.75% GG (Fig. 7A) was lower than the 1.25% GG hydrogels (Fig. 7B). Moreover,itwasobservedthatindependentlyof theprocessing conditions,thepresenceofthemicroparticulatesystemwithinthe hydrogelincreaseditsviscosity,nevertheless,wasnotsignificantly. 3.3.2.BSAreleasefromtheinjectableGG/PHBVMPssystem

TheimpactoftheincorporationoftheMPsinthehydrogelover BSA release profile, showed that the presence of the matrix surroundingtheMPs(Fig.8A)almosteliminatedthereleaseburst observedforthesingleMPssuspensions(Fig.4A).Thiseffectwas more notorious for the 1.25% (w/v) of GG; 35% of the BSA incorporatedintheMPsproducedintheabsenceofethanolwas released within the first 8h. In contrast, about 40% of the incorporatedBSA was releasedfrom the loaded-MPs produced inthepresenceofethanolwithinthefirst8h.Boththosevalues werethemaximumreleasedafter21days.Incontrast,therelease

profileofBSAfromtheMPsembeddedinthe0.75%GGhydrogel indicated that 47% of proteinwas releasedwithin the _first 8h reaching a release of approximately 60% and 63% of the incorporatedBSAatday21,respectivelyfortheBSAloaded-MPs produced in the absence and in the presence of ethanol. Interestingly, independently of the concentration of the GG hydrogel,therelease kineticssuggestedthattherelease ofBSA fromthetestedformulationsfollowedazero-orderrelease.

Fig.7. Rheologicalbehaviorof(A)0.75%and(B)1.25%GGhydrogels,containingornotPHBVMPsproducedunderdifferentconditions.

Fig.8. (A)BSAreleasefromloaded-PHBVMPsproducedunderdifferentconditions andembeddedin0.75%and1.25%GGhydrogels.(B)SEMmicrographsoftheGG hydrogelcontainingtheBSAloaded-MPsafter21daysofinvitroreleasestudies.The sequentialimagesrepresent(i)anoverviewofthedistributionoftheparticles withinthehydrogel,(ii)acloseuponthesurface,and(iii)across-section.

After21daysofimmersioninPBSthesystemremainedstable. The PHBV MPs were still embedded in the system (Fig. 8C) althoughthe hydrogelrevealed somesignsof degradation.The surfaceerosionwasapparentresultingintheformationofsome cavities.Noparticlesmigrationwasobservedafterthe21daysof releasestudies.

3.3.3.DexreleasefromtheinjectableGG/PHBVMPssystem InaccordancetowhatwasobservedfortheBSA-loadedPHBV MPsembeddedintheGGhydrogels,theDexreleaseprofile(Fig.9) confirmedthattheGGhydrogelsurroundingtheDex-loadedMPs eliminatedtheinitialburstobservedforthesingleMPs suspen-sions(Fig.5A).Again,ahighereffectobservedforthe1.25%GG systemsbutinthiscase,thereleaseprofilewasnotdependenton theMPsprocessing conditionsasnodifferences wereobserved betweenthesystemsproducedinthepresenceandintheabsence ofethanol.Bothsystemsattainedareleaseofaround26%ofthe incorporatedDexintheendofthe21days.Regardingtherelease profileofDexfrom0.75%(w/v)GGhydrogels,dependencewiththe processing conditions was observed. The MPs produced in the absenceofethanolreleasedaround20%oftheincorporatedDex within the first 8h reaching approximately 43% of at day 21. Similarly,theDexloaded-MPsproducedinthepresenceofethanol releasedabout28%oftheincorporatedDexwithinthefirst8h. However,after21daysof incubationapproximately52%ofDex wasreleasedfromthis system.The releasekineticsof allthese formulationsfollowedaKorsmeyer–Peppasmodel,suggestinga classicalFickianactiveagentdiffusionsincen<0.43.

Moreover, in similarity to what was observed during BSA releasefromGGhydrogels,noparticlemigrationfromthehydrogel wasnoticed.

4. Discussion

Aimingatdevelopingaminimally-invasiveDDSwedesigneda biphasicinjectableDDSthatcanprovidealocalizeddeliveryof MPsavoidingtheiruncontrolledmigration throughthe human body,alongwithaprolongedreleaseofanactiveagent.Forthis purpose,BSA-andDex-loadedPHBVMPs,processedbyadouble emulsification–solventevaporation methodwithmodifications, were incorporatedinto injectable GG hydrogels with different concentrations.

The solvent type, the concentration of the emulsifier, the activeagent/polymerratioandthestirringrateare parameters knowntoaffectthephysicochemicalpropertiesoftheMPs(Bazzo

et al., 2008).In thisregard, a mixture of different solvents,as organic phase, is herein proposed aiming totailor the release profile ofhydrophilic and hydrophobicagents bychangingthe propertiesoftheparticles.Whilechloroformwasselectedasthe solventforthePHBV,ethanolwasaimedtoimprovethemutual solubility oftheorganicandtheaqueousphases, basedonthe type of molecular interactions between the organic solvents (dipole–dipole)andbetweenethanolandwater(hydrogenbond)

(Poletto et al., 2008). The obtained PHBV MPs had sizes that

ranged from 41.9 to 78.5

m

m, depending on the presence of ethanol,aswellasonthetypeofmoleculeincorporated.Thesize oftheMPsinthepresenceofethanolintheorganicphasewas higherthaninitsabsence.Theseresultsarenotinaccordanceto whatwaspreviouslyreportedbyPolettoetal.(2008)thatshowed adecreaseinthesizeofPHBVnanoparticleswhenethanolwas usedasasurfaceagent.Thedistinctresultsareprobablydueto the differences between the methods used to produce the particles;whiletheyfollowedanemulsification–diffusion meth-od,thedoubleemulsification–solventevaporationmethodherein proposed involveslesshigh-speedhomogenization, whichwas previouslyprovedtohaveatremendouseffectontheparticlesize (Mukherjeeetal.,2008).Furthermore,theconcentrationofPHBV usedPolettoetal.(2008)wasmuchsmallerthantheoneweused. Higher polymer concentrations in the organic phase lead to increased viscosity of the solution and consequently lower dispersion thus larger MPs are produced (Kassab et al.,1997; Yangetal.,2001).

In addition toaneffect overparticlesize, thesolvents used duringprocessingalsoimpactthemorphologyandtopographyof theparticles. It was previouslyreportedthat whentheorganic phase has increased miscibility with water, the interaction between emulsifier and the aqueous phase increases, thus particles with rougher surfaces are obtained (Sahoo et al., 2002).TheroughnessoftheMPssurfacehasbeenalsoassociated to the high crystallinity and fast precipitation of PHBV, after solvent evaporation from the internal phase of the emulsion

(Bidoneet al., 2009).The produced PHBVMPs showed a

well-definedsphericalshapeindependentoftheprocessingconditions. However,thetopographyofthesurfaceoftheMPswasinfluenced by the presence of ethanol in the process. This finding is in accordance with what was stated by others that showed the ethanol improvesthemiscibilityofbothphasesand thereforea better interactionwiththe emulsifier (Sahooet al., 2002).The presenceofethanolalsoinfluencedtheinnercoreporosityofthe particlescontrarilytothemicro-sizeporosityfoundonthePHBV MPs surfaces observed for all the preparation conditions. The presenceofethanolontheorganicphaseledtolargercorepores, whichcouldbeattributedtothehydrophobicityofPHBV.During the first emulsification, water droplets (first aqueous phase) interact withtheethanol presentin theorganicphase,but the sameisnotobservedwhentheorganicphaseiscomposedonlyby chloroform.Inthissense,thepresenceofethanolintheorganic phaseleadstotheformationoflargerdropletsinsidethedispersed phaseandconsequentlytolargerporesafterremovalofthewater dropletsbyfreeze-drying,inaccordancetowhatwaspreviously reportedintheliterature(Zhuetal.,2007).

TheincorporationofBSAand DexwithinthePHBVMPsalso affectedtheirproperties.Anincreaseinthemeansizewasobserved for the loaded-PHBV MPs, proportional to the MW of the incorporatedactiveagent.TheBSA-loadedMPs,incorporatingthe proteinwithaMWof66kDa,exhibitedlargermeansizethan Dex-loadedMPsthathaveentrappedamoleculewithaMW of392.46Da. Theincorporationofactiveagentsishighlydependentonthe chemical structureandMWof themolecules,aswellasonthe hydrophobicityofthepolymers(Parketal.,2005).Inthisstudy, BSA IE revealed to be independent of the organic phase

Fig.9.Dexreleasefromloaded-PHBVMPsproducedunderdifferentconditionsand embeddedin0.75%and1.25%GGhydrogels.

composition,whichwasalsoobservedfortheIEofDex.However, the active agent IE was affected by its MW, since Dex had a significantlyhigherIEthanBSA.Theseresultsareinaccordance withLionzo etal. (Lionzoet al.,2007)thatproposeda general singleemulsification–solventevaporationmethodforthe incor-porationofDexwithinPHBVMPs.

The in vitro release profile of a molecule entrapped in polymericMPs/matrixis controlledbyavarietyoffactors,such asthesolubilityof themolecules inthesurrounding media,its MWandmobilitywithintheswollenpolymericnetwork,aswell astheinteractionsofthepolymerwiththeactiveagent(Embleton andTighe,1993;Sahetal.,1994;Giunchedietal.,1994;Igartua etal.,1997;Yangetal.,2001).Moreover,otherfeaturessuchasthe morphology, size and size distribution of the particles also influencethereleasekinetics.Theserepresenttheonesthatcan be controlled by varying the processing conditions. Although PHAshavebeendescribedashighcrystallinepolymers,thereason foritsslowdegradation,excessivereleaseratesfromPHAsMPs havebeenpreviouslyandconsistentlyreported,andassumedto be mainly dependent on active agent dissolution rather than matrix degradation (Bidone et al., 2009). This is in agreement withourresultsthatconfirmedstructurallyintactandspherical shapeMPsafterthereleasestudies.WealsofoundthattheDex amountreleasedfromtheDex-loadedPHBVMPsproducedinthe absenceofethanolwasminorthanfortheMPsproducedinthe presenceofethanol.Thisfindingwasexpectedsincetheaddition ofethanoltotheorganicphaseledtotheformationofparticles with larger inner pores that facilitated the water penetration. Regarding the BSA, the release studies revealed no significant differenceswhenethanolwasaddedtotheorganicphase,anda releaseofaround78%oftheincorporatedproteinwasobserved after21daysof immersion.Thus,it wasalsoobservedthatthe MWoftheincorporatedactiveagentsdidnotonlyaffecttheIEbut alsotheamountofactiveagentreleased.DuetoitslowerMW, Dex had higher ability to diffuse across the polymeric matrix, corroborating the fact that the amount of small molecules releasediscommonlyhigherthanproteins.

In this sense, MPs composed by PHBV with tuned surface topography, size distribution and core porosity were obtained usingsimple variationsofthecompositionoftheorganicphase duringprocessing.TheinvitroreleaseprofileofDexbutnotofBSA, wasalso affectedallowingthecontrolover theamountof Dex releasealongthetime.

Nonetheless,abiphasicsystemswithauniformdistributionof PHBVMPsacrosstheGGmatrix,andtheabilitytobeinjectedinto theorganismandwithgelation occurringapproximatelyatthe bodytemperature, aspreviouslyreported(Oliveiraet al.,2010), wasdeveloped.

Thecombinationofactiveagent-loadedMPswithinsitugelling systemwastriedbyafewgroupswithvariousparticle-embedded hydrogels(Lagarceetal.,2005;Defailetal.,2006).Inthesestudies, itwasnoticeablethatduringtheinvitroreleasestudiestheamount ofactive agentrelease was decreasing, having a higherimpact duringthe_firsthours,whosebursteffectwasalmosteliminated. The incorporation of the MPs into the GG hydrogels strongly influenced both the profile and the amount of BSA and Dex released.Asexpected,hydrogelsofhigherGGconcentrationleadto a slower and more sustained release, independently of the processingconditionsofthemicroparticulatesystemduetothe higherdensity of thehydrogel which reducesthe activeagent diffusion. Moreover, the initial burst effect and the previously observed differences with the processing conditions, more notoriouslyforthe1.25%GGhydrogels,wereattenuated.Active agentsentrappedinsideMPshavetodiffusethroughtwobarriers (particulate system and hydrogel) to be released into the surroundingmedia,whichcanreducethebursteffectandprolong

thedelivery(Chenetal.,2004;Sivakumaranet al.,2011).Thus, alongwiththetime,theamountofreleasedactiveagentwillbe mainlydependentonthedegradationofthehydrogel. After21 days of incubation, the GGhydrogels presented some signs of degradation after immersedin PBS for 3 weeks, which can be explainedbyhydrolyticreactions(Coutinhoetal.,2010).However, astrongintegrationofthemicroparticulatesystemwithintheGG matrixwasobserved,evenafterthereleasestudies,whichshows the superior ability of these systems to maintain the particles withinthedefectsite.

In summary, this work confirmed the capacity of PHBV microparticulatesystemstodeliverbothhydrophilicand hydro-phobic active agents, and determined how the organic phase composition can be defined to tailor their incorporation and respective release profile. Moreover, the embedding of these systems in injectable hydrogels confirmed that the proposed system provides a localized delivery avoiding uncontrolled migration,alongwithasustainedreleaseoftheactiveagents. 5. Conclusion

Themaingoalofthisworkwastoproposeabiphasicinjectable DDS based on the combination of microparticulate systems andaninjectablehydrogel forthelocalizeddeliveryand long-term retention of MPs carrying hydrophilic and hydrophobic modelactiveagentsrelevantforRMpurposes,thusguaranteeing a prolonged release. In this sense, injectable GG hydrogels entrapping Dex- or BSA-loaded PHBV MPs were developed. Suchplatformwascharacterizedbyaprolongedrelease,which profilecanbetailoredaccordingtotheMPsandhydrogelfeatures. The proposed system also offers the advantage of a high spatialcontrol over the defectfilling by a fast gelation at the injectionsite.Thus,thecombinationofaninjectableGGhydrogel with Dex- or BSA-loaded PHBV MPs represents a promising strategytodelivermultipleactiveagentswithdistinctkinetics, which is likely required to drive tissue development to regeneration.

Acknowledgments

The authors would like to acknowledge the Project RL1 – ABMR – NORTE-01-0124-FEDER-000016 co-financed by North PortugalRegionalOperationalProgramme(ON.2–ONovoNorte), under the National Strategic Reference Framework (NSRF), throughtheEuropeanRegionalDevelopmentFund(ERDF).This workwaspartiallysupportedbyEuropeanResearchCouncilgrant agreement ERC-2012-ADG 20120216-321266 for project Com-plexiTE.

References

Abraham,G.,Gallardo,A.,SanRomán,J.,Fernández-Mayoralas,A.,Zurita,M., Vaquero,J.,2003.PolymericmatricesbasedongraftcopolymersofPCLonto acrylicbackbonesforreleasingantitumoraldrugs.J.Biomed.Mater.Res.A64, 638–647.

Amass,W.,Amass,A.,Tighe,B.,1998.Areviewofbiodegradablepolymers:uses, currentdevelopmentsinthesynthesisandcharacterizationofbiodegradable polyesters,blendsofbiodegradablepolymersandrecentadvancesin biodegradationstudies.Polym.Int.47,89–144.

Balmayor,E.,Tuzlakoglu,K.,Azevedo,H.,Reis,R.,2009.Preparationand characterizationofstarch-poly-e-caprolactonemicroparticlesincorporating bioactiveagentsfordrugdeliveryandtissueengineeringapplications.Acta Biomater.5,1035–1045.

Bazzo,G.,Lemos-Senna,E.,Gonçalves,M.,Pires,A.,2008.Effectofpreparation conditionsonmorphology,drugcontentandreleaseprofilesofpoly

(hydroxybutyrate)microparticlescontainingpiroxicam.J.Braz.Chem.Soc.19, 914–921.

Bidone,J.,Meloa,A.P.P.,Bazzo,G.C.,Carmignan,F.,Soldi,M.S.,Pires,A.T.N., Lemos-Senna,E.,2009.Preparationandcharacterizationofibuprofen-loaded microspheresconsistingofpoly(3-hydroxybutyrate)andmethoxypoly

(ethyleneglycol)-b-poly(D,L-lactide)blendsorpoly(3-hydroxybutyrate) andgelatincompositesforcontrolleddrugrelease.Mater.Sci.Eng.C29, 588–593.

Biondi,M.,Ungaro,F.,Quaglia,F.,Netti,P.A.,2008.Controlleddrugdeliveryintissue engineering.Adv.DrugDeliveryRev.60,229–242.

Calandrelli,L.,DeRosa,G.,Errico,M.E.,LaRotonda,M.I.,Laurienzo,P.,

Malinconico,M.,Oliva,A.,Quaglia,F.,2002.NovelgraftPLLA-basedcopolymers: potentialoftheirapplicationtoparticletechnology.J.Biomed.Mater.Res.62, 244–253.

Champa,A.,Bhat,A.,2010.Mesenchymalstemcellfunctiononhybridorganic/ inorganicmicroparticlesinvitro.J.TissueEng.Regen.Med.4,340–348. Chan,B.P.,So,K.F.,2005.Photochemicalcrosslinkingimprovesthe

physicochemical propertiesofcollagenscaffolds.J.Biomed.Mater.Res.A75, 689–701.

Chen,B.H.,Lee,D.J.,2001.Slowreleaseofdrugthroughdeformedcoatingfilm:

effectsofmorphologyanddrugdiffusivityinthecoatingfilm.J.Pharm.Sci.90, 1478–1496.

Chen,P.C.,Kohane,D.S.,Park,Y.J.,Bartlett,R.H.,Langer,R.,Yang,V.C.,2004. Injectablemicroparticle-gelsystemforprolongedandlocalizedlidocaine release.II.Invivoanestheticeffects.J.Biomed.Mater.Res.A70,459–466. Chen,W.,Tong,Y.W.,2012.PHBVmicrospheresasneuraltissueengineeringscaffold

supportneuronalcellgrowthandaxon–dendritepolarization.ActaBiomater.8, 540–548.

Ciçek,H.,Tuncel,A.,Tuncel,M.,Pişkin,E.,1995.Degradationanddrugrelease

characteristicsofmonosizepolyethylcyanoacrylatemicrospheres.J.Biomater. Sci.Polym.Ed.6,845–856.

Correia,C.R.,Reis,R.L.,Mano,J.F.,2013.Multilayeredhierarchicalcapsules providing celladhesionsites.Biomacromolecules14,743–751.

Costa,P.,SousaLobo,J.M.,2001.Modellingandcomparisonofdissolutionprofiles. Eur.J.Pharm.Sci.13,123–133.

Coutinho,D.F.,Sant,S.V.,Shin,H.,Oliveira,J.T.,Gomes,M.E.,Neves,N.M., Khademhosseini,A.,Reis,R.L.,2010.Modifiedgellangumhydrogelswith tunable physicalandmechanicalproperties.Biomaterials31,7494–7502.

daSilva,L.P.,Cerqueira,M.T.,Sousa,R.A.,Reis,R.L.,Correlo,V.M.,Marques,A.P.,2014. Engineeringcell-adhesivegellangumspongy-likehydrogelsforregenerative medicinepurposes.ActaBiomater.10,4787–4797.

Defail,A.J.,Edington,H.D.,Matthews,S.,Lee,W.C.,Marra,K.G.,2006.Controlled releaseofbioactivedoxorubicinfrommicrospheresembeddedwithingelatin scaffolds.J.Biomed.Mater.Res.A79,954–962.

Dhule,S.S.,Penfornis,P.,Frazier,T.,Walker,R.,Feldman,J.,Tan,G.,He,J.,Alb,A., John,V.,Pochampally,R.,2012.Curcumin-loadedg-cyclodextrinliposomal nanoparticlesasdeliveryvehiclesforosteosarcoma.Nanomedicine8, 440– 451.

Donbrow,M.,Samuelov,Y.,1980.Zeroorderdrugdeliveryfromdouble-layered porousfilms:releaserateprofilesfromethylcellulose,hydroxypropylcellulose and polyethyleneglycolmixtures.J.Pharm.Pharmacol.32,463–470. Embleton,J.K.,Tighe, B.J.,1993.Polymersforbiodegradablemedicaldevices.X.

Microencapsulationstudies:controlofpoly-hydroxybutyrate-hydroxyvalerate microcapsulesporosityviapolycaprolactoneblending.J.Microencapsul.10, 341–352.

Fitzgerald,P.,Hadgraft,J.,Wilson,C.G.,1987.Agammascintigraphicevaluation of theprecornealresidenceofliposomalformulationsintherabbit.J.Pharm.

Pharmacol.39,487–490.

Francis,L.,Decheng,M.,Knowles,J.,Keshavarz,T.,Boccaccini,A.R.,Roy,I.,2011. Controlleddeliveryofgentamicinusingpoly(3-hydroxybutyrate) microspheres.Int.J.Mol.Sci.12,4294–4314.

Fulzele,S.V.,Satturwar,P.M.,Kasliwal,R.H.,Dorle,A.K.,2004.Preparationand evaluationofmicrocapsulesusingpolymerizedrosinasanovelwallforming material.J.Microencapsul.21,83–89.

Gibaldi,M.,Feldman,S.,1967.Establishmentofsinkconditionsindissolution ratedeterminations.Theoreticalconsiderationsandapplicationto

nondisintegratingdosageforms.J.Pharm.Sci.56,1238–1242.

Giunchedi,P.,Conti,B.,Maggi,L.,Conte,U.,1994.Celluloseacetatebutyrateand polycaprolactoneforketoprofenspray-driedmicrospherepreparation.J. Microencapsul.11,381–393.

Griffith,L.G.,2000.Polymericbiomaterials.ActaMater.48,263–277.

Higuchi,T.,1961.Rateofreleaseofmedicamentsfromointmentbasescontaining drugsinsuspension.J.Pharm.Sci.50,874–875.

Higuchi,T.,1963.Mechanismofsustained-actionmedication.Theoreticalanalysis ofrateofreleaseofsoliddrugsdispersedinsolidmatrices.J.Pharm.Sci.52, 1145–1149.

Igartua,M.,Hernández,R.M.,Esquisabel,A.,Gascon,A.R.,Calvo,M.B.,Pedraz,J.L., 1997.Influenceofformulationvariablesonthein-vitroreleaseofalbuminfrom biodegradablemicroparticulatesystems.J.Microencapsul.14,349–356. Iwanaga,S.,Saito,N.,Sanae,H.,Nakamura,M.,2013.Facilefabricationofuniform

size-controlledmicroparticlesandpotentialityfortandemdrugdeliverysystem ofmicro/nanoparticles.ColloidsSurf.B109,301–306.

Jin,R.,2012.In-situformingbiomimetichydrogelsfortissueregeneration.In:Lin,C. (Ed.),Biomedicine,vol.2.InTech,Croatia,pp.35–58.

Kabilan,S.,Ayyasamy,M.,Jayavel,S.,Paramasamy,G.,2012.Pseudomonassp.asa sourceofmediumchainlengthpolyhydroxyalkanoatesforcontrolleddrug

delivery:perspective.Int.J.Micr.Res.2012,1–10.

Kassab,A.C.,Xu,K.,Denkbaş,E.B.,Dou,Y.,Zhao,S.,Pişkin,E.,1997.Rifampicin

carryingpolyhydroxybutyratemicrospheresasapotentialchemoembolization agent.J.Biomater.Sci.Polym.Ed.8,947–961.

Korsmeyer,R.W.,Gurnya,R.,Doelkera,E.,Buria,P.,Peppas,N.A.,1983.Mechanisms ofsolutereleasefromporoushydrophilicpolymers.Int.J.Pharm.15,25–35.

Korsmeyer,R.W.,Peppas,N.A.,1981.Macromolecularandmodelingaspectsof swelling-controlledsystems.In:Roseman,T.,Mansdorf,S.(Eds.),Controlled ReleaseDeliverySystems.MarcelDekker,NewYork,pp.77–90.

Kost,J.,Langer,R.,2001.Responsivepolymericdeliverysystems.Adv.Drug Delivery Rev.46,125–148.

Kozielski,K.L.,Tzeng,S.Y.,Green,J.J.,2013.Bioengineerednanoparticlesfor siRNA delivery.WileyInterdiscip.Rev.Nanomed.Nanobiotechnol.5,449–468. Lagarce,F.,Faisant,N.,Desfontis,J.C.,Marescaux,L.,Gautier,F.,Richard,J.,Menei,P.,

Benoit,J.P.,2005.Baclofen-loadedmicrospheresingelsuspensionsfor intrathecaldrugdelivery:invitroandinvivoevaluation.Eur.J.Pharm. Biopharm.61,171–180.

Lemperle,G.,Morhenn,V.B.,Pestonjamasp,V.,Gallo,R.L.,2004.Migrationstudies andhistologyofinjectablemicrospheresofdifferentsizesinmice.Plast. Reconstr.Surg.113,1380–1390.

Liang,C.Z.,Li,H.,Tao,Y.Q.,Peng,L.H.,Gao,J.Q.,Wu,J.J.,Li,F.C.,Hua,J.M.,Chen, Q.X.,2013.Dualreleaseofdexamethasoneandtransforminggrowthfactorb3from

polymericmicrospheresforthestemcellmatrixaccumulationinaratdisc degenerationmodel.ActaBiomater.9,9423–9433.

Lionzo,M.I.,Ré,M.I.,Guterres,S.S.,Pohlmann,A.R.,2007.Microparticlesprepared withpoly(hydroxybutyrate-co-hydroxyvalerate)and

poly(epsilon-caprolactone)blendstocontrolthereleaseofadrugmodel.J.Microencapsul. 24,175–186.

Liu,J.,Xiao,Y.,Allen,C.,2004.Polymer-drugcompatibility:aguidetothe developmentofdeliverysystemsfortheanticanceragent,ellipticine.J.Pharm. Sci.93,132–143.

Mi,F.L.,Lin,Y.M.,Wu,Y.B.,Shyu,S.S.,Tsai,Y.H.,2002.Chitin/PLGAblend microspheresasabiodegradabledrug-deliverysystem:phase-separation, degradationandreleasebehavior.Biomaterials23,3257–3267.

Monteiro,N.,Martins,A.,Ribeiro,D.,Faria,S.,Fonseca,N.A.,Moreira,J.N.,Reis,R.L., Neves,N.M.,2013.Ontheuseofdexamethasone-loadedliposomestoinduce theosteogenicdifferentiationofhumanmesenchymalstemcells.J.TissueEng. Regen.Med.doi:http://dx.doi.org/10.1002/term.1817(inpress).

Mukherjee,B.,Santra,K.,Pattnaik,G.,Ghosh,S.,2008.Preparation,characterization andin-vitroevaluationofsustainedreleaseprotein-loadednanoparticlesbased onbiodegradablepolymers.Int.J.Nanomed.3,487–496.

Ogawa,Y.,Yamamoto,M.,Takada,S.,Okada,H.,Shimamoto,T.,1988.

Controlled-releaseofleuprolideacetatefrompolylacticacidorcopoly(lactic/glycolic)acid microcapsules:influenceofmolecularweightandcopolymerratioofpolymer. Chem. Pharm.Bull.36,1502–1507.

Oliveira,J.M.,Sousa,R.A.,Malafaya,P.B.,Silva,S.S.,Kotobuki,N.,Hirose,M., Ohgushi,H.,Mano,J.F.,Reis,R.L.,2011.Invivostudyofdendronlikenanoparticles for stemcellstune-up:fromnanototissues.Nanomedicine7,914–924. Oliveira,J.T.,

Santos,T.C.,Martins,L.,Picciochi,R.,Marques,A.P.,Castro,A.G.,Neves, N.M.,Mano,J.F.,Reis,R.L.,2010.Gellanguminjectablehydrogelsforcartilage tissueengineeringapplications:invitrostudiesandpreliminaryinvivo evaluation.TissueEng.PartA16,343–353.

Park,J.,Ye,M.,Park,K.,2005.Biodegradablepolymersformicroencapsulationof drugs.Molecules10,146–161.

PintoReis,C.,Neufeld,R.J.,Ribeiro,A.J.,Veiga,F.,2006.NanoencapsulationI. Methodsforpreparationofdrug-loadedpolymericnanoparticles. Nanomedicine2,8–21.

Poletto,F.S.,Fiel,L.A.,Donida,B.,Ré,M.I.,Guterres,S.S.,Pohlmann,A.R.,2008. Controllingthesizeofpoly(hydroxybutyrate-co-hydroxyvalerate) nanoparticles preparedbyemulsification–diffusiontechniqueusingethanolas asurfaceagent. ColloidsSurf.A324,105–112.

Rodríguez-Contreras,A.,Canal,C.,Calafell-Monfort,M.,Ginebra,M.P.,Julio-Moran,G., Marqués-Calvo,M.S.,2013.Methodsforthepreparationof doxycycline-loaded phbmicro-andnano-spheres.Eur.Polym.J.49,3501–3511. Sah,H.K.,Toddywala,R.,Chien,Y.W.,1994.Theinfluenceofbiodegradable

microcapsuleformulationsonthecontrolledreleaseofaprotein.J.Controlled Release30,201–211.

Sahoo,S.K.,Panyam,J.,Prabha,S.,Labhasetwar,V.,2002.Residualpolyvinylalcohol associatedwithpoly(D,L-lactide-co-glycolide)nanoparticlesaffectstheir physicalpropertiesandcellularuptake.J.ControlledRelease82,105–114. Shishatskaya,E.I.,Voinova,O.N.,Goreva,A.V.,Mogilnaya,O.A.,Volova,T.G.,2008.

Biocompatibilityofpolyhydroxybutyratemicrospheres:invitroandinvivo evaluation.J.Mater.Sci.Mater.Med.19,2493–2502.

Sivakumaran,D.,Maitland,D.,Hoare,T.,2011.Injectablemicrogel–hydrogel compositesforprolongedsmall-moleculedrugdelivery.Biomacromolecules 12,4112–4120.

Soppimath,K.S.,Aminabhavi,T.M.,Kulkarni,A.R.,Rudzinski,W.E.,2001. Biodegradablepolymericnanoparticlesasdrugdeliverydevices.J.Controlled Release70,1–20.

Tunón,A.,Gråsjö,J.,Alderborn,G.,2003.Effectofintragranularporosityon compressionbehaviourofanddrugreleasefromreservoirpellets.Eur.J.Pharm. Sci.19,333–344.

Wagner,J.G.,1969.Interpretationofpercentdissolved-timeplotsderivedfromin vitrotestingofconventionaltabletsandcapsules.J.Pharm.Sci.58,1253–1257. Wen,Y.,Gallego,M.R.,Nielsen,L.F.,Jorgensen,L.,Møller,E.H.,Nielsen,H.M.,2013.

Designandcharacterizationofcore–shellmPEG–PLGAcomposite microparticlesfordevelopmentofcell-scaffoldconstructs.Eur.J.Pharm. Biopharm.85,87–98.

Xu,L.,Xu,X.,Chen,H.,Li,X.,2013.Ocularbiocompatibilityandtolerancestudyof biodegradablepolymericmicellesintherabbiteye.ColloidsSurf.B112,30–34. Yang,Y.Y.,Chung,T.S.,Ng,N.P.,2001.Morphology,drugdistribution,andinvitro

fabricatedbydouble-emulsionsolventextraction/evaporationmethod. Biomaterials22,231–241.

Yeo,Y.,Baek,N.,Park,K.,2001.Microencapsulationmethodsfordeliveryof protein drugs.Biotechnol.BioprocessEng.6,213–230.

Zhang,Y.,Chu,C.C.,2002.Invitroreleasebehaviorofinsulinfrombiodegradable hybridhydrogelnetworksofpolysaccharideandsyntheticbiodegradable polyester.J.Biomater.Appl.16,305–325.

Zhu,X.H.,Wang,C.H.,Tong,Y.W.,2007.Growingtissue-likeconstructswithHep3B/ HepG2livercellsonPHBVmicrospheresofdifferentsizes.J.Biomed.Mater.Res. BAppl.Biomater.82,7–16.

Zonari,A.,Cerqueira,M.T.,Novikoff,S.,Goes,A.M.,Marques,A.P.,Correlo,V.M.,Reis, R.L.,2014.Poly(hydroxybutyrate-co-hydroxyvalerate)bilayerskintissue engineeringconstructswithimprovedepidermalrearrangement.Macromol. Biosci.14,977–990.