lable at ScienceDirect

Biomaterials 35 (2014) 2322e2335

Contents lists avai

Biomaterials

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

Star-branched amphiphilic PLA-b-PDMAEMA copolymers forco-delivery of miR-21 inhibitor and doxorubicin to treat glioma

Xiaomin Qian a, Lixia Long a, Zhendong Shi b, Chaoyong Liu a, Mingzhe Qiu b, Jing Sheng a,Peiyu Pu b, Xubo Yuan a,*, Yu Ren c,***, Chunsheng Kang b,**

a Tianjin Key Laboratory of Composite and Functional Materials, School of Materials Science & Engineering, Tianjin University, Tianjin 300072, ChinabDepartment of Neurosurgery, Tianjin Medical University General Hospital, Laboratory of Neuro-Oncology, Tianjin Neurological Institute, Key Laboratory ofNeurotrauma, Variation and Regeneration, Ministry of Education and Tianjin Municipal Government, Tianjin 300052, Chinac Tianjin Research Center of Basic Medical Science, Tianjin Medical University, Tianjin 300070, China

a r t i c l e i n f o

Article history:Received 13 September 2013Accepted 14 November 2013Available online 12 December 2013

Keywords:Star-branched copolymerPolylactic acidPolydimethylaminoethyl methacrylatemiR-21Co-deliverySynergistic effects

* Corresponding author. Tel.: þ86 22 87401832; fax** Corresponding author. Tel.: þ86 22 60817499; fax*** Corresponding author. Tel./fax: þ86 22 8333686

E-mail addresses: xbyuan@tju.edu.cn (X. Yuan)(Y. Ren), kang97061@yahoo.com (C. Kang).

0142-9612/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.biomaterials.2013.11.039

a b s t r a c t

The combined treatment of chemotherapeutant and microRNA (miR) has been proven to be a viablestrategy for enhancing chemosensitivity due to its synergistic effect for tumor therapy. However, theco-delivery of drugs and genes remains a major challenge as they lack efficient co-delivery carriers. Inthis study, three amphiphilic star-branched copolymers comprising polylactic acid (PLA) and poly-dimethylaminoethyl methacrylate (PDMAEMA) with AB3, (AB3)2,and (AB3)3 molecular architectureswere synthesized respectively by a combination of ring-opening polymerization, atom transfer radicalpolymerization, and click chemistry via an “arm-first” approach. The star copolymers possessed a lowcritical micelle concentration (CMC) and formed nano-sized micelles with positive surface charges inwater as well as exhibiting a much lower cytotoxicity than PEI 25 kDa. Nevertheless, their genetransfection efficiency and tumor inhibition ability showed a remarkable dependence on their mo-lecular architecture. The (AB3)3 architecture micelle copolymer exhibited the highest transfection ef-ficiency, about 2.5 times higher than PEI. In addition, after co-delivering DOX and miR-21 inhibitor(miR-21i) into LN229 glioma cells, the micelles could mediate escaping miR-21i from lysosomedegradation and the release of DOX to the nucleus, which significantly decreased the miR-21 expres-sion. Moreover, co-delivery of DOX and miR-21i surprisingly exhibited an anti-proliferative efficiencycompared with DOX or the miR-21i treatment alone. These results demonstrated that amphiphilic star-branched copolymers are highly promising for their combinatorial delivery of genes and hydrophobictherapeutants.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Accumulating evidences indicate ‘one target strategy’ remainssuboptimal in cancer chemotherapy [1]. Recent clinical trials usingcombined administration of multiple-target chemotherapeutantsprovide promising results in glioblastomas, lung cancers and breastcancers [2,3]. Moreover, RNA-based drug development, includingRNAi based strategy and miRNA-based strategy are still in its in-fancy. Many efforts have been devoted to siRNA-based combinationtherapy [4]. However, the intrinsic drawbacks of siRNA

: þ86 22 27404704.: þ86 22 27813550.6., renyu_1212@yahoo.com.cn

All rights reserved.

methodology, such as the off-target effects and elicitation of theinterferon response, greatly hamper its therapeutic use [5].

MiRNAs are a class of short, non-coding RNAs with post-transcriptional regulatory functions [6]. “One shot, multiple tar-gets” signature of miRNA allow it to be exempt from the problem ofsiRNA therapy, whichmightmore closely resemble the action of theso-called “dirty drugs” used in the clinic today, thereby, makemiRNA a better tool for gene therapy [7]. However, miRNA-baseddelivery still has another balk that makes it difficult to deliverytherapeutic miRNA and chemo-drugs into tumor cells while keepsurrounding normal cell away from inappropriate therapy [8]. It isexpected the chemotherapeutic drug and miRNA should besimultaneously delivered to the same tumor cell after in vivoadministration and, ideally, be distributed in the cells at their cor-responding points to maximize intracellular cooperation [9]. Inspite of the progress in liposomal [10] and silica-based [11] cationic

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2323

nanoparticles, more attention is focused on polymer supports dueto the easy synthesis of polymers, their strong stability, and theirbiodegradability [12].

The amphiphilic star copolymer, as a result of its unique struc-ture and distinct physical properties, serves as a good support forgenes and medication. Star polymers are three-dimensional glob-ular or branched polymers containingmultiple arms connected to acentral core [13]. As compared with the linear polymers, starpolymers have very high molecular weights but still possess asolubility and viscosity similar to that of linear or branched poly-mers of relatively lowmolecular weights [14]. The core cross-linkedamphiphilic star copolymer has a lower critical micelle concen-tration (CMC), which avoids the problem of being disassemble as aresult of dilution in linear amphiphilic copolymers. As the innercore of a star polymer is degradable, the star polymer micelles willrelease medicines through degradation when they reach the focusof the infection. Compared with dendrimers, star polymers enjoythe advantages of facile synthesis, flexible compositions, andtunable sizes (10e100 nm), which allow them to be able to carrymore “cargo” within one molecule [15]. Up to now, star polymersare widely applied in the delivery of chemotherapy and genes[16,17]. However, there is no report on the simultaneous support ofgenes and medicine, nor is there a report on the synergic effectscaused by the star polymer structure.

In this study, the amphiphilic hyper branched star macromole-cules of the AB3, (AB3)2, and (AB3)3 architecture by using PLA as ahydrophobic branch and PDMAEMA as a hydrophilic core weresynthesized. The aim of this work was to investigate the feasibilityof using the star branched micelles as a drug and gene co-carriertreating the glioma as shown in Scheme 1. Thus, miRNA conden-sation ability and physiochemical properties of star copolymers,including size and zeta potential, were characterized. Meanwhile,the cytotoxicity of star branched copolymers and gene transfectionefficiency were evaluated. The endosome escape ability of DOX-loaded and miR-21i-binded polyplex was also measured. Futher-more, the suppressive effects of co-delivery of DOX and miR-21iusing these nanoparticles were investigated via colony formationand flow cytometry assays in LN229 glioma cells in vitro as well asin subcutaneous nude mouse models.

2. Materials and methods

2.1. Materials

3-Butyn-1-ol, 2-bromo isobutyryl bromide, CuBr, sodium azide, 2,20-Bipyridine,and 2-(N,N-dimethylaminoethyl) methacrylate (DMAEMA, 97%) were purchasedfrom Alfa Aesar. Stannous octoate, and 3-Bromo-2,2-bis(bromomethyl) propanolwere purchased from SigmaeAldrich. Trimesoyl chloride and 1,4,7,10,10-hexamethyltriethylenetetramine were supported by J&K. All other chemicals were of reagentgrade and were used as received. Doxorubicin hydrochloride was also purchasedfrom SigmaeAldrich. Human glioma cell lines LN229 was obtained from the ChinaAcademia Sinica cell repository (Shanghai, China). The LysoTracker Blue DND-22wasobtained from Molecular Probes (Invitrogen). The 20-O-methyl (20-O-Me) miR-21inhibitor (sequences: miR-21i: 50-GTC CAC TCT TGT CCT CAA TG-30 , scrambled se-quences: 50-AAG GCA AGC UGA CCC UGA AGU-30) and the FITC-labeled miR-21 in-hibitor were chemically synthesized by Shanghai Gene Pharma (Shanghai, China).They were then dissolved in diethylpyrocarbonate (DEPC) water and frozenat �20 �C.

2.2. Synthesis and characterization of amphiphilic star-branched copolymers

2.2.1. Synthesis of pentaerythriol triazide-terminated PLA (PLA-(N3)3)

2.2.1.1. Synthesis of pentaery thrioltriazide. 3-Bromo-2,2-bis(bromomethyl) prop-anol (5.0 g,15.4 mmol) and sodium azide (6.0 g, 92.3 mmol) were added into 30 mLof DMF. The solution was flushed with dry N2 and stirred at 70 �C for three (3) days.Afterwards, the reaction mixture was filtered, and the solvent was removed underreduced pressure using a rotary evaporator. The crude product was dissolved in50 mL of double distilled water, and then extracted with 100 mL of dichloromethanetwice. The solutionwas dried over CaCl2 and filtered, and pentaerythriol triazidewasobtained after removing dichloromethane by rotary evaporation. The structure was

determined by 1H NMR and 13C-HMR (Varian 500 MHz spectrometer, USA) (Sup-plementary information, Figs. S1 and S2).

2.2.1.2. Synthesis of PLA-(N3)3 by ROP. First, 6.0 g of lactide were added to a reactiontube. Next, a solution of stannous octoate and pentaerythriol triazide (mole ratio,monomer and initiator to catalyst, M/I/C ¼ 500/1/1) in dry chloroform was added.The solvent was removed in vacuo, and the tube was sealed and immersed in asilicone oil bath at 130 �C. At the end of polymerization (12 h), the product wasdissolved in a small amount of chloroform and precipitated in an excess of methanol.The purification was repeated three (3) times, and the final product was dried in avacuum oven at 40 �C for 48 h. Then, the weight-average molecular weight (Mw)and polydispersity index (PDI) were determined by gel permeation chromatography(GPC). The elution solvent was tetrahydrofuran (THF), and the flow rate was 1 mL/min (30 �C). Calibration was based on polystyrene standards. The number averagemolecular weight (Mn) and the chemical structure were confirmed by 1H NMR. 1HNMR spectra were recorded in CDCl3 with a Varian 500 MHz spectrometerinstrument.

2.2.2. Synthesis of monocapped PDMAEMA with an alkynyl group (PDMAEMA-C^CH)

2.2.2.1. Synthesis of 3-Butyn-1-ol-2-bromoisobutyrate (BOBiB) ATRP initiator.To begin, 1.0 mL (13.2 mmol) of 3-Butyn-1-ol and 1.84 mL (13.2 mmol) of triethyl-amine (TEA) were dissolved in 15 mL of dichloromethane. The flask was cooled in awater/ice bath. Next, 1.8 mL (14.5 mmol, 1.1 equiv) of 2-bromoiso-butyryl bromide(BiBB) was diluted with 10 mL dichloromethane and added drop-wise to the solu-tion under nitrogen while stirring. The reaction was left to proceed for approxi-mately 24 h. The reaction products were filtered and extracted with H2O. Theorganic phase was dried over CaCl2 and filtered, and the solvent was removed byrotary evaporation. The product was purified by silica gel chromatography withdichloromethane as an eluant. The structure was determined by 1H NMR (Supple-mentary information, Fig. S3).

2.2.2.2. Synthesis of PDMAEMA-C^CH by ATRP. Initiator BOBiB (83 mg, 0.38 mmol)was dissolved in DMAEMA (3.0 g, 19 mmol) in a glass tube equipped with a Rotafloat ambient temperature. The catalyst CuBr (54 mg, 0.38 mmol) and ligand1,1,4,7,10,10-Hexamethyl triethylenetetramine (1.0 g, 0.45 mmol) were added, andthe mixture was degassed by three (3) vacuum/nitrogen cycles. Polymerization wascarried out at 60 �C. The product was then dissolved in THF and recovered byprecipitation in cold heptane. After drying to constant mass, conversion wasdetermined by gravimetry. The copper catalyst was removed by passing a solutionof the copolymer in THF through a column of basic alumina, before characteriza-tion. The molecular weight (Mn) of the polymer was calculated by comparing theintegrals of the alkynyl protons and the peaks of the PDMAEMA backbone. The Mwand PDI were determined by GPC. THF was used as the eluant at a flow rate of1 mL/min (30 �C).

2.2.3. Synthesis of star-branched PLA-b-PDMAEMA copolymersPLA-PDMAEMA3 was synthesized by direct click chemistry of PLA-N3 and

PDMAEMA-C^CH, while (PLA-PDMAEMA3)2 and (PLA-PDMAEMA3)3 were preparedvia first-coupled PLA-N3 onto terephthaloyl chloride and trimesoyl chloriderespectively, and then clicked with PDMAEMA-C^CH (Scheme 2).

Choosing (PLA-PDMAEMA3)3 as an example (Scheme 2), 2.0 g PLA-N3 and TEA(1.1 equiv.) was dissolved in 10 mL dry THF, and the solution was cooled in an ice/water bath. Then, 27 mg (0.1 mmol) of trimesoyl chloride was dissolved in 10 mL ofTHF and was added dropwise to the solution under nitrogen. The reaction mixturewas continually stirred at 0 �C for 3 h and then left stirring at room temperatureovernight. The white precipitate was filtered, and the filtrate was concentratedbefore precipitated in methanol. The crude product was redissolved in dichloro-methane and precipitated in methanol three (3) times. The star PLA-(N3)3 polymer(sPLA-(N3)3) was obtained after being dried at 40 �C for 48 h in a vacuum oven. Theaverage number of arms was calculated through 1H NMR by comparing the integralsof the aromatic protons and the appropriate peaks related to the PLA backbone. TheMw and PDI of sPLA-(N3)3 were measured by GPC.

For the click coupling reaction, 1.0 g (0.051 mmol) of star PLA-(N3)3 was reactedwith an excess of PDMAEMA-C^CH (1.2 � based on NMR determined functionality)in the presence of CuBr (21.5 mg, 0.15mmol) and 80 mg (0.51mmol) of 2,2-dipyridylin 10 mL of THF. After stirring at 35 �C for 12 h under nitrogen, the reaction mixturewas diluted with THF and then passed through an alumina column to remove excesscopper. The reaction products were concentrated before precipitation into hexanesand diethylether (V/V ¼ 3). The precipitate collected was dissolved in THF and thensubmitted to dialysis against water (Cut ¼ 10 kDa) for 48 h to remove excessPDMAEMA-C^CH. The star-branched copolymers were obtained after freeze-drying. The synthesis was confirmed by the 1H NMR measurement, and the armnumber ratio of PLA to PDMAEMA was calculated from the integration ratios of thePLA proton peak and PDMAEMA proton peak.

2.3. Determination of critical micelle concentration

The critical micelle concentration (CMC) was determined by a spectro-fluorophotometer (F24500, Hitachi, Japan) with pyrene as a hydrophobic probe. The

Scheme 1. Illustration of co-delivery of DOX and miR-21i by star-branched amphiphilic copolymer micelles and the mechanism for enhancement of chemosensitivity of gliomacells.

X. Qian et al. / Biomaterials 35 (2014) 2322e23352324

concentration of copolymers varied from 1 � 10�5 to 1 mg/mL, and the concentra-tion of pyrene was fixed at 6.0 � 10�7

M. Sample suspensions were prepared byadding a known amount of pyrene in acetone to a series of 20 mL vials. Then, theacetone was evaporated under a stream of nitrogen to a final concentration. Variousconcentrations (10 ml) of star-branched copolymers suspension were then added toeach vial in order to equilibrate with pyrene overnight at room temperature. Anemission spectra were obtained at an excitation wavelength of 339 nm, and theexcitation spectrawere obtained at an emissionwavelength of 395 nm. The intensityratio of peaks at 336 nm to those at 334 nm from pyrene excitation spectra versusthe logarithm of the copolymer concentration was used to measure CMC.

2.4. Preparation and characterization of star-branched copolymer micelles (sCPM)

The micelles of PLA-PDMAEMA3, (PLA-PDMAEMA3)2, and (PLA-PDMAEMA3)3(denoted as sCPM1, sCPM2 and sCPM3 respectively) were prepared by a solventevaporation method above their CMC concentrations. Briefly, 10 mg star-branchedpolymers were dissolved into 1 mL of dichloromethane. The mixture was theninjected into 10 mL of deionized water by a syringe and stirred for 8 h to evaporatethe dichloromethane.

The sizes of the self-assembled micelles were determined by dynamic lightscattering (DLS) (BI-90Plus, Brookhaven Instruments Ltd, USA) and zeta potentialswere determined by a zeta potential analyzer (Zeta PALS, Brookhaven InstrumentsLtd, USA) at 25 �C.For each sample, data obtained from the five measurements wereaveraged to yield a mean particle size and a z-potential. The morphologies of mi-celles were observed by TEM (JEOL-100CXⅡ, Japan).

2.5. Cytotoxicity of sCPM

The cytotoxicity of three sCPM was evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, using PEI 25 kDa as a control.4000 GES-1 cells were seeded in a 96-well plate at 37 �C for 24 h in 100 mL DMEM(Gibco), which was supplemented with 10% FBS (Gibco), 2 mM of glutamine (Sigma),100 mg/ml of penicillin (Sigma), and 100 mg/ml of streptomycin (Sigma). Differentconcentration of sCPM solutions were added and incubated for 48 h. The reactionwas then stopped by dissolving the cells in 200 mL of dimethyl sulfoxide (DMSO) for15 min. Quantification measurements (optical density) were obtained at a wave-length of 570 nm using a spectrophotometric analysis.

2.6. Transfection efficiency assess with reporter gene

pEGFP plasmid was used to study the gene transfection efficiency of polymerswith different architectures. PEI was used as a control group. The polyplexes ofvarious groups were prepared by incubating pEGFP plasmid and sCPM at an N/P of 2

according to the results of the agarose gel electrophoresis for miR-21i. After thepolyplexes were added to each well (2 mL) with a serum-free DMEM medium, thecells were incubated at 37 �C for 4 h in a humidified atmosphere with 5% CO2. Then,the media was replaced with fresh DMEM containing10% FBS and incubated for anadditional 44 h at 37 �C. Afterwards, the cells were observed on a Carl Zeiss Aviox-1inverted fluorescence microscope equipped with appropriate filters. Fluorescentimages were captured and recorded. The number of cells transfected with the GFPgene was counted under the fluorescence microscope.

2.7. DOX loading and release

For drug loading, 1 mg of DOX was dissolved in 2 mL of dichloromethane, andthe solution was mixed with a polymer solution before being injected. The DOXcontents in three sCPM were measured by a UVeVIS spectrophotometer (PerkinElmer, USA) at 485 nm. Before drug loading into polymeric micelle, DOX$HCl wasstirred with a 2 mole ratio of TEA in distilled water overnight to remove HCl fromDOX$HCl. The free base of drug DOX were extracted by dichloromethane and thenpurified via rotary evaporation, fully washing and freeze-drying.

Next, 4 mg of DOX-loaded micelles (D-sCPM1, D-sCPM2 and D-sCPM3) wereseparately dispersed in 5 mL of PBS buffer (pH 7.4) and placed in a dialysis mem-brane tube with MWCO of 3500 Da in a shaking water bath at 37 �C. The externalphase (PBS, pH 7.4, 500 mL) was stirred continuously. At scheduled intervals of time,500 mL samples were collected from the dialysis bag for measurement of theabsorbance at 485 nm using a spectrophotometer to characterize the concentrationof DOX that remained in the micelles. In addition, the external phase was replacedwith 5 ml distilled water to ensure the sink conditions for the drug release until the35th hour. The concentrations of DOX in the dialysis bag were determined accordingto the standard curves at corresponding buffer solutions. This allowed the releaseamount of the drugs to be calculated.

2.8. Preparation of DOX-loaded and miR-21i-complexed sCPM (D-sCPM/miR-21i)

A predetermined amount of miR-21i (20 mm/L) was added to vials containingappropriate amounts of D-sCPM solutions. The mixture was vortexed for 15s andthen kept still at room temperature for 30 min. A series of D-sCPM/miR-21i wereformed at different N/P ratios (2/1, 4/1, 8/1). The condensation ability of D-sCPMwasassessed by agarose gel electrophoresis. Electrophoresis was carried out on a 1%agarose gel with a current of 80 V for 15 min in a TAE buffer solution (40 mM TriseHCl, 1 v/v % acetic acid, and 1 mM EDTA). The retardation of the complexes wasvisualized through an ethidium bromide stain. The images were acquired by a geldoc system (G:BOX Chemi XR5, Syngene, USA).

Scheme 2. Synthesis of star-branched PLA-PDMAEMA block copolymers.

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2325

Fig. 1. (a)1H NMR spectra of PLA-(N3)3, (b) PDMAEMA-C^C, (c) sPLA-(N3)3, and (d) (PLA-PDMAEMA3)3.

X.Q

ianet

al./Biom

aterials35

(2014)2322

e2335

2326

Table 1Structure parameters of star-branched copolymers.

Sample NAPLAa NAPDMAEMA

b

Theoreticalvalue

Actualvalue

Theoreticalvalue

Actualvalue

PLA-PDMAEMA3 1 1 3 2.96(PLA-PDMAEMA3)2 2 1.96 6 5.80(PLA-PDMAEMA3)3 3 2.85 9 8.64

a NAPLA, the average number of arms of star PLA.b NAPDMAEMA, the arm number of PDMAEMA clicked to star PLA.

Table 2Characterization of sCPM and D-sCPM/miR-21i.

Sample Sizea

(nm)PDIa Zetab

(mV)CMCc

(�10�3

mg/ml)

Drugloadingd

(%)

sCPM1 84.7 � 0.8 0.973 16.6 � 0.4 4.63sCPM2 82.1 � 0.1 0.747 22.2 � 0.1 1.54sCPM3 38.4 � 0.2 0.486 35.0 � 0.5 0.32D-sCPM1/miR-21i 93.2 � 0.3 0.832 10.4 � 0.7 10.0 � 0.5D-sCPM2/miR-21i 91.7 � 0.4 0.719 13.6 � 0.7 12.2 � 0.9D-sCPM3/miR-21i 65.0 � 0.8 0.433 26.8 � 0.6 17.3 � 0.2

a Determined by a dynamic light scattering (DLS) at 25 �C.b Determined by a zeta potential analyzer at 25 �C.c Determined using pyrene as a fluorescence probe.d DOX loading efficiency was analyzed using the UVeVIS spectrophotometer

(Perkin Elmer, USA) at 485 nm and the following expressions: Drug loading effi-ciency (%w/w) ¼ (Weight of drug found loaded/Weight of drug�loadednanoparticles) � 100.

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2327

2.9. Cellular uptake studies and flow cytometry analysis

D-sCPM1/FITC-miR-21i, D-sCPM2/FITC-miR-21i and D-sCPM3/FITC-miR-21iwere added to LN229 cells at 80e90% confluency and incubated for 4 h at 37 �C.Then, the samples were washed thrice with PBS of the corresponding temperature,and uptake rates were detected by flow cytometry (Becton Dickinson, USA).

2.10. Intracellular fate of DOX and miR-21i

Confocal fluorescent microscopy was used to assess the intracellular traffickingof miR-21i and DOX (1.8 mm/L), which was mediated by sCPM3. Cells grown on theglass coverslips of a 6-well plate were incubated with the D-sCPM3/FITC-miR-21icomplexes for 2 h, 4 h, and 8 h. At the end of the incubation period, the cells werewashed three times with PBS and fixed in paraformaldehyde in PBS for 15min. Then,the cells were incubated with the LysoTracker blue (50 nM, Molecular Probe, Invi-trogen Co, OR, USA) for 0.5 h at the end of an uptake study for endosome/lysosomelabeling. The cells were then washed three times with PBS and stored at4 �C.Localization of FITC-miR-21i and DOX in cells was visualized by a Zeiss 510LSMNLO confocal microscope (Carl Zeiss Microscope systems, Jena, Germany) withidentical settings for each confocal study.

2.11. Colony formation assay

A colony formation assay was used to measure the proliferative capacity of thecancer cells to anticancer drugs. Cells treated with D-sCPM3, sCPM3/miR-21i, D-sCPM1/miR-21i, D-sCPM2/miR-21i, D-sCPM3/miR-21i (DOX concentration 1.8 mm/L)were seeded separately in 6-well plates (0.5 � 103 cells per well) and cultured for 2weeks. For the miR-21i transfection, D-sCPMwere mixed with miR-21i (0.29 mg/mL)respectively in DMEM culture medium and incubated for 15 min at room temper-ature .The colonies were fixed with methanol for 10 min and stained with 1% ofcrystalviolet (Sigma) for 1 min. Each group was measured in triplicate.

2.12. RNA extraction and real-time PCR

The glioma cells LN229were treated with various groups including sCPM, sCPM/miR-21i, PEI/miR-21i, or D-sCPM/miR-21i for 72 h. The scramble sequence was usedas a control. Then, the RNA of different groups was extracted using a trizol reagent

Fig. 2. Excitation fluorescence spectra of pyrene in aqueous solutions (Inset, plot ofI336/I334 from pyrene excitation spectra versus log C for concentration of (PLA-DMAEMA3)3 copolymers).

(Invitrogen) according to the standard protocol. The quantitative analysis of thechange in expression levels was calculated by a real-time PCR machine (750 ABI,USA). Furthermore, the TaqMan MicroRNA assay kit (Applied Biosystems) was usedaccording to the manufacturer’s instructions.

2.13. Evaluation of cell apoptosis

Apoptosis was evaluated by flow cytometry analysis. Briefly, after treatmentwith different groups as described in the colony formation assay for 48 h, LN229cells were collected and subjected to an annexin V/PI stain using an annexin V-FITCApoptosis Detection Kit (BioVision, Palo Alto, CA) in accordance with the manu-facturer’s protocol. Finally, fluorescence was measured by flow cytometry using aFACS flow cytometer (Becton Dickinson, San Jose, CA).

2.14. Protein extraction and western blotting

Glioma cells were treated with different groups as shown in the cell apoptosisassay. After two days, each group of cells was washed with pre-chilled phosphate-buffered saline (PBS) three times and then solubilized in 1% of Nonidet P-40 lysisbuffer. Then, total proteins were extracted as previously described [18]. The proteinswere separated by SDS-PAGE on 8% of SDS-acrylamide gels, transferred to PVDFmembranes (Millipore), and then incubated with primary antibodies detectingcaspase-3 and BCL-2 (Zhongshan Bio Corp, Beijing, China) that were followed byincubation with an HRP-conjugated secondary antibody (Zhongshan Bio Corp).GAPDH was selected as a housekeeping gene. The protein was detected using aSuper Signal protein detection kit (Pierce, Rockford, IL).

2.15. In vivo antitumor activity study

BALB/c-A nude mice, four to six weeks old, were purchased from the AnimalCenter of the Cancer Institute, Chinese Academy of Medical Science. For inoculation,glioma LN229-luc cells were suspended in 100ml of serum-free DMEMmedium andinoculated subcutaneously to the flanks of unanesthetized mice (2 � 106 cells permouse). Tumor establishment and growth were monitored by transduced cells withluciferase lentivirus stereotactically. Mice were randomly divided into six (6) groups(n ¼ 6): Saline, sCPM3, DOX solution, sCPM3/miR-21i, D-sCPM3, D-sCPM3/miR-21i.The solutions were given by the intratumoral injections every two (2) days (5 mgDOX-equiv. per kg body weight for each dose) for three (3) weeks. Tumor volumewas measured using the formula: volume ¼ length � width2/2.

2.16. Histology and immunohistochemistry

The paraffin-embedded tissue sections were used for examination of PTEN,pAKT, BCL-2, and caspase-3 as well as HE stains. Sections were incubated withprimary antibodies (1:100 dilutions) overnight at 4 �C, followed by a biotin-labeledsecondary antibody (1:100 dilutions) for 1 h at 37 �C, and then incubated with ABC-peroxidase and diaminobenzidine (DAB), counterstained with hematoxylin, andvisualized using a light microscope.

2.17. MiR-21 detection by fluorescence in situ hybridization

MiR-21expression was detected by fluorescence in situ hybridization (FISH) aspreviously described [19]. Locked nucleic acid (LNA)-modified oligonucleotideprobes and LNA/DNA oligos contained locked nucleic acids at eight consecutivecentrally located bases (indicated by the underlined sequence) and had thefollowing sequences: LNA-miR-21 50-TCAACATCAGTCTGATAAGCTA-3’.

2.18. Statistical analysis

SPSS 16.0 (SPSS, Chicago, IL) was used for ANOVA, c2 test. All data were rep-resented by mean � SD. Statistical significance was determined at P < 0.05.

Fig. 3. (a) TEM image of sCPM3 (Scale bar ¼ 100 nm). (b) Size distribution of sCPM3 measured by DLS. (c) Cytotoxicity of the star copolymers in GES-1 cells. The cells were treatedfor 48 h, and metabolic activity was measured by a MTT assay (mean � S. D, n ¼ 3).

X. Qian et al. / Biomaterials 35 (2014) 2322e23352328

3. Results

3.1. Synthesis of PLA-PDMAEMA amphiphilic star-branchedpolymers

PLA-PDMAEMA star-branched copolymers were synthesized forpreparation of the assemble micelles for co-delivery of miR-21i andthe anticancer drug DOX. As shown in Scheme 2, the hydrophobicarm PLA-(N3)3 was synthesized by ROP of lactide with Sn(Oct)2 asthe catalyst and pentaerythriol triazide as the co-initiator (Scheme

Fig. 4. (a) Fluorescent images of EGFP expression of sCPM1/pEGFP, sCPM2/pEGFP and sCPM3of 2.(b) Transfection efficiency was counted under the fluorescence microscope. Incubation

2, ①). The hydrophilic arm PDMAEMA-C^C was synthesized byATRP using BOBiB as an initiator and CuCl/bpy as the catalyst sys-tem (Scheme 2, ②). Then PLA-PDMAEMA3 was synthesized bydirectly clicked chemistry of PLA-(N3)3 and PDMAEMA-C^C(Scheme 2, ③), while (PLA-PDMAEMA3)2 and (PLA-PDMAEMA3)3were obtained by first coupling PLA-(N3)3 onto terephthaloylchloride and trimesoyl chloride respectively, and then clicked withPDMAEMA-C^C (Scheme 2, ④ and ⑤).

The 1H NMR spectra of PLA-(N3)3, sPLA-(N3)3, PDMAEMA-C^C,and typical star copolymer (PLA-PDMAEMA3)3 shown in Fig. 1aed,

/pEGFP complexes obtained by fluorescence microscope in LN229 cells at the N/P ratiotime: 48 h. Each value represents the mean � SD from triplicate determinations.

Fig. 5. (a) Gel retardation assay of D-sCPM1/miR-21i, D-sCPM2/miR-21i, and D-sCPM3/miR-21i at different N/P ratios. (b)TEM images and (d) size distribution of D-sCPM3/miR-21i.(c) Release of DOX from D-sCPM micelles.

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2329

respectively agree with the expected structures of the polymers.For PLA-(N3)3 (Fig. 1a), characteristic resonance peaks from PLAwere clearly seen at 1.58 ppm (methane proton) and 5.17 ppm(methyl protons), while the resonance peaks from pentaerythrioltriazide end group appeared at 3.37ppm. After coupling PLA-(N3)3onto trimesoyl chloride and purification by repeated methanolprecipitation, the polymer sPLA-(N3)3 reserved distinct resonancepeaks of PLA and pentaerythriol triazide end group in its 1H NMRspectrum (Fig. 1c). In addition, the resonance peak of benzene coreappeared at 8.92ppm, demonstrating the success of the couplingreaction. The 1H NMR spectrum of hydrophilic cationic armPDMAEMA-C^C was shown in Fig. 1b. All of the characteristicresonance peaks (2.28ppm-N-CH3, 2.55ppm -N-CH2-, 4.05ppm-CH2-O-, 1.81ppm, and 1.90ppm -CH2-C-) that belonged toPDMAEMA protons were clearly displayed. The resonance peak ofthe chain-end alkynyl proton at 2.0ppm was covered, but the peakat 1.25ppm, which was assigned to eCH3 protons of the terminalBOBiB group, could be clearly seen. Fig. 1d showed the 1H NMRspectrum of (PLA-PDMAEMA3)3. All of the characteristic peaksassigned to PLA and PDMAEMA protons were clearly displayed. Theprotons formed by click chemistry appeared at 7.536 ppm,demonstrating the successful synthesis of the star copolymer.

The molecular weights of the PLA arm and the PDMAEMA armwere determined by 1H NMR and GPC. The Mn of PLA was 6500Dwith a corresponding degree of polymerization (DP) of 90, whilethe Mn of PDMAEMA arm was 3925D with a corresponding DP of25 (Table S1, Supplementary information). The average arm

number of PLA and PDMAEMA determined by 1H NMRwas listed inTable 1.

3.2. Critical micelle concentration

The CMC of star-branched copolymers was measured before thepreparation of sCPM. Fig. 2 showed the fluorescence spectra ofpyrene in presence of various concentration of (PLA-PDMAEMA3)3.A remarkable red-shift could be observed, indicating that the hyperbranched star copolymers had a strong trend of assembling mi-celles. By calculating the contrast of peaks, the CMC of three co-polymers was determined as 4.63 � 10�3 mg/ml, 1.54 � 10�3 mg/ml, and 3.2 � 10�4 mg/ml, respectively (Table 2).

3.3. Preparation and characterization of sCPM

The sCPM were prepared by a solvent evaporation method abovethe CMC concentration. The morphologies of sCPM were observedvia TEM. At the same time, the size and zeta potential were deter-mined through DLS and zeta potential analyzer. The results wereshown in Fig. 3 and Table 2. It was apparent that the sCPM weredispersed as individual nanoparticles with a well-defined sphericalshape and homogeneously distributed (Fig. 3a,b). In addition, as thenumber of arms of the copolymers increased, the size of their sCPMdecreased from 84.7 nm to 38.4 nm, whereas the values of zeta po-tentials increased from þ16.6 mV toþ35.0 mV (Table 2).

Fig. 6. (a, b) Uptake of the D-sCPM/FITC-miR-21i polyplex. LN229 cells were treated with the D-sCPM/FITC-miR-21i polyplex for 4 h at 37 �C. Then, the samples were detected byflow cytometry. Each experiment was done three times. (*p ＜ 0.05). (c) Colocalization of the D-sCPM3/FITC-miR-21i polyplex with LysoTracker Blue, as a function of incubationtime. LN229 cells were labeled with lysosomal probe (blue) for 30 min after the D-sCPM3/FITC-miR-21i treatment. Then, cells were fixed and prepared for LSCM. DOX (red) andFITC-miR-21i (green) were seen. White arrow head points the FITC labled miR-21i escaped from the lysosome. (For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)

Fig. 7. Real-time PCR analysis detected relative miR-21 expression levels followingdifferent groups. *P < 0.05, **P < 0.01.

X. Qian et al. / Biomaterials 35 (2014) 2322e23352330

3.4. Cytotoxicity

The cytotoxicity of threes sCPMs were evaluated by the MTTassay using GES-1 cells. The commonly used polymeric transferagent, 25 kDa PEI, was used as a control. The results (Fig. 3c)showed that all vectors displayed a dose dependent cytotoxicity,and the toxicity increased with the number of arms of star-branched copolymers. However, these copolymers exhibitedmuch lower cytotoxicity compared with PEI over the tested con-centration range (between 13 and 400 mg/mL). At the concentra-tions of 50 and 200 mg/ml, PEI 25 kDa lead to the reduction of cellviability by 53% and 89%, respectively; while at 200 mg/ml, only 19%and 36% cells were killed when incubationwith sCPM1 and sCPM3,demonstrating the well non-cytotoxic of the star-polymersmicelles.

3.5. Gene transfection mediated by sCPM using the reporter geneassay

The sCPM mediated gene transfection was evaluated usingpEGFP plasmid in LN229 cell lines at N/P ratio of 2 according toagarose gel electrophoresis (Fig. 5a). Cells were treated withpEGFP complexes for 48 h and 25 kDa PEI were employed as

Fig. 8. (a) Colony formation assay in LN229 cell lines treated with the various polyplexes. (b) Flow cytometric analysis of cell death, and (c) a western blot analysis of BCL-2 andcaspase-3 protein expression, which were performed as described in the “Materials and Methods”. Each value represents the mean � SD from triplicate determinations. *P < 0.05,**P < 0.01.

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2331

positive controls. As shown in Fig. 4, all polyplexes based onPLA-PDMAEMA exhibited certain transfection activity. Thetransfection efficiency increased with the increasing number ofarms of the star copolymers, reaching from 12% to 45%. Incontrast, 25 kDa PEI, showed only 18% transfection efficiency atits optimal N/P ratio of 10/1 [20], which was much lower thansCPM3. The largely enhanced gene transfection of sCPM ascompared to PEI was most likely due to their increasing surfacecharge density and improved gene condensation as well asendosomal escaping ability.

3.6. Co-encapsulation of DOX and miR-21i

The DOX loading content of the three nanoparticles sCPM1,sCPM2, and sCPM3 were 10.0 wt%, 12.2wt%, and 17.3wt%,respectively, determined by UVeVIS spectrophotometer at485 nm (Table 2). Then, the drug-loaded sCPM was submitted tocondense miR-21i. The condensation capability of D-sCPM withmiR-21i was assessed by agarose gel electrophoresis at N/P ratiosranging from 0 to 8. As shown in Fig. 5a, the mobility of miR-21iwas changed with the N/P ratio. Furthermore, the miR-21i wascompletely retarded at the ratio of 2, especially in the case of D-sCPM3, which indicates that the D-sCPM could strongly bind miR-21i. The complex of miR-21i by D-sCPM was further confirmed bya zeta potential analysis. The surface charge decreased due to thebinding of electro negative miRNA (Table 2). The morphology ofthe D-sCPM3/miR-21i polyplex was observed by TEM. As shownin Fig. 5b, the complex were well dispersed as individual nano-particles with a spherical shape though the sizes were slightlylarger than sCPM3. The increase in size was confirmed by DLS

measured, which indicates that the size of the complexesincreased by between 11 and 71%, depending on the architectureof the copolymers. However, the size distributions were keptnarrow as can be seen from Fig. 5d.

All D-sCPMs exhibited two-phase release kinetics: a quick burstrelease, followed by a slow prolonged release (Fig. 5c). However, D-sCPM3 showed a relative slower sustained-release rate comparedwith other drug-loaded micelles, suggesting awell retention abilityof this copolymer, probably due to its micelle-like unimer statearchitecture.

3.7. Cellular uptake and intracellular fate

In order to identify the uptake efficiency of the miR-21i medi-ated by D-sCPM, miR-21i was labeled with FITC fluorescence, andthe polyplexes D-sCPM1/miR-21i, D-sCPM2/miR-21i, and D-sCPM3/miR-21i were added to the LN229 cells, respectively. Inaddition, their uptake efficiency by LN229 tumor cells was assessedthrough flow cytometry. As shown in Fig. 6a and b, after incubationin the cells for 4 h, the uptake efficiency increased in the followingorder: D-sCPM1/miR-21i (78.3%), D-sCPM2/miR-21i (80.5%), and D-sCPM3/miR-21i (89.1%) respectively, and the statistical analysisshowed the uptake efficiency of D-sCPM3/miR-21i was significantlyhigher than the other two polyplexes D-sCPM1/miR-21i and D-sCPM2/miR-21i, which was consistent with the results of the eGFPreporter gene assay.

To track the D-sCPM/miR-21i polyplex following its uptake, thelysosomal compartment of the cultured LN229 cells was stainedwith the LysoTracker blue probe after being treated with D-sCPM3/miR-21i. The intracellular localization of the DOX (red)

X. Qian et al. / Biomaterials 35 (2014) 2322e23352332

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2333

and miR-21i (green) was traced at different time points (Fig. 6c).At 2 h after D-sCPM3/miR-21i treatment, many complexes weredispersed in the cytoplasm and some aggregates were foundentrapped within the lysosomal vesicles. Magnified imagesclearly showed the colocalization of the D-sCPM3/miR-21i poly-plex with lysosomal vesicles in cells. Interestingly, as the incu-bation time increased to 4 h, some of the DOX and miR-21i fell offthe D-sCPM3/miR-21i micelles. At the same time, DOX and miR-21i entrapped in the lysosomal compartments graduallydisperse into cytosol and escaped from the lysosome. At 8 h in-cubationwith D-sCPM3/miR-21i, significantly higher intracellularDOX fluorescence intensity was observed in the nucleus of thecells. Simultaneously, much of miR-21i fluorescence was seen inthe cytosol, indicating that the D-sCPM3/miR-21i polyplex couldescape the lysosome successfully.

3.8. MiR-21 expression in LN229 cells

MiR-21 expression suppression was investigated by RT-PCR af-ter transfection of miR-21i to LN229 cells by star-copolymer mi-celles. Scramble miR-21i was used as a control. As shown in Fig. 7,the miR-21 expression was significantly reduced after the treat-ment of sCPM/miR-21i complexes. In the three star copolymers,miR-21 was reduced mostly, about 79% quantified by real-time PCRwhen delivered by sCPM3 compared with the other twomicelles ofsCPM1 and sCPM2 as well as PEI, reducing only 38%, 60%, and 41%,respectively. Notably, after treating the cells with D-sCPM/miR-21i,the miR-21 expression showed more reduction than sCPM/miR-21ialone. For the sCPM3, the expression of miR-21 was only 13% leftwhen added with D-sCPM3/miR-21i complex, possibly becauseDOX alone could decrease the activation of the miR-21 (26%reduced).

3.9. In vitro antiproliferative effect

The colony formation assay was performed to evaluate thein vitro cytotoxicity of D-sCPM/miR-21i complexes. As illustrated inFig. 8a, the viability of LN229 cells was only 21% when treated withthe D-sCPM3/miR-21i polyplex while 78.4% and 70.3% whentreatedwith the D-sCPM3 or sCPM3/miR-21i alone, which indicatesa significant synergetic effectiveness for the combination of miR-21i and DOX. In addition, the combined treatment delivered bysCPM3 produced significantly higher cytotoxicity than the otherstar copolymer micelles sCPM1 (49.2% survival) and sCPM2 (38.0%survival).

We also analyzed the effect of D-sCPM/miR-21i on apoptosis byflow cytometry and a western blot. Similarly, the combination ofDOX and miR-21i exhibited more apoptosis of the cells (Fig. 8b,c).Moreover, the Annexin V-positive early-phase apoptotic cellsshowed the largest increase (32%) with the treatment of D-sCPM3/miR-21i comparedwith cells culturedwith D-sCPM2/miR-21i (26%)or D-sCPM1/miR-21i (19%). Consistent with our hypothesis, in thewestern blot assay, the anti-apoptotic protein BCL-2 and apoptosis-related protein caspase-3 showed different levels of reduction andan increase with the treatment of a series of D-sCPM/miR-21i.Simply, the greatest reduction of BCL-2 and the fastest increase ofcaspase-3 appeared when treated with D-sCPM3/miR-21i, whichfollows the flow cytometry assay.

Fig. 9. LN229 cells transduced with a luciferase reporter were implanted into the flanks ofdose. Tumor mass was determined by bioluminescent imaging using an IVIS 100 and represfour tumors from each arm. (a) Representative images and (b) fluorescence quantity are showday 1 before drug injection treatment. (d) In situ examination of miR-21 expression and (e) HLN229 xenograft tumors.

3.10. In vivo anti-glioblastoma effect

Following up on the cell line data, we wished to evaluate thein vivo efficacy of D-sCPM3/miR-21i complexes. LN229 cells werefirst engineered to stably express luciferase before subcutaneousimplantation into mice. Bioluminescent imaging showed distincttumor growth with different treatments (Fig. 9a,b). When the DOXwas injected into the mouse, the tumor was inhibited as comparedwith no treatment. More importantly, the combined therapy of D-sCPM3/miR-21i resulted in a more obvious stasis of tumor growththan the DOX or miR-21i loaded in sCPM3 alone treatment, espe-cially after Day 14. Furthermore, the volumes of mice after differenttreatments were presented in Fig. 9c. The tumor volume of D-sCPM3/miR-21i group reached 0.168 cm3 by Day 22, while thecontrol treatment reached 1.52 cm3, which were nine times that ofthe D-sCPM3/miR-21i group.

The FISH assessment displayed different miR-21 expressionsafter the D-sCPM3/miR-21i or the alone treatment of a single agent(Fig. 9d), which was consistent with the in vitro experimentsmeasured by the PCR. The histological changes of the tumorinduced by different regimens were compared using the ex vivo HEstaining assay. As shown in Fig. 9e, tumors derived from PBS orsCPM3 treatment groups showed an obvious nuclear increase anddeeper chromatin staining. Among the four therapeutic groups,tumor tissues from animals receiving varying degrees of tumornuclear decreases and the D-sCPM3/miR-21i treatment showed thefewest and lightest tumor nuclear staining, indicating, once again,the best therapeutic effect resulted from the combined therapy viaco-delivery of the sCPM.

In addition, an immunohistochemistry was done for furtherstudies at the molecular level (Fig. 9f). The expressions of PTEN andcaspase-3 were significantly increased in the combined therapy ofDOX and miR-21i using sCPM3 co-delivery strategy whereas theBCL-2 and pAKT decreased the most as compared with othertreatments. These results indicated that the PTEN-pAKT pathwayplayed key roles in the increased sensitivity of the chemotherapydrug DOX and induced of apoptosis when combined with miR-21iand DOX.

4. Discussion

Co-delivery of drugs and genes to the same tumor cell viapolymer nanoparticles has been proposed to achieve the maximalsynergistic/combined effect of drug and gene therapies [21]. PLAand PDMAEMA are the two of the most studied polymers amongthe variety of synthetic particles that have been utilized for drugand gene delivery. PDMAEMA has excellent “proton sponge” ef-fects, and PLA is biodegradable [22,23]. Proton buffering effect iscorrelated with high gene transfection efficiency, while biode-gradability is a key criterion determining whether a vector is suit-able for in vivo use [24]. Therefore, we hope that a combination ofPLA and PDMAEMA fulfills the requirements for the construction ofa nonviral vector with high transfection and drug-loading effi-ciency and minimal side effects. To deliver miR-21i and DOXsimultaneously, we synthesized three types of molecular archi-tectures based on PLA-PDMAEMA hyper-branched star-copolymer,i.e., AB3, (AB3)2, and (AB3)3, by conjugating different number of lowmolecular weight PDMAEMA arms onto the star-like PLA core. In

male mice at four to six weeks of age. Mice were then treated with the indicated drugentative bioluminescent images are presented. The graph represents the mean � SD ofn in. (c) The tumor volume was measured and is plotted in. The tumor size measured atE staining assay were performed. (f) represents the immunohistochemistry analysis of

X. Qian et al. / Biomaterials 35 (2014) 2322e23352334

the structural design, on the one hand by making use of the hy-drophobic of PLA chain, the formation of micelles containing hy-drophobic kernel could package hydrophobic drugs and ensurebiodegradability; on the other hand, the low molecular weight ofPDMAEMA was used because it has demonstrated a very lowtoxicity and can be readily cleared by renal excretion [25]. It wasnoted that very recently PDMAEMA with low molecular weightswere grafted onto biodegradable polymer such as PCL [26], pHEMA[27], and PMMA [28] to improve the safety of gene delivery.However, due to the decreased amount of primary amino groups,the ability to complex pDNA also has decreased. The use of a highly-branched structure not only reduces the toxicity of PDMAEMA butalso makes the micelle surface have the same high charge densityas the high molecular weight of PDMAEMA, which favors electro-static binding with negatively charged genes.

In constructing of gene carriers, the N/P ratio plays an importantrole in influencing the degree of the polyplexes particle size,transfection efficiency, and cytotoxicity of carriers [29]. High N/Pratio induces not only increased cellular uptake through charge-mediated interactions, but also disadvantageous higher cytotox-icity. While reduction of the N/P ratio will make the particle sizeincrease and decrease the transfection efficiency. Thus, the balancebetween toxicity and transfection efficiency must be sought [30]. Inpresent study, the star PLA-PDMAEMA micelles could condensemiR-21i at the N/P ratio of 2, which was much lower than generalgene delivery systems [31]. Therefore, the polyplexes displayed amoderate positive charge and ideally smaller sizes (<100 nm) infavor of achieving ideal transfection, demonstrating the vector hasthe advantage of high-load capacity of drugs and genes whencompared with other carrier systems. The transfection study car-ried out in the LN229 cell line using reporter gene assay (Fig. 4)strongly demonstrated that grafting a proper amount of PDMAEMAchains to PLA can remarkably improve the vector’s ability ofmediating efficient gene transfection. sCPM3 with nine PDMAEMAarms displayed the best performance among the three synthesizedsCPM. The most likely reason for this is the highest charge densityand the smallest particle size for sCPM3, which facilitates an easyuptake and endosome/lysosome escape of polyplexes and releaseof the loaded gene into cytosol, which are known to be essential forthe transfection of delivered genes [32].

The tracking of the intracellular fate of polyplexes by confocallaser scanning fluorescent microscopy demonstrated how miR-21iescaped from endosome (Fig. 5), while DOX diffused out of theendosome and finally accumulated in the nucleus where theyintercalated DNA and inhibited the biosynthesis of the macromo-lecular. Furthermore, bothmiR-21i and DOX suppressed themiR-21expression in tumor cells; and in all cases, co-delivery resulted inthe lowest expression level of miR-21 (Fig. 7). BCL-2, being a keyanti-apoptotic protein in the outer mitochondrial membrane, canprotect cells from apoptosis by acting at a point downstream from arelease of mitochondrial cytochrome c, thereby preventing acaspase-3-dependent proteolytic cascade [33]. Down regulation ofmiR-21 suppressed the expression of BCL-2 by upregulating theexpression of PTEN, which blocks the PI3K-AKT pathway. As aresult, the expression of caspase-3 was upregulated, leading to theapoptosis of tumor cells. On the other hand, the DOX, whichwas co-delivered with miR-21i, also inhibited tumor cells, resulting in asynergistic effect. The synergistic effect was proved by using animalmodels. Compared with DOX or miR-21i alone, co-delivery of DOXand miR-21i by sCPM3 to the tumor reduced the BCL-2 expressionand decreased the tumor volume noticeably, exhibiting a thera-peutic advantage for the combination of miR-21i and DOX in vivoinhibition of the tumor. Furthermore, the gene transfection effi-ciency and synergistic inhibitory effect showed a significant posi-tive correlation, illustrating that the hyper branched structure of

star polymers is critical for the synergy of genes and chemotherapydrugs. The more number of arms, the higher the transfection effi-ciency, and the more obvious the synergic effect. In summary, thesecationic star-branched micelles capable of efficiently co-deliveringmiRNA and hydrophobic anti-cancer drugs into cancer cells arehighly promising for their combination with cancer therapy.

5. Conclusion

To enhance the chemosensitivity of drugs for curing cancer bytaking the advantage of simultaneously delivery of miR-21i andchemotherapeutants, three amphiphilic star-branched copolymersPLA-PDMAEMA3, (PLA-PDMAEMA3)2 and (PLA-PDMAEMA3)3 weresynthesized. All three star-branched copolymers exhibited lowcytotoxicity and CMC concentration, however, the gene trans-fection efficiency and tumor inhibition ability of the combinationcopolymer micelles showed a remarkable dependence on theirmolecular architecture. The micelles of (PLA-PDMAEMA3)3, whopossessed a snowflake-like symmetric structure, exhibited smallersize, higher surface charge density and superior eGFP transfectionefficiency compared to PEI. In addition, the micelles could entraphydrophobic DOX, and at the same time, bind miR-21i efficiently.More importantly, the micelles could mediate the escaping of miR-21i from lysosome degradation and the release of DOX to nucleus,leading to the significantly synergistic inhibition of tumor throughregulation of expression of BCL-2 apoptosis by PI3k/AKT signalpathway. Our results revealed the great potential of amphiphilicstar-branched copolymer as an advanced gene and hydrophobictherapeutants co-delivery system for the chemosensitivity oftumor.

Acknowledgments

This project was financially supported by the National NatureScience Foundation of China (Grant Nos. 51073118, 51103107, and81172406); the Program for New Century Excellent Talents inUniversities (Grant No. NCET-08-0393); the National High Tech-nology Research and Development Program 863 (2012AA02A508),and Specialized Research Fund for the Doctoral Program of HigherEducation (20111202110004).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2013.11.039.

References

[1] Dong J, Zhao YP, Zhou L, Zhang TP, Chen G. Bcl-2 upregulation induced bymiR-21 via a direct interaction is associated with apoptosis and chemo-resistance in MIA PaCa-2 pancreatic cancer cells. Arch Med Res 2011;42:8e14.

[2] Reardon DA, Wen PY. Therapeutic advances in the treatment of glioblastoma:rationale and potential role of targeted agents. Oncologist 2006;11:152e64.

[3] So A, Sinnemann S, Huntsman D, Fazli L, Gleave M. Knockdown of the cyto-protective chaperone, clusterin, chemosensitizes human breast cancer cellsboth in vitro and in vivo. Mol Cancer Ther 2005;4:1837e49.

[4] Spankuch B, Kurunci-Csacsko E, Kaufmann M, Strebhardt K. Rational combi-nations of siRNAs targeting Plk1 with breast cancer drugs. Oncogene 2007;26:5793e807.

[5] Shafi G, Aliya N, Munshi A. MicroRNA signatures in neurological disorders. CanJ Neurol Sci 2010;37:177e85.

[6] Wu Y, Crawford M, Mao Y, Lee RJ, Davis IC, Elton TS, et al. Therapeutic deliveryof microRNA-29b by cationic lipoplexes for lung cancer. Mol Ther NucleicAcids 2013;2:e84.

[7] Mishra PJ, Merlino G. MicroRNA reexpression as differentiation therapy incancer. J Clin Invest 2009;119:2119e23.

[8] Aagaard L, Rossi JJ. RNAi therapeutics: principles, prospects and challenges.Adv Drug Deliv Rev 2007;59:75e86.

X. Qian et al. / Biomaterials 35 (2014) 2322e2335 2335

[9] Borel F, Konstantinova P, Jansen PL. Diagnostic and therapeutic potential ofmiRNA signatures in patients with hepatocellular carcinoma. J Hepatol2012;56:1371e83.

[10] Zhou J, Shum KT, Burnett JC, Rossi JJ. Nanoparticle-based delivery of RNAitherapeutics: progress and challenges. Pharmaceuticals (Basel) 2013;6:85e107.

[11] Cao N, Cheng D, Zou S, Ai H, Gao J, Shuai X. The synergistic effect of hierar-chical assemblies of siRNA and chemotherapeutic drugs co-delivered intohepatic cancer cells. Biomaterials 2011;32:2222e32.

[12] Moghimi SM, Hunter AC, Murray JC. Nanomedicine: current status and futureprospects. FASEB J 2005;19:311e30.

[13] Cho HY, Srinivasan A, Hong J, Hsu E, Liu S, Shrivats A, et al. Synthesis ofbiocompatible PEG-Based star polymers with cationic and degradable core forsiRNA delivery. Biomacromolecules 2011;12:3478e86.

[14] Sulistio A, Lowenthal J, Blencowe A, Bongiovanni MN, Ong L, Gras SL, et al.Folic acid conjugated amino acid-based star polymers for active targeting ofcancer cells. Biomacromolecules 2011;12:3469e77.

[15] Gao H. Development of star polymers as unimolecular containers for nano-materials. Macromol Rapid Commun 2012;33:722e34.

[16] Park S, Cho HY, Yoon JA, Kwak Y, Srinivasan A, Hollinger JO, et al. Photo-cross-linkable thermoresponsive star polymers designed for control of cell-surfaceinteractions. Biomacromolecules 2010;11:2647e52.

[17] Etrych T, Kovar L, Strohalm J, Chytil P, Rihova B, Ulbrich K. Biodegradable starHPMA polymer-drug conjugates: biodegradability, distribution and anti-tumor efficacy. J Control Release 2011;154:241e8.

[18] Shi Z, Qian X, Liu C, Han L, Zhang K, Chen L, et al. Aspirin-/TMZ-coloadedmicrospheres exert synergistic antiglioma efficacy vial inhibition of b-catenintransactivation. CNS Neurosci Ther 2013;19:98e108.

[19] Qian X, Shi Z, Ren Y, Liu C, Ji Y, Long L, et al. Synergistic inhibition of humanglioma cell line by temozolomide and PAMAM-mediated miR-21i. J ApplPolym Sci 2013;127:570e6.

[20] Hu C, Zhang L, Wu D, Cheng S, Zhang X, Zhuo R. Heparin-modified PEIencapsulated in thermosensitive hydrogels for efficient gene delivery andexpression. J Mater Chem 2009;19:3189e97.

[21] Janat-Amsbury MM, Yockman JW, Anderson ML, Kieback DG, Kim SW. Com-bination of local, non-viral IL12 gene therapy and systemic paclitaxelchemotherapy in a syngeneic ID8 mouse model for human ovarian cancer.Anticancer Res 2006;26:3223e8.

[22] Majewski AP, Stahlschmidt U, Jerome V, Freitag R, Muller AH, Schmalz H.PDMAEMA-grafted core-shell-corona particles for nonviralgene delivery andmagnetic cell separation. Biomacromolecules 2013;14:3081e90.

[23] Rancan F, Papakostas D, Hadam S, Hackbarth S, Delair T, Primard C, et al.Investigation of polylactic acid (PLA) nanoparticles as drug delivery systemsfor local dermatotherapy. Pharm Res 2009;26:2027e36.

[24] Liang B, He ML, Chan CY, Chen YC, Li XP, Li Y, et al. The use of folate-PEG-grafted-hybranched-PEI nonviral vector for the inhibition of glioma growthin the rat. Biomaterials 2009;30:4014e20.

[25] Jiang X, Lok MC, Hennink WE. Degradable-brushed pHEMA-pDMAEMA syn-thesized via ATRP and click chemistry for gene delivery. Bioconjug Chem2007;18:2077e84.

[26] Zhu C, Jung S, Luo S, Meng F, Zhu X, Park TG, et al. Co-delivery of siRNA andpaclitaxel into cancer cells by biodegradable cationic micelles based onPDMAEMA-PCL-PDMAEMAtriblock copolymers. Biomaterials 2010;31:2408e16.

[27] Olga S, Christian P, Markus H, Olivia MM, Thomas K. Low molecular weightpDMAEMA-block-pHEMA block-copolymers synthesized via RAFT-polymerization: potential non-viral gene delivery agents? Polymers 2011;3:693e718.

[28] Shen JN, Ye YF, Zeng GN, Qiu JH. Preparation and characterization of PMMA-b-PDMAEMA/polysulfone composite membranes by RAFT polymerization andtheir permeation performance of carbon dioxide. Adv Mater Res 2011;284:1717e23.

[29] Zhao QQ, Chen JL, Lv TF, He CX, Tang GP, LiangWQ, et al. N/P ratio significantlyinfluences the transfection efficiency and cytotoxicity of a polyethylenimine/chitosan/DNA complex. Biol Pharm Bull 2009;32:706e10.

[30] Zheng M, Pavan GM, Neeb M, Schaper AK, Danani A, Klebe G, et al. Targetingthe blind spot of polycationicnanocarrier-based siRNA delivery. ACS Nano2012;6:9447e54.

[31] Ma D, Liu ZH, Zheng QQ, Zhou XY, Zhang Y, Shi YF, et al. Star-shaped polymerconsisting of a porphyrin core and poly(L-lysine) dendron arms: synthesis,drug delivery, and in vitro chemo/photodynamic therapy. Macromol RapidCommun 2013;34:548e52.

[32] Cheng D, Cao N, Chen J, Yu X, Shuai X. Multifunctional nanocarrier mediatedco-delivery of doxorubicin and siRNA for synergistic enhancement of gliomaapoptosis in rat. Biomaterials 2012;33:1170e9.

[33] Adams JM, Cory S. The Bcl-2 apoptotic switch in cancer development andtherapy. Oncogene 2007;26:1324e37.