lable at ScienceDirect

Biomaterials 31 (2010) 4194–4203

Contents lists avai

Biomaterials

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

Polycation/DNA complexes coated with oligonucleotides for gene delivery

Yi-Chen Chung a, Wen-Yuan Hsieh b, Tai-Horng Young a,c,*

a Institute of Polymer Science and Engineering, National Taiwan University, Taipei 106, Taiwanb Material and Chemical Research Laboratories, Industrial Technology Research Institute, Hsin-Ju 300, Taiwanc Institute of Biomedical Engineering, College of Medicine and College of Engineering, National Taiwan University, Taipei 100, Taiwan

a r t i c l e i n f o

Article history:Received 8 December 2009Accepted 19 January 2010Available online 16 February 2010

Keywords:Ternary nanoparticlesPolyallylamineHistidineOligonucleotidesGene delivery

* Corresponding author. Institute of Biomedical EngNational Taiwan University, Taipei 100, Taiwan. Tel.: þþ886 2 23940049.

E-mail address: thyoung@ntu.edu.tw (T.-H. Young

0142-9612/$ – see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.biomaterials.2010.01.116

a b s t r a c t

Ternary nanoparticles with negatively charged surface were prepared by coating single-strandedoligonucleotides (50–C10A20–30) on histidine-conjugated polyallylamine (PAA-HIS)/DNA complexes forgene delivery. Characterization of PAA-HIS/DNA/oligonucleotide complexes demonstrated that nano-particles possessed the negative surface charge �27 mV and size of around 100 nm when the molar ratioof oligonucleotide/PAA-HIS exceeded 1.5. The negatively charged oligonucleotide-coated PAA-HIS/DNAcomplexes could be entirely internalized by the living HeLa cells to exhibit high gene expression withlow cytotoxicity and the resistance against erythrocyte agglutination and serum inhibition. Since thegene expression of PAA-HIS/DNA complexes was significantly inhibited by coating other polyanions andoligonucleotides, the ternary PAA-HIS/DNA/deoxyadenosine-rich oligonucleotide complexes wereuptaken by specific receptor-mediated process. Additionally, the deposition of a layer of oligonucleotidesonto the binary PAA-HIS/DNA complexes could effectively transfect various types of cells including HEK-293, HepG2 and Hs68 cells, indicating the technology of coating specific oligonucleotides on PAA-HIS/DNA complexes or other cationic binary DNA complexes might facilitate the use of nanoparticles for safeand efficient gene delivery and eventual therapy.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Gene therapy offers a potential method to cure inherited oracquired diseases by transferring exogenous nucleic acids into cellsto alter protein expression profiles. Although recombinant virusesare currently used for this purpose because of their high trans-fection activity, the use for nonviral vectors such as cationic poly-mers is still desired from the viewpoint of safety. A variety ofpolyamines have been widely investigated in DNA delivery such aspolyethylenimine (PEI), poly(L-lysine) (PLL), poly (2-dimethylamino ethyl) methacrylate (PDMAEMA) and so on [1,2]. Polyallyl-amine (PAA), also a synthetic cationic polymer, possesses highdensity of primary amino groups which are suitable for binding andpackaging negatively charged DNA. However, the utility of PAA forgene delivery application is limited by its cytotoxicity of too strongpolycationic character [3]. Different chemical modifications havebeen used to decrease cytotoxicity and enhance transfection effi-ciency of PAA. For example, Boussif et al. showed glycolated PAA notonly decreased the cytotoxicity of PAA/DNA complex but alsoincreased the transfection efficiency of PAA/DNA complex [4].

ineering, College of Medicine,886 2 23123456x81455; fax:

).

ll rights reserved.

Pathak et al. demonstrated that PAA modified with imidazole andpoly ethylene glycol (PEG)-bis (phosphate) could reduce the posi-tive surface charge to decrease cytotoxicity and to achieveenhanced transfection efficiency [3]. In addition, Nimesh et al.showed PAA complexed with dextran and DNA could simulta-neously improve upon transfection efficiency and cell viability [5].

On the other hand, to achieve desired therapeutic goals, efficientdelivery of a therapeutic gene to target cells is an important issue.Cationic polymers can be readily modified with antibodies, aptam-ers, peptides, or small molecules, which induce endocytosis uponbinding to their cognate cell surface receptors to enhance cellselectivity and delivery [6–9]. To this end, the technique DNAhybridization was used to prepare nanoparticles containing tumor-targeting antibodies in our laboratory. At first, oligonucleotides werecoated onto the cationic polyplexes led to the surface incorporationof negatively charged single-stranded DNA. The cationic polymersused were PAA and histidine-conjugated PAA (PAA-HIS). Subse-quently, we employed complementary oligonucleotides to conju-gate with specific antibodies, and assembled them witholigonucleotide-coated polyplexes into multilayered nanoparticlesby nucleic acid–base-paired interaction, which allowed for thesurface conjugation of biomolecules to mediate internalization ina cell-specific manner. Generally, the uptake of polycation/DNAcomplexes is mediated by electrostatic interactions of positivelycharged complexes with negatively charged cellular surface, so the

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–4203 4195

anionic oligonucleotide-coated polyplexes should show minimalcellular transfection if specific targeting ligands were not exposed onthe multilayered nanoparticle surfaces. Accidentally, we discoveredthat the negatively charged oligonucleotide-coated PAA-HIS/DNAcomplexes could enter cells to exhibit high gene expression withoutinducing cytotoxicity and agglutination of erythrocytes even theydid not conjugate any specific targeting molecules. Therefore, thepurpose of this work was to study whether the technology of coatingoligonucleotides on PAA-HIS/DNA complexes could facilitate the useof nanoparticles for safe and efficient gene delivery.

Fig. 1. Schematic representation of (a) PAA-HIS, (b) binary poly

2. Materials and methods

2.1. Synthesis of PAA-HIS

All chemicals were reagent grade and used as received unless noted otherwise.PAA-HIS was synthesized according to Fig. 1(a). Briefly, 62.6 mg/mL N-(3-dimethy-laminopropyl)-N0-ethyl-carbodiimide hydrochloride (EDC) (Aldrich, St. Louis, MO)and 24.3 mg/mL N-hydroxysuccinimide (NHS) (Fluka, St. Louis, MO) were added to53.7 mg/mL N-Boc histidine (Fluka, St. Louis, MO) in dimethyl sulfoxide (DMSO)(Aldrich, St. Louis, MO). Reaction was carried out under stirring for 30 min at roomtemperature. EDC/NHS-activated N-Boc histidine was then gently added dropwise toa stirred PAA (Mw¼ 17 kDa, Aldrich, St. Louis, MO) solution (2 mg/mL MilliQ water)

plex and (c) ternary polyplex for the gene delivery system.

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–42034196

(NH2: COOH¼ 1:1). After 24 h, 3 mL trifluoroacetic acid (TFA) was used to removethe Boc groups from N-Boc histidine. Subsequently, the PAA-HIS solution was dia-lyzed against MilliQ water at room temperature using Cellu SepH1 membrane(molecular weight cut off: 3500) for 5 days. The purified PAA-HIS were harvestedafter lyophilization. The chemical structure of PAA-HIS was characterized by ATR-FTIR (UMA 600, Varian, USA) and 1H-NMR (AVANCE-500 MHz FT-NMR, Bruker,Germany). The extent of the substitution of histidine and the buffering capacity ofPAA-HIS were determined by ninhydrin method and acid–base titration, respec-tively [3].

2.2. Preparation of binary complexes

PAA-HIS/DNA complexes were formulated by mixing 20 mg/mL DNA solution(enhanced green fluorescent protein (EGFP) plasmid (4.7 kbp; Clontech Laborato-ries, Palo Alto, CA) or luciferase plasmid (4.1 kbp; Clontech Laboratories, Palo Alto,CA)) with 100 mg/mL PAA-HIS solution in MilliQ water. For cellular uptake assay,plasmids were labeled with YOYO-1 dye (Invitrogen Co., USA) at 0.1 mM, followingthe method reported by Gebhart et al. [10].

2.3. Preparation of oligonucleotide-coated PAA-HIS/DNA complexes

Single-stranded oligonucleotides (Bio Basic Inc., Canada) dissolved in MilliQwater at 100 mM was added to binary DNA complexes at oligonucleotide/(PAA-HIS)molar ratio of 0.5–3 [11]. All oligonucleotides mentioned in this study are 50–C10A20–30 , unless indicated otherwise. The particle size and surface charge of oligonucleo-tide-coated PAA-HIS/DNA complexes were determined by using a Zetasizer (ZS-90,Malvern, UK).

For comparison, poly (styrenesulfonic acid) (PSS)- and poly (acrylic acid) (PAAc)-coated PAA-HIS/DNA complexes were formed by the similar method mentionedabove. Briefly, binary PAA-HIS/DNA complexes were mixed with various amount ofPSS or PAAc solution (100 mM) at polyanion/(PAA-HIS) molar ratio of 0.5–3 to reversethe surface charge of polyplexes.

2.4. Cell line

The model cell line used in this study was a human cervix carcinoma cell line,HeLa cell (BCRC number: 60005). The used medium was Dulbecco’s Modified EagleMedium (DMEM) supplemented with 10% fetal bovine serum (FBS, BiologicalIndustries, Israel). In general, transfection efficiency is affected strongly by the celltype. Therefore, for comparison, various human cell lines, such as HEK-293, HepG2and Hs68 cells, which are derived, respectively, from embryonic kidney, hepatomaand normal fibroblasts, were incubated with nanoparticles under conditions rec-ommended by ATCC.

2.5. Cellular toxicity

Cytotoxicity of oligonucleotide-coated ternary complexes was determined bythe MTT (3-(4,5-dimethylthiazol-2-yl)-diphenyl tetrazolium bromide; Sigma)colorimetric assay. For comparison, commercial nonviral vector, ExGEN 500 (bymixing 20 mg/mL DNA solution with 1.45 mg/mL ExGEN 500 solution) was used ascontrol in this study. HeLa cells were cultured in 12-well tissue culture polystyreneplates (Costar, USA) at a density of 105 cells/well. After incubation for 24 h, theDMEM medium was removed and the cells were incubated with polyplexes con-taining 2 mg plasmid DNA in OPTI-MEM medium. After incubation for another 24 h,the cells were incubated with 200 mL of MTT solution (5 mg/mL) for 3 h at 37 �C.After incubation, the MTT solution was aspirated and the formazan reaction prod-ucts were dissolved in DMSO and shaken for 20 min. The optical density of theformazan solution was read on an ELISA plate reader (SpectraMax M2e, MolecularProbe) at 570 nm. The results are shown as a percentage of untreated cells with 100%viability.

2.6. Transfection efficiency assay

HeLa cells were cultured in 12-well plates at a density of 105 cells/well. Afterincubation for 24 h, the DMEM medium was removed and the cells were incubatedwith PAA-HIS/DNA/oligonucleotide complexes containing 2 mg plasmid DNA inOPTI-MEM medium. After incubation for another 24 h, the EGFP expression wasobserved with fluorescent microscope (IX71, Olympus, Japan).

The luciferase activity assay was performed according to the manufacturer’sprotocol. Briefly, cells were lysed by lysis reagent (Promega, USA) for 20 min andthen centrifuged at 12,000 g at 4 �C for 30 min. The relative luminescence and totalprotein of supernatant were determined using a chemiluminometer (MonolightLuminometer 2010, Becton-Dickinson, San Jose, CA) and Bio-Rad protein assay(USA), respectively. Luciferase activity was indicated as relative light units per mgprotein (RLU/mg protein). For comparison, HEK-293, HepG2 and Hs68 cells werealso used to evaluate transfection efficiency at the same condition mentioned above.

2.7. Cellular uptake

To visualize the uptake of PAA-HIS/DNA/oligonucleotide complexes, HeLa cells(105 cells/well) were incubated with PAA-HIS/DNA/oligonucleotide complexescontaining YOYO-1-labeled DNA for 4 h and then FM4-64FX dye (Invitrogen Co.,USA) was added to mark HeLa cells for 30 min at 25 �C. Furthermore, to studywhether oligonucleotides and cationic polyplexes were separated before ternarycomplexes were uptaken by cells, 1% Nulight (537)–50–C10T20–30 oligonucleotide(ThermoFisher, Germany) was mixed with 50–C10A20–30 oligonucleotides to markoligonucleotide distribution. Cellular uptake was visualized with a confocal scan-ning laser microscope (Leica TCS SP5).

2.8. Erythrocyte agglutination study

To evaluate erythrocyte agglutination induced by oligonucleotide-coatedternary complexes, 2% human erythrocyte solution in phosphate buffered saline(PBS) was mixed with PAA-HIS/DNA/oligonucleotide, PAA-HIS/DNA and ExGEN 500/DNA complexes in polyplex/erythrocyte solution volume ration of 1:1 and incubatedfor 15 min at room temperature. Samples (10 mL) on a glass plate was observed bymeans of optical microscopy.

2.9. The serum inhibition to the transfection efficiency

To evaluate the effect of serum inhibition to the transfection efficiency ofoligonucleotide-coated PAA-HIS/DNA complexes, HeLa cells were incubated withPAA-HIS/DNA/oligonucleotide complexes in the OPTI-MEM medium without or with10% FBS for 24 h. The transfection efficiency was determined as described above.

2.10. Statistical analysis

All data are presented as the mean� standard deviation (SD) of three to sixindependent cultures. Statistical significance was calculated using one-way analysisof variance (ANOVA) followed by Duncan test (p< 0.05 was considered significant).

3. Results

3.1. The physicochemical properties of oligonucleotide-coatedPAA-HIS/DNA complexes

The buffering capacity of gene delivery system has beendemonstrated to improve the transfection efficiency during endo-cytosis, so histidine was conjugated with PAA to improve thebuffering capacity of the polycation in this study [3,12]. Thechemical structure of imidazole-containing PAA-HIS was charac-terized by ATR-FTIR and 1H-NMR (Fig. 2). The ATR–FITR spectrashows the carbonyl group (C]O) from amide was at 1645 cm�1.The characteristic peaks of imidazole at 1000–1200 cm�1 show thesuccessful conjugation of PAA with histidine. The peak of aminogroup (–NH2) at 1537 cm�1 indicates the amino groups on PAAwere not completely replaced by histidine. The 1H-NMR spectrashows the peak of –CH2– and –CH– at 1.20–1.26 ppm, the peak of–NCH2– at 2.6 ppm, and the peak of imidazole at 6.88 ppm and7.68 ppm. The amount of histidine conjugated on PAA wasapproximately 80.4%, determined by ninhydrin method. Further-more, the buffering capacity of PAA-HIS analyzed by acid–basetitration was between pH 4.5 and 6.5 (Fig. 3), indicating PAA-HIScould buffer protons to rupture endosomal compartment duringendocytosis.

As shown in Fig. 1(b), binary complexes could be formed bymixing plasmid DNA with various amount of PAA-HIS. In this study,binary DNA complexes were formed with 10 mg of PAA-HIS and 2 mgof plasmid DNA because they could form stable nanoparticles whenthe weight ratio of PAA-HIS/DNA was greater than 2–3 (data notshown). The binary DNA complexes at 5 of PAA-HIS/DNA weightratio possessed nano-size and positive surface charge (Fig. 4) andshowed low cytotoxicity in HeLa cells (Fig. 5). Subsequently, single-stranded oligonucleotides, which have been reported to be able toconjugated with ligand or other molecules by nucleic acid–base-paired interaction [13–16], was coated on the binary DNAcomplexes to reverse the surface charge, as shown in Fig. 1(c). Fig. 4

Fig. 2. (a) The ATR-FTIR spectra of PAA and PAA-HIS, (b) The 1H-NMR of PAA-HIS.

Fig. 4. Particle sizes (black dot) and surface charges (hollow dot) of PAA-HIS/DNA/oligonucleotide complexes with different oligonucleotides/PAA-HIS molar ratios.

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–4203 4197

shows the variation of surface charges and particle sizes of ternaryoligonucleotide-coated PAA-HIS/DNA complexes as a function ofthe molar ratio of oligonucleotide relative to PAA-HIS. Withincreasing the amount of oligonucleotides, the surface charge ofternary complexes gradually approached to zero from positivevalues without oligonucleotides and reversed the sign at 0.5 ofoligonucleotide/PAA-HIS ratio. Reasonably, the formation of the

Fig. 3. Buffering capacity of PAA and PAA-HIS. Acid titration curves of aqueous solutionof PAA (10 mg/mL) and PAA-HIS (10 mg/mL) dissolved in 150 mM NaCl were obtainedby adding equal aliquots of 0.1 N HCl.

Fig. 5. Cytotoxicity assay. HeLa cells were treated with (a) PAA-HIS/DNA/oligonucleotidecomplexes and (b) PAA/DNA/oligonucleotide complexes with different oligonucleotide/PAA-HIS molar ratios for 24 h. The cell viability was determined by MTTassay. The resultsare expressed as percentage of cell viability relative to untreated cells. Asterisk (*)denotes significant differences (p< 0.05) compared with binary PAA/DNA complexes.

Fig. 6. Fluorescent microscopy images of HeLa cells treated with PAA-HIS/DNA/oligonucleotide complexes for 24 h at (a) 0; (b) 0.5; (c) 1; (d) 1.5; (e) 3 of oligonucleotide/PAA-HISmolar ratio. (f) ExGEN 500 was used as control. (scale bar¼ 100 mm).

Fig. 7. (a) Transfection into HeLa cells treated with PAA-HIS/DNA/oligonucleotidecomplexes containing different oligonucleotide/PAA-HIS molar ratios for 24 h. ExGEN500 was used as control. (b) Transfection into HEK-293, HepG2 and Hs68 cells treatedwith binary PAA-HIS/DNA complexes and ternary PAA-HIS/DNA/oligonucleotidecomplexes at 3 of oligonucleotide/PAA-HIS molar ratio for 24 h. The cells were analyzedfor luciferase activity. Asterisk (*) denotes significant differences (p< 0.05) comparedwith binary PAA-HIS/DNA complexes.

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–42034198

particle aggregation was observed at this time. Furthermore, whenthe molar ratio of oligonucleotide/PAA-HIS exceeded 1.5, thesurface charge of ternary complexes would change to negativecharge, –27 mV approximately and remained essentially constant.Similarly, the particle size reduced to around 100 nm. These indi-cate that the binary PAA-HIS/DNA complexes had been completelycovered by the additional layer of oligonucleotides (Fig. 1(c)).

3.2. Cytotoxicity of oligonucleotide-coated polyplexes

The ternary PAA-HIS/DNA/oligonucleotide complexes wereexamined for their cytotoxicity on HeLa cells using MTT assay (Fig. 5(a)) The viability of HeLa cells treated with ternary oligonucleotide-coated PAA-HIS/DNA complexes was estimated as 80–90%,regardless of the oligonucleotide/PAA-HIS ratio. Compared tobinary DNA complexes, the ternary negatively charged complexespossessed slightly higher cell viability, which was similar to theprevious reports [17,18]. For comparison, Fig. 5(b) shows theviability of HeLa cells treated with ternary oligonucleotide-coatedPAA/DNA complexes. As expected, PAA without further chemicalmodification for gene delivery application caused obvious celldamage. When oligonucleotides were coated on the binary PAA/DNA complexes, ternary PAA/DNA/oligonucleotide complexesexhibited only very slight cytotoxicity (more than 85% cellviability), regardless of the oligonucleotide/PAA ratio. Therefore,although modification of PAA with imidazole-containing histidinecould decrease cytotoxicity, the cytotoxicity of PAA/DNA/oligonu-cleotide and PAA-HIS/DNA/oligonucleotide complexes was similar.This indicates that the oligonucleotide layer on the polycation/DNAcomplex surface could govern the cytotoxic effects of polyplexproperties.

3.3. Transfection efficiency of oligonucleotide-coated polyplexes

The transfection efficiency of PAA-HIS/DNA/oligonucleotidecomplexes was analyzed by using EGFP expression. In contrast toprevious studies that anionic nanocomplexes with minimal celltransfection [18–20], Fig. 6 shows fluorescence of HeLa cells derivedfrom EGFP still could be observed at very high negative surfacecharge and even slightly increased with increasing the

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–4203 4199

oligonucleotide/PAA-HIS ratio. Differences of the transfection effi-ciency among these ternary nanoparticles were further confirmedby luciferase activity assay. Similarly, as shown in Fig. 7(a), higherlevel of luciferase expression was observed when the oligonucle-otide/PAA-HIS ratio is greater than 1. However, it should be notedthat expression levels obtained in the oligonucleotide-coated PAA(without imidazolyl substitution)/DNA complexes were very low(data not shown), indicating that the modification of PAA isrequired.

It is well known that the different cell lines possess quitedifferent cell uptake characteristics leading to different transfectionresults. Therefore, in addition to HeLa cells, we measured theluciferase activity of PAA-HIS/DNA/oligonucleotide complexes inother three kinds of cell lines (Fig. 7(b)). Oligonucleotide-coatedPAA-HIS/DNA complexes at 3 of oligonucleotide/PAA-HIS molarratio showed higher transfection efficiencies in HEK-293 andHepG2 cells than in Hs68 cells, which may be due to the normalfibroblast characteristics. Compared with binary PAA-HIS/DNAcomplexes, ternary PAA-HIS/DNA/oligonucleotide complexesshowed two-fold higher transfection efficiency in HEK-293 cells,30-fold higher transfection efficiency in HepG2 cells, and 40-foldhigher transfection efficiency in Hs68 cells. These results imply thatthe deposition of a layer of oligonucleotides onto the binary PAA-HIS/DNA complexes could promote introduction of plasmid DNAinto cells and present higher transfection efficiency.

3.4. Cellular uptake

In order to assess whether the observed gene expression isaccompanied by similar uptake results, plasmid DNA was firstlabeled with YOYO-1 fluorescent dye in binary PAA-HIS/DNAcomplexes, and then oligonucleotides were coated on the binaryDNA complexes. The confocal microscopy images presented in

Fig. 8. Confocal microscopy images of HeLa cells treated with PAA-HIS/DNA/oligonucleotideof oligonucleotide/PAA-HIS molar ratio for 4 h. FM4–64FX dye (red) was used to mark HeL

Fig. 8 illustrate the differences in the quantity of YOYO-1 in HeLacells with various oligonucleotide/PAA-HIS ratios at 4 h posttransfection. All PAA-HIS/DNA/oligonucleotide complexes indi-cated the localization of YOYO-1 in the cytoplasma (FM4-64FXmarker) and the ternary polyplexes with the oligonucleotide/PAA-HIS ratio greater than 1 delivered a higher YOYO-1-labeled plasmidDNA in HeLa cells. Therefore, gene expression of ternary polyplexeswith a higher negative surface charge was accompnied by anincrease of cellular uptake. In contrast, Fig. 4 shows the stableparticle size of the ternary complexes could not be obtained untilthe ratio of oligonucleotide/PAA-HIS greater than 1.5. Therefore, theeffect of coating oligonucleotides on gene expression and cellularuptake of PAA-HIS/DNA complexes with the oligonucleotide/PAA-HIS ratio of 0.5 and 1 was not obvious.

Generally, polycation/DNA complex uptake is mediated byelectrostatic interactions of positively charged complexes withnegatively charged cellular surface. In addition, it is known thataddition of certain polyanions to polyplexes will result in unpack-aging due to competitive binding between polyanions and plasmidDNA with polycations [17]. Therefore, it is possible that oligonu-cleotides and binary cationic DNA complexes were separated beforeternary nanoparticles were uptaken by cells. To address this issue,cellular uptake was further performed with the addition of Nulight(537)–50–C10T20–30 oligonucleotide, a red fluorescence marker, tothe 50–C10A20–30 oligonucleotide. Fig. 9 shows the confocalmicroscopy image of HeLa cells transfected with PAA-HIS/DNA/oligonucleotide complexes containing YOYO-1-labeled DNA andNulight oligonucleotide at 3 of oligonucleotide/PAA-HIS ratio for4 h. As expected, almost all oligonucleotides were visualized to becolocalized with YOYO-1-labeled DNA (yellow dots) in cytoplasm.This demonstrates that PAA-HIS/DNA/oligonucleotide complexescan be completely taken into cells without dissociation. Therefore,the internalization of entire ternary nanoparticles into living HeLa

complexes containing YOYO-1-labeled DNA (green) at (a) 0; (b) 0.5; (c) 1; (d) 1.5; (e) 3a cells. (scale bar¼ 50 mm).

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–42034200

cells was not prevented in the presence of the negatively chargedoligonucleotide surface.

3.5. The effect of polyanions

Subsequently, HeLa cells were transfected with PAAc- and PSS-coated polyanionic complexes containing PAA-HIS/DNA to deter-mine the polyanionic effect on transfection efficeincy (Fig. 10).Clearly, almost luciferase activities of all ternary anionic polyplexeswere significantly lower than the binary PAA-HIS/DNA complexes(p< 0.05). Especially, when the ratio of polyanion/PAA-HIS wasgreater than 1, transfection efficiencies of PAAc and PSS-coatedpolyplexes with negative surface potential only show very lowluciferase activity (below 106 RLU/mg protein). This indicates thatthe oligonucleotide is a unique polyanion to promote introductionof plasmid DNA into cells for effective gene expression.

Furthermore, the nucleotide effect on transfection efficiency wasdetermined by delivering PAA-HIS/DNA complexes coated withdifferent constituent oligonucleotides into HeLa cells (Fig. 11).Comparing with 50–C10A20–30, 50–C10T20–30 exhibited an about 1000-fold drop in luciferase activity (p< 0.05). Since these two ternarycomplexes showed similar anionic charges (w�30 mV), thedecreased gene expression of PAA-HIS/DNA/50–C10T20–30 oligonu-cleotide complexes was attributed to the difference of chemicalstructure between thymidine and adenosine. These results indicatecell transfection is quite senstive to the chemical structure of topmostlayer of the ternary PAA-HIS/DNA/oligonucleotide complexes.

Fig. 9. Confocal microscopy image of HeLa cells transfected with PAA-HIS/DNA/oligonucleoratio. 50–C10A20–30 oligonucleotides were mixed with Nulight (537)–50–C10T20–30 oligonuc(b) YOYO-1-labeled DNA (green), (c) Nulight (537)–50–C10T20–30 oligonucleotide (red), andlabeled DNA (yellow) indicates that the ternary complexes could be uptaken by cells witho

3.6. Erythrocyte agglutination study

Erythrocyte agglutination due to cationic polymeric character isa severe event since aggregates can block capillaries andinduce thrombogenesis. Fig. 12 shows PAA-HIS/DNA/oligonulceo-tide complexes at 3 of oligonucleotide/PAA-HIS molar ratio causedno agglutination. In contrast, there was an immediate formation ofaggregates when PAA-HIS/DNA and ExGEN 500/DNA complexesinteracted with erythrocyte solution due to the characteristics ofcationic polyplexes.

3.7. In vitro serum compatibility

Fig. 13 shows the effect of serum inhibition to the transfectionefficiency of oligonucleotide-coated PAA-HIS/DNA complexes.Generally, negatively charged serum protein such as albumin canbind to positively charged complexes to limit their in vivo efficacy.Trubetskoy et al. found the inclusion of polyanions with DNA/polycations complexes could prevent the serum inhibition duringcell transfection [18]. Similarly, transfection in OPTI-MEM mediumcontaining 10% serum did not affect the luciferease activity ofoligonucleotides-coated PAA-HIS/DNA complexes. In contrast,transfection by binary PAA-HIS/DNA complexes was significantlyinhibited by an approximatedly 10-fold drop in the prsence of 10%serum (p< 0.05). This indicates that the favorable serum perfor-mance of ternary oligonucleotide-coated PAA-HIS/DNA complexesresulted from their negative surface charge.

tide complexes containing YOYO-1-labeled DNA at 3 of oligonucleotide/PAA-HIS molarleotide to mark oligonucleotides. Four hours after transfection, (a) bright field image,(d) merged image are indicated. Colocalization of oligonucleotides and and YOYO-1-ut dissociation. (scale bar¼ 25 mm).

Fig. 11. Transfection into HeLa cells treated with PAA-HIS/DNA complexes coated by50–C10A20–30 and 50–C10T20–30 oligonucleotides at 3 of oligonucleotide/PAA-HIS molarratio for 24 h. The cells were analyzed for luciferase activity. Asterisk (*) denotessignificant differences (p< 0.05) between these two complexes.

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–4203 4201

4. Discussion

Generally, naked DNA or oligonucleotides is hardly transferredinto cells because of enzymatic degradation and electrostaticrepulsion. Thus, cationic polymers, such as PEI, and PLL have beenemployed to condense and protect DNA to efficient gene delivery[1,2]. However, high density amino groups of cationic moleculesare well known for their cytotoxicity to limit the application [4].Previous reports have described various techniques to decreasecytotoxicity and enhance transfection efficiency of cationicpolymers [3,6,7,9,21]. One of promising methods is to coatanionic polymers on the surface of cationic complexes todecrease the cationic characters. Although polyanions on the DNAcomplex surface could diminish the disturbance of the positivecharge to cellular membrane to reduce cytotoxicity, the trans-fection efficiency of anionic complexes would be decreased dueto the same effect [18,21]. Only few studies reported anionicpolyplexes, without conjugating with additional ligands, could beeffectively taken up well by cells [17,22]. To our knowledge,negatively charged oligonucleotides have never been used toenvelop cationic nanoparticle (Fig. 1). Interestingly, oligonucleo-tide-coated polyplexes could enter cells to exhibit high geneexpression (Fig. 7) even they presented strong anionic surfacecharges (Fig. 4).

Fig. 10. Transfection into HeLa cells treated with (a) PAA-HIS/DNA/PSS complexes and(b) PAA-HIS/DNA/PAAc complexes with different polyanion/PAA-HIS molar ratios for24 h. ExGEN 500 was used as control. The cells were analyzed for luciferase activity.Asterisk (*) denotes significant differences (p< 0.05) compared with binary PAA-HIS/DNA complexes.

In the present study, the cationic polymers used to complexwith DNA in the core were PAA and PAA-HIS. Modification of PAAwith imidazole-containing histidine could not only decrease cyto-toxicity by diminishing the intrinsic property of PAA, but couldenhance the release of the DNA–polymer complex into the cyto-plasm by virtue of its proton sponge property following endocy-tosis, leading to improve the efficacy of transfection [3]. As shownin Figs. 3 and 5, PAA-HIS had the buffering capacity between pH 4.5and 6.5 and binary PAA-HIS/DNA complexes exhibited only veryslight toxicity, respectively.

The surface charges and particle sizes of ternary oligonucleo-tide-assembled PAA-HIS/DNA complexes were shown in Fig. 4.When the molar ratio of oligonucleotide/PAA-HIS exceeded 1.5, theternary complexes have been found to form stable nanoparticleswith negative surface charge,�27 mV and nanometer size, 100 nm,indicating an anionic oligonucleotide layer successfully coated onthe ternary complex surface. Due to the negative surface charge,both PAA/DNA/oligonucleotide and PAA-HIS/DNA/oligonucleotidecomplexes possessed high cell viability (Fig. 5). However, comparedto traditional anionic polyplexes, apparent negative surface chargedid not attenuate the transfection efficiency of ternary PAA-HIS/DNA/oligonucleotide complexes. As shown in Fig. 7(a), luciferaseactivity of ternary polyplexes with negative surface charge in HeLacells were still slightly higher than the level obtained from binarypolyplexes with positive surface charge. Furthermore, Fig. 7(b)shows the deposition of a layer of oligonucleotides onto the binaryPAA-HIS/DNA complexes could effectively transfect various types ofcells including HEK-293, HepG2 and Hs68 cells. Especially, thetransfection efficiency of HepG2 and Hs68 cells substantiallyincreased when the binary complexes were coated with oligonu-cleotide, indicating the oligonucleotide-coated technique might beapplied to improve transfection efficiency of cells that are hardlytransfected.

Generally, anionic complexes cannot be taken up by cells due tothe electrostatic repulsion, so it is reasonable to assume oligonu-cleotides and binary cationic polyplexes have been dissociatedbefore they enter cells. Figs. 8 and 9 confirm that PAA-HIS/DNA/oligonucleotide complexes could be completely taken into cells andlocated in the cytoplasm without dissociation by using fluorescencemarkers under confocal microscope. Therefore, when oligonucle-otides were coated on PAA-HIS/DNA complexes, the interactionbetween oligonucleotides and PAA-HIS/DNA complexes was strongenough to prevent the entire ternary polyplexes from dissociation,against competition with cell membrane-bound proteoglycans.

Fig. 12. Agglutination of erythrocytes by (a) PAA-HIS/DNA/oligonucleotide complexes at 3 of oligonucleotide/PAA-HIS molar ratio, (b) PAA-HIS/DNA and (c) ExGEN 500/DNAcomplexes in polyplex/2% erythrocyte solution weight ratio of 1:1 and incubated for 15 min at room temperature. (scale bar¼ 100 mm).

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–42034202

After observing the effect of oligonucleotides on the transfectionactivity of PAA-HIS/DNA complexes, we studied the influence ofother polyanions on activity of ternary complexes with negativesurface charge. Consistent with the common viewpoint, the addi-tion of PSS and PAAc to the PAA-HIS/DNA complexe surfaceinhibited the gene transfer activity (Fig. 10), demonstrating theenhanced transfection activity are associated with the character ofoligonucleotides. Thus, it would be interesting to study whetherother oligonucleotides coated on PAA-HIS/DNA complexes couldachieve efficient gene transfer to permissive cells. However, similarto PSS and PAAc, deoxythymidine-rich oligonucleotide (50–C10T20–30) coated on PAA-HIS/DNA complexes showed very low geneexpression (Fig. 11). Because both deoxythymidine-rich oligonu-cleotide (50–C10T20–30) and deoxyadenosine-rich oligonucleotide(50–C10A20–30) have the same DNA backbone structure, the geneexpression of ternary complexes is strongly dependent on the kindof nitrogenous base bonded to the sugar-phosphate backbone.Therefore, not all oligonucleotides like the deoxyadenosine-richoligonucleotide (50–C10A20–30) could be well utilized for designingthe special transfection agent.

A successful gene delivery system includes cellular entry of thepolyplexes and effective release carried gene into cytosol forfurther transferring gene into nuclei [23]. Therefore, the genedelivery carrier system developed in this study still requiresfacilitating the escape from endosome. Fig. 3 shows the bufferingcapacity of PAA-HIS should improve the release behavior of

Fig. 13. Transfection into HeLa cells treated with PAA-HIS/DNA complexes and PAA-HIS/DNA/oligonucleotide complexes at 3 of oligonucleotide/PAA-HIS molar ratiowithout and with the addition of 10% serum for 24 h. The cells were analyzed forluciferase activity. Asterisk (*) denotes significant differences (p< 0.05) for binary PAA-HIS/DNA complexes in the presence and absence of 10% serum.

PAA-HIS/DNA/oligonucleotide complexes from endosomalcompartment. Hence, expression levels obtained in the oligonu-cleotide-coated PAA (without imidazolyl substitution)/DNAcomplexes were very low, even they could be taken up by cells(data not shown). According to our cell viability studies, this is notdue to residual cytotoxicity (Fig. 5). Consequently, it is importantto design the appropriate cationic polymer to complex with DNAin the core for developing oligonucleotide-coated polycation/DNAcomplexes. Polycations with the buffering capacity such as PEI[21] or polyamidoamine (PAMAM) [24] might be suitable toimprove the transfection efficiency of oligonucleotide-coatedpolycation/DNA complexes, which needs further investigation.

In addition to low cytotoxicity, the anionic polyplexes have beendemonstrated to prevent serum inhibition by their decreasedtendency to opsonization with serum proteins [18]. Likewise, theinclusion of oligonucleotides with PAA-HIS/DNA complex not onlyreduced agglutination of erythrocytes (Fig. 12), but also preventedthe serum inhibition of the transfection in cultured cells (Fig. 13).Therefore, the oligonucleotide-coated PAA-HIS/DNA complexes stillpossessed the characteristics of anionic complexes to reduce seruminhibition, indicating our ternary anionic polyplexes have potentialfor further used in vivo gene transfer application.

5. Conclusion

Negatively charged oligonucleotide (50–C10A20–30) was firstlyused to prepare a ternary anionic gene delivery system with lowcytotoxicity, high transfection efficiency and the resistance againsterythrocyte agglutination and serum inhibition. The techniqueof assembling oligonucleotides with binary polycation/DNAcomplexes developed in this study could be applied to other deliverysystem, such as drug delivery system. Furthermore, the ternarypolyplexes can be easily modified with specific molecules to thepolyplex surface by nucleic acid–base-paired interaction. Therefore,the oligonucleotide-coated PAA-HIS/DNA complex seems to bea feasible approach to develop new anionic gene delivery systems.

Acknowledgements

The authors thank National Science Council of the Republic ofChina and Industrial Technology Research Institute for theirfinancial support of this research. We are also grateful to the staffsof TC5 Bio-Image Tools, Technology Commons, College of LifeScience, NTU for help with the confocal laser scanning microscopy(CLSM).

Appendix

Figures with essential colour discrimination. Figs. 6,8,9 and 12 inthis article have parts that may be difficult to interpret in black andwhite. The full colour images can be found in the on-line version, atdoi:10.1016/j.biomaterials.2010.01.116.

Y.-C. Chung et al. / Biomaterials 31 (2010) 4194–4203 4203

References

[1] Herweijer H, Wolff J. Progress and prospects: naked DNA gene transfer andtherapy. Gene Ther 2003;10:453–8.

[2] Niidome T, Huang L. Gene therapy progress and prospects: nonviral vectors.Gene Ther 2002;9:1647–52.

[3] Pathak A, Aggarwal A, Kurupati RK, Patnaik S, Swami A, Singh Y, et al. Engi-neered polyallylamine nanoparticles for efficient in vitro transfection. PharmRes 2007;24:1427–40.

[4] Boussif O, Delair T, Brua C, Veron L, Pavirani A, Kolbe HVJ. Synthesis of poly-allylamine derivatives and their use as gene transfer vectors in vitro. Bio-conjugate Chem 1999;10:877–83.

[5] Nimesh S, Kumar R, Chandra R. Novel polyallylamine-dextran sulfate-DNAnanoplexes: highly efficient non-viral vector for gene delivery. Int J Pharm2006;320:143–9.

[6] Ogris M, Brunner S, Schuller S, Kircheis R, Wagner E. PEGylated DNA/trans-ferrin-PEI complexes: reduced interaction with blood components, extendedcirculation in blood and potential for systemic gene delivery. Gene Ther1999;6:595–605.

[7] Kleemann E, Neu M, Jekel N, Fink L, Schmehl T, Gessler T, et al. Nano-carriersfor DNA delivery to the lung based upon a TAT-derived peptide covalentlycoupled to PEG-PEI. J Control Release 2005;109:299–316.

[8] Cao ZH, Tong R, Mishra A, Xu WC, Wong GCL, Cheng JJ, et al. Reversible cell-specific drug delivery with aptamer-functionalized liposomes. Angew ChemInt Edit 2009;48:6494–8.

[9] 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.

[10] Gebhart CL, Sriadibhatla S, Vinogradov S, Lemieux P, Alakhov V, Kabanov AV.Design and formulation of polyplexes based on pluronic-polyethylenimineconjugates for gene transfer. Bioconjugate Chem 2002;13:937–44.

[11] Lee JS, Han MS, Mirkin CA. Colorimetric detection of mercuric ion (Hg2þ) inaqueous media using DNA-functionalized gold nanoparticles. Angew Chem IntEdit 2007;46:4093–6.

[12] Park JS, Han TH, Lee KY, Han SS, Hwang JJ, Moon DH, et al. N-acetyl histidine-conjugated glycol chitosan self-assembled nanoparticles for intracytoplasmicdelivery of drugs: endocytosis, exocytosis and drug release. J Control Release2006;115:37–45.

[13] Ivanova GD, Arzumanov A, Abes R, Yin H, Wood MJA, Lebleu B, et al. Improvedcell-penetrating peptide-PNA conjugates for splicing redirection in HeLacells and exon skipping in mdx mouse muscle. Nucleic Acids Res 2008;36:6418–28.

[14] Lytton-Jean AKR, Han MS, Mirkin CA. Microarray detection of duplex andtriplex DNA binders with DNA-modified gold nanoparticles. Anal Chem2007;79:6037–41.

[15] Roth L, Zagon J, Ehlers A, Kroh LW, Broll H. A novel approach for the detectionof DNA using immobilized peptide nucleic acid (PNA) probes and signalenhancement by real-time immuno-polymerase chain reaction (RT-iPCR).Anal Bioanal Chem 2009;394:529–37.

[16] Selvaraju SB, Kapoor R, Yadav JS. Peptide nucleic acid-fluorescence in situhybridization (PNA-FISH) assay for specific detection of mycobacteriumimmunogenum and DNA-FISH assay for analysis of pseudomonads in metal-working fluids and sputum. Mol Cell Probes 2008;22:273–80.

[17] Trubetskoy VS, Loomis A, Hagstrom JE, Budker VG, Wolff JA. Layer-by-layerdeposition of oppositely charged polyelectrolytes on the surface of condensedDNA particles. Nucleic Acids Res 1999;27:3090–5.

[18] Trubetskoy VS, Wong SC, Subbotin V, Budker VG, Loomis A, Hagstrom JE, et al.Recharging cationic DNA complexes with highly charged polyanions for invitro and in vivo gene delivery. Gene Ther 2003;10:261–71.

[19] Sakaguchi N, Kojima C, Harada A, Koiwai K, Kono K. The correlation betweenfusion capability and transfection activity in hybrid complexes of lipoplexesand pH-sensitive liposomes. Biomaterials 2008;29:4029–36.

[20] Yuba E, Kojima C, Sakaguchi N, Harada A, Koiwai K, Kono K. Gene deliveryto dendritic cells mediated by complexes of lipoplexes and pH-sensitivefusogenic polymer-modified liposomes. J Control Release 2008;130:77–83.

[21] Saul JM, Wang CHK, Ng CP, Pun SH. Multilayer nanocomplexes of polymer andDNA exhibit enhanced gene delivery. Adv Mater 2008;20:19–25.

[22] Kurosaki T, Kitahara T, Fumoto S, Nishida K, Nakamura J, Niidome T, et al.Ternary complexes of pDNA, polyethylenimine, and g-polyglutamic acid forgene delivery systems. Biomaterials 2009;30:2846–53.

[23] Wong SY, Pelet JM, Putnam D. Polymer systems for gene delivery – past,present, and future. Prog Polym Sci 2007;32:799–837.

[24] Harada A, Kawamura M, Matsuo T, Takahashi T, Kono K. Synthesis and char-acterization of a head-tail type polycation block copolymer as a nonviral genevector. Bioconjugate Chem 2006;17:3–5.