lable at ScienceDirect

Biomaterials 30 (2009) 4014–4020

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

The use of folate-PEG-grafted-hybranched-PEI nonviral vector for the inhibitionof glioma growth in the rat

Bing Liang a,1, Ming-Liang He c,1, Chu-yan Chan c, Yang-chao Chen c, Xiang-Ping Li d, Yi Li a, Dexian Zheng f,Marie C. Lin e, Hsiang-Fu Kung c, Xin-Tao Shuai b,*, Ying Peng a,**

a Department of Neurology, The Second Affiliated Hospital, Sun Yat-sen University, No. 107 West Road of Riverside, Guangzhou 510120, Chinab Biomedical Engineering Center, School of Chemistry and Chemical Engineering, Sun Yat-Sen University, Guangzhou 510275, Chinac The Center for Emerging Infectious, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kongd Department of Otolaryngology, Nanfang Hospital, Nanfang Medical University, Guangzhou, Chinae Department of Chemistry, The University of Hong Kong, Hong Kongf Department of Biochemistry and Molecular Biology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, China

a r t i c l e i n f o

Article history:Received 18 February 2009Accepted 13 April 2009Available online 8 May 2009

Keywords:In vivo gene transferNonviral vectorCombined gene therapyPEI PEGylationGliomaTumor targeting

Abbreviations: CD, cytosine deaminase; FA, folagrafted-hyperbranched-PEI; FR, folate-receptor; pDN(ethylene imine); PEG, poly(ethylene glycol); TRAIL, Tligand; 5-FC, 5-fluorocytosine; 5-FU, 5-fluorouracil.

* Corresponding author. Tel.: þ86 20 8411 0365; fa** Corresponding author. Tel.: þ86 20 8133 2095; fa

E-mail addresses: shuaixt@mail.sysu.edu.cn (X.-T. Sh(Y. Peng).

1 These two authors contributed equally to the ma

0142-9612/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.biomaterials.2009.04.011

a b s t r a c t

Combined treatment using nonviral agent-mediated enzyme/prodrug therapy and immunotherapy hadbeen proposed as a powerful alternative method of cancer therapy. The present study was aimed toevaluate the cytotoxicity in vitro and the therapeutic efficacy in vivo when the cytosine deaminase/5-fluorocytosine (CD/5-FC) and TNF-related apoptosis-inducing ligand (TRAIL) genes were jointly usedagainst rat C6 glioma cells. The potency of the FA-PEG-PEI used as a nonviral vector was tested in the FR-expressed C6 glioma cells and Wistar rats. The C6 glioma cells and animal model were treated by thecombined application of FA-PEG-PEI/pCD/5-FC and FA-PEG-PEI/pTRAIL. The antitumor effect was eval-uated by survival assays and tumor volume. This study revealed a significant increase of cytotoxicityin vitro following the combined application of FA-PEG-PEI/pCD/5-FC and FA-PEG-PEI/pTRAIL treatmentsin C6 glioma cells. Animal studies showed a significant growth inhibition of the C6 glioma xenograftsusing the combined treatment. These results demonstrated that the combined treatment generatedadditive cytotoxic effect in C6 glioma cells in both in vitro and in vivo conditions, and indicated that suchtreatment method using both enzyme/prodrug therapy and TRAIL immunotherapy might be a promisingtherapeutic strategy in treating glioma.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Glioma is the most common primary malignancy of the brain. Itis known as a highly chemoresistant and radioresistant cancer withhigh morbidity, mortality and extremely grim prognosis. Themedian survival time of glioma patients is generally less than 2years, despite multi-modality treatments with extensive surgicalresection, radiotherapy, chemotherapy or immunotherapy. Recentadvances in the neurosurgical technique, radiation therapy and

te; FA-PEG-PEI, folate-PEG-A, plasmid DNA; PEI, poly-

NF-related apoptosis-inducing

x: þ86 20 8411 2245.x: þ86 20 8133 2833.uai), docpengy@yahoo.com.cn

nuscript.

All rights reserved.

chemotherapy have all failed to improve the survival rate of thisgroup of patients [1]. Therefore, novel strategy is urgently neededfor promoting the survival rate of the glioma patients. In recentstudies, gene therapy has been proposed as one of the potentialstrategies which bears several advantages over conventional drugtherapies [2,3]. A long-term expression of high dosage therapeuticeffect could be easily achieved using gene therapy. The genetic agentcould be delivered locally, and specifically to target tissues, whichreduce the risk of nonspecific toxicity and ineffective dosing. In thisstudy, the gene therapeutic strategies being investigated werebased on some previously established anti-neoplastic principles,which included the use of prodrug/suicidal genes, tumor suppressorgenes and immune-enhancing cytokine genes [3–6].

Among all gene therapeutic strategies now being investigated,suicide gene/prodrug system has been recognized as one of the mosteffective method in treating tumors. It has been revealed as a highlypotent agent in treating most chemoresistant and radioresistanttumors [7–9]. One of the most widely investigated suicide gene/pro-drug systems is the cytosine deaminase/5-fluorocytosine (CD/5-FC)

B. Liang et al. / Biomaterials 30 (2009) 4014–4020 4015

which has been studied extensively during the last decade. Cytosinedeaminase is an enzyme that could be found in bacteria and fungi. Itcould deaminate the nontoxic prodrug 5-fluorocytosine (5-FC) to formthe highly cytotoxic 5-fluorouracil (5-FU). The latter compound thenmetabolite to inhibit thymidylate synthase or to act as false bases inDNA and RNA, thereby killing cells that are in the S-phase of the cellcycle [10]. By expressing the CD gene, and by administering water-soluble and low-toxicity 5-FC systemically, 5-FU is self-regenerative intumor cells. More importantly, 5-FC prodrug is membrane permeablewhich has high bioavailability that could penetrate easily via theblood–brain barrier into the cerebrospinal fluid [11]. The application of‘localized’ 5-FU chemotherapy might therefore avoid the toxicityassociated with systemic 5-FU therapy, leading to a higher intra-tumoral concentration. In addition, the characteristic that the 5-FUcould diffuse into adjacent cells via the cellular membrane would exerta powerful bystander effect [12]. This effect is essential in maximizingthe therapeutic effectof gene therapy, since it is currentlya challenge todirect therapeutic gene to target tumor cells [13]. As a result, regressioncan occur when a tumor is treated with a nontoxic level of 5-FC, even ifonly a small percentage of cells expresses CD.

Another effective antitumor strategy that is proposed currently isthe apoptosis-inducing gene therapy. A wide variety of apoptosis-inducing molecules have been identified to combat tumor cells.Among all, the ligand-type cytokine molecules of the tumor necrosisfactor (TNF) family are recognized as the best candidate. TNF-relatedapoptosis-inducing ligand (TRAIL) is a type II transmembranemolecule, in which the carboxyl-terminus of the receptor-bindingdomain protrudes extracellularly [14]. Recombinant soluble humanTRAIL has already been employed in clinical investigation as cancertherapy for it has been shown to induce apoptosis in various humancancers. It functions by triggering the apoptotic signal cascadesthrough binding cognate receptors displayed on the cell surface. Itwas also noted to exhibit potent antitumor activity without inducedtoxicity in healthy tissue in various cancer xenograft models [15].

To date, the major challenge in gene therapy is to developa highly effective gene delivery system with low toxicity. Nonviralvector is still an attractive option although the current agents beingused displayed disadvantages, e.g. low transfection efficiency andhigh toxicity [16]. To overcome the problem of high cationic toxicity[i.e. polyethylene imine (PEI)] and low transfection efficiency [i.e.PEGylated PEI (PEG-PEI)], our team has linked a cell specific tar-geting molecule folate (FA) on polyethylene glycol (PEG). The FA-PEG was then grafted onto the hyperbranched PEI (25 kD). Folate isa common targeting ligand used for anti-cancer agents, since itstarget (i.e. folate receptor) is often overexpressed in tumor cells (i.e.C6 cell line) yet, rarely found in normal tissue, especially in thenormal brain tissue [17]. Therefore, the FA has been used to test itsenhancing effect on vector delivery in FR-enriched tumor cells suchas C6 glioma cells [18]. In our pervious study, the FA-PEG-grafted-hyperbranched-PEI (FA-PEG-PEI) could effectively condenseplasmid DNA (pDNA) into nanoparticles with a positive surfacecharge under a suitable N/P ratio of 15 [19]. In the present study, thepotency of the FA-PEG-PEI which could be used as a nonviral vectorwas tested in the FR-expressed C6 glioma cells and Wistar rats.

2. Materials and methods

2.1. Plasmids and chemicals

Plasmids pCMVCD was kindly provided by Dr. W. Walther (Max-Delbruck-Center for Molecular Medicine, Berlin, Germany) and the pCMVTRAIL was productedby our lab. Plasmid DNAs were amplified in Escherichia coli and were purifiedaccording to the manufacturer’s instructions (QIAGEN, CA, USA). The quantity andquality of the purified pDNA were assessed by measuring its optical density at260 nm and 280 nm, and by electrophoresis in 1% argrose gel, respectively. Thepurified plasmid DNA was kept in aliquots at a concentration of 1 mg/ml. In this study,all chemicals including PEI 25,000 Da, monomethoxy PEG (mPEG-OH) 3400 Da, 5-

fluorocytosine and (3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyl tetrazolium bromide(MTT) were products from Sigma–Aldrich (St Louis, MO, USA). Polyplexes, i.e. thedelivery agent/pDNA complexes, used throughout the present study were preparedat N/P 15 according to our previous in vitro results that the polyplexes received thehighest transfection efficiency in C6 glioma cells whilst a low cytotoxicity at this N/Pvalue [19].

2.2. Synthesis of delivery agents

a-Hydroxy-3-amino-poly(ethylene glycol) (HO-PEG-NH2) (Mn¼ 3.4 kDa, Mw/Mn¼ 1.15) was prepared according to a report by Kataoka et al. [20]. To conjugatefolate to HO-PEG-NH2, folic acid (2 mmol) was dissolved in anhydrous DMSO(20 ml). N-hydroxysuccinimide (NHS, 4 mmol) and dicyclohexylcarbodiimide (DCC,4 mmol) were added and the mixture was stirred overnight at room temperature.The mixture was mixed with a DMSO solution of HO-PEG-NH2, and TEA solution (pH8.0) was added. The mixture was then filtrated, dialysized against deionized water(MWCO: 1000 Da), and lyophilized. FA-PEG-OH thus prepared was converted intoFA-PEG-COOH by reaction with succinic anhydride (SA). FA-PEG-OH and SA (1:5 inmolar ratio) were dissolved in 20 ml anhydrous chloroform and refluxed at 70 �C for48 h. After chloroform was removed by distillation, polymer was re-dissolved in20 ml deionized water and dialyzed against water for two days to remove smallmolecular succinic acid and succinic anhydride. Polymer solution was freeze-driedto yield pure FA-PEG-COOH. FA-PEG-PEI was synthesized and characterized aspreviously described [18]. In brief, FA-PEG-COOH (1 mmol) was activated with NHS(2 mmol) for 24 h in dry dichloromethane (20 ml) containing dicyclohexyl-carbodiimide (1.2 mmol) as a catalyst. The precipitated 1,3-dicyclohexylurea (DCU)was removed by filtration. The filtrate was added to diethyl ether and cooled at 4 �Cfor 2 h. The precipitate was collected by filtration and dried under vacuum at roomtemperature. Hyperbranched PEI 25 kDa and the NHS activated PEG were dissolvedin PBS (pH 7.4) and magnetically stirred for 24 h at room temperature to produce FA-PEG-PEI. The mixture was purified by membrane dialyses (MWCO: 8000 Da) indistilled water for 1 day and the solution was lyophilized. Nontargeting PEG-PEI wassynthesized by the same approach using the NHS/DCC chemistry. PEG graftingdensity: PEG(mol)/PEI(mol)¼ 3:1.

2.3. Cell culture

Glioma C6 cells were obtained from American Type Culture Collection (ATCC)and maintained in high glucose Dulbecco’s modified Eagle’s medium (DMEM)supplemented with 10% fetal bovine serum (FBS) and 1% antibiotics (penn/strep,Invitrogen Corporation) in a humidified atmosphere of 5% CO2 at 37 �C. When thecell confluence of 90% was reached, they were trypsinized and subcultured. All cellculture reagents were purchased from Invitrogen Corporation (Carlsbad, CA, USA).

2.4. Western blotting analysis

The C6 glioma cells (1�105) were seeded in 24-well plate a day before trans-fection. They were bathed in DMEM with 10% FBS complete media and were incu-bated in a humidified atmosphere with 5% CO2 at 37 �C until the cell confluence wasaround 70%. Four hours prior to transfection, the media was removed and replacedwith fresh DMEM with 10% FBS. To test the expression of plasmids, pCMVCD andpCMVTRAIL, 1 mg of DNA was diluted in 50 ml of serum-free DMEM in an Eppendorftube and mixed thoroughly. Based on an N/P ratio of 15, corresponding quantity ofFA-PEG-PEI was added to a 50 ml serum-free DMEM in another sterilized tube andthe sample was vortex mixed immediately. Both mixtures were then left to incubateat room temperature for 5 min. Samples of both tubes were then vortex mixedtogether and was left for incubation at room temperature for 30 min. The originalcell culture media was replaced with a 100 ml complex solution whilst a 200 mlserum-free DMEM was added on top for each well. They were incubated at 37 �C for4 h. Thereafter, the transfection medium was changed with fresh and completeDMEM culture media. At 72 h after the transfection, cells were washed twice withPBS and lysed with SDS sample buffer. Protein samples (20 mg) were separated usingSDS-PAGE and transferred onto polyvinylidene difluoride (PVDF) membranes. Themembranes were incubated at room temperature for 30 min in a blocking buffer (5%low fat milk, 150 mm NaCl, and 20 mm Tris–HCl, pH 7.5), and were then incubatedwith a sheep primary antibody (1:200 dilution) against the CD or a rabbit antibody(1:500 dilution) against the TRAIL and caspase-3 (Covance, Richmond, CA, USA).Secondary antibody, horseradish peroxidase (HRP)-conjugated anti-sheep or anti-rabbit IgG, was used to amplify the signal. The blots were developed using chem-iluminescence system (New Life Science Products, Boston, MA, USA) and the resultswere photo-documented. The membranes were then washed again with buffer, andwere rehybridized with a primary antibody for b-actin (1:500 dilutions). HRP-conjugated antibody and chemiluminescence system were used for the detection ofb-actin as described earlier. The proteolysis of caspase-3 was evaluated by Westernblotting by using antibodies against caspase-3 that could detect both unprocessedproenzyme and active forms after cleavage.

B. Liang et al. / Biomaterials 30 (2009) 4014–40204016

2.5. In vitro 5-FC sensitivity

The C6 cells were grown in 96-well plates at an initial density of 6000 cells/well.They were transfected with the same concentration as that used for the 24-wellplates. Transfection was performed by using 300 ng of pDNA in 150 ml of serum-freegrowth medium. Four hours later, the transfection medium was changed with freshand complete DMEM culture media, and 5-FC (0–160 mg/ml) was added concur-rently. After 72 h of incubation, 150 ml of serum-free growth medium was replaced,and 20 ml of MTT solution (5 mg/ml) was added. The cells were then incubated foraround 4 h before a 100 ml of DMSO was added. After gentle agitation for 5 min, theabsorbance at 570 nm of each well was measured using the FLUOstar microplate UVspectrometer.

2.6. In vivo studies

C6 glioma cells were stereotactically implanted into the right caudate nucleus ofmale Wistar rats (250–280 g). Briefly, rats were anesthetized with ketamine (20 mg/kg) and placed in a stereotactic frame (Wood Dale, IL, USA). With a Hamilton syringe,C6 glioma cells (5�105) in 20 ml serum-free DMEM were injected through a burrhole into the right caudate nucleus (3 mm lateral and 1 mm anterior to the bregma,5 mm deep from the dura) over 10 min. The syringe was then retracted over 5 min.The burr hole was filled with bone wax.

Five days after the injection, the rats were divided into four groups (9 rats pergroup) for immunocytochemical and survival rate analyses. Different injectionswere carried out on the right striatum of each rat according to grouping. Details arelisted as follows: Group 1: PBS buffer (50 ml); Group 2: 40 mg of pCMVTRAIL /FA-PEG-PEI (50 ml, N/P¼ 15); Group 3: 40 mg of pCMVCD/FA-PEG-PEI (50 ml, N/P¼ 15); Group4: 40 mg of pCMVTRAILþ pCMVCD/FA-PEG-PEI (50 ml, N/P¼ 15); and four normalrats were kept as control. The injection was performed at 5 different points along theneedle tract (6, 5.5, 5, 4.5, 4 mm deep from the dura), over a period of 10 min;afterwards, the needle was left in place for 5 min and was then retracted over 5 min.The burr hole was blocked with bone wax. Two days after injection, 5-FC wasadministrated intraperitoneally at 250 mg/kg/day (the serum 5-FC consentrationmight be 100 mg/ml) for 14 consecutive days [7].

All animals were kept under the same laboratory conditions with no steroids orantibiotics. When symptoms including severe paresis and/or ataxia, or more than20% of body weight were lost, the animal was sacrificed. The total volume of thetumor (cubic millimeters) was calculated by summing up the cross-sectional areas.Survival rate was analyzed by a log-rank test based on the Kaplan–Meier survivalanalysis by using MedCalc statistical software.

2.7. Immunohistochemistry study

Rats were killed and perfused intracranial with warm normal saline (150 ml)followed by 4% paraformaldehyde (150 ml). The brains were removed, and allsections were cut at 3–5 mm thick from buffered formalin-fixed, paraffin-embeddedtissue. After deparaffinization, sections were stained with haematoxylin/eosin(H&E). For histopathological analysis, at least five paraffin sections from each animalwere used for hematoxylin/eosin staining. Immunohistochemisty studies wereperformed in serial sections obtained from the paraffin blocks which were meant forhistological diagnosis. They were immunostained with CD and TRAIL monoclonalantibodies (dilution 1:500). The one step Envision polymer (Dako A/S, Glostrup,Denmark) was used as a secondary link to DAB chromogen. Prior to the applicationof the primary antibody, antigen retrieval was carried out by three-minute incu-bation with a pressure boiler in citrate buffer solution (pH 6.0).

5-FC Concen

1

0.8

0.6

0.4

0.2

00 10 20

Cell viab

ility

Fig. 1. The cell viability of the C6 glioma cells changed with various 5-FC concentrations (mePEI/pCMVTRAIL complexes.

2.8. Statistical analysis

All data were analyzed with SPSS 13.0. The results are expressed as mean� SEand the statistical significance was defined as P< 0.05.

3. Results

3.1. Combination of CD/5-FC and TRAIL enhanced cytotoxicityin vitro

The antitumor effect of the combined treatment using CD/5-FCand TRAIL was evaluated at various 5-FC concentrations. In thestudy, C6 cells were transfected with FA-PEG-PEI/pCMVCD alone orin combination with FA-PEG-PEI/pCMVTRAIL with various 5-FCconcentrations. As illustrated in Fig. 1, the viability of C6 cells was94.2% when treated with FA-PEG-PEI/pCMVCD alone (i.e. without5-FC). Such cell viability was similar to that when cells weretransfected with FA-PEG-PEI/pEGFP. Cell viability was significantlylower (50.2%) for samples treated with the combined FA-PEG-PEI/pCMVTRAIL agent. At a 5-FC concentration of 80 mg/ml, MTT assayrevealed a 42.5% cell viability for samples treated with FA-PEG-PEI/pCMVCD alone while 23.4% for the combined treatment. Based onthe data, the combined treatment using CD/5-FC with TRAILproduced significantly higher cytotoxicity than the case when theywere used as a single agent. When the 5-FC concentration reached80 mg/ml or higher (i.e. 160 mg/ml), no significant difference wasfound between the cytotoxicity induced in different treatmentgroups. High cell viability constantly remained in the control cells,which were transfected with FA-PEG-PEI/pEGFP, despite any vari-ation in the 5-FC concentration. In general, the level of cytotoxicityof the 5-FC was neglectable in the controls.

3.2. The expression of the CD, TRAIL and activated caspase-3protein in FA-PEG-PEI/pDNA complexes transfected C6 glioma cells

In this study, C6 glioma cells were transfected with FA-PEG-PEI/pCMVCD complexes at an N/P ratio of 15, and similar procedureswere carried out for FA-PEG-PEI/pCMVCD. The expression of theCD, TRAIL and activated caspase-3 protein was examined usingWestern blotting. In Fig. 2, the expression of CD and TRAIL proteinsas well as the caspase-3 activation in C6 cells after FA-PEG-PEI/pCMVCD and FA-PEG-PEI/pCMVTRAIL transfection is shown. Therewere obvious protein bands in the transfected cells and the acti-vation of caspase-3 confirmed that the CD and TRAIL proteins hadbeen expressed and activated.

trations (ug/ml)

40 80 160

EGFPCDC/T

an� SD, n¼ 4) after transfection of FA-PEG-PEI/pEGFP, FA-PEG-PEI/pCMVCD or FA-PEG-

Caspase-3

actin

CDControl C/TTRAIL

TRAIL

CD

TRAILControl C/TCD

actin

Fig. 2. Western blot analysis of the expression of CD, TRAIL, and the caspase-3 acti-vation in the C6 glioma cells after FA-PEG-PEI/pCMVCD or pCMVTRAIL transfection.CD: cytosine deaminase; TRAIL: TNF-related apoptosis-inducing ligand; C/T: CD/TRAIL.

10 20 30 40 50 60 70 80

100

80

60

40

20

0

Time(Day)

Su

rv

iv

al p

ro

ba

bility

(%

)

Control (PBS)TRAILCD/5-FCCD/TRAIL

Fig. 4. The survival probability of the rats in different groups after C6 glioma cellsimplantation. They include the PS-control, CD/5-FC single therapy, TRAIL singletherapy, and combined therapy (i.e. CD/5-FC with TRAIL) groups.

B. Liang et al. / Biomaterials 30 (2009) 4014–4020 4017

3.3. Effect of combined therapy of CD/5-FC with TRAIL genes on C6glioma cells in vivo

The rat C6 glioma models were employed in our study forexamining the combined treatment effect in vivo. Three weeks afterthe initial injection, all animals in the control group died due toexcessive tumor burden. The average tumor size for the PBS-controlgroups was 172.52� 8.02 mm3, while 53.13� 3.72 mm3 noted forthe combined therapy (i.e. CD/5-FC and TRAIL) group. For thegroups treated with CD/5-FC or TRAIL single gene, the averagetumor sizes were 90.37� 5.01 mm3 and 87.24�7.28 mm3,respectively (Fig. 3). According to our results, the combined therapycould significantly diminish the glioma tumor size when comparedwith the controls or single gene therapy (p< 0.01).

In this study, the survival time course of the tumor-bearing ratswas also recorded. As seen in Fig. 4, all Wistar rats in the controlgroups died before day 20. Among the other three treatmentgroups, rats belonged to the combined therapy group demon-strated a longer survival time course when compared with the CD/5-FC and TRAIL single therapy groups (p< 0.01). No significantstatistical difference was found between the two single therapygroups (p> 0.05). On the 35th day after C6 glioma cells’ implan-tation, about 2/3 of the CD/5-FC and TRAIL single treated ratssurvived but the rest died in the next 15 days. For the combined

Tu

mo

r S

ize (m

m3)

Groups

p < 0.01

p < 0.01

p < 0.01

250

200

150

100

50

0Control (PBS) CD/5-FC TRAIL CD/TRAIL

Fig. 3. The average tumor size in different groups after C6 glioma cells’ implantation.They include the PS-control, EGFP-control, CD/5-FC single therapy, TRAIL singletherapy, and combined therapy (i.e. CD/5-FC with TRAIL) groups.

treatment group, 80% of the animals survived until day 35 and 2 ofthem were still alive after 80 days of implantation.

During necropsy, tumors excised from the control group werefairly large in size, and hypervascularized with petechia and centralnecrosis. As shown in Fig. 5B, the histopathological analysis oftumors excised from the control showed generally larger tumors(6–8 mm) with scattered necrosis, haemorrhage with cerebraledema (data not shown). Microscopic view showed that they weremostly hypercellular with nuclear pleomorphism and scatteredhaemorrhage (Fig. 6a(B)). Gliomas isolated from the combinedtherapy group were generally small and pale with fewer visiblesuperficial blood vessels (Fig. 6a(E)). For rats that had a survivaltime course longer than 60 days (i.e. from the combined therapygroup), the tumors were tiny like nodules in the brains. For thosetwo rats that survived up to 80 days, they were found tumor-freewhilst microscopic examination disclosed cyst formation withmacrophage infiltration (Fig. 6a(E)). These indicated that thecombined application of CD/5-FC and TRAIL was effective in sup-pressing and treating glioma growth in vivo.

Immunohistochemisty studies were performed to confirm theexpression of the CD and TRAIL proteins in the rat brains. Negativecontrol is shown in Fig. 6b(A) and the cells were stained blue as wellas in the PBS-control group. Marked membranous CD and TRAILdistribution with nuclear immunostaining was partly noted in theCD/5-FC and TRAIL single treated rats (Fig. 6b(C, D). Diffuse, cyto-plasmic and focally membranous distribution of CD and TRAILimmunoreactivity was revealed by DAB immunostaining, as seen inrats from the combined therapy group (Fig. 6b(E)).

4. Discussion

Nowadays, no treatment is yet found to be effective in treatingmalignant glioma. The use of conventional (i.e. chemotherapeutic)agents is often limited by the systemic toxicity. Surgical removal oftumor mass together with postoperative radiotherapy and/orchemotherapy is the most common strategy applied in gliomapatients, but the survival rate is usually less than 2 years after diag-nosis. The recurrent rate is very high, with around 90% of the patientshave tumor recurs within 2 cm from the primary site [21]. Recentprogress in the molecular and cellular biology has promoted genetherapy as a promising treatment strategy in brain tumors [2].Therapeutic vectors which could directly injected into the glioma viathe brain cortex, has became a less risky option when compared withthe conventional surgical approach [22]. This is because lesser

Fig. 5. The brain sections of the C6 glioma cell implanted rat brains. (A) PS-control, (B) CD/5-FC single therapy, (C) TRAIL single therapy, and (D) combined therapy (i.e. CD/5-FC withTRAIL) groups.

Fig. 6. a: Histological characteristics of the C6 gliomas in the C6 glioma cell implanted rat brains (sections were stained with H&E). b: Immunohistological characteristics of the C6gliomas in the C6 glioma cell implanted rat brains (sections were immunostained with CD and TRAIL monoclonal antibodies respectively). (A) Normal rat, (B) PS-control, (C) CD/5-FCsingle therapy, (D) TRAIL single therapy, and (E) combined therapy (i.e. CD/5-FC with TRAIL) groups.

B. Liang et al. / Biomaterials 30 (2009) 4014–40204018

B. Liang et al. / Biomaterials 30 (2009) 4014–4020 4019

normal tissues would be affected in the new approach, while somenormal brain tissues might also be excised during surgical procedure.

The suicide gene/prodrug treatment strategy is originally devel-oped based on the concept of the chemotherapeutic agent-inducedsystemic toxicity. It was postulated that the overall toxicity might beminimized if the tumor is targeted correctly or the surroundingcellular enzymes are capable to convert nontoxic prodrugs to toxicagents. In this study, the CD/5-FC system makes use of the CD genewhich could encode for an enzyme that works to convert 5-FC pro-drug to 5-FU, thus results in tumor cell death and strong bystandereffect [8]. This approach produces intense toxicity exclusively in thevicinity of tumor, which could maximize the therapeutic efficacy.Although such approach has been proposed and developed for years,novel combinations that could work well with other therapies arestill being investigated [9,23,24].

In previous studies, researchers had tried to improve the thera-peutic efficacy of CD/5-FC by combining other genetic agentsincluding IL2 and TRAIL [13,25], in which inspiring results have beenshowed. In the present study, we have tried to enhance the sensitivityof the C6 cells to 5-FC by introducing TRAIL. A number of recentstudies have demonstrated that the combined treatment with TRAILand 5-FU could lead to increased suppression and regression oftumor growth [13,26]. The cross-sensitization between TRAIL and 5-FU could induce apoptotic pathway through caspase activation,which was found to depend on the expression of proapoptotic baxgene [26,27]. In addition, researchers have demonstrated that severalchemotherapeutic agents could induce upregulation of deathreceptor 5 (DR5) expression [28]. Therefore, TRAIL applicationtogether with the CD/5-FC system might be an effective therapeuticstrategy for C6 gliomas.

The development of a highly efficient, safe and cost-effectivevector is the major challenge of gene therapy. Although viral vectorsshowed very high transfection efficiency, their use was limited by therisk of immunogenicity, tumorigenicity and cytotoxicity [29].Nonviral vectors which bear comparatively lower gene transfectionefficiency became a better option for they induce lesser immunereaction and thus, are more cost-effective. Moreover, they have thecapability to deal with large DNA plasmids. They are easy to prepareand flexible to use. Finally, the cell-type specificity which theypossess after chemical conjugation of a targeting ligand are alladvantages over the viral vectors [30]. In our study, the high toxicityof PEI and the low transfection efficiency of PEG-PEI have beenovercome by the production of FA-PEG-PEI for gene delivery. Suchvector has the characteristics of good biocompatibility, potentialbiodegradability, and relatively high gene transfection efficiency.Folate as a targeting ligand was ligated on the vector, since folatereceptors (FR) were generally overexpressed in human cancer cells.Enhanced expression of human FR has also been reported in malig-nant cell lines (>20-fold), in human brain tumors, and in lung tumors.While in normal human tissues, FR has a limited expression mainlyfound in kidney, lung, choroid plexus, and placenta [17].

Initially, transfection efficacy of FA-PEG-PEI mediated delivery ofpCMVCD and pCMVTRAIL in C6 glioma cells was compared (at an N/Pratio ranged from 5 to 30; data not shown). An N/P ratio of 15 wasfound to be the best ratio of which an effective killing of C6 cells couldbe carried out. This agreed with our previous study that carried out indifferent cell lines including HEK 293T, glioma C6 and hepatomaHepG2 cells [19]. The advantages of 5-FC/CD gene therapy had beendescribed. First, the mechanism of the antitumor effect was inde-pendent of the cell cycle. Second, the prodrug 5-FC was highlypermeable and readily crossed the BBB. Cerebrospinal fluid distri-bution from the blood sera of 5-FC was w60–80%. Drug delivery tothe tumor was one of the most important factors for successfulchemotherapy and could be the largest factor limiting clinicalapplications. Third, the bystander effect of 5-FC/CD gene therapy was

strong and did not require direct cell-to-cell contact [7]. Theconcentration of 5-FC was found to be a crucial factor for the CD/5-FCsystem to sustain effective killing of glioma cells. Therefore, the C6cells were transfected with FA-PEG-PEI/pCMVCD alone at an N/Pratio of 15 whilst the 5-FC concentration was varied accordingly. Ourresults indicate that the optimal concentration of 5-FC for maximumcytotoxicity to generate was 80 mg/ml, which agreed with previousreport [7]. The combined transfection of FA-PEG-PEI/pCMVCD withFA-PEG-PEI/pCMVTRAIL at the N/P ratio of 15 was also tested againstvarious 5-FC concentrations. As seen in Fig. 1, combined therapygenerated additive cytotoxic effect in the C6 cells than when theagent was applied alone in vitro. The dose of i.p. 5-FC for in vivo genetherapy ranges from 250 to 500 mg/kg/day, and which was about100–200 mg/ml the achieved dose of 5-FC in the serum of injectedanimals [7].

In order to investigate the underlying mechanism of the additivecytotoxic effect, the expression of CD and TRAIL proteins wereanalyzed using Western blotting. The level of activation of caspase-3 was also examined, for it is a good indicator of apoptotic level inmammalian cells. In a recent study, TRAIL was found to induceapoptosis via interaction with DR5, which was mediated by cas-pase-3-initiated pathways in sensitive glioma cells [31]. In ourstudy, both the CD and TRAIL proteins have demonstrated theirantitumor efficacy through the apoptotic process as shown in Fig. 2.The synergistic antitumor effects performed by the combined CD/5-FC and TRAIL proteins/prodrug might partly, if not all, wererelated to the activation of apoptotic process in the glioma samples.

On the other hand, in vivo study has demonstrated a significantdelay in human tumor growth with the combined gene therapeuticapproach. The results were relatively correlated with the in vitrocondition, which indicate a delay in tumor cell proliferation (i.e.MTT assay) and acceleration in apoptosis. The combined FA-PEG-PEI/pCMVCD/5-FC and pCMVTRAIL treatment has revealed signif-icant tumor suppression in Wistar rats by decreasing the tumorvolume and increasing the survival rate of the animal models.These results suggested that the combined treatment strategyusing FA-PEG-PEI/pCMVCD/5-FC and FA-PEG-PEI/pCMVTRAIL wasan effective strategy that deserved further investigation in bothin vitro and in vivo conditions.

5. Conclusion

The present study demonstrated an effective antitumor effectinduced by CD gene transfected TRAIL gene, which was mediatedby the FA-PEG-PEI. Combined treatment with CD/5-FC and TRAILhad synergistic inhibitive ability, when compared with the single-gene therapy with either CD or TRAIL gene. Our data suggested thatan effective antitumor efficacy might be achieved by using FA-PEG-PEI vector bearing the CD and TRAIL genes. This combined treat-ment might be a treatment strategy that would help promote thesurvival rate of glioma patients. Nevertheless, further investiga-tions and more in vivo studies would be required to confirm itsoptimal dosage and safety in clinical application.

Acknowledgements

This research work was supported by the 863 Programs of China(2007AA021101 to YP) and the National Science Foundation ofChina (NSFC) (30672411 to YP and 50673103 and 50830107 to XS),the Ph.D. Programs Foundation of Ministry of Education of China(20050558084 to YP and 20060558083 to XS) and Research GrantCouncil (RGC) of Hong Kong (CUHK7394/04M, to MLH). Thanks toDr. W. Walther (Max-Delbruck-Center for Molecular Medicine,Berlin, Germany) for kindly providing the plasmids pCMVCD. Thereare no potential conflicts of interest.

B. Liang et al. / Biomaterials 30 (2009) 4014–40204020

References

[1] Levy ML, Tung H, Couldwell WT, Hinton DR, Apuzzo ML. Neurosurgery,molecular medicine, and the pandora-panacea continuum: future implica-tions for glioma therapy? Clin Neurosurg 1992;39:421–62.

[2] Avgeropoulos NG, Batchelor TT. New treatment strategies for malignantgliomas. Oncologist 1999;4:209–24.

[3] Niranjan A, Moriuchi S, Lunsford LD, Kondziolka D, Flickinger JC, Fellows W,et al. Effective treatment of experimental glioblastoma by HSV vector-medi-ated TNF alpha and HSV-tk gene transfer in combination with radiosurgeryand ganciclovir administration. Mol Ther 2000;2:114–20.

[4] Benns JM, Mahato RI, Kim SW. Optimization of factors influencing the trans-fection efficiency of folate-PEG-folate-graft-polyethylenimine. J Control Release2002;79:255–69.

[5] Marconi P, Tamura M, Moriuchi S, Krisky DM, Niranjan A, Goins WF, et al.Connexin 43-enhanced suicide gene therapy using herpesviral vectors. MolTher 2000;1:71–81.

[6] Moriuchi S, Oligino T, Krisky D, Marconi P, Fink D, Cohen J, et al. Enhancedtumor cell killing in the presence of ganciclovir by herpes simplex virus type 1vector-directed coexpression of human tumor necrosis factor-alpha andherpes simplex virus thymidine kinase. Cancer Res 1998;58:5731–7.

[7] Ichikawa T, Tamiya T, Adachi Y, Ono Y, Matsumoto K, Furuta T, et al. In vivoefficacy and toxicity of 5-fluorocytosine/cytosine deaminase gene therapy formalignant gliomas mediated by adenovirus. Cancer Gene Ther 2000;7:74–82.

[8] Lawrence TS, Rehemtulla A, Ng EY, Wilson M, Trosko JE, Stetson PL. Prefer-ential cytotoxicity of cells transduced with cytosine deaminase comparedto bystander cells after treatment with 5-flucytosine. Cancer Res 1998;58:2588–93.

[9] Stackhouse MA, Pederson LC, Grizzle WE, Curiel DT, Gebert J, Haack K, et al.Fractionated radiation therapy in combination with adenoviral delivery of thecytosine deaminase gene and 5-fluorocytosine enhances cytotoxic and anti-tumor effects in human colorectal and cholangiocarcinoma models. Gene Ther2000;7:1019–26.

[10] Pederson LC, Buchsbaum DJ, Vickers SM, Kancharla SR, Mayo MS, Curiel DT,et al. Molecular chemotherapy combined with radiation therapy enhanceskilling of cholangiocarcinoma cells in vitro and in vivo. Cancer Res 1997;57:4325–32.

[11] Stegman LD, Rehemtulla A, Hamstra DA, Rice DJ, Jonas SJ, Stout KL, et al.Diffusion MRI detects early events in the response of a glioma model to theyeast cytosine deaminase gene therapy strategy. Gene Ther 2000;7:1005–10.

[12] Chen JK, Hu LJ, Wang D, Lamborn KR, Deen DF. Cytosine deaminase/5-fluo-rocytosine exposure induces bystander and radiosensitization effects in hypoxicglioblastoma cells in vitro. Int J Radiat Oncol Biol Phys 2007;67:1538–47.

[13] Kaliberov SA, Chiz S, Kaliberova LN, Krendelchtchikova V, Della Manna D,Zhou T, et al. Combination of cytosine deaminase suicide gene expression withDR5 antibody treatment increases cancer cell cytotoxicity. Cancer Gene Ther2006;13:203–14.

[14] Naumann U, Waltereit R, Schulz JB, Weller M. Adenoviral (full-length) Apo2L/TRAIL gene transfer is an ineffective treatment strategy for malignant glioma.J Neurooncol 2003;61:7–15.

[15] Chul Cho K, Hoon Jeong J, Jung Chung H, Joe CO, Wan Kim S, Gwan Park T.Folate receptor-mediated intracellular delivery of recombinant caspase-3 forinducing apoptosis. J Control Release 2005;108:121–31.

[16] Godbey WT, Wu KK, Mikos AG. Poly(ethylenimine) and its role in genedelivery. J Control Release 1999;60:149–60.

[17] Gabizon A, Horowitz AT, Goren D, Tzemach D, Mandelbaum-Shavit F,Qazen MM, et al. Targeting folate receptor with folate linked to extremities ofpoly(ethylene glycol)-grafted liposomes: in vitro studies. Bioconjug Chem1999;10:289–98.

[18] Gottschalk S, Cristiano RJ, Smith LC, Woo SL. Folate receptor mediated DNAdelivery into tumor cells: potosomal disruption results in enhanced geneexpression. Gene Ther 1994;1:185–91.

[19] Liang B, He ML, Xiao ZP, Li Y, Chan CY, Kung HF, et al. Synthesis and charac-terization of folate-PEG-grafted-hyperbranched-PEI for tumor-targeted genedelivery. Biochem Biophys Res Commun 2008;367:874–80.

[20] Cammas S, Nagasaki Y, Kataoka K. Heterobifunctional poly(ethylene oxide):synthesis of a-methoxy-x-amino and ahydroxy-x-amino PEOs with the samemolecular weights. Bioconjug Chem 1995;6:226–30.

[21] Legler JM, Ries LA, Smith MA, Warren JL, Heineman EF, Kaplan RS, et al. Cancersurveillance series [corrected]: brain and other central nervous systemcancers: recent trends in incidence and mortality. J Natl Cancer Inst 1999;91:1382–90.

[22] Rubinchik S, Yu H, Woraratanadharm J, Voelkel-Johnson C, Norris JS, Dong JY.Enhanced apoptosis of glioma cell lines is achieved by co-delivering FasL-GFPand TRAIL with a complex Ad5 vector. Cancer Gene Ther 2003;10:814–22.

[23] Yazawa K, Fisher WE, Brunicardi FC. Current progress in suicide gene therapyfor cancer. World J Surg 2002;26:783–9.

[24] Patterson AV, Saunders MP, Greco O. Prodrugs in genetic chemoradiotherapy.Curr Pharm Des 2003;9:2131–54.

[25] Colombo F, Barzon L, Franchin E, Pacenti M, Pinna V, Danieli D, et al. CombinedHSV-TK/IL-2 gene therapy in patients with recurrent glioblastoma multi-forme: biological and clinical results. Cancer Gene Ther 2005;12:835–48.

[26] von Haefen C, Gillissen B, Hemmati PG, Wendt J, Guner D, Mrozek A, et al.Multidomain Bcl-2 homolog Bax but not Bak mediates synergistic induction ofapoptosis by TRAIL and 5-FU through the mitochondrial apoptosis pathway.Oncogene 2004;23:8320–32.

[27] Keane MM, Ettenberg SA, Nau MM, Russell EK, Lipkowitz S. Chemotherapyaugments TRAIL-induced apoptosis in breast cell lines. Cancer Res 1999;59:734–41.

[28] Wen J, Ramadevi N, Nguyen D, Perkins C, Worthington E, Bhalla K. Antileu-kemic drugs increase death receptor 5 levels and enhance Apo-2L-inducedapoptosis of human acute leukemia cells. Blood 2000;96:3900–6.

[29] Mulligan RC. The basic science of gene therapy. Science 1993;260:926–32.[30] Fominaya J, Uherek C, Wels W. A chimeric fusion protein containing trans-

forming growth factor-alpha mediates gene transfer via binding to the EGFreceptor. Gene Ther 1998;5:521–30.

[31] Nagane M, Pan G, Weddle JJ, Dixit VM, Cavenee WK, Huang HJ. Increased deathreceptor 5 expression by chemotherapeutic agents in human gliomas causessynergistic cytotoxicity with tumor necrosis factor-related apoptosis-inducingligand in vitro and in vivo. Cancer Res 2000;60:847–53.