Acta Biomaterialia

ContentslistsavailableatScienceDirect

Acta Biomaterialia

journalhomepage:www.elsevier.com/locate/actbio

Full length article

Evaluation of the anti-oxidative and ROS scavenging properties of biomaterials coated with epigallocatechin gallate for tissue

engineering

Sangmin Lee

^a,b

, Jinkyu Lee

^a

, Hayeon Byun

^a,b

, Se-jeong Kim

^a,b

, Jinmyoung Joo

^c

, Hee Ho Park

^d

, Heungsoo shin

^a,b,e,∗

aDepartment of Bioengineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul, 04763, Republic of Korea

bEducation and Research Group for Biopharmaceutical Innovation Leader, Hanyang University, Republic of Korea

cDepartment of Biomedical Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan, 44919, Republic of Korea

dDepartment of Biotechnology and Bioengineering, Kangwon National University, Chuncheon, Gangwon-do, 24341, Republic of Korea

eInstitute of Nano Science and Technology, Hanyang University, Republic of Korea

a rt i c l e i nf o

Article history:

Received 2 November 2020 Revised 2 February 2021 Accepted 2 February 2021 Available online 6 February 2021 Keywords:

Tissue engineering Oxidative stress Reactive oxygen species Antioxidant

Polyphenol coating

a b s t r a c t

Intissueengineering,excessivelygenerated reactiveoxygenspecies(ROS)duringbiomaterialimplanta- tionorcelltransplantationisaoneofmajorcausesofdiminishingtherapeuticeffects.Inthisstudy,we preparedbiomaterialsurfacescoatedwithantioxidantepigallocatechingallate(EGCG)andmetalions,and evaluatedtheiranti-oxidativeand ROSscavengingproperties.We revealedthatEGCG-coatingonpoly- caprolactone(PCL)filmsurfaceincreasedhydrophilicityandanti-oxidativepropertiesasafunctionofto- talphenolcontent(TPC)potentiallyduetotheincreaseinphenolic-OHandπ^{-electronsfromstructural} maintenanceanddirectlyremovedthehydrogenperoxide(H₂O₂)byresonance-stabilization.Furthermore, EGCG-coatedPCLfilmincreasedattachment,spreadingarea,andviabilityofhumanadipose-derivedstem cells(hADSCs)againstH₂O₂ treatmentwhilestimulatedthecellularsignalingtoreduceapoptoticgene andenhanceanti-oxidativeenzymeexpression.Further,weappliedEGCGcoatingonthesurfaceofpoly- L-lacticacid(PLLA)fibers.SpheroidsincorporatingEGCG-coatedPLLAfiberswereabletomaintaintheir shapeand showed improvedviability and anti-oxidativeactivities inresponsetoH₂O₂-induced oxida- tivestressthancontrolspheroids.Therefore,metal-phenolicnetwork(MPN)coatingofEGCGisasuitable methodtoimparttheanti-oxidativepropertiestobiomaterialsbyevaluatingthestructuralpropertiesand biologicaleffects.

StatementofSignificance

This manuscriptdescribesan antioxidant coating for biomaterialsto control reactive oxygen species (ROS). Antioxidant epigallocatechin gallate (EGCG) was coated on the PCL film sur- face via metal-phenolic network (MPN). We revealed that the phenolicfunctional groupsof EGCGare structurallymaintained asconfirmedby quantitative reducing power,radical scav- engingassays.EGCGcoatingnotonlyremovedROSdirectly,butalsoisinvolvedincellsignal- ingtoenhancetheanti-apoptoticgene,anti-oxidativeenzymeexpression.Furthermore,human adipose-derived stem cells (hADSCs) spheroid produced by self-assembly with EGCG-coated poly-L-lacticaicd(PLLA)fibersshowedanti-oxidativepropertiesfrominsideofspheroids.Thus, ourevaluationoftheanti-oxidativepropertiesofEGCGcoatingcanbeappliedtovarioustissue engineeringapplications.

© 2021ActaMaterialiaInc.PublishedbyElsevierLtd.Allrightsreserved.

∗Corresponding author.

E-mail address: hshin@hanyang.ac.kr (H. shin).

1. Introduction

Reactiveoxygenspecies(ROS),includingsuperoxide(•O₂⁻),hy- drogenperoxide(H₂O₂),andhydroxylradicals(•OH),aregenerally produced during implantation of biomaterials or cells for tissue

https://doi.org/10.1016/j.actbio.2021.02.005

1742-7061/© 2021 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

engineeringtherapy.ExcessiveROScauseapoptosisandactivation of inflammatory signalingcascades, leading to incomplete recov- ery of damaged tissue [1,2]. For example, the high level of ROS surrounding dentalimplantsresultsinuncontrolledinflammation with bone loss, and elevation of ROS during autologous or allo- genichematopoieticstemcelltransplantation(HSCT)canresultin serioustoxicproblems[3–5].Ithasbeenreportedthatpro-oxidant methemoglobin (Fe³⁺) released from lysed blood cellsand H₂O₂ derivedfromthelysisofepithelialorimmunecellsduringsurgical incisionofimplantationprocessareassociatedwithROS-mediated failure oftissue engineeringapproaches [6,7]. In addition, serum proteins adsorbed to the surfaces of biomaterials are known to bind to Fcdomains of transmembranereceptors of inflammatory cells, stimulatingthe synthesisofROS bythesecells[8,9].There- fore,itisimportanttocontrolROSforsuccessfulregenerativeout- comesintissueengineering.

Various strategies based on the use of antioxidants and anti- oxidative enzymeshavebeeninvestigated tocontrol ROS.Forex- ample, delivery of the solubleantioxidant N-acetylcysteine (NAC) with mesenchymal stem cell (MSC) aggregates into ischemic tis- sue successfullyremoved 80% ofH₂O₂ with enhancement ofcell adhesion and survival of the engrafted cells [10]. Furthermore, delivery of the soluble antioxidant edaravone with human um- bilical cord mesenchymal stem cells (hUMSCs) in an acute liver failure model increased the viability of transplanted hUMSCs by 1.5-fold and reduced the area of hepatic necrosis by approxi- mately60%[11].However,co-injectedsolubleantioxidantsoranti- oxidative enzymes were rapidly cleared (8–90% within 24 hr), andtherefore,ineffectiveforcontinuouscontrolofROSgenerated by damaged tissue [12,13]. As an alternative, biomaterials have been mixed or surface-coated with antioxidant or anti-oxidative enzymes to achieve sustained anti-oxidative activity. For exam- ple, polycaprolactone (PCL)/polyethylene glycol (PEG) electrospun nanofibers mixed with antioxidant chrysin significantly reduced chrysin releasewith1,000-foldgreateroxygenradical decomposi- tioncapacitythancontrolnanofibers,whilehyaluronicacidhydro- gelsconjugatedwithanti-oxidativedopamineshowed40%and65%

improvedhydroxylradicalandDPPHscavengingactivityovernon- conjugated hydrogels, respectively, with increased SOD and GPX enzymeactivity[14,15].However,mixingofROS-controllingagents withbiomaterialsmayalterthecharacteristicsofbiomaterialssuch as their degradation rate, swelling ratio, andmechanical proper- ties, while thesurface conjugatingstrategy mayrequirecomplex chemical processes andcould decreasethe biological activities of antioxidantsoranti-oxidativeenzymes[14–16].

Polydopamine and polyphenol coating of biomaterials have beenactivelystudied.Polydopamineformsamaterial-independent coating onthebiomaterial surfaceata basicpHbyself-oxidation of catechol groups indopamine [17]. Similarly, polyphenols such as tannic acid (TA), catechin (CA), and epigallocatechin gallate (EGCG) can readily form a metal-phenolic network (MPN) with cationic metal ions due to hydrogen bonding,

π

^-

π

^{stacking, and}

cation-

π

interactions by incubationofthe sampleina bufferso- lution containing metal ions such as Na⁺, Fe³⁺ [18]. In particu- lar, an MPN can form on the surfaces of various types of mate- rials, facilitatingsimple and effectivecoating andconferring that material with the biological properties of polyphenols including their anti-inflammatory, anti-cancer,andanti-oxidativeproperties [19]. Forexample, goldnanoparticles coatedwithpolyphenols in the presence of HAuCl₄•3H₂O showed 46%, 51%, and 55% ABTS radical scavenging activity when coated with EGCG, resveratrol (RSV), and fisetin (FS), respectively [20].Furthermore, tissue cul- ture polystyrene(TCP) coatedwithTAmodified withNa⁺ coated silicananoparticlesreducedtheintracellularROSlevelsignificantly morethanuncoatedTCP[21].Inaddition,MPNcoatingshavebeen reported to have anti-cancer activity(coating of RSV withAu³⁺)

andto enhanceendothelialization (coatingof TAwith Sr²⁺) [22–

24]. Despitethe focus on usingpolyphenols assurfacemodifiers ofbiomaterials,detailedinformationontheeffectsofcoatingcon- centrationandstructuralfeatureof polyphenolsonanti-oxidative activityandcellularsignalinghavenotyetbeenreported.

Previously, our group reported the EGCG coating on Ti sur- faceinthepresence ofMg²⁺,andinvestigatedtheir effectonos- teogenicdifferentiationofhumanadipose-derivedstemcells(hAD- SCs) and suppression of osteoclast functions [25]. Compared to thispreviouswork,thecurrentstudyisfocused ontheevaluation of the anti-oxidative properties of EGCG-coated biomaterial sur- facesandtheireffectsoncells,basedonthestructuralandchem- icalpropertiesofEGCGthrough2D and3D cultureconditionsfor hADSCs. Giventhat,we prepared PCLfilms andpoly-L-lacticacid (PLLA)fiberscoatedwithEGCGasmodelsubstrates,andthen,an- alyzed the effect of coating in the presence of cations (Na⁺ or Mg²⁺)onthetotalphenolcontent(TPC)depositedonthesurfaces ofthesubstrates,surfaceproperties,andanti-oxidativeproperties.

Second, we culturedhADSCs onEGCG-coated surfacesandinves- tigated theeffect ofthe EGCGcoating on thesurvival ofhADSCs under a ROS environment induced by H₂O₂ treatment. In addi- tion,PLLAfiberscoatedwithEGCGwerecombinedwithhADSCsto prepare3DspheroidsandweexaminedwhethertheEGCG-coated fibers were able to protect stem cells within the spheroidsfrom oxidativestress.

2. Materialsandmethods

Polycaprolactone (PCL) (M_w=80,000) and poly-L-lactic acid (PLLA, 1479) were obtained from Sigma-Aldrich (St. Louis, MO, USA) and Samyang (Jeollabuk-do, Korea), respectively. 2,2,2- trifluoroethanol(TFE)andethylenediamine(EDM)werepurchased fromSigma-Aldrich (St.Louis,MO, USA).Dichloromethane(DCM) and isopropyl alcohol (IPA) were purchased from Junsei Chemi- cal Co., Ltd (Chuo-ku, Tokyo, Japan) and EDM Millipore (Darm- stadt, Germany), respectively. BIS-TRIS wasobtained fromSigma- Aldrich (St. Louis, MO, USA), and EGCG was purchased from TCI America. Magnesiumchloride hexahydrate (MgCl₂^.6H₂O) and sodium chloride (NaCl) were obtained fromJunsei Chemical Co., Ltd. Distilled water (D.W.) was acquired from an Elix Advan- tageSystem(Millipore,MA,USA).Phosphatebufferedsaline(PBS), trypsin/EDTA,fetalbovineserum(FBS),andpenicillin-streptomycin (PS)werepurchasedfromWisent(St.Bruno,QC,Canada). Mesen- Pro RS medium and L-glutamine were obtained from Gibco BRL (Rockville,MD,USA).Humanadipose-derivedstemcells(hADSCs) were purchased from Invitrogen (Carlsbad, CA, USA). 3% hydro- genperoxide(H₂O₂)waspurchasedfromSigma-Aldrich(St.Louis, MO,USA).LIVE/DEAD assay kitand AlexaFluor^TM 488Phalloidin were purchased from Invitrogen (Carlsbad, CA, USA). Mounting medium containing4,6-diamidino-2-phenylindolewas purchased fromVectashield® (Burlingame,CA,USA).TheRNeasyMiniKitwas obtainedfromQiagen(Valencia,CA,USA).MaximeRTPremixwas purchasedfromIntron(Seoul,Korea),andSYBRPremixExTaqwas acquiredfromTAKARA(Otsu,Shiga,Japan).

2.1. PreparationandcharacterizationofEGCG-coatedPCLfilms

TopreparePCL films,PCL granules(4 g) were dissolved in50 mlofdichloromethane(DCM)and2,2,2-trifluoroethanol(TFE)(8:2, v/v)for24hr,andthissolutionwasthen pouredinto a100 mm glassdishandthoroughlydriedina60°Cdryovenfor6hr.Circles werepunchedfromthedriedPCLfilm(diameter,1.91cm²).1mM EGCG solution wasprepared in BIS-TRIS buffer (100 mM of BIS- TRISinD.W.atpH7).Weaddedeither300mMofNaClorMgCl2

tothebuffersolutionforMPNformation.PCLfilmwasplacedona 24-wellplateandincubatedwith1mlofEGCGsolutionfor24hr

at37°C.Thetotal phenolcontent (TPC) onthePCL filmswasde- terminedusingthe Folin-Ciocalteumethod.Briefly,films were in- cubatedwith300μLFolin-Ciocalteusolutionfor10min,andthen 700μLofsodiumcarbonatesolution(Na₂CO₃,2%w/v)wasadded.

Aftera2-hrincubationatRT,absorbancewasmeasuredat780nm.

The TPConthe PCLfilm wascalculatedusingan EGCGstandard.

The same concentration (300 mM)of buffer solutions containing different monovalent (NaCl, KCl) and divalent (MgCl₂, CaCl₂, and SrCl₂)metalionswereusedfortheEGCGcoatingonPCLfilmsur- face withsamemethoddescribedabove.The relativeTPConPCL filmwascalculatedbyusingthesameFolin-Ciocalteumethodand normalization with the values from the use of NaCl and MgCl₂, respectively. The EGCG-coated PCL films were incubated in 1 ml of D.W. for 1, 2, 4, and 7 days to measure the remaining TPC on the surface. The relative TPC of Surface functional groupson the PCL filmwere analyzed by attenuated totalreflection-Fourier transform infrared spectroscopy (ATR-FTIR, Nicolet 6700, Thermo Scientific, Waltham, MA, USA) and the surface atomic composi- tion ofPCL film wasanalyzed by X-ray photoelectron spectrom- etry (XPS) (ThetaProbe Base System, Thermo Fisher Scientific) at the Hanyang Center for Research Facilities (Seoul) using analysis software (Thermo VG Scientific, MA, USA).Surface morphological changes wereobservedusingscanningelectronmicroscopy(SEM) (Hitachi S-4800, Hitachi, Ltd, Tokyo, Japan). Water contactangle wasmeasured by capturing images(Phoenix300, Surface Electro OpticsCo.,Suwon,Korea)ofa1μl dropletofD.W.onthesurface ofthePCLfilmsandanalyzingthemwithImage-Proplussoftware (MediaCybernetics,SilverSpring,MD,USA).

Anti-oxidative propertiesof EGCG-coated PCL films were ana- lyzed using a ferric reducing antioxidant power assay and ABTS radicalscavengingassay.Fortheferricreducingantioxidantpower assay,1mg/mlof1,10-phenanthrolineand1mMofFeCl₃(Sigma- Aldrich, St. Louis, MO, USA) were dissolved in D.W. to obtain an Fe(III) solution.Prewetted EGCG-coated PCLfilms were mixed with 1 ml of Fe(III) solution, and the absorbance of the solu- tion was measured at 510 nm and used to quantify the con- version ratio of Fe(III) to Fe(II) (%) calculated using an ascorbic acidstandard.ABTSradicalswere generatedinasolutioncontain- ing 7.0 mM 2,2-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt and 2.4 mM potassium persulfate (K₂S₂O₈) (Sigma-Aldrich, St. Louis, MO,USA) in D.W. for 24hr. The ABTS solutionwasdilutedwithPBSuntil anabsorbancevalueof0.7at 732 nm wasreachedandthenmixedwith50μl ofascorbic acid standard orEGCG-coated PCLfilmprewettedwith50μl D.W. Af- ter10minofreaction,thepercentageABTSradicalsscavengedwas calculatedbymeasuringtheabsorbanceofthesolutionat732nm.

2.2. AdhesionofhADSCsandROS-scavengingactivityofEGCG-coated PCLfilms

hADSCs were maintained under standard culture conditions withMesenProRSmedium (with1%L-glutamine and1%PS) and cells from passage number 4 were used in all experiments. To determine the effect of EGCG on the viability of hADSCs, hAD- SCs were seeded ina tissue culture plateat 0.5× 10⁴ cells/cm² and cultured for24hr. Then, the media wasreplaced withfresh medium containingEGCG(0,12.5,25,50μM)andtheviabilityof hADSCswasmeasuredbyMTTassayaftera24-hrincubation.hAD- SCswereseededat0.5×10⁴ cells/cm² onthesurfaceofPCLfilm andEGCG-coatedPCLfilmandculturedfor24hr.hADSCscultured onPCLfilmwerefixedwith4%paraformaldehyde,andthenincu- batedwith1:200AlexaFluor^TM488Phalloidinfor1hrat37°Cfor F-actin staining.hADSCsstained forF-actin werepreservedusing mounting medium withDAPI. The numberof attachedcells was calculatedbycountingthenumberofnucleiperarea(/mm²),and spread area was calculated by measuring the existing area (x10³

μm²) ofactin filamentsper individual cellsusingImage-Pro plus software.

We firstinvestigated theeffect ofthe concentration ofexoge- nousH₂O₂ on theviability ofhADSCs; hADSCswere cultured on a tissue culture plate at 0.5 × 10⁴ cells/cm² for 24 hr and the mediumwasreplenishedbyfreshmediumcontaining0,100,200, or400 μMH₂O₂.Cellviabilitywasthen measured by MTTassay after4,8,and24hr.ToinvestigatetheeffectofEGCGonROSscav- enging,hADSCsweresimilarlyculturedonatissuecultureplateat 0.5×10⁴ cells/cm²,andthe culturemedia wasthenreplaced by onecontainingH₂O₂(100μM)and0,12.5,25,or50μMofEGCG.

After 24hr ofculture, cell viabilitywasmeasured by MTTassay.

Wethen cultured hADSCsonthesurfacesof PCLfilmandEGCG- coatedPCLfilm for24hrandexposedthehADSCsonthefilmto 200μM ofH₂O₂-containing culture mediafor12 hr. Thenumber ofcellsbefore andafter H₂O₂ treatmentwere counted andmor- phologicalchangeswereobservedafterF-actinstaining.

2.3. Anti-apoptoticandanti-oxidativeactivitiesofEGCG-coatedPCL filmagainstH₂O₂

Apoptotic DNA fragmentation ofhADSCs cultured on PCL and EGCG-coated PCL films in response to H₂O₂ treatment wasana- lyzedbyTUNELassay(ApopTag® fluoresceininsituapoptosisde- tectionkit,Sigma-Aldrich,St.Louis,MO,USA).hADSCsat0.5×10⁴ cells/cm² were cultured on the surfacesof PCLand EGCG-coated PCL films for 24 hr, and the films were then exposed to 200 μM of H₂O₂ for 12 hr. H₂O₂-exposed cells on PCL films were fixedwith1%paraformaldehydefor10minandthentreatedwith ethanol:acetic acid (2:1, v:v) for 5 min. After two washes with PBS, equilibration buffer was added for 10 sec. Then, the cells were incubated with terminal deoxynucleotide transferase (TDT) for 1 hr in a 37°C incubator. Immediately after TDT treatment, stop/wash buffer was addedfor 10 minat RT, followed by addi- tionofanti-digoxigeninconjugate(fluorescein)for30min.Stained cellswere preservedinmounting medium withDAPI. Percentage TUNEL-positive nucleiwascalculated by counting thenumber of fluorescein-stainednuclei(positive)andnormalizingthistotheto- talnumberofnuclei.

Toexamine theabilityof EGCG-coatedfilm toscavenge H₂O_2, PCLandEGCG-coatedPCLfilmswereincubatedinD.W.containing 100μMH₂O₂ for2hrandtheremainingH₂O₂wasmeasuredus- ing an H₂O₂ detectionkit (Fluorimetric HydrogenPeroxide Assay Kit,Sigma-Aldrich,St.Louis,MO,USA).Cellswerealsoculturedas describedabove andthen treated withH₂O₂ for12 hr, followed by incubation with 10 μM of 2,7-dichlorofluorescein diacetate (DCFH-DA)(Sigma-Aldrich,St.Louis,MO,USA)for30min.Intracel- lular ROSlevelwasassessed by fluorescencemicroscopy, andthe relativemeanfluorescenceintensity(MFI)valuewasquantifiedby calculating fluorescence intensity within areas of individual cells using Image-Proplus software andnormalized to that measured in the PCL group. We also analyzed the expression of apoptotic andoxidative genesby real-timereversetranscriptionpolymerase chainreaction (RT-PCR). TotalRNA ofhADSCscultured on EGCG- coatedPCLfilmwascollectedusingTRizolreagent(LifeTechnolo- gies,Carlsbad,CA,USA),andcDNAwassynthesizedfromcollected totalRNAusingtheMaxime RTPreMixKit(IntronBiotechnology, Gyeonggi-do,Korea).Then, thecDNAwasmixedwithSYBRGreen PCR Mastermix (TAKARA, Otsu, Shiga, Japan), and PCR was per- formed ina StepOnePlusthermocycler (LifeTechnology, Carlsbad, CA,USA). Thesequences of theprimers usedforRT-PCR wereas follows:GAPDH(Fw:5’-GTC AGTGGTGGACCTGAGCT-3’,Rv:5’- TGCTGTAGCCAAATTCGTTG-3’),BAX(Fw:5’-TTTGCTTCAGGG TTTCATCC-3’,Rv:5’-CAGTTGAAGTTGCCGTCAGA-3’),BCL2(Fw:

5’-GAGGATTGTGGCCTTCTTTG-3’,Rv:5’-ACAGTTCCACAAAGG CATCC-3’),BCL2L1(Fw:5’-CTGAATCGGAGATGGAGACC-3’,Rv:

5’-TGG GAT GTCAGG TCACTG AA-3’), Catalase (Fw: 5’-GCC TGG GACCCAATTATCTT-3’,Rv:5’-GAATCT CCGCACTTC TCCAG-3’), FOXO3(Fw:5’-GCAAGCACAGAGTTGGATGA-3’,Rv:5’-CAGGTC GTC CAT GAG GTT TT-3’), GPX-1 (Fw: 5’-CTC TTC GAG AAG TGC GAGGT-3’,Rv:5’-TCGATGTCAATGGTCTGGAA-3’).

2.4. PreparationandcharacterizationofEGCG-coatedPLLAfibers

TopreparePLLAfibers,anelectrospunnanofibersheetwaspre- paredbyejecting10mlof3%ofPLLAsolutiondissolvedinamix- tureofDCMandTFE(8:2,v/v)usingasyringepump(KDS200,KD Scientific, New Hope, PA, USA)through a 23-gauge needle under 16 kVatarateof2ml/h. Scatteredfibers werecollectedon alu- minum foilcovered arotatingmandrel (SPG,Incheon,Korea),and the sheetwasdried overnight. The nanofiber sheetwaschopped intosmallpiecesanddispersedinethylenediamine(EDM)solution (10%,v/v inIPA)foraminolysis.Aftervigorous shakingat90rpm for 30min at37°C, theaminolyzed PLLA fiberswere centrifuged at4,000rpmfor10min andwashed in70%ethanol. 50–100 μm sizedPLLAfiberswerecollectedusingasievegridandlyophilized after3D.W.washes.TocoatPLLAfibers(PF)withEGCG,1mg/ml ofpreparedPF(50–100μm) wasdispersedinEGCGsolution(100 mMofBIS-TRISwith300mMofMgCl₂ dissolvedinD.W.atpH7) andincubatedfor24hrat37°Conarotator.TheTPConthePLLA fiberswasmeasuredbytheFolin-Ciocalteumethodasdescribedin Section2.1.

The anti-oxidative activity of EGCG-coated PLLA fibers (E-PF) wasmeasuredby aferric reducingantioxidantpower assay,ABTS radical scavenging assay, and H₂O₂ scavenging assay. PF and E- PF were dispersedin D.W.(1 mg/ml)andserially diluted.500 μl of the PF sample, E-PF sample, andascorbic acid standard were reacted with Fe(III) solution for30 min andthe absorbance was measured at510 nm.Similarly,50μlofthePFsample,E-PF sam- ple, and ascorbic acidstandard were reacted withABTS solution and the absorbance was measured at 732 nm. To analyze H₂O₂ scavenging activity,PF andE-PF were dispersed in100 μM H₂O₂ solution (0 to 1 mg/ml), dissolved in D.W., and incubated for 2 hr at 37°C. The remaining concentration of H₂O₂ in the solution was measured using an H₂O₂ detection kit. To evaluate the bio- compatibility ofPF andE-PF andhADSCs, hADSCs were cultured ona24-wellplateat1×10⁴ cells/cm² for24hr.Then,themedia waschanged toonecontaining5μg ofPForE-PF.After24hr of culture, theviabilityofhADSCswasmeasured byMTTassay.The intracellularROSlevelofhADSCswasmeasuredbyincubatingthe cellswith10μMDCFH-DAfor30minafterthesameculturepro- cess asdescribedabove. Fluorescence intensityofindividual cells wascalculatedbyImage-Proplussoftwareandnormalizedtothat ofthecontrol(-)group.

2.5. hADSCspheroidformationwithfibersandanti-oxidativeeffectof EGCGcoatedfibersonH2O2-treatedhADSCs

To fabricate hADSC spheroids (cells only), 40,000 hADSCs in 100 μl of culture media were placed in a 0.2 mltube andcen- trifugedat1,200rpmfor3min.Cellsinthetubegraduallyformed a spheroid duringincubationfor24hr at37°C, andthespheroid was then transferred to an ultra-low attachment plate (Costar®

ultra-lowattachmentmultiple96-wellplate,CorningIncorporated, New York,USA).PF-incorporatingspheroidswerepreparedasde- scribedaboveexceptthat10μgofPF(1μg/μl)wasaddedthe100 ofculturemediumcontaining40,000cells.Cell-onlyspheroidsand PF-incorporatingspheroidswereculturedinmediacontaining400 μMH₂O₂for12hr.Changesinthemorphologyofspheroidswere monitored by phase contrast microscopy and the numberof de- formed spheroids wasquantified by image analysisand reported as a percentage. Using the same cell culture process described

above,thedistributionofliveanddeadcellsinspheroidswasde- termined by confocal microscopy (TE2000 andEclipse C1, Nikon, Tokyo,Japan)after30minofincubationinLIVE/DEADstainingso- lution(Invitrogen,Carlsbad,CA,USA).

Forhistologicalanalysis,H₂O₂-exposedspheroidswerefixedin 4% paraformaldehyde. Fixed individual spheroids were then im- mersed inoptimalcutting temperaturecompound (OCT) solution andfrozen in a deep freezer (-80°C)for more than 2hr to pro- duceanOCTblock.ThefrozenOCTblockcontainingspheroidswas cut into10 μm sectionsusing acryostat microtome andsections werecollectedonslideglass.Toobservetheinternal morphology of spheroids, H&E staining of cross-sectionedspheroids wasper- formed.First,cross-sectionedsampleswereincubatedtwiceinxy- lenefor10min.After ahydrationstep(rinsingwith100%to70%

ethanolfollowedbyrunningwaterfor10min),sampleswereincu- batedinhematoxylinfor2minandeosinfor8min.Followingde- hydration,sampleswerepreservedinmountingmediumforobser- vation. TheDNA integrity ofhADSCswithin spheroidswasdeter- minedbyTUNELassay.First,cross-sectionedsamplesofspheroids from the OCT block were incubated twice in xylene for 10 min.

Followinghydration,samplesweretreatedwithethanol:aceticacid (2:1, v:v) for 5 min. After two washes with PBS, samples were incubated in equilibration buffer for 10 sec. Then, samples were incubated with terminal deoxynucleotide transferase (TDT) for 1 hr at 37°C. Immediately after TDT treatment, stop/wash buffer was added for 10 min at RT, followed by anti-digoxigenin con- jugate (fluorescein) for 30 min. Stained samples were preserved in mounting medium with DAPI. TUNEL-positive nuclei (%) were quantified by counting the number of fluorescein-stained nuclei (positive) normalized to the total number of nuclei. In addition, spheroidsexposed to 400 μMH₂O₂ for12 hr were incubatedin 10μMDCFH-DAfor30min,andthenintracellularROSlevelswere quantifiedbyfluorescencemicroscopy.RelativeMFIwasquantified by calculating the fluorescence intensity of individual spheroids andnormalizingthis tothat ofcell-only spheroids. Forapoptotic andoxidativegeneexpressionanalyses,spheroidsexposedtoH₂O₂ wereimmersedinTRizolreagent, andwere thenbroken apartby pipetting several times to extract RNA. Subsequent steps for RT- PCRwerethesameasthosedescribedinSection2.3.

2.6. Statisticsandanalyses

Quantitative data are presented as means ± standard devia- tions. One-way analysis of variance (ANOVA)using Tukey’s hon- estlysignificantdifferencetestandtwo-wayANOVA(fortwovari- ables)usingtheBonferronipost-testwereperformedusingGraph- Pad Prism 5 software (La Jolla, CA, USA). A pvalue < 0.05was consideredtodenotestatisticalsignificance.

3. Resultsanddiscussion

3.1. CharacterizationofEGCG-coatedPCLfilms

The overall scheme of EGCG coating on a biomaterial surface through MPNformation andthe potential effects on 2D, 3D cul- tured cell underlying its anti-oxidative activity are illustrated in Fig. 1. According to the EGCG coating, the white color PCL sur- face turned slightly brown in Na-E and Mg-E, and small aggre- gates were observed in Mg-E (Supplementary Fig. S1). The TPC onthePCLfilm increasedsignificantlyinthepresence ofcations, with Mg²⁺ increasing TPC on the PCL firm to a greater extent thanNa⁺(17± 3μg/cm² vs8±1μg/cm²,respectively)(Fig.2a).

When various types of metal ions were used for EGCG coating, TPCfromdivalent wassignificantlygreaterthan thatfrommono- valentions whileno significant difference wasfound indifferent ionsof eachgroup (Supplementary Fig.S2). The TPConPCL film

Fig. 1. Schematic diagram of the anti-oxidative properties of EGCG coating to reduce intracellular ROS and anti-apoptosis on 2D, 3D cultured cells.

Fig. 2. Characterization of EGCG-coated PCL film. (a) The TPC on the PCL film surface using NaCl- or MgCl 2-containing buffer. Spectra of PCL film, Na-E, and Mg-E obtained by (b) ATR-FTIR analysis and (c) XPS analysis. (d) SEM images of the surfaces of PCL film, Na-E, and Mg-E. (e) Water contact angles of PCL film, Na-E, and Mg-E. The anti-oxidative activity of EGCG-coated PCL film as measured by (f) ferric reducing antioxidant power assay and (g) ABTS radical scavenging assay ( ^∗, # p < 0.05).

surface remained80% and60% in Na-E (6 ± 1 μg/cm²) andMg- E(10±1μg/cm²)for7days(SupplementaryFig.S3).ATR-FTIRre- sults demonstrated that a broad O-H stretch at3000-3700 cm⁻² and phenolicC=C stretchat 1630cm⁻¹ were presentaftercoat- ing withEGCG(Fig.2b),indicating successfulcoating withEGCG.

These resultsare consistent witha previous report [26].In addi- tion,specificpeaksofNa1sandMg1sweredetectedbyXPSanal- ysis on the surface of Na-E and Mg-E, respectively (Fig. 2c). In the XPS C1s high-resolution spectra,peak intensity ofC-Obond- ing (286.4eV)was elevated in Na-E andmore in Mg-E than PCL (SupplementaryFig.S4). SEM resultsshowedthat barePCLhada

smooth surfacewhile agglomerated particles were found on the Mg-Esurface(Fig.2d). FollowingEGCGcoating,bare PCLfilm(67

±2^o)became hydrophilicand completewettingwasobservedon thePCLsurfacecoatedwiththegreaterTPC(Fig.2e).Asshownin Fig.2fandg, theconversion ratioofFe(%/cm²) wassignificantly elevatedafterEGCGcoatingascomparedtobarePCL(2±0%/cm²), andMg-E (16±1%/cm²)hadasignificantly greaterFeconversion ratiothanNa-E(6±1%/cm²).Similarly,Mg-EinhibitedABTSradi- calssignificantlymorethanbarePCLandNa-E(Fig.2g).

The divalent cationMg²⁺ led to a greaterlevel ofMPN-based coatingthan themonovalentcation Na⁺.MPNisthesupramolec-

ular network structure, in which metal ions are coordinated to phenolic ligands with

π

^-cationinteractions playing acrucial role [19,27]. Inaprevious study,nocatechin (CA)coating wasformed on a PCL nanofiber surface in the absence of sodium ion (Na⁺) [28]. Similar to our findings, more pyrogallol (PG) coating was achieved on the titanium surface with magnesium ion (Mg²⁺) than Na⁺ [29]. In the XPS C1s high-resolution results, the grad- ual improvement in C-Obonding inNa-E andMg-E than in PCL can be estimated by the improvement of phenolic C-O contents due to the increase in TPC ofthe surface. The structure of MPN with non-covalent interactions between metal ions and pheno- lic moleculeshas beenindirectly estimatedby stoichiometry and XPSanalysis[19,30].Forexample,mono/bis/tris-complexwasesti- matedthroughtherelativeratioofFe³⁺andTA,andthepeakshift tohigherbindingenergythroughXPSanalysiswouldhavecaused electron transferfromTA toFe ions.Inourresults,the increased phenolic C-Ocontent indirectlyshows that Mg²⁺ can formmore MPNthanNa⁺.SEManalysisconsistently supportedmorecoating with Mg²⁺ than Na⁺. The formationof small aggregateswasre- portedduringpolyphenolcoatingwhencoatingexceededacertain level [29]. The surface hydrophilicity increased with increase in TPC,indicatingthatthe-OHgroupsofcatechinandgallolinEGCG after coating mayhavecontributedto the formationofhydrogen bonds withH₂Oon the surfaceandthe increased hydrophilicity.

Consistently, the Mg-E surface exhibited2-3-fold greater Fe con- version/ABTSscavengingactivitythanNa-Esurface,suggestingthat phenoliccontents onthesurfaceappearto beresponsibleforan- tioxidantactivitytoreduceFe(Fe³⁺→Fe²⁺)andtoscavengeradi- cals[31].Ithasbeenreportedthatthehydroxylgroupsinpheno- licringstructureofEGCGcanscavengeradicalsandchelatemetal ionduringself-oxidation,andtheelectrontransferinphenolic

π

^-

electrons isdependentontheamountofEGCGcoatedonthesur- faceofthesubstrate[32].

3.2. AdhesionofhADSCsandROSscavengingactivityofEGCG-coated PCLfilms

Toconfirm theviabilityofhADSCsafterEGCG treatment,cells wereobservedafterLIVE/DEADstainingandcellularmetabolicac- tivity wasquantified by MTTassay (Fig.3a andb).Cellmorphol- ogy and viability were not significantly affected by EGCG treat- ment. EGCG has been reported to be cytotoxic to mesenchymal cellsonlyatconcentrationhigherthan50μM[33].Consistentwith thispreviousstudy,EGCGatconcentrationslowerthat50μMwas not cytotoxic to hADSCs. The number of attached cellson films cultured for 24 hr wassignificantly greater on the EGCG-coated surfaces (Na-E,177 ±21/mm² andMg-E, 128± 7/mm²) thanon the PCL film (87 ± 20/mm²) (Fig. 3c). In addition, hADSCs be- came narrower and thinner (spreading area 5.2 ± 1.6 10³ μm² per cell) on the PCL film, while they had a more broad polygo- nal shape onNa-E andMg-E films (spreadingarea 9.0± 3.510³ μm² and7.3±1.510³ μm² per cell,respectively) (Fig.3dande).

According to previous reports, surfacecharacteristics such ashy- drophilicity androughnessarealteredbypolyphenolcoating, and cellattachment andspreadingareachangeaccordingly.Forexam- ple,theFe³⁺/TAcoatedsurfaceofPCLnanofibermeshbecamehy- drophilic and smallaggregates formed on the surface, similar to our findings [34].Furthermore,human umbilical vein endothelial cells (HUVECs) cultured on PCL nanofiber mesh showed greater adhesion and spreading than HUVECs cultured on uncoated PCL nanofibers[34].ThereducednumberofadherentcellsontheMg- EsurfacemaybeduetotheroughnessoftheMg-Esurface.Ithas been reported that the adhesion force of cellsdecreases as sur- faceroughnessincreasesaboveacertainlevel(adhesioninhibition) [35].

Viability of hADSCs after H₂O₂ treatment decreased signifi- cantlyastheconcentration ofH₂O₂ increasedandexposuretime increased(Fig.4a). Forexample,after24hr,theviabilityofhAD- SCsdecreasedfrom83±2%at200μMH₂O₂to11±0%at400μM H₂O₂ comparedtothe control.However, whencellsweretreated with both H₂O₂ and EGCG (0, 12.5, 25, 50 μM), the viability of hADSCsafter24hrimprovedsignificantly (by~ 10–20%) ascom- pared to that without EGCG (74 ± 5%) (Fig. 4b). Changes in the morphologyofhADSCsculturedonPCLfilmandEGCG-coatedPCL filmwereobservedafter12hrofexposureto200μMH₂O₂.Cells culturedonPCLfilmwereshrunkenandhadalowspreadarea(3

±210³ μm² per cell)while cellscultured onNa-EandMg-Eex- hibited anincreasedspread area relativeto cellsculturedon PCL film(12± 310³ μm² and11± 310³ μm² percell, respectively) (Fig.4candd).

EGCGtreatmentprotectedhADSCsagainstH₂O₂ mediated-ROS damage, consistentwith a previous studythat reportedthat 75%

of hMSCs treated with 200 μM of H₂O₂ for 24 hr were senes- centbased on

β

-galactosidase staining whereas pre-treatment of hMSCs with 50 μM EGCG for 6 hr reduced senescence to 50%

[36].Controlofoxidativestressgeneratedduringbiomaterialtrans- plantation or cell transplantation is important,and several stud- ies have confirmed that tissue regeneration improves when an- tioxidantsareused[37,38].Forexample,silkfibroinnanofibersin- corporating the antioxidant fenugreek (1:1) had67% DPPH radi- cal scavengingeffectafter 24hr andimprovedwound healingin a rat full thickness excision model as compared to control and silk fibroin only [39]. In addition, anti-oxidative polyorganophos- phazene (PATGP) microspheres fabricated using anilinetetramers and glycine ethyl ester co-substitution showed more than twice the DPPH, •OH scavenging activity and bone regeneration activ- ity in a rat calvarial defect model after 8 and 16 weeks com- pared to PLGA and PAGP microspheres [40]. We found that the spread area of hADSCs cultured on Na-E andMg-E surfaces was significantly greaterthan that onPCLfilm. H₂O₂ canformhighly reactive hydroxyl radicals either spontaneouslyor through afen- ton reactionthat causesoxidation,which directlyattacksthe cell walls(lipids)andDNA ofhADSCs[41]. Therefore,itcan beinter- pretedthat the reducedspreadarea ofdamaged hADSCsmay be due to the contraction of apoptotic cells, whereas cell adhesion and spread area was improved on EGCG-coated PCL film due to protective effect by ROS scavengingproperties of EGCG. Phenols canreactwiththefreeradicalspeciestodelocalizeunpairedelec- tronswithinthearomaticring,givingphenolradicalsthatundergo resonance-stabilization withinthemolecule andthus,formstable intermediates[42].H₂O₂ isanon-radicaloxidizingagentthat can decomposeinphysiologicalconditionstoformhighlyreactive hy- droxylradicals(•OH)[43,44].Thehydroxylradicalscausecytotoxic- ity,andtheresonance-stabilizationbythearomaticringofphenol can remove thishydroxyl radicals by aforementioned resonance- stabilization[42].Given that,the delocalization ofunpaired elec- tronswithinthephenolicstructureofMPNcoatedonthesurface mayhaveremovedcytotoxichydroxylradicalsandproducedstable H₂OandO₂,improvingcellviability.

3.3. Anti-apoptoticandanti-oxidativeactivitiesofEGCG-coatedPCL filmagainstH₂O₂

H₂O₂ treatment did not change the number of cellsattached tothesurfacesofNa-E andMg-E,whileitsignificantly decreased the numberof cells attachedto bare PCL film (Fig. 5b). In addi- tion,thepercentageofTUNEL-positivenuclei(%)wassignificantly greaterforcellsattachedtoPCL (75± 7%)thanNa-E (35± 12%) andMg-E(10 ± 10%)(Fig. 5aandc).H2O2 causesapoptoticDNA fragmentation, andEGCG has beenshown to reduce the damage causedbyH₂O₂.Forexample,treatmentofpancreaticalphaTC1-6

Fig. 3. EGCG cytotoxicity and attachment of hADSCs to PCL film. (a) LIVE/DEAD staining of hADSCs after 24 hr of EGCG treatment (0, 12.5, 25, 50 μM) and (b) MTT assay results. (c) The number of hADSCs attached to PCL, Na-E, and Mg-E after 1 day of culture. After F-actin staining, (d) cell spread area and (e) fluorescence were quantified ( ^∗,

# p < 0.05).

Fig. 4. Viability and morphological change of hADSCs after H 2O 2treatment. (a) Viability of hADSCs after treatment with 0, 10 0, 20 0, or 40 0 μM H 2O 2for 4, 8, and 24 hr. (b) Viability of hADSCs after 1 day of culture with 200 μM H 2O 2+ EGCG (0, 12.5, 25, 50 μM). After 12 hr of 200 μM H 2O 2treatment of hADSCs cultured for 1 day on the PCL film surface, changes in (c) cell spread area and (d) fluorescence were quantified ( ^∗, # p < 0.05).

Fig. 5. EGCG-coated PCL film protects against H 2O 2-induced apoptosis. (a) Apoptotic signals of hADSCs cultured on the surface of PCL film after H 2O 2treatment as measured by fluorescence microscopy after the TUNEL assay. (b) Change in relative cell number of hADSCs on PCL film after H 2O 2treatment and (c) percentage of TUNEL-positive nuclei. Relative expression of the apoptosis-related genes (d) BAX, (e) BCL2, and (f) BCL2L1 in hADSCs ( ^∗, # p < 0.05).

cells for4hr with 100 μMH₂O₂ resulted in 90%TUNEL-positive nuclei, whichwasreducedto70%and5%after1hpre-treatment withEGCG30μMand100μM,respectively[45].Ourresultssug- gestthatEGCGcoatedonthepolymersurfacewasalsoabletopro- tectagainstH₂O₂-inducedapoptosisinadose-dependentmanner.

TheexpressionoftheapoptoticgeneBAXinculturedhADSCswas significantly decreased in cellscultured on Na-E and Mg-E rela- tive to barePCL film,withthe greatestdecrease inexpression of thisgeneobservedforcellsculturedonMg-E(Fig.5d).Expression oftheanti-apoptotic genesBCL2 andBCL2L1 wassignificantly in- creased in Na-E andMg-E withthe highest expression observed forcellsgrownonMg-EcomparedtobarePCLfilm(Fig.5eandf).

EGCGhaspreviouslybeenreportedtohaveamodulatoryeffecton theexpressionofapoptosis-relatedgenes;mousevascularsmooth muscle cells cultured in a conditioned environment with H₂O₂ showedsignificantlyreducedexpressionofthepro-apoptoticgenes caspase-3,8,9andBAXandasignificantincreaseintheexpression of the anti-apoptotic gene BCL-2 after pre-treatment with EGCG [46].OurresultsconfirmedthattheEGCGcoatingonthesurfaceof PCL filmwasbiologicallyactive andinfluencedanti-oxidativesig- nalingprocesseswithinhADSCsculturedonEGCG-coatedfilm.

To investigatetheH₂O₂ scavenging activityofthe EGCGcoat- ing,PCLfilmandEGCG-coatedPCLfilmwereincubatedin100μM H2O2 solution for 2 hr. H2O2 level was not reduced by the un- coatedPCLfilmwhilelevels ofH₂O₂ were significantlydecreased insolutions containingNa-E andMg-Efilm to97± 1%and87± 2%ofcontrollevels,respectively(Fig.6a).TherelativeMFIpercell intheDCFH-DAassaywassignificantlylowerforcellsculturedon Na-E andMg-E filmat43±10% and24±5%oftheMIFpercell

observedforcellsculturedonPCLfilm(Fig.6b).Fluorescencemi- croscopyrevealedthathADSCsculturedonPCLfilmwithoutH₂O₂ showedlittlefluorescencesignal(Fig.6c),whiletherewasasignif- icantincreaseinfluorescencesignalafterH₂O₂treatment(Fig.6d).

Inaddition,fluorescenceintensitywasreducedinNa-E andMg-E comparedtoPCLfilm(Fig.6d).Theexpressionofcatalase,FOXO3, and GPX-1 was significantly increased in cells cultured on Na-E andhighestforcellsculturedonMg-Ecomparedtocellscultured on bare PCL film (Fig. 6e, f, and g). Polyphenols have been re- portedtoeliminateH₂O₂ andrelatedROS.Forexample,real-time H2O2 monitoring of a skin membrane-covered oxygen electrode (SCOE)revealedthat 1mM H₂O₂ wasrapidlyreducedaftertreat- ment with0.1mM polyphenolssuch ashydroquinone, quercetin, piceatannol,andresveratrol[47].AsdescribedinSection3.2,EGCG isexpectedtobe ableto removeH₂O₂ byresonance-stabilization mechanism,andourresultssuggestthatEGCGretaineditspheno- licstructureaftercoating.Furthermore,theEGCGcoatingappeared not only to remove ROS through direct anti-oxidative processes andresonance-stabilization,butalsotoregulateanti-oxidativeen- zymeexpressioninproportiontotheTPCdepositedonthefilm.

3.4. CharacterizationofEGCG-coatedPLLAfibers

Wenextinvestigatedtheanti-oxidativeprotectiveeffectsofan EGCGcoatingon3Dspheroidalculturesofstemcellsbypreparing EGCG-coatedPLLAfibers.AccordingtotheEGCGcoating,thewhite color of PF turned brown in E-PF (Supplementary Fig. S5). SEM photographs(Fig.7a)demonstratedthatthesmoothsurfacesofPF becamerough after coating with48± 2 μg/mgEGCG, indicating

Fig. 6. Anti-oxidative activity of EGCG-coated PCL film in response to H 2O 2treatment. (a) The amount of H 2O 2remaining in D.W. after a 2 hr incubation with PCL film. (b) Intracellular ROS level of hADSCs cultured on PCL, Na-E, and Mg-E as determined by DCFH-DA and fluorescence images (c) without H 2O 2treatment (d) with H 2O 2treatment.

Relative expression of anti-oxidative genes (e) catalase, (f) FOXO3, and (g) GPX-1 in hADSCs ( ^∗, # p < 0.05).

homogeneous deposition of EGCG(Fig. 7b).Ferric reducing activ- ity[conversion rate(%)]increasedupto 28± 5%asthe percent- ageoffibersincreased(Fig.7c).Similarly,E-PF showedABTSradi- cal scavengingactivitythat increasedinproportiontotheamount of fibers, as shown in Fig. 7d. In an aqueous environment con- taining H₂O₂, E-PF effectively reduced H₂O₂ and thus decreased the amountofH₂O₂ significantly (85±2%)asafunction offiber concentration (Fig.7e).EGCG-coatedfibershadnodetrimentalef- fect oncellviability(Fig.7f).IntracellularROSlevel(relativeMFI) increased rapidlyin the positive control (143 ± 17) treated with H2O2 compared to the negativecontrol (100 ± 1) without H2O2

treatment. In the PF-treated group,fluorescence intensityper in- dividual cell wasnot statisticallydifferent than that in the posi- tive control(140±8),butwassignificantlydecreasedintheE-PF treatedgroup(110±4)(Fig.7gandh).

TheaforementionedresultssuggestthatEGCGdirectlyremoves ROS through phenolic

π

^{-electron and} resonance-stabilization of catechin and gallol groups, and/or stimulates anti-apoptotic and anti-oxidative cell signaling pathways. Similarly, EGCG-coated fibers exhibitedanti-oxidativepropertiesasconfirmedbyFe con- version,ABTSradicalscavenging,andeliminationofH₂O₂,indicat- ing maintenanceofthebiologicalactivityofEGCGindependentof the surfacechemistry oftheunderlying materials. Inaddition, E- PF effectively reduced the intracellular ROS level of hADSCs cul- tured ontissueculture plates.Ithasbeenreported thatthelevel of intracellularROS isindicativeof cellulardamage. Forexample,

thegreatertheintracellularROSlevelofprimarycardiacstemcells andembryonicstemcells,themoreDNAdamagewasfound[48]. Theantioxidantcurcumin(0-4μM)concentration-dependentlyde- creasedtheDCFfluorescenceintensityofL-6myoblasts[49],while the antioxidant edaravone decreased the intracellular ROS level ofhuman umbilical cord mesenchymalstem cells[11], indicating thattheseantioxidants candecreasethelevelofintracellularROS.

However,directlydeliveredsolubleantioxidantswererapidlyelim- inated(8–90%removedwithin24hr)andshoweddecreasedanti- oxidativeactivityin vivo[12].Ourresultsindicatecoating thesur- faces of biomaterials withEGCG can effectively remove ROS and potentially protect tissues against oxidative damage. It is known thatsolubleEGCGexhibitsanti-oxidativepropertiesinproportion to the concentration, but its use hasbeen very limited in tissue engineeringapplicationsduetopotentialcytotoxicityatconcentra- tionsabove50μM[33,50].WhenEGCGwascoatedonbiomaterial, EGCGcanbestablycoatedonthesurfaceviaMPNformationand couldbemaintainedforalongtimewithoutcytotoxicity.

3.5. SpheroidformationwithEGCG-fibersandtheirabilitytoprotect hADSCswithinthespheroidsagainstH₂O₂

We next fabricatedPLLA fibersfor incorporationin stemcells spheroidsandinvestigatedtheir effectson3D culturedstemcells.

Similar to the aforementioned experiments, spheroids were ex- posedto400 μMH₂O₂ for12hr. Structural deformation wasob-

Fig. 7. Characterization of PLLA fibers (PF) and EGCG-coated PLLA fibers (E-PF). (a) SEM image of PF and E-PF surfaces. (b) The TPC on E-PF. The anti-oxidative activity of PF and E-PF was measured by (c) ferric reducing antioxidant power assay, and (d) ABTS radical scavenging assay. (e) The amount of H 2O 2remaining after incubation of PF and E-PF in 100 μM of H 2O 2in D.W. for 2 hr. (f) Viability of hADSCs cultured with PF or E-PF for one day. (g) Fluorescence images of intracellular ROS in hADSCs cultured with PF or E-PF in the presence of H 2O 2for 1 day and (h) relative MFI per individual cell based on the DCFH-DA assay ( ^∗p < 0.05).

Fig. 8. Structural deformation of hADSCs spheroid in response to H 2O 2exposure. (a) Phase contrast image of cell-only, PF, and E-PF spheroids and (b) percentage of deformed spheroids (%). (c) Survival of hADSCs in spheroids after LIVE/DEAD staining. (d) Internal morphology of spheroids as determined by H&E staining and (e) high resolution imaging of PF-incorporating spheroids ( ^∗, # p < 0.05).

Fig. 9. Anti-apoptotic activity of EGCG-coated PLLA fibers incorporated in hADSC spheroids in response to H 2O 2exposure. (a) Apoptotic signals of cell-only, PF, and E-PF spheroids and (b) percentage of TUNEL-positive nuclei (%). (c) Relative expression of (d) BAX and (f) BCL2L1 in cell-only, PF and E-PF spheroids ( ^∗p < 0.05).

Fig. 10. Anti-oxidative activity of EGCG-coated PLLA fibers incorporated in hADSCs spheroids against H 2O 2. (a) Intracellular ROS in spheroids were observed by fluorescence microscopy and (b) levels were quantified. Relative expression of anti-oxidative genes (e) catalase, (f) FOXO3, and (g) GPX-1 in cell-only, PF and E-PF spheroids ( ^∗p < 0.05).

served in the cell-only (45 ± 13%) and PF (85 ± 6%) spheroids, while there wasno deformation of spheroids incorporating E-PF (3 ± 5%) (Fig. 8a and b). There were more dead cells in cell- only and PFspheroids than inE-PF spheroids(Fig. 8c). Histolog- ical analysisofcross-sectionedspheroidsrevealedthat cellswere homogeneously distributed andtightly integrated throughoutthe spheroids(Fig.8d)whileinthePFgroup,deformationofspheroids at theperiphery wasobserved(Fig.8e).Ourresults indicatethat cell damage duetodiffusion ofH₂O₂ inthe media waslocalized at theperiphery of thePF spheroids. Deformation(%) waslarger in the PF group than the cell-only group because the diffusion

ofH₂O₂ wasmitigated bythe relativelyloosely formednetworks between cells and fibers. In fact, histological images revealed a thickercellularlayerattheperipheryofcell-onlyspheroidswhile theouter celllayerbecamethinnerinspheroidscontainingfibers (Fig.8d).TheprotectivefunctionofE-PFagainstH₂O₂wasevident.

Necrotic coreformation or external damage can lead to contrac- tionanderuptionofspheroidsandthusstructuraldeformationofa spheroidcanserveasanindicatoroftheintegrityofthespheroid.

Forexample,extensiveDNAdamagewasobservedinmulticellular tumorspheroids(MCTSs)culturedwithH₂O₂ (0.1–5mM)for1hr, resultinginstructuralchanges[51],whilelaserirradiation ofMSC

spheroidscauseddestructionfromtheoutsideofthespheroid[52]. Our results indicate that the presence of E-PF protected against ROSdamage.

There was a large number of TUNEL-positive nuclei in the PF group in the region of deformation (Fig. 9a), and the num- ber of TUNEL-positive nuclei was significantly decreased in E-PF spheroids (24 ± 2%) compared to cell-only spheroids (35 ± 4%) andPFspheroids(43±7%)(Fig.9b).Expressionleveloftheapop- totic gene BAX wassignificantly lower in E-PF spheroids than in cell-onlyandPFspheroids(Fig.9c),andconversely,theexpression oftheanti-apoptoticgeneBCL2L1wassignificantly higherinE-PF spheroidsthanincell-onlyandPFspheroids(Fig.9d).Intracellular ROSlevelsinspheroidsweredeterminedafterH2O2treatment.The relative MFI perspheroid was45 ±11for E-PF spheroids,which wassignificantlylowerthanthatinPFspheroids(90±7)andcell- only spheroids(100± 11)(Fig.10aandb). Catalase,FOXO-3,and GPX-1 expression wassignificantly higherin E-PF spheroidsthan cell-only and PFspheroids (Fig. 10c,d, ande).Consistent witha previous report that intraperitonealinjectionof EGCG(70mg/kg) improved expression ofthe anti-oxidativeenzyme GPX-1 in hep- atictissue[36],EGCG-coatedfibersappeartomodulatehADSCsig- naling in spheroids, resulting in improved anti-oxidative enzyme expression.Ourresearchgrouphasbeeninvestigatingstrategiesto deliverinstructivesignalstocellsinsidespheroidsbyincorporating biofunctionalized fibersintothespheroidsduringspheroidgener- ation. Thisapproachcouldefficientlydirectandmodulatecellular activitieswithinspheroids.Forexample,fibershavebeenfunction- alized with PDGF to improve the proliferation of hADSCs within spheroids [53] andto enhance osteogenic differentiationthrough adenosine immobilization andmineralization [54,55]. Our results suggestthatincorporatingEGCG-coatedfibersinspheroidsconfers the spheroidswithsustainedanti-oxidativepropertiesandmodu- latessignalingbycellsinsidethespheroid.

4. Conclusions

Inthisstudy,PCLsurfaceswerecoatedwithEGCGinthepres- ence of Na⁺ or Mg²⁺ cations via MPN formation, and the TPC wasgreater inthe presenceofMg²⁺ than Na⁺. EGCGcoating in- creased surface hydrophilicityand roughness proportional to the TPC.EGCGcoatingmaintaineditsanti-oxidativepropertiesascon- firmed by Fe conversion and ABTS radical scavengingactivity of hADSCsculturedonEGCG-coatedsurfacesinthepresenceofH₂O₂. EGCG coatingdirectly removedsurroundingH₂O₂ andmodulated signaling in hADSCs, leading to a decrease in expression of the apoptoticgeneBAXandan increaseintheexpressionoftheanti- apoptotic genes BCL2 and BCL2-L1 as well as the anti-oxidative enzymes catalase, FOXO3, and GPX-1. Furthermore, EGCG-coated PLLAfibersexhibitedanti-oxidativepropertiesasevidenced byFe conversion,ABTSradicalscavengingactivity,andremovalofH₂O₂. EGCG-coated PLLA fibers incorporated in stem cell spheroidsre- duced apoptotic signals in the presence of H₂O₂-induced oxida- tive stress andinduced anti-oxidativeenzyme expression relative to cell-only spheroidsand spheroidsincorporating fiberswithout an EGCGcoating.Inconclusion,EGCGcoating caneffectivelypro- tectagainstROS-inducedoxidativedamage,suggestingitspotential versatilityinvarioustissueengineeringapplications.

DeclarationofCompetingInterest

Theauthorsdeclarethattheyhavenoknowncompetingfinan- cialinterestsorpersonalrelationshipsthatcouldhaveappearedto influencetheworkreportedinthispaper.

Acknowledgements

Thisresearch wassupported by National ResearchFoundation of Korea (NRF) grants funded by the Korean government (MEST) (NRF-2019R1A2C2084965,NRF-2020R1A4A3078645).

Supplementarymaterials

Supplementary material associated with this article can be found,intheonlineversion,atdoi:10.1016/j.actbio.2021.02.005.

References

[1] M. Dizdaroglu, P. Jaruga, Mechanisms of free radical-induced damage to DNA, Free Radic. Res. 46 (2012) 382–419, doi: 10.3109/10715762.2011.653969 . [2] R. Félix, P. Valentão, P.B. Andrade, C. Félix, S.C. Novais, M.F.L. Lemos, Evaluat-

ing the in vitro potential of natural extracts to protect lipids from oxidative damage, Antioxidants 9 (2020) 1–29, doi: 10.3390/antiox9030231 .

[3] E. Mijiritsky, L. Ferroni, C. Gardin, O. Peleg, A . Gultekin, A . Saglanmak, L.G. De- logu, D. Mitrecic, A. Piattelli, M. Tatullo, B. Zavan, Presence of ROS in inflam- matory environment of peri-implantitis tissue: in vitro and in vivo human ev- idence, J. Clin. Med. 9 (2019) 38, doi: 10.3390/jcm9010038 .

[4] D. Pietropaoli, E. Ortu, M. Severino, I. Ciarrocchi, R. Gatto, A. Monaco, Glycation and oxidative stress in the failure of dental implants: a case series, BMC Res.

Notes. 6 (2013) 1–6, doi: 10.1186/1756- 0500- 6- 296 .

[5] A.M. Evens, J. Mehta, L.I. Gordon, Rust and corrosion in hematopoietic stem cell transplantation: the problem of iron and oxidative stress, Bone Marrow Transplant 34 (2004) 561–571, doi: 10.1038/sj.bmt.1704591 .

[6] G.G.A. Silkstone, M. Simons, B.S. Rajagopal, T. Shaik, B.J. Reeder, C.E. Cooper, O.

Thews, J.C. LaManna, D.K. Harrison, G.G.A. Silkstone, M. Simons, B.S. Rajagopal, T. Shaik, B.J. Reeder, C.E. Cooper, Novel Redox Active Tyrosine Mutations En- hance the Regeneration of Functional Oxyhemoglobin from Methemoglobin:

Implications for Design of Blood Substitutes., United States, 2018. 10.1007/978- 3-319-91287-5_70.

[7] W. Razzell, I.R. Evans, P. Martin, W. Wood, Calcium flashes orchestrate the wound inflammatory response through duox activation and hydrogen perox- ide release, Curr. Biol. 23 (2013) 424–429, doi: 10.1016/j.cub.2013.01.058 . [8] S. Franz, S. Rammelt, D. Scharnweber, J.C. Simon, Immune responses to im-

plants - a review of the implications for the design of immunomodulatory bio- materials, Biomaterials 32 (2011) 6692–6709, doi: 10.1016/j.biomaterials.2011.

05.078 .

[9] H. Blaser, C. Dostert, T.W. Mak, D. Brenner, TNF and ROS Crosstalk in Inflam- mation, Trends Cell Biol. 26 (2016) 249–261, doi: 10.1016/j.tcb.2015.12.002 . [10] C.C. Huang, W.Y. Pan, M.T. Tseng, K.J. Lin, Y.P. Yang, H.W. Tsai, S.M. Hwang,

Y. Chang, H.J. Wei, H.W. Sung, Enhancement of cell adhesion, retention, and survival of HUVEC/cbMSC aggregates that are transplanted in ischemic tissues by concurrent delivery of an antioxidant for therapeutic angiogenesis, Bioma- terials 74 (2016) 53–63, doi: 10.1016/j.biomaterials.2015.09.043 .

[11] W. Zeng, J. Xiao, G. Zheng, F. Xing, G.L. Tipoe, X. Wang, C. He, Z.Y. Chen, Y. Liu, Antioxidant treatment enhances human mesenchymal stem cell anti- stress ability and therapeutic efficacy in an acute liver failure model, Sci. Rep.

5 (2015) 1–17, doi: 10.1038/srep11100 .

[12] L. Zhang, Y. Ma, X. Pan, S. Chen, H. Zhuang, S. Wang, A composite hydrogel of chitosan/heparin/poly ( γ-glutamic acid) loaded with superoxide dismutase for wound healing, Carbohydr. Polym. 180 (2018) 168–174, doi: 10.1016/j.carbpol.

2017.10.036 .

[13] D.B. Zorov, M. Juhaszova, S.J. Sollott, Mitochondrial reactive oxygen species (ROS) and ROS-induced ROS release, Physiol. Rev. 94 (2014) 909–950, doi: 10.

1152/physrev.0 0 026.2013 .

[14] Y. Deldar, Y. Pilehvar-Soltanahmadi, M. Dadashpour, S. Montazer Saheb, M. Rahmati-Yamchi, N. Zarghami, An in vitro examination of the antioxidant, cytoprotective and anti-inflammatory properties of chrysin-loaded nanofibrous mats for potential wound healing applications, Artif. Cells, Nanomed. Biotech- nol. 46 (2018) 706–716, doi: 10.1080/21691401.2017.1337022 .

[15] S. Zhang, J. Hou, Q. Yuan, P. Xin, H. Cheng, Z. Gu, J. Wu, Arginine derivatives as- sist dopamine-hyaluronic acid hybrid hydrogels to have enhanced antioxidant activity for wound healing, Chem. Eng. J. 392 (2019) 123775, doi: 10.1016/j.cej.

2019.123775 .

[16] T.Y. Yen, Z.R. Stephen, G. Lin, Q. Mu, M. Jeon, S. Untoro, P. Welsh, M. Zhang, Catalase-functionalized iron oxide nanoparticles reverse hypoxia- induced chemotherapeutic resistance, Adv. Healthc. Mater. 8 (2019) 1–8, doi: 10.10 02/adhm.20190 0826 .

[17] S.K. Madhurakkat Perikamana, J. Lee, Y. Bin Lee, Y.M. Shin, E.J. Lee, A.G. Mikos, H. Shin, Materials from Mussel-inspired chemistry for cell and tissue engi- neering applications, Biomacromolecules 16 (2015) 2541–2555, doi: 10.1021/

acs.biomac.5b00852 .

[18] D.G. Barrett, T.S. Sileika, P.B. Messersmith, Molecular diversity in phenolic and polyphenolic precursors of tannin-inspired nanocoatings, Chem. Commun. 50 (2014) 7265–7268, doi: 10.1039/C4CC02961E .

[19] H. Ejima, J.J. Richardson, F. Caruso, Metal-phenolic networks as a versatile plat- form to engineer nanomaterials and biointerfaces, Nano Today 12 (2017) 136–

148, doi: 10.1016/j.nantod.2016.12.012 .