(参考)Chitosan-graft-polyethylenimine as a gene carrier n

Chitosan-graft-polyethylenimine as a gene carrierHu-Lin Jiang a ,You-Kyoung Kim a ,Rohidas Arote a ,Jae-Woon Nah b,⁎,Myung-Haing Cho c ,Yun-Jaie Choi a ,Toshihiro Akaike d ,Chong-Su Cho a,⁎aSchool of Agricultural Biotechnology,Seoul National University,Seoul 151-921,South KoreabDepartment of Polymer Science and Engineering,Sunchon 540-742,South KoreacLaboratory of Toxicology,College of Veterinary Medicine,Seoul National University,Seoul 151-742,South KoreadDepartment of Biomolecular Engineering,Tokyo Institute of Technology,Yokohama 226-8501,JapanReceived 20July 2006;accepted 24October 2006Available online 28October 2006AbstractChitosans have been proposed as biocompatible alternative cationic polymers that are suitable for non-viral delivery.However,the transfection efficiency of chitosan-DNA nanoparticles is still very low.To improve transfection efficiency,we prepared chitosan-graft-polyethylenimine (CHI-g-PEI)copolymer by an imine reaction between periodate-oxidized chitosan and polyethylenimine (PEI).The molecular weight and composition of the CHI-g-PEI copolymer were characterized,using multi-angle laser scattering (GPC-MALS)and 1H nuclear magnetic resonance (1H NMR),respectively.The copolymer was complexed with plasmid DNA (pDNA)in various copolymer/DNA (N/P)charge ratios,and the complex was characterized.CHI-g-PEI showed good DNA binding ability and high protection of DNA from nuclease attack.Also,with an increase in charge ratio,the sizes of the CHI-g-PEI/DNA complex showed a tendency to decrease,whereas the zeta potential of the complex showed an increase.The CHI-g-PEI copolymer had low cytotoxicity,compared to PEI 25K from cytotoxicity assays.At high N/P ratios,the CHI-g-PEI/DNA complex showed higher transfection efficiency than PEI 25K in HeLa,293T and HepG2cell lines.Our results indicate that the CHI-g-PEI copolymer has potential as a gene carrier in vitro .©2006Elsevier B.V .All rights reserved.Keywords:Non-viral gene delivery;Chitosan;Polyethylenimine;Chitosan-graft-PEI;Cytotoxicity;Transfection efficiency;Gene therapy1.IntroductionGene therapy refers to the transmission of DNA encoding a therapeutic gene of interest into the targeted cells or organs with consequent expression of the transgene.In the past several years,gene therapy has received significant attention due to its potential application in the replacement of dysfunctional genes and treatment of acquired diseases [1,2].In biomedical research,gene delivery is an essential tool for elucidating gene structure,regulation,and function and is the technological basis for gene therapy [3].While most gene therapy protocols presently inclinical trials employ recombinant viral vectors,safety concerns have led to the pursuit of non-viral alternatives.Viral vectors are mostly constructed from adenovirus,retrovirus,adeno-associ-ated virus,herpes simplex virus,and lentivirus.They are all highly efficient in specific circumstances,but the potential risks of undesired immune and toxic side reactions have raised concerns [4,5].Non-viral vectors can be classified into lipoplex-and polyplex-forming carriers [6],and they have several advantages over viral vectors,including safety,lower immuno-genicity,and the ability to transfer larger DNA molecules [7].Therefore,the research and development of non-viral vectors are a very important undertaking.Chitosans,a family of linear binary polysaccharides,com-prised of beta (1–4)linked 2-amino-2-deoxy-β-D -glucose (GlcN;D-unit)and the N -acetylated analogue (GlcNAc;A-unit),have been proposed as biocompatible alternative cationic polymers that are suitable for non-viral gene delivery [8].However,this system has a significant limitation,owing toitsJournal of Controlled Release 117(2007)273–280⁎Corresponding authors.Cho is to be contacted at School of Agricultural Biotechnology,Seoul National University,Seoul 151-921,South Korea.Tel.:+8228804636;fax:+8228752494.Nah,Department of Polymer Science and Engineering,Sunchon 540-742,South Korea.Tel.:+82617503566;fax:+82617503508.E-mail addresses:jwnah@sunchon.ac.kr (J.-W.Nah),chocs@plaza.snu.ac.kr (C.-S.Cho).G E N E D E L I V E R Y0168-3659/$-see front matter ©2006Elsevier B.V .All rights reserved.doi:10.1016/j.jconrel.2006.10.025low transfection efficiency [9].The transfection efficiency may depend on several factors,such as the chemical structure of polycations,cell type,nanoparticle size and composition,and interactions with cells [10].One of the primary causes of poor gene delivery is the inefficient release from endosomes into the cytoplasm [9].To improve transfection efficiency,several derivatives of chitosan have been designed,based on reactions with free amino groups.Park et al.prepared modified galactosylated chitosan (GC)/DNA complexes,and these systems could efficiently transfect liver cells,expressing asialoglycoprotein receptors,which specifically recognize the galactose ligands on chitosan,although the transfection was still low [11–13].Kim et al.also reported chitosan coupled to urocanic acid (UA),which bears an imidazole ring that can play a crucial role in endosomal rupture through a proton sponge mechanism [14].Wong et al.synthesized PEI-graft-chitosan by performing a cationic polymerization of aziridine to water-soluble chitosan [15].The results indicated that PEI-g-chitosan had a lower cytotoxicity and showed higher transfection efficiency than PEI 25K both in vitro and in vivo .Among non-viral vectors,polyethylenimine (PEI)has been shown to effectively condense plasmids into colloidal particles that can effectively transfect a variety of cells both in vitro and in vivo due to its buffering capacity [16].PEI exists as a branched polymer,as well as in linear form.The branched form of PEI shows a theoretical ratio of primary to secondary to tertiary amine groups of 1:2:1.These amines have p K a values spanning the physiological pH range,resulting in buffering capacity [17].On the other hand,many studies have addressed concerns about the toxicity of conventional PEI.The cytotoxi-city of PEI is dependent on its molecular weight.A lower molecular weight PEI has a lower cytotoxicity [18,19].In this study,we designed the chitosan-graft-PEI (CHI-g-PEI)copolymer by an imine reaction between periodate-oxidized chitosan and low molecular weight PEI to reduce cytotoxicity and enhance the transfection efficiency,a design which has not been reported elsewhere.Physicochemical properties of CHI-g-PEI/DNA complex were analyzed.Cy-totoxicity,transfection efficiency,and serum resistance abil-ity of the CHI-g-PEI/DNA complexes were also investigated in vitro .2.Materials and methods 2.1.MaterialsChitosan (molecular weight,100kDa;deacetylation degree,87.7%)was kindly supplied from Jakwang (Ansung,Korea).Branched PEI 25K,potassium periodate,and DNA (sodium salt from calf thymus)were obtained from Sigma-Aldrich (St.Louis,MO,USA).Branched PEI 1800Da was purchased from Wako (Osaka,Japan).A 5.3kb expression vector,pGL3-control (Promega,Madison,USA),contained a luciferase gene,driven by an SV40promoter and enhancer.pEGFP-N2(4.7kb),encoding a green fluorescent protein,driven by an immediate early promoter of CMV ,was purchased from Clontech Laboratories (Palo Alto,CA,USA).The plasmidswere propagated in Escherichia coli ,extracted by the alkali-lysis technique,and purified by a QIAGENR kit (Chatsworth,CA,USA).The purity of the plasmids (consisting of super-coiled and open circular forms)was checked by electrophoresis on a 1%agarose gel,and the concentration of DNA was determined by measuring UV absorbance at 260nm.DNA from calf thymus was used for the measurement of particle size and zeta potential.2.2.Synthesis of copolymerCHI-g-PEI copolymer was synthesized in two steps.In the first step,periodate-oxidized chitosan was prepared by a modified method of V old and Christensen [20].Briefly,chitosan 100K (0.1M)and potassium periodate (0.01M)were each dissolved in sodium acetate buffer (pH 4.5).Prior to mixing the solutions,each solution was degassed with N 2and adjusted to 4°C.After mixing the solutions,the reaction was performed for 48h and stopped by adding ethylene glycol (10%v/v).Then the solution was dialyzed (Spectra/Por®membrane:MWCO=3500)against NaCl (0.2M,pH 4.5)and,thereafter,against deionized water (pH 4.5).In the second step,PEI 1800(20mmol)was reacted with the periodate-oxidized chitosan solution (10mmol)with magnetic stirring for 2days at 4°C.Subsequently,the solution was treated with sodium borohydride (2g NaBH 4/g chitosan),dialyzed against NaCl (0.2M,pH 4.5)(MWCO=12–14K),and thereafter,against deionized water (pH 4.5)at 4°C to remove unreacted PEI.After dialysis,the copolymer was lyophilized.The reaction scheme is shown in Fig.1.2.3.Characterization of copolymerThe composition of the prepared CHI-g-PEI copolymer was estimated by measuring 1H nuclear magnetic resonance (1H NMR)(Avance ™600,Bruker,Germany).The molecular weight of the CHI-g-PEI copolymer was measured,using a gel permeation chromatography column with multi-angle laser scattering (GPC-MALS)at 690nm laser wavelength (Dawn Eos,Wyatt,USA).The column temperature was maintained at 25°C.The flow rate was 0.5ml/min,and the mobile phase was 0.5M ammonium acetate (pH5.5).Fig.1.Proposed reaction scheme for synthesis of CHI-g-PEI.G E N E D E L I V E R Y274H.-L.Jiang et al./Journal of Controlled Release 117(2007)273–2802.4.Cell lines,cell culture and cell viability assaysHeLa(human cervix epithelial carcinoma cells),293T (human kidney cells),and HepG2(human hepatoblastoma cells)were cultured in Dulbecco's modified Eagle medium (DMEM,Gibco BRL,Paris,France),supplemented with10% fetal bovine serum(FBS,HyClone,Logan,Utah),streptomycin at100μg/ml,and penicillin at100U/ml.All cells were incubated at37°C in humidified5%CO2atmosphere.Cells were split,using trypsin/EDTA medium when almost confluent.In vitro cytotoxicity tests were evaluated by Cell Titer96®A queous One Solution Cell Proliferation Assay(Promega).Cells were seeded in96-well plates at an initial density of1×104cells/ well in0.2ml of growth medium and incubated for24h prior to the addition of filtered polymers.Growth media were replaced by fresh,serum-free media,containing various amounts of polymers. After an additional incubation for24h,the media were changed with growth media containing20μl of Cell Titer96®A queous One Solution Reagent.Finally,after further incubation for3h,the absorbance was measured at570nm,using an ELISA plate reader (GLR1000,Genelabs Diagnostics,Singapore)to measure the metabolic activity of the cells.Cell viability(%)was cal-culated according to the following equation:cell viability(%)= (OD570(sample)/OD570(control))×100,where OD570(sample)represents a measurement from a well treated with polymer and OD570(control) from a well treated with PBS buffer only.2.5.Preparation of CHI-g-PEI/DNA complexesAll CHI-g-PEI/DNA complexes were freshly prepared before use.Charge ratio(N/P)of the CHI-g-PEI/DNA complexes was expressed as the ratio of moles of the amine groups of copolymer to moles of phosphates of DNA.For calculation of N/P ratios,330Da was used as an average mass per charge for DNA.Complexes were prepared by adding copolymer solution to equal volumes of calf thymus DNA(Sigma)(for size and zeta potential measurements),to pGL3-control solution(for lucifer-ase assay),or to pEGFP-N2(for in vitro GFP transfection)with gentle vortexing and incubated at room temperature for30min.2.6.Gel retardation assayTo confirm DNA condensation ability of the copolymer, electrophoresis was plex formation was induced at various N/P ratios from0.1to5,and the final volume with the 6×agarose gel loading dye mixture(Biosesang,Korea)was 12μl.The complexes were loaded onto1%agarose gels with EtBr(0.1μg/ml)and run with Tris-acetate(TAE)buffer at100V for40min.DNA retardation was observed by irradiation with UV light and assayed with Cam2com software.2.7.Observation with energy-filtering transmission electron microscopy(EF-TEM)The morphology of CHI-g-PEI/DNA(pGL3)complex was observed using EF-TEM(LIBRA120,Carl Zeiss,Germany).One drop of CHI-g-PEI/DNA complex was placed on a coppergrid and stained with1%uranyl acetate solution for10s.Thegrid was allowed to dry for10min further and was then examined under the electron microscope.2.8.Measurement of particle size and zeta potentialThe size and surface charge of the CHI-g-PEI/DNA complexwere measured,using an electrophoretic light scattering spectrophotometer(ELS8000,Otsuka Electronics,Osaka, Japan),with90and20scattering angles at25°C,respectively.CHI-g-PEI/DNA complexes were prepared in water at N/Pratios of0.1–35.The volume of each sample was2ml, containing a final DNA concentration of40μg/ml.2.9.Protection and release assay of DNAProtection and release of DNA in complexes were carriedout by electrophoresis,according to the modified method ofPark et al.[21].Briefly,1μl of DNase I(2units)or PBS in DNase/Mg2+digestion buffer(50mM Tris–Cl,pH7.6and10mM MgCl2)was added to4μl of polyplex solution or to0.1μg of naked plasmid DNA(N/P ratio:5),and incubated at37°C with shaking at100rpm for30min.For DNase inactivation,all samples were treated with4μl of EDTA(250mM)for10min and mixed with1%sodium dodecyl sulfate(SDS),dissolved in1M NaOH(pH7.2)at a final volume of18μl.The final samples were incubated for2h,and electrophoresis was performed in1%agarose gel with TAE running buffer for1h at50V.2.10.Cell transfection—luciferase activity assayCells were seeded in24-well plates at an initial density of10×104cells/well in1ml of growth medium.After incubationfor18h(to reach70%confluence at the time of transfection),the media were replaced with serum-free or10%serum-containing media with polymer/pGL3-control(1μg)complexesat various N/P ratios and additionally incubated for4h.Thenthe media were exchanged for fresh media,containing serum,and allowed to incubate for24h at37°C.The luciferase assaywas performed according to the manufacturer's protocols. Relative light units(RLUs)were measured with a chemilumin-ometer(Autolumat LB953,EG and G,Berthold,Germany). Protein quantification was determined by the BCA method,andRLUs were normalized to protein concentration in the cell extracts[22].Each transfection experiment was carried out in triplicate,and transfection activity was expressed as relativelight units.2.11.Confocal laser scanning microscopy and flow cytometryThe HeLa cells were plated on collagen-coated glass coverslips in24-well plates(Iwaki Glass Co.,Tokyo,Japan)at5×104cells/well.The cells were incubated for18h.Then the media were replaced with serum-free media,containing polymer/pEGFP-N2(1μg)complex at N/P ratio35.AfterGENEDELIVERY275H.-L.Jiang et al./Journal of Controlled Release117(2007)273–280incubation for 4h,serum-free media were changed with fresh media,containing serum.After 24h incubation,cells were washed in PBS,and the coverslips were mounted on microslides.Cells were directly observed with a confocal laser scanning microscope (MRC-1024,Bio-Rad,UK).For flow cytometry,transfected cells were prepared by the same method as for confocal microscopy.After transfection,cells were washed once with PBS and detached with 0.25%trypsin/EDTA.Transfection efficiency was evaluated by scoring the percentage of cells expressing GFP,using a FACS Calibur System from Becton-Dickinson (San Jose,CA).Fluorescence parameters from 10,000cells were acquired,and transfection was carried out in triplicate.3.Results and discussion3.1.Synthesis and characterization of copolymerWe succeeded in synthesizing CHI-g-PEI copolymer by an imine reaction between periodate-oxidized chitosan and anamine group of PEI (Fig.1).The periodate ion,IO 4−,attacks vicinal diols to cleave the carbon –carbon bond by an oxidation reaction,leading to the formation of a dialdehyde [20].Com-position of synthesized copolymer was analyzed by 1H NMR (Fig.2and Table 1).After oxidation,the proton peak of the 2-carbon of chitosan was decreased (the part shown in dotted line)due to cleavage of the carbon –carbon bond by the periodate ion,leading to the formation of a dialdehyde.In addition to the vicinal diols,other 1,2-dioxygenated groups and 1,2-amino alcohols are also oxidatively cleaved by the periodate.N -acetylation groups in the chitosan,however,are not cleaved [20].The degree of oxidized chitosan determined from integral values from the proton peak of the 2-carbon was about 45mol%(Table 2).The assignment of chemical shifts of PEI was determined by the same method previously reported [21].Asshown in Fig.2,the proton peaks of PEI (–NHCH 2CH 2–)appeared at 3.3–2.5ppm (Table 1),indicating that PEI was grafted to the chitosan chain.The characteristics of synthesized copolymer are shown in Table 2.The molecular weight of CHI-g-PEI (24.99kDa)was similar to the molecular weight of PEI 25K,which was used as a control [23].Wong et al.also synthesized PEI-graft-chitosan using water-soluble chitosan [15].However,in our study,we used commercial chitosan with low water-solubility.After PEI was grafted,CHI-g-PEI was completely water-soluble at physiological pH conditions due to the hydrophilic property of PEI,although chitosan itself is only soluble in acidic conditions [24].3.2.Characterization of CHI-g-PEI/DNA complexesOne prerequisite of a polymeric gene carrier is DNA condensation.The condensation capability of CHI-g-PEI with DNA was evaluated,using agarose gel electrophoresis.Fig.3demonstrates that the migration of DNA was completely retarded when the N/P ratio of CHI-g-PEI/DNA complex was around three.The formation of CHI-g-PEI/DNA complexes was also monitored by observation of the morphology and changes in particle size and zeta potential of the complex as a function of N/P ratio.Fig.4(a)shows a representative morphology of a CHI-g-PEI/DNA complex.The result indicated that the complex had a well-formed spherical shape and compacted structure.Surface properties,such as particle size and surface charge of the complex,are necessary to assure its uptake by cells.In particular,the particle size of a complex is an important factor that influences the access and passage of the complex through the targeting site.As shown in Fig.4(b),all sizes of the complexes were less than 250nm,and the particle sizes of the complex decreased with an increase in the N/P ratio.It is thought that the polymer complexes with DNA throughionicFig.2.Representative 1H NMR spectrum of chitosan-g-PEI in D 2O:δ=3.2–3.1(–CH 2–,2-carbon of chitosan)and 3.3–2.5(–NHCH 2CH 2–,PEI ethylene).Table 1Assigned proton peaks Polymer Assigned peak Chitosan –CH –3.1ppm PEI–NHCH 2CH 2–3.3–2.5ppmTable 2Characteristics of prepared chitosan-g-PEI Initial MW (kDa)Degree of oxidation (mol%)MW of periodate-oxidized chitosan (kDa)MW of chitosan-g-PEI (kDa)Degree of grafted PEI (mol%)1004513.425.07.9Fig. 3.Agarose gel electrophoresis of copolymer/DNA (pGL3-control)complexes at various N/P ratios.G E N E D E L I V E R Y276H.-L.Jiang et al./Journal of Controlled Release 117(2007)273–280interactions,and at high N/P ratios,there are net electrostatic repulsive forces to prevent aggregation among complexes.The relatively homogenous size distributions of complexes,mea-sured by dynamic light scattering,were unimodal [Fig.4(c)].Upon self-assembly of DNA and cationic polymer,the highly negative charge of DNA is neutralized rapidly,and the surface charge of the complex becomes positive at higher N/P ratios.A positive surface charge of polyplexes is necessary for binding to anionic cell surfaces,which consequently facilitates uptake by the cell [19,25].Fig.4(b)shows the zeta potentials of the complexes at various N/P ratios.At N/P ratio 0.1,where thecomplex could not form completely,the zeta potential of the copolymer/DNA complex was negative.With increasing N/P ratios,the zeta potential of the complex rapidly increased to positive values.For effective gene expression,the DNA in the gene vehicle should be protected from degradation by enzymes [21].As shown in Fig.5,in contrast to the control naked plasmid DNA,DNA in the complex was protected from DNase I.These results suggested that more integral DNA could transfer to cells without degradation byenzymes.Fig.4.EF-TEM images of CHI-g-PEI/DNA complex at N/P ratio 35(a),particle sizes and surface charges of copolymer/DNA complexes at various N/P ratios (b)and size distribution of complex prepared at N/P ratio 35(c)(n =3,error bars represent standarddeviation).Fig.5.Protection and release assay of DNA.DNA was released by adding 1%SDS to the copolymer/DNA complex at N/P ratio 5.(1)DNA marker,(2)plasmid DNA (pGL3-control)alone,(3)plasmid DNA treated with DNase I,(4)CHI-g-PEI/DNA complex and (5)CHI-g-PEI/DNA complex treated with DNaseI.Fig.6.Cytotoxicity of copolymer at various concentrations in different cell lines.(a)293T,(b)HeLa and (c)HepG2(n =3,error bars represent standard deviation).G E N E D E L I V E R Y277H.-L.Jiang et al./Journal of Controlled Release 117(2007)273–2803.3.Cytotoxicity of CHI-g-PEITo investigate the cytotoxicity of CHI-g-PEI copolymer,cell viability was determined by the Cell Titer96®AQ ueous One Solution Cell Proliferation Assay at various concentrations of copolymer.Fig.6shows that the copolymer has low cytotoxicity,compared to PEI 25K,in three different cell lines.The cytotoxicity of cationic polymers is probably caused by polymer aggregation on cell surfaces,impairing importantmembrane functions.Also,the cationic polymers may interfere with critical intracellular processes of cells:in particular,the primary amine was reported to disrupt PKC function through disturbance of protein kinase activity [15,26].It has been reported that high molecular weight PEI is significantly more toxic than low molecular weight PEI [18,19,27].Our results also showed that,by increasing the molecular weight of grafted PEI,the cell viability decreased (figure not shown).Further-more,cell viabilities were significantly higher in all cell lines exposured to CHI-g-PEI (1800)(M.W.of PEI:1800Da)copolymer than to CHI-g-PEI (10K and 25K)(M.W.of PEI:10and 25K)at a concentration of 20μg/ml (figure not shown).3.4.Cell transfection —luciferase activity assayTo investigate CHI-g-PEI copolymer transfection efficien-cy,we performed luciferase activity assays in vitro ,using three different cell lines (293T,HeLa,and HepG2),as shown in Fig.7.Transfection efficiency of the copolymer was dependent on the cell line,and the highest transfection efficiency was observed in the 293T cell line.Transfection efficiency of the copolymer increased with increasing N/P ratios,whereas the transfection efficiency of PEI 25K decreased due to cytotoxicity.Interestingly,higher transfec-tion efficiency was observed for the copolymer than for PEI 25K at a high N/P ratio of 35,which is similar to what Wong et al.have reported [15].They attributed the higher transfection efficiency of PEI-g-chitosan to higher amine content in the complex.The higher amine content and an amine composition in the copolymer similar to PEI allow the complexes to escape the endosome easily due to a higher buffering capacity.In addition,the copolymer may have a weaker condensation capability than PEI 25K,which would result in an easy release of DNA from the complexes after passing through the cell membrane.This also would facilitate the higher transfection efficiency of DNA.It is reasonable to assume that,withFig.8.Effect of bafilomycin A1on gene transfection.Copolymer/DNA complex at N/P ratio 35in HeLa cell line.Bafilomycin A1(200nM)diluted in DMSO were put into wells.After a 10min incubation period,transfection solutions were added into the wells for 4h.Then the cells were incubated in the growth medium for 24h.Fig.7.Transfection efficiency of copolymer/DNA (pGL3-control)complex at various N/P ratios and in various cell lines.(a)293T,(b)HeLa and (c)HepG2(n =3,error bars represent standard deviation).G E N E D E L I V E R Y278H.-L.Jiang et al./Journal of Controlled Release 117(2007)273–280increasing N/P ratios,the buffering capacity of CHI-g-PEI/DNA complex increased,whereas,in the case of the PEI (25K)/DNA complex,not only the buffering capacity,but also its cytotoxicity,was increased.Also,interestingly,the CHI-g-PEI/DNA complexes had almost similar transfection efficien-cy as lipofectamine in the 293T cell line.To further elucidate the mechanism of transfection,we checked the buffering capacity of the copolymer.To determine the buffering capacity of CHI-g-PEI,cells were treated with bafilomycin A1during transfection.As a specific inhibitor of vacuolar type H +ATPase,bafilomycin A1has been reported to decrease PEI-mediated transfection by inhibiting the endo-/lysosomal proton pump [28].Transfection of CHI-g-PEI/DNA com-plexes was drastically decreased after bafilomycin treatment,as shown in Fig.8.The fact that the efficiency of CHI-g-PEI transfection declined in the presence of bafilomycin suggests the involvement of the proton sponge effect in PEI-mediated transfection.On the other hand,when we checked the effect of bafilomycin on the transfection of the chitosan/DNA complex into the HeLa cell line,the efficiency was not significantly decreased in the presence of bafilomycin.Also,we checked expression of GFP by confocal laser scanning microscopy and FACS (Fig.9).Fig.9(a)shows confocal images of GFP expressed in the HeLa cell line transfected with PEI 25K/DNA (N/P ratio:7)and CHI-g-PEI/DNA (N/P ratio:35).More GFP expression was observed with the CHI-g-PEI/DNA complexes,compared with the PEI 25K/DNA complexes,indicating that CHI-g-PEI is more effective as a gene carrier.Furthermore,the transfection efficiency was checked by viewing GFP expression,using FACS [Fig.9(b)].The transfection efficiency of CHI-g-PEI was higher than that of PEI 25K,which confirms the confocal [Fig.9(a)]and luciferase assay (Fig.7)results.It has been reported that one of the practical problems for in vivo gene delivery,mediated by cationic liposomes,is that gene expression is inhibited by serum [29].The development of gene delivery systems that are stable even in serum is very important for the improvement of gene therapy by non-viral vectors [30].We investigated the effect of serum on the transfection efficiency of the CHI-g-PEI/DNA complexes.As shown in Fig.10,the transfection efficiency of the CHI-g-PEI/DNA complexes was slightly decreased in the presence of serum,whereas the transfection efficiencies of PEI/DNA and lipofectamine/DNA complexes were drastically decreased.It was already reported that the transfection efficiency of the chitosan/plasmid complex did not change substantially in the presence of serum [31].Therefore,it is thought that the minimal change in transfection efficiency of the CHI-g-PEI/DNA complexes in the presence of serum is due to the chitosan.The above highlights make CHI-g-PEI more attractive for further characterization.We are focusing currently on more comprehensive studies to characterize CHI-g-PEI,especially lung cancer therapy by aerosoladministration.Fig.9.GFP expressed in a HeLa cell line transfected with copolymer/DNA complex.(a)Confocal laser scanning microscopy (scale bar=50μm).(b)FACS analysis (n =3,error bars represent standarddeviation).Fig.10.Effect of serum on gene transfection efficiency.HeLa cells were incubated in the absence or presence of 10%serum with copolymer/DNA (pGL3-control)complex (n =3,error bars represent standard deviation).G E N E D E L I V E R Y279H.-L.Jiang et al./Journal of Controlled Release 117(2007)273–280。