Alkylcarboxylate grafting to polyethylenimine: a simple approach to producing a DNA nanocarrier with...

THE JOURNAL OF GENE MEDICINE R E S E A R C H A R T I C L EJ Gene Med 2009; 11: 921–932.Published online 24 July 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/jgm.1374

Alkylcarboxylate grafting to polyethylenimine: asimple approach to producing a DNA nanocarrierwith low toxicity

Reza K. Oskuee1,2

Ali Dehshahri1

Wayne T. Shier3

Mohammad Ramezani1*

1Pharmaceutical and BiotechnologyResearch Centers, School ofPharmacy, Mashhad University ofMedical Sciences, Mashhad, Iran2Department of Modern Sciences andTechnologies, School of Medicine,Mashhad University of MedicalSciences, Mashhad, Iran3Department of Medicinal Chemistry,College of Pharmacy, University ofMinnesota-Twin Cities, Minneapolis,MN, USA

*Correspondence to:Mohammad Ramezani,Pharmaceutical and BiotechnologyResearch Centers, School ofPharmacy, Mashhad University ofMedical Sciences, PO Box91775-1365, Mashhad, Iran. E-mail:[email protected]

Received: 19 February 2009Revised: 2 June 2009Accepted: 16 June 2009

Abstract

Background Various strategies have been examined to improve bothtransfection efficiency and cytotoxicity of polyethylenimine (PEI), a widelyused polycationic nonviral gene vector. In the present study, we sought toimprove PEI transfection efficiency by combining the osmotic burst mechanismfor lysing endocytotic vesicles with the lipid depletion mechanism, which wasaccomplished by maintaining buffering capacity at the same time as adding alipid-absorbing hydrophobic shell.

Methods PEI was altered via the substitution of various percentages of itsprimary amines with carboxylate-terminated short, moderate and long alkylchains, by reaction with bromoacetic, 6-bromohexanoic, 10-bromodecanoicand 16-bromohexadecanoic acids. Modified polymers were complexed withplasmid and the particle size and zeta potential of the polyplexes weredetermined. Ethidium bromide dye exclusion was used to show the DNA-binding ability of the polymers and their transfection activity and cytotoxicitywas evaluated in Neuro2A mammalian cells.

Results Decreased DNA-binding ability resulted from increases in either thedegree of substitution or hydrocarbon chain length. Particle size and zetapotential measurements demonstrated that modified PEI polymers were ableto form nanoparticles in the size range 60–195 nm, and surface chargedecreased with an increasing degree of substitution. Higher degrees ofsubstitution resulted in decreased cytotoxicity of polymers. Alkylcarboxylatesubstitution of primary amines on PEI enhanced transfection efficiencies byup to approximately five-fold relative to underivatized PEI, with the greatestincreases occurring with 6-bromohexanoic acid derivatives at degrees ofsubstitution below 10%.

Conclusions The results obtained suggest that an appropriate balancebetween cationic and hydrophobic regions of alkylated PEI yields the optimalnonviral vector with high transfection efficiency and low toxicity. Copyright 2009 John Wiley & Sons, Ltd.

Keywords gene delivery; hydrophobic chain; nonviral vector; polyethylenimine;primary amine grafting

Introduction

Gene therapy could potentially be used to prevent, treat or cure a variety ofconditions, including several inheritable or acquired diseases [1]. All gene

Copyright 2009 John Wiley & Sons, Ltd.

922 R. K. Oskuee et al.

therapy strategies depend on introducing a gene oroligonucleotide into target cells. However, delivery ofnucleic acid therapeutic agents into cells remains theprimary limiting factor in gene therapy [1,2]. Thetransport of nucleic acids into cells can be achieved usingvectors, which generally fall into two categories: viraland nonviral vectors [3,4]. Viral vectors have been usedmost to deliver therapeutic genes into living cells in mostclinical trials because they achieve a high transfection rate,stable transfection and a rapid transcription of the foreignmaterial inserted in the viral genome. However, they haveserious limitations with respect to the amount of geneticmaterial that can be delivered, as well as severe safety risks[5–7]. Problems with viral vectors have led to increasedinterest in the development of nonviral vectors, for whichsafety is more easily defined and the amount of geneticmaterial is not inherently limited. However, nonviralvectors also have serious problems, most notably a lowtransfection efficiency. Cationic polymers and cationiclipids are by far the most widely used nonviral vectors forgene and oligonucleotide delivery [8].

Among the cationic polymers, polyethylenimine (PEI),which was introduced as a gene vector more than adecade ago [9], has been the most extensively studied.PEI has the disadvantage of being cytotoxic to manymammalian cell lines. Many factors affect both thetransfection efficiency and cytotoxicity of PEI includingmolecular weight, degree of branching, ionic strengthof the solution, zeta potential and particle size [10].Different strategies have been investigated in an attemptto improve the gene transfection efficiency of PEIwithout a corresponding increase in cytotoxicity. Theeffect of conjugating different residues to the nitrogenatoms of PEIs has been investigated for the ability toimprove transfection efficiency, including grafting ona surface coating of hydrophilic poly(ethylene glycol)[11,12], ligand tagging [13,14], dextran grafting [15] andhydrophobic modifications [16]. Increasing the molecularweight of PEI improves transfection efficiency, but it alsoincreases cytotoxicity [17], which begins to limit theusefulness of PEI larger than 25 kDa [18]. The toxicity isrelated to strong positive charge of the polymer, whichleads to electrostatic interaction with the negatively-charged cell surface membrane and causes damage[19]. Various approaches to reducing cytotoxicity havebeen investigated, including PEGylation, which has beenreported to enhance the transfection efficiency alongwith a reduction in cytotoxicity [20,21]. Putnam et al.[22] synthesized a series of polymers that optimized thebalance between polymer cationic density and endosomal-escape moieties, which led to effective gene transfer withlow cytotoxicity. It was demonstrated that acetylationdecreased the positive charge of the resulting polyplexes,and reduced the cytotoxicity of the modified PEI [23,24].In a study on preparing dextran grafted to 25 kDaPEI, Tseng et al. [25] reported a significant decrease incytotoxicity as the degree of grafting increased.

Surface charge is one of the important factors thataffects the transfection efficiency and cytotoxicity of

PEI vectors. It has been shown that conjugation ofhydrophobic residues to PEI can alter: (i) its abilityto condense DNA; (ii) the type of interaction withcell membranes, channels and nuclear pores; and (iii)the unpacking of complexes with DNA. The result ofconjugating PEI with hydrophobic residues has beeneither higher transfection efficiency [16,23] or lowertransfection efficiency [26], depending on the nature ofthe hydrophobic residue and how it is attached.

In the present study, the goal of developing improvednonviral DNA vectors based on branched PEI has beeninvestigated by starting with 10 kDa branched PEI, totake advantage its low toxicity, and modifying it bycarboxyalkylation of primary amines. Among differentapproaches for decreasing PEI cytotoxicity, reducing thesurface charge of PEI without substantially reducingits overall buffering capacity or ‘proton sponge’ [27]capability appeared to be a strategy worth investigating.The simplest model for the structures of alkylcarboxylatederivatives of PEI is that of dendrimers consisting of apolycationic PEI core surrounded by a hydrophobic alkylshell, which is in turn surrounded by an outer anionicshell of carboxylate moieties. This system provides anopportunity to combine the osmotic burst mechanism[27] with the lipid depletion mechanism [28] for lysingendocytotic vesicles after endocytotic uptake of theDNA-containing complex, if a hydrophobic shell can beconstructed that is an effective sink into which lipidsreleased from endosome membranes will dissolve. Theapproach used in the present study was to prepare a seriesof derivatives of 10 kDa branched PEI with negligible (twocarbons thick), intermediate (six to ten carbons thick) orlarge (16 carbons thick) hydrophobic shells and to testthe effect of hydrophobic shell thickness on transfectionefficiency.

Materials and methods

Materials

Branched polyethylenimine with an average molecu-lar weight of approximately 10 000 Da (PEI 10 kDa),were purchased from Polysciences, Inc. (Warrington,PA, USA). Branched polyethylenimine (b-PEI; averageMW = 25 kDa), bromoacetic acid, 6-bromohexanoic acid,10-bromodecanoic acid, 16-bromohexadecanoic acid,6-bromohexanoic acid ethyl ester, N-[2-hydroxythyl]piperazine-N′-[2-ethanesulfonic acid] (HEPES) and glu-cose were obtained from Sigma-Aldrich (Germany). Plas-mid pRL-CMV (Renilla luciferase under control of thecytomegalovirus enhancer/promoter), cell proliferationand reporter gene expression assays were obtained fromPromega (Madison, WI, USA). Fetal bovine serum (FBS)and Dulbecco’s modified Eagle’s medium (DMEM) werepurchased from Gibco (Gaithersburg, MD). All solventswere purchased from Sigma-Aldrich (Germany), and wereof the highest purity available. Dialyses were carried out

Copyright 2009 John Wiley & Sons, Ltd. J Gene Med 2009; 11: 921–932.DOI: 10.1002/jgm

Gene delivery of alkylcarboxylate-grafted PEI 923

by using Spectra/Por dialysis membranes (Spectrum Lab-oratories, Houston, TX, USA).

Synthesis and characterizationof alkylcarboxylate-grafted PEI

Alkyation reactions were carried out on 10 kDa branchedPEI. The strategy behind selecting 10 kDa branchedPEI, rather than the widely used 25 kDa branched PEI,was to start with a relatively less cytotoxic polymerand attempt to increase its transfection efficiency bystructural modifications. With 10 kDa PEI, some increasein cytotoxicity can be tolerated along with anticipatedincreases in transfection efficiency. However, with 25 kDaPEI cytotoxicity is at or near the limit of acceptability,so that structural alterations must improve transfectionefficiency with no increase in cytotoxicity, which hasproven difficult to achieve in previous studies [10,16,17].

The alkylcarboxylate chain sources (1-bromoaceticacid, 6-bromohexanoic acid and 10-bromodecanoic acid)and 6-bromohexanoic acid ethyl ester were dissolved indimethylformamide, and added dropwise over a periodof 2 h to a vigorously stirred solution of 10 kDa branchedPEI in the same solvent. The reaction was allowed toproceed at room temperature for another 24 h thendiluted into doubly distilled water (DDW). The reactionof 16-bromohexadecanoic acid with 10 kDa branched PEIwas carried out in a similar manner in chloroform, exceptthe solvent was removed under reduced pressure and theresidue was dissolved in DDW.

The resulting solutions were poured into a dialysismembrane (10 000 cut-off, Spectra/Por membrane) anddialysed against 150 mM NaCl for 1 day and then DDWfor 2 days to remove unreacted alkylcarboxylates. Afterthe dialysis, the aqueous solution was lyophilized toyield fluffy powders that were characterized by 1H NMR(D2O) spectroscopy using a Bruker Avance DRX-500 MHzNMR spectrometer (Bruker Daltonik GmbH, Bremen,Germany).

Determination of the degree of graftingusing 2,4,6-trinitrobenzenesulfonicacid

The amounts of accessible primary amines on unmodifiedPEI and on modified-PEIs were determined using 2,4,6-trinitrobenzenesulfonic acid (TNBS) using the standardmethod [29]. The degrees of grafting were calculatedby the differences in the amounts of primary amines onunmodified PEI compared to modified PEIs.

Measurements of the size andzeta-potential of the DNA–polymercomplexes

The sizes and zeta-potentials of the DNA-polymercomplexes were examined in a sodium chloride-free

buffer; HBG (20 mM HEPES, 5.2% glucose, pH 7.2). Thedesired amounts of cationic polymers were diluted in125 µl of buffer and mixed with an equal volume ofthe same buffer containing DNA. The surface chargeand particle size of the DNA/polymer complexes wereanalysed using a Malvern Zetasizer nano ZS (MalvernInstruments, Malvern, UK).

Measurement of the buffering capacityof the polymers

Each polymer was diluted to a final concentration of0.4 mg/ml. The solution was adjusted to a pH ofapproximately 12 using 1 M NaOH. Five microliter aliquotsof 1 M HCl were subsequently added until a pH of 2.5 wasreached. The solution pH was measured with a pH meter(Mettler Todelo, Greifensee, Switzerland). The reciprocalslopes of graphs plotting pH versus the amount of HCladded were used to determine the buffering capacity ofPEI derivatives [23,30].

Measurement of the binding strengthof the modified polymer to DNA

A DNA-intercalating dye (ethidium bromide), was usedto examine the association of DNA with the cationicpolymers [31]. A solution of 400 ng/ml ethidium bromidein HBG was prepared with further addition of 10 µg/mlof pDNA. The fluorescence intensity of ethidium bromidewas measured at an excitation wavelength 510 nm andemission wavelength 590 nm with a 5-nm slit using aJasco FP-6200 spectrofluorimeter (Jasco Global, Hachioji,Tokyo, Japan) and fluorescence was set to 100%.Equimolar amounts of modified polymers were addedstepwise to the pDNA-ethidium bromide solution and thefluorescence intensity was recorded. All measurementswere performed in triplicate and a graph was constructedby plotting the relative fluorescence intensity (%) againstthe polymer/plasmid DNA ratio (w/w).

Preparation of plasmid DNA

Plasmid DNA (pRL-CMV; Promega) was purified using aQiagen endotoxin-free Mega Plasmid kit (Qiagen, Hilden,Germany) according to the manufacturer’s instructions.The purity and identity of the plasmid was confirmed byagarose gel electrophoresis.

Cell line and cell culture

Neuro2A murine neuroblastoma cells (catalog no. CCL-131; ATCC, Manassas, VA, USA), were grown in DMEM(1 g/l glucose, 2 mM glutamine) supplemented with 10%FBS, streptomycin at 100 µg/ml and penicillin at 100U/ml. All cells were incubated at 37 ◦C in a humidified5% CO2 atmosphere.



Transfection procedure

Cells were seeded at a density of 1 × 104 cells/well in 96-well plates 1 day prior to transfection experiments, andgrown in the appropriate medium with 10% fetal bovineserum. Transfections were carried out when cultureswere 60–90% confluent. Polymer/pRL-CMV complexeswere prepared in HBG. Different polymer/plasmid DNAweight ratios, in the range 2 : 1 to 8 : 1 (w/w), wereused to prepare the polycation/plasmid complexes (i.e.polyplexes). Polyplexes were prepared by adding 50 µlof a solution of polycation at varying concentrations inHBG to 50 µl of a solution of plasmid DNA (40 µg/ml) inHBG with mixing by pipetting up and down followed byincubating for 30 min at room temperature. Transfectionwas performed by adding 10 µl of polyplex solution tothe wells of 96-well plates containing 60–90% confluentcultures of cells in complete medium containing 10% FBS.After 4 h, the medium was replaced with a fresh completemedium and gene expression was assayed 24 h later.Control transfections were performed by using branchedPEI 25 kDa. All transfection and cytotoxicity assays werecarried out in triplicate.

Luciferase reporter gene expression

Twenty-four hours post-transfection, the medium wasremoved and cells were lysed with 50 µl of Promegacell lysis buffer. Luciferase activity was measuredusing a Sirius Tube Luminometer (Berthold DetectionSystems, Pforzheim, Germany). Luciferase light unitswere recorded from a 20-µl aliquot of the cell lysatewith a 10-s integration time after automatic injection offreshly prepared luciferin using the Renilla LuciferaseAssay System (Promega). Transfection efficiency wasreported as relative light units per number of seededcells.

Cytotoxicity of polyplexes and freepolycations

Neuro2A cells were seeded at 10 000 cells/well in 100 µlof complete culture medium in 96-well plates and culturedfor 24 h. Wells were treated in triplicate with the sameamounts of polyplex used for transfection experimentfor 4 h, then the medium was replaced by 100 µl offresh culture medium. Cell survival in each well wasestimated after 24 h as the metabolic activity, which wasdetermined with a CellTiter 96 AQueous One SolutionCell Proliferation Assay (MTS) (Promega). MTS reagent(20 µl) was added to each well, the plate was incubatedfor 2 h at 37 ◦C and the absorbance was measuredat 490 nm in a microplate reader. Cell viability wasexpressed as a percentage absorbance relative to thatin untreated control cells. Values of metabolic activity arepresented as the mean ± SD of triplicates. Cytotoxicity offree polycations at increasing concentrations in the range

0–100 µg/ml during a 4-h incubation was assayed by thesame method.

Hemolytic activity of polymers

Freshly blood isolated human erythrocytes were washedin phosphate-buffered saline (four times centrifugationat 800 g for 10 min at 4 ◦C). The erythrocyte pelletwas diluted ten-fold in 150 mM NaCl. Unmodified andmodified polymers were serially diluted in 150 µl ofthe HEPES-buffered saline per well using triplicatesin a V-bottom 96-well plate. The resulting polymerconcentrations were 0.25–1.0 mg/ml. Control wellscontained 150 µl of buffer with 1% Triton-X-100. Trwentymicroliters of erythrocyte suspension was added to eachwell and the plates were incubated at 37 ◦C for 45 minunder constant shaking. After centrifugation at 2200r.p.m. (800 × g) for 10 min, 70 µl of supernatant wasanalysed for haemoglobin release at 405 nm using amicroplate plate reader.

Statistical analysis

Multiple measurement comparisons were performedusing Student’s t-test. p < 0.05 was considered statisti-cally significant.

Results and Discussion

Synthesis and characterizationof modified polymers

In the present study, derivatives of PEI were pre-pared by reacting accessible primary amines with ω

-bromoalkylcarboxylic acids having four different alkylchain lengths to investigate the effect of hydrocarbonchain lengths on the surface of the branched PEI den-drimer (Scheme 1).

Moreover, the effect of the degree of substitutionof primary amines on PEI by alkylcarboxylate chainswas evaluated by conjugating varying amounts of thealkylating agents reacting with PEI to correspond to 10,30, 50 and 100 mol% of the estimated number of primaryamines. The modified polycations were labelled as PEI-n-%, where n is the number of carbons in the alkyl chainand % is the percentage of primary amines substituted.

The compositions of the modified polymers wereconfirmed by 1H NMR. As an example, the obtaineddata from 1H NMR (D2O) spectrum of PEI-6 was:δ (p.p.m.) 1.29 (b,–(CH2)–), 2.1 (b,–(CH2)–), 2.60(b,–CH2CH2N–).

Branched PEI has been reported [32] to contain 25%primary, 50% secondary and 25% tertiary amines inits structure. By determining the amount of modifiableprimary amines using TNBS, it was possible to estimatethe percentage substitution of primary amines, assuming



Scheme 1. Synthesis of alkylcarboxylate grafted PEI derivatives

Table 1. Degree of primary amine substitution in 10 kDa PEI estimated by TNBS assay and size and zeta potential of the polymer/DNAcomplexes prepared with either modified or unmodified PEI

Polymers Initial reaction (mol%)

Substitution ofprimary amines by

alkyl carboxylates (mol%)Size of DNA/polymercomplex (nm) (± SD)

Zeta potential of DNA/polymercomplex (mV) (± SD)

PEI-2–8 10 8.4 61.7 ± 4 18.7 ± 1.7PEI-2–30 50 29 63 ± 11.5 16.5 ± 1.4PEI-2–40 100 39 126 ± 6.2 13.8 ± 1.7PEI-6–9 10 8.7 137 ± 5.6 20.5 ± 1.2PEI-6–17 50 17 146 ± 6.9 19.6 ± 0.6PEI-6–35 100 35.4 195 ± 17.4 15.7 ± 1.6PEI-10–7 10 7 193.3 ± 5.7 28.9 ± 0.7PEI-10–21 30 20.7 190.3 ± 35.8 27.2 ± 1PEI-10-%∗ 100 ND ND NDPEI-16–8 10 7.8 178 ± 9.6 26.4 ± 2.5PEI-16–18 30 18 140 ± 40.1 28 ± 5.3PEI-16-%∗ 100 ND ND ND10 kDa PEI 0 0 72 ± 10.1 34.2 ± 1.1

The modified polymers were coded as PEI-n-%, in which n represents the alkylcarboxylate chain (2 = acetic, 6 = hexanoic, 10 = decanoic, 16 =hexadecanoic), and % represents the percentage of primary amines substituted with alkylcarboxylate chains. ND, not determined. ∗The product wasinsoluble.

that 25% of the nitrogens in branched PEI were primaryamines. The TNBS assay indicated that the achievabledegrees of substitution were in the range of 7–40 mol%,which is lower than the calculated values based on theinitial molar ratios of alkylating agents to PEI, indicatingthat complete substitution of primary amines was notpossible (Table 1).

According to the ‘proton sponge’ hypothesis [27],buffering capacity plays a central role in the release ofpolycation-containing polyplexes from the endosomes.The buffering capacity that is important for driving theosmosis and bursting the endosome can be defined as themolar amount of protons required for the reduction of pHthrough the range that can reasonably be encountered inthe endosome, specifically from 7.5–5.

The effect of primary amines substitution withalkylcarboxylate chains on the buffering capacity of thepolymers was estimated by measuring the change inpH of the polymer solution (0.4 mg/ml) upon titrationwith 1M HCl (Figure 1). Alkylation of primary amines,which typically have pKa values in the pH range 9–11,creates secondary amines. Secondary amines typicallyhave pKa values in a similar pH range [33]. PEI (25 kDa)showed substantial buffering capacity in the pH range9–11, which can be attributed to primary plus secondaryamines. However, 25 kDa PEI also exhibited substantialbuffering capacity in the pH range 5.5–7, and it isbuffering capacity in this pH range that is expected to

be important in driving the osmotic burst mechanismduring transfection. It is not obvious which structuralfeatures of 25 kDa PEI are responsible for the bufferingcapacity in the pH 5.5–7 range, but they are also presentin unmodified 10 kDa PEI, which exhibits approximatelytwo-thirds of the buffering capacity in this pH range. Lowlevels of alkylation of 10 kDa PEI (in PEI-2-8, PEI-6-9 andPEI-10-7) resulted in little or no significant reduction inbuffering capacity in the pH range 5.5–7. However, higherlevels of alkylation (in PEI-2-40, PEI-6-35, PEI-10-21 andPEI-16-19) or longer alkyl chains (PEI-16-8 and PEI-16-19) resulted in negligible buffering capacity in the pHrange 10–4. If buffering capacity in the pH range 5.5–7is a determining factor for the transfection activity ofpolyplexes, analysis of buffering capacity would suggestthat relatively low levels of carboxyalkylation (<10%)with alkyl chains of ten carbons or less should have littleadverse effect on the transfection activity of 10 kDa PEIderivatives.

By grafting of different moieties on PEI, there is alwaysa possibility of reducing buffering capacity. As reported byTseng et al. [25], 1.84% grafting of PEI 25 kDa primaryamines with dextran led to a loss of one-third of thebuffering capacity. Other studies have reported a decreasein buffering capacity by an increase in the percentage ofacetylation [23] or acylation [34] of PEI.

Hydrophobic modification of PEI has been accom-plished by different approaches, including conjugation



Figure 1. pH titration of aqueous polymer solutions (0.4 mg/ml). The solution was adjusted to a pH of approximately 12 using 1 M

NaOH. Five microliter aliquots of 1 M HCl were subsequently added until a pH of 2.5 was reached

with cholesterol and myristate [16], acetylation [24],amide linkage of saturated hydrocarbon chains [35]and palmitic acid [26]. In most of these studies, theeffects of change in the degree of conjugation have beeninvestigated. Doody et al. [35] used a water-soluble car-bodiimide reagent to conjugate acetate, butanoate andhexanoate to the primary amine groups of branchedpolyethylenimine through amide linkages that simultane-ously remove the positive charge on an accessible primaryamine and add a hydrophobic element. However, con-jugation of PEI with hydrocarbon chains longer than sixcarbons was not achieved because of the limited solubilityof long-chain fatty acids in the aqueous buffer in whichthe reaction was being performed [35].

We have succeeded in synthesizing a series of novelpolycations with grafted hydrophobic regions that varyin the alkyl chain length as well as the percentage ofgrafting. The grafted polymer system used in this studyhas the advantage that carboxyalkylated PEI derivativesretain water solubility even in the case of longerchain hydrocarbons, presumably because the negatively-charged carboxylate groups at the ends of the alkyl chainsincrease solubility of the complex, or because micelle-like structures are formed, which shield the hydrophobicmoieties from the aqueous environment. The hydrophobicregions have the potential to: (i) improve transfectionefficiency by facilitating release of the negatively-chargedplasmid DNA from the polyplex after internalizationand (ii) reduce cytotoxicity by reducing the electrostaticinteraction of the grafted polymer with the negatively-charged cell surface [35]. The negative charges on theends of alkyl chains should also prevent aggregationof polycations grafted to longer chain hydrocarbonsinto large micelles with lipophilic cores. The resultingseries of grafted polymers provides the possibility ofgaining better insight into the effect of hydrocarbon chain

length on transfection efficiency and cytotoxicity of themodified PEI.

Characterization of polymer/DNAcomplexes

The first requirement for cationic polymers to act asgene delivery vectors is the ability to condense DNAinto polyplexes with optimal size ranges. Changes inthe fluorescence intensity of ethidium bromide, a DNA-intercalating dye, can be used to measure the extent ofassociation of DNA with cationic polymers [31]. Figure 2shows that, at a cationic polymer to plasmid DNA (C/P)ratio of 2, all vectors were able to optimally condensethe plasmid DNA. Unmodified, 10 kDa PEI was ableto fully condense the DNA at the lowest concentrationtested, and PEI derivatives with the lowest levels ofcarboxyalkylation (PEI-2-8, PEI-6-9 and PEI-10-7) wereindistinguishable from it in effectiveness at condensingDNA. Higher concentrations of polycation were requiredfor optimal DNA condensation by PEI derivatives withhigher levels of carboxyalkylation (PEI-10-21 > PEI-16-18 > PEI-2-30 > PEI-2-40 > PEI-6-35 > PEI-6-17), orwith longer alkyl chains (PEI-16-8) and they resultedin less complete condensation of DNA. Because thepolycation/DNA complexes were prepared on a weight-to-weight basis in the present study, rather than on aphosphorous-to-nitrogen molar ratio basis, the increasein molecular weight of the carboxyalkylated polycationscould be part of the reason why the amount of vectorrequired to completely condense DNA increased withthe extent of alkyl chain substitution. Furthermore, theconversion of primary amines to secondary amines andthe addition of negatively-charged carboxylate anions onthe surface of the polycation would be expected to reduce



Figure 2. DNA binding of modified polymers measured by the ethidium bromide exclusion assay in HBG

the electrostatic binding affinity of DNA. Altered bindingaffinity can lead to changes in the polyplex stability andits properties such as size, density, surface charge and,more importantly, unpacking ability. In other studies withhydrophobic moieties grafted onto PEI, more polymerwas needed to form highly compact polyplexes as thepercentage of grafting increased [15,35,36]. Gabrielsonet al. [23] in their study of acetylated PEI reportedlooser polyplexes at a polymer/DNA ratio of 1 : 1, but,at higher mass ratios, there was no significant differencebetween unmodified and modified PEI [23]. From thedata shown in Figure 2, the weakest condensations ofDNA were associated with higher levels of alkylation andlonger alkyl chains. By contrast, Doody et al. reportedthat increasing hydrophobic chain length from acetylatedPEI to butanoic and hexanoic derivatives increased DNAbinding affinity [35]. However, they did not study theeffects of longer hydrocarbon chains. The polycationsthat were most effective at DNA condensation in Figure 2were also the most effective at transfection (vide infra).The observations suggest that facilitating dissociation ofpolycation–DNA complexes is not a promising strategyfor improving transfection efficiency of polycations.Presumably, dissociation of polycation-DNA complexesafter the complex is inside the cytoplasm or nucleuscomprises an essential step in the transfection mechanism,but facilitating it will not improve transfection if it is notthe rate limiting step in the process.

In order for cationic polymers to act as efficient genedelivery vectors, the polyplexes they form with DNAshould fall within the optimal size range and retain a netpositive charge. The particle size measured by dynamiclight scattering and the zeta potential measurementresults are presented in Table 1 for various polyplexesformed at their optimum transfection weight ratios.These results demonstrate that alkylcarboxylate graftedPEI derivatives are able to form nanoparticles in the60–195 nm size range and that alkylcarboxylate graftingresulted in larger polyplex size compared to unmodifiedPEI. The zeta potential values were found to decreaseon increasing the degree of substitution in all the

nanoparticles. As also shown in Table 1, addition ofalkylcarboxylate chains to PEI resulted in a reductionin the zeta potential of all polyplexes, ranging from34.2 mV in the case of unmodified 10 kDa PEI to 13.8 mVand 15.7 mV for PEI-2-40 and PEI-6-35, respectively.However, all the polyplexes in the study retained sufficientpositive charge to bind to cell surfaces. Previous studieson modification of the surface positive charge of PEI alsoresulted in nanoparticles with decreasing charge withincreasing substitution [23].

Transfection efficiency and cytotoxicity

The transfection efficiency of the polyplexes was evaluatedin Neuro2A cells using the plasmid pRL-CMV as reportergene. The cytotoxicity of polyplexes was evaluatedunder the same culture conditions by measuring theviability of Neuro2A cells with the MTS assay. Figure 3compares the cytotoxicity and transfection efficiency ofpolyplexes prepared with 200 ng of plasmid pRL-CMVDNA at polycation/plasmid (C/P) ratios (w/w) of 2 : 1,4 : 1, 6 : 1 and 8 : 1. The polycations tested consistedof 10 kDa PEI grafted with four different lengths ofalkylcarboxylate chains, namely acetate (two carbons,PEI-2 series), hexanoate (six carbons, PEI-6 series),decanoate (ten carbons, PEI-10 series) and hexadecanoate(16 carbons, PEI-16 series). Each alkylcarboxylate chainlength series was examined with two or three percentagesof the primary amine groups in the PEI carboxyalkylated,specifically 8%, 30% and 40% carboxyalkylation in thePEI-2 series; 9%, 17% and 35% carboxyalkylation in thePEI-6 series; 7% and 21% carboxyalkylation in the PEI-10series; and 8% and 18% carboxyalkyation in the PEI-16series. Transfection efficiencies were also determined fortwo reference cationic polymers, underivatized 10 kDaPEI over the same range of C/P ratios as the derivatives,and 25 kDa PEI at a C/P ratio of 0.8, which wasobserved (data not shown) to be its optimal transfectionto cytotoxicity ratio.



Figure 3. Transfection efficiency and cytotoxicity of polyplexes formed with alkylcarboxylate-grafted polyethylenimine and 200 ngof plasmid pRL-CMV DNA. Neuro2A cells were transfected with polyplexes formed from PEI-2 series (A, D), PEI-6 series (B, E),PEI-10 and PEI-16 series (C, F) in HBG over the indicated range of cationic polymer to plasmid DNA (C/P) (w/w) ratios. Polyplexesprepared with 10 kDa PEI in HBG were included as controls at the indicated C/P ratios, as well as the polyplex prepared with25 kDa PEI in HBG at C/P 0.8. Luciferase activity (A–C) and metabolic activity (MTS assay) (D–F) are presented as the mean ± SDof triplicates. a: p < 0.01, b: p < 0.001, modified PEI compared to PEI10, c: p > 0.05 modified PEI compared to PEI25

Underivatized 25 kDa PEI exhibited a higher transfec-tion efficiency than 10 kDa PEI and all alkylcarboxylatederivatives tested, but 25 kDa PEI also exhibited greatercytotoxicity than 10 kDa PEI and all alkylcarboxylatederivatives tested except one possibly anomalous sample(PEI-2-8 at C/P 6; Figure 3D). Alkylcarboxylate sub-stitution of primary amines on 10 kDa PEI enhancedtransfection efficiencies by up to almost five-fold, with thegreatest increases occurring for hexylcarboxylate deriva-tives at degrees of primary amine substitution below20% (p < 0.001; Figure 3B). Optimal transfection effi-ciency was observed with PEI-6-17 at a C/P ratio of 6 : 1,and it was approximately equal to the transfection effi-ciency of the reference polycation, 25 kDa PEI (p > 0.05)(Figure 3B). Increasing the C/P ratio from 2 : 1 to 8 : 1 con-sistently improved the transfection efficiency of 10 kDaPEI, but carboxyalkylated 10 kDa PEI exhibited optimumtransfection efficiency at a C/P ratio of 6 : 1. Increasing thedegree of primary amine substitution with hexylcarboxy-late to 35% resulted in a markedly reduced transfectionefficiency (Figure 3B). Although the optimal transfec-tion efficiency was obtained with polyplexes formed

with the hexylcarboxylate-grafted 10 kDa PEI, acetylate-and decanylcarboxylate-grafted 10 kDa PEI also signifi-cantly enhanced the transgene expression levels by up totwo- and three-fold, respectively, compared to unmod-ified 10 kDa PEI, when the degree of primary aminesubstitution was <10% (p < 0.01) (Figures 3A and 3C).Hexadecanoate derivatives of 10 kDa PEI were the leastefficient transfection agents.

The effect of hydrocarbon chain length appears tobe complex. One plausible explanation is in terms of amodel in which carboxyalkylated PEI derivatives possessdendrimeric structures with 10 kDa PEI cores, surroundedby various different thicknesses of hydrophobic shellsand an outer anionic shell. In this model, increasedtransfection efficiency is determined by hydrophobic shellthickness: (i) PEI-2 series with no significant hydrophobicshell did not exhibit increased transfection activity, andcytotoxicity was reduced by the outer anionic shell; (ii)PEI-6 series and PEI-10 with a low percentage alkylationof primary amines formed almost optimal hydrophobicshells, which resulted in increased transfection efficiency,whereas the outer anionic shell prevented increases in



Figure 4. Comparison of the transfection efficiency of polyplexesformed from 6-bromohexanoic acid (PEI-6) and 6-bromohex-anoic acid ethyl ester (PEI-6Et) grafted polyethylenimine and200 ng of plasmid pRL-CMV DNA in Neuro2A cells

cytotoxicity; whereas (iii) in derivatives with longer alkylchains or higher densities of 10-carbon alkyl chains, thedendrimeric structure was lost, resulting in negligibletransfection activity.

A stable complex of a cationic polymer with DNA canprotect the DNA from enzymatic degradation. As men-tioned above, factors that affect the physicochemical prop-erties of a polyplex can influence transfection efficiency aswell as cytotoxicity. Although reducing the surface chargeof a polycationic vector could reduce unwanted inter-actions with cell surface negative charges, it could alsoweaken the stability of the polyplex formed with DNA.The fact that hexylcarboxylate-grafted 10 kDa PEI wasmore effective at delivery of DNA than longer or shorteralkyl chains could be the result of the interaction betweenseveral factors, including hydrophobic–hydrophilic bal-ance, surface charge, condensation ability, particle sizeand unpacking ability.

Increased transfection efficiency through the acetyla-tion of branched PEI has previously been reported by Packet al. [23,24]. Enhanced unpacking of the modified poly-mer/DNA complexes, along with reduced positive surfacecharge on polymers, was considered the reason to be thereason for this. However, reducing the positive surfacecharge can also reduce the transfection ability of polyca-tions, as previously reported [37,38]. The results obtainedin the present study show that reducing the net surfacepositive charge by adding negatively-charged carboxylategroups increased transfection efficiency (Figure 3), whichis consistent with the observations of Pack et al. [23,24].

The effect of polymer lipophilicity on the transfectionefficiency has been addressed in numerous studies[39–41]. Moderate enhancement of the hydrophobicityof 25 kDa PEI by grafting of Ala residues increasedthe transfection efficiency [41]. Doody et al. [35],in their study on substituting primary amines ofbranched polyethylenimine with saturated hydrocarbons,observed enhanced transfection efficiency at a degreeof substitution below 25%, but reduced transfectionefficiency at higher degrees of substitution [35]. Acylationof high molecular weight PEI (750 kDa) by up to 30% wasalso reported to enhance its transfection efficiency [34].

A few studies have investigated the effect of hydropho-bic chain length on the transfection efficiency, especially

Figure 5. Cell viability of Neuro2A cells measured by theMTS assay. Cells were incubated for 4 h with the indicatedconcentrations of the polycations: (A) PEI-2 series, (B) PEI-6series and (C) PEI-10 and PEI-16 series

longer chains. N-acylation of 25 kDa PEI with alanine hasbeen reported to double its transfection efficiency [41].In addition, dodecylation and hexadecylation of primaryamines of 25 kDa PEI has led to a five-fold increase intransfection efficiency compared to underivatized 25 kDaPEI [41]. Nimesh et al. [34] reported that transfection effi-ciency of acylated 750 kDa PEI was dependent on the acylchain length, with optimal transfection efficiency observedwith 30% propionylated 750 kDa PEI. Doody et al. [35]



Figure 6. Hemolytic activity of PEIs and carboxyalkylated PEI. Human erythrocytes were incubated for 45 min at 37 ◦C with serialdilutions of 10 kDa PEI, 25 kDa PEI and carboxyalkylated 10 kDa PEI. haemoglobin release was measured by absorption at 405 nmin the supernatant after centrifugation

compared the effect of conjugating acetate, butanoate andhexanoate to branched PEI on its transfection efficiency.No general trends in transfection efficiency were observedwith different cell lines. It was suggested that each lipidmight have its own mechanism for enhancing transfec-tion efficiency [35]. The results obtained in the presentstudy are generally consistent with the above-mentionedstudies, although the observation of optimal transfec-tion efficiency with hexylcarboxylate modified polymersis novel.

In an effort to evaluate the effect of the terminal neg-ative charges on the hydrophobic regions on transfectionproperties, 6-bromohexanoic acid ethyl ester derivativesof 10 kDa PEI were synthesized and compared with 6-bromohexanoic acid derivatives. The results showed noconsistent or significant difference in the transfection effi-ciency, especially at lower C/P ratios (Figure 4). In C/Pratios higher than 6, the transfection efficiency of ethylester derivatives decreased, indicating higher cytotoxicity,which was consistent with the results of MTS assay data(data not shown). This result is consistent with surfacenegative charges not being critical factors in enhancingtransfection activity of carboxyalkylated 10 kDa PEI. How-ever, it should be recognized that replacing carboxylategroups with ethyl ester groups increases the hydropho-bicity of the attached moiety, as well as eliminating thecharge.

The cytotoxicity of polyplexes prepared with alkyl-carboxylate-modified 10 kDa PEI was evaluated by theireffects on Neuro2A cell cultures measured with the MTSmetabolic activity assay. Most modified PEIs inducedless cytotoxicity than 10 kDa or 25 kDa PEI, so poly-plexes prepared with them were essentially nontoxicunder conditions used to evaluate transfection efficiency(Figures 3D, 3E and 3F). The results obtained also demon-strated that cytotoxicity was strongly dependent on degreeof substitution of primary amines on PEI with alkylcar-boxylate residues. Results of in vitro cytotoxicity studies

after 4 h of incubation with different concentrations ofthe polymers, in the range 0–100 µg/ml, are shown inFigure 5. In each study, the greatest cytotoxicity wasexhibited by the positive controls, underivatized 25 kDaPEI, followed by underivatized 10 kDa PEI. Among thealkylcarboxylate derivatives of 10 kDa PEI, cytotoxicitywas observed only with acetate (C-2) at the lowest per-centage alkylation of primary amines examined (8%;Figure 5A), and with the longest alkyl chains studied(ten carbons and 16 carbons) (Figure 5C). The hexanoate(C-6) carboxyalkylation derivatives of 10 kDa PEI, whichexhibited the largest enhancements in transfection effi-ciency relative to 10 kDa PEI, were not cytotoxic withNeuro2A cells at any concentration tested (Figure 5B).

For use in vivo, an ideal vector should have minimalinteraction with blood components. Polycations can causehemolysis by destabilizing the erythrocyte membrane[42,43]. To evaluate the membrane destabilizing activityof carboxyalkylated PEI polymers at neutral pH, theirlytic activity was measured on erythrocytes (Figure 6).Carboxyalkylation of PEI reduced lytic activity comparedto 25 kDa and 10 kDa PEIs for all alkyl chain lengths andall degrees of carboxyalkylation tested. Other studies havereported that polycation charge reduction or hydrophobicmodification reduces haemolytic activity of polycations[26,44].

In conclusion, we have prepared a small library of10 kDa PEI derivatives modified by substitution of variouspercentages of the primary amines with alkylcarboxylatemoieties with different alkyl chain lengths. The simpleststructural model for PEI derivatives prepared in thisstudy is that of dendrimers with polycationic PEIcores surrounded by hydrophobic alkyl shells of variousthicknesses (two, six, ten or 16 carbon chain lengths)and densities (2%, 6%, 10% or 16% primary aminesubstitution), each of which is surrounded by an anionicouter shell of carboxylate moieties at physiological pH. PEIderivatives with alkyl chain lengths of ten carbons or less



exhibit good solubility and little loss of buffering capacityin the endosomal pH range. Substantially increasedtransfection activity (two- to almost five-fold) wasobserved with alkyl chain lengths of six and ten carbons atdensities of <20% primary amine substitution. Alkyl chainlength and density had no significant effect on cytotoxicityof the polyplexes formed from carboxyalkylated 10 kDaPEI and plasmid DNA. Cytotoxicity was not observed withNeuro2A cells for 10 kDa PEI carboxyalkylated with 6-carbon alkyl chains at any concentration tested. Solubilityand transfection activity in polyplexes were not retainedafter carboxyalkylation of 10 kDa PEI with a longer, 16-carbon alkyl chain. Carboxylate groups on the surfaceof dendrimeric PEI-based nonviral vectors provide aconvenient functional group for the design of the nextgeneration of vectors.

Acknowledgements

This work was funded by Mashhad University of Medical Sciences(MUMS), Mashhad, Iran. Financial support provided by theIranian Nanotechnology Initiative is gratefully acknowledged.

References

1. Cavazzana-Calvo M, Thrasher A, Mavilio F. The future of genetherapy. Nature 2004; 427: 779–781.

2. Verma IM, Somia N. Gene therapy – promises, problems andprospects. Nature 1997; 389: 239–242.

3. Liu D, Ren T, Gao X. Cationic transfection lipids. Curr Med Chem2003; 10: 1307–1315.

4. Tietze N, Pelisek J, Philipp A, et al. Induction of apoptosis inmurine neuroblastoma by systemic delivery of transferrin-shielded siRNA polyplexes for downregulation of Ran.Oligonucleotides 2008; 18: 161–174.

5. Lehrman S. Virus treatment questioned after gene therapy death.Nature 1999; 401: 517–518.

6. Liu Q, Muruve DA. Molecular basis of the inflammatory responseto adenovirus vectors. Gene Ther 2003; 10: 935–940.

7. Sun JY, Anand-Jawa V, Chatterjee S, et al. Immune responses toadeno-associated virus and its recombinant vectors. Gene Ther2003; 10: 964–976.

8. Zhang S, Xu Y, Wang B, et al. Cationic compounds used inlipoplexes and polyplexes for gene delivery. J Control Release2004; 100: 165–180.

9. Boussif O, Lezoualc’h F, Zanta MA, et al. A versatile vector forgene and oligonucleotide transfer into cells in culture andin vivo: polyethylenimine. Proc Natl Acad Sci USA 1995; 92:7297–7301.

10. Kunath K, von Harpe A, Fischer D, et al. Low-molecular-weight polyethylenimine as a non-viral vector for DNAdelivery: comparison of physicochemical properties, transfectionefficiency and in vivo distribution with high-molecular-weightpolyethylenimine. J Control Release 2003; 89: 113–125.

11. Kursa M, Walker GF, Roessler V, et al. Novel shieldedtransferrin-polyethylene glycol-polyethylenimine/DNA com-plexes for systemic tumor-targeted gene transfer. BioconjugChem 2003; 14: 222–231.

12. Kircheis R, Schuller S, Brunner S, et al. Polycation-based DNAcomplexes for tumor-targeted gene delivery in vivo. J Gene Med1999; 1: 111–120.

13. Ogris M, Walker G, Blessing T, et al. Tumor-targeted genetherapy: strategies for the preparation of ligand-polyethyleneglycol-polyethylenimine/DNA complexes. J Control Release2003; 91: 173–181.

14. Kircheis R, Wightman L, Schreiber A, et al. Polyethylen-imine/DNA complexes shielded by transferrin target geneexpression to tumors after systemic application. Gene Ther 2001;8: 28–40.

15. Tseng WC, Jong CM. Improved stability of polycationic vector bydextran-grafted branched polyethylenimine. Biomacromolecules2003; 4: 1277–1284.

16. Kim S, Choi JS, Jang HS, et al. Hydrophobic modification ofpolyethylenimine for gene transfectants. Bull Korean Chem Soc2001; 22: 1069–1075.

17. Fischer D, Bieber T, Li Y, et al. A novel non-viral vector forDNA delivery based on low molecular weight, branchedpolyethylenimine: effect of molecular weight on transfectionefficiency and cytotoxicity. Pharm Res 1999; 16: 1273–1279.

18. Moghimi SM, Symonds P, Murray JC, et al. A two-stage poly(ethylenimine)-mediated cytotoxicity: implications for genetransfer/therapy. Mol Ther 2005; 11: 990–995.

19. Hunter AC. Molecular hurdles in polyfectin design andmechanistic background to polycation induced cytotoxicity. AdvDrug Deliv Rev 2006; 58: 1523–1531.

20. Petersen H, Fechner PM, Fischer D, et al. Synthesis, Charac-terization, and Biocompatibility of Polyethylenimine-graft-poly(ethylene glycol) block copolymers. Macromolecules 2002; 35:6867–6874.

21. Kunath K, von Harpe A, Petersen H, et al. The structure of PEG-modified poly (ethylene imines) influences biodistribution andpharmacokinetics of their complexes with NF-kappaB decoy inmice. Pharm Res 2002; 19: 810–817.

22. Putnam D, Gentry CA, Pack DW, et al. Polymer-based genedelivery with low cytotoxicity by a unique balance of side-chaintermini. Proc Natl Acad Sci USA 2001; 98: 1200–1205.

23. Gabrielson NP, Pack DW. Acetylation of polyethylenimineenhances gene delivery via weakened polymer/DNAinteractions. Biomacromolecules 2006; 7: 2427–2435.

24. Forrest ML, Meister GE, Koerber JT, et al. Partial acetylation ofpolyethylenimine enhances in vitro gene delivery. Pharm Res2004; 21: 365–371.

25. Tseng WC, Tang CH, Fang TY. The role of dextran conjugationin transfection mediated by dextran-grafted polyethylenimine.J Gene Med 2004; 6: 895–905.

26. Brownlie A, Uchegbu IF, Schatzlein AG. PEI-based vesicle-polymer hybrid gene delivery system with improvedbiocompatibility. Int J Pharm 2004; 274: 41–52.

27. Behr JP. The proton sponge: A trick to enter cells the viruses didnot exploit. Chimia 1997; 51: 34–36.

28. Xu Y, Szoka FC Jr. Mechanism of DNA release from cationicliposome/DNA complexes used in cell transfection. Biochemistry1996; 35: 5616–5623.

29. Snyder SL, Sobocinski PZ. An improved 2,4,6-trinitrobenzene-sulfonic acid method for the determination of amines. AnalBiochem 1975; 64: 284–288.

30. Tang MX, Szoka FC. The influence of polymer structure on theinteractions of cationic polymers with DNA and morphology ofthe resulting complexes. Gene Ther 1997; 4: 823–832.

31. Geall AJ, Blagbrough IS. Rapid and sensitive ethidium bromidefluorescence quenching assay of polyamine conjugate–DNAinteractions for the analysis of lipoplex formation in genetherapy. J Pharm Biomed Anal 2000; 22: 849–859.

32. von Harpe A, Petersen H, Li Y, et al. Characterization ofcommercially available and synthesized polyethylenimines forgene delivery. J Control Release 2000; 69: 309–322.

33. Hall HK Jr. Correlation of the base strengths of amines. J AmChem Soc 1957; 79: 5441–5444.

34. Nimesh S, Aggarwal A, Kumar P, et al. Influence of acyl chainlength on transfection mediated by acylated PEI nanoparticles.Int J Pharm 2007; 337: 265–274.

35. Doody AM, Korley JN, Dang KP, et al. Characterizing thestructure/function parameter space of hydrocarbon-conjugatedbranched polyethylenimine for DNA delivery in vitro. J ControlRelease 2006; 116: 227–237.

36. Jones SP, Gabrielson NP, Pack DW, et al. Synergistic effects ingene delivery – a structure-activity approach to the optimisationof hybrid dendritic-lipidic transfection agents. Chem Commun(Camb) 2008; 39: 4700–4702.

37. Brissault B, Kichler A, Guis C, et al. Synthesis of linearpolyethylenimine derivatives for DNA transfection. BioconjugChem 2003; 14: 581–587.

38. Thomas M, Lu JJ, Ge Q, et al. Full deacylation ofpolyethylenimine dramatically boosts its gene delivery efficiencyand specificity to mouse lung. Proc Natl Acad Sci USA 2005; 102:5679–5684.

39. Han S, Mahato RI, Kim SW. Water-soluble lipopolymer for genedelivery. Bioconjug Chem 2001; 12: 337–345.



40. Wang DA, Narang AS, Kotb M, et al. Novel branched poly(ethylenimine)-cholesterol water-soluble lipopolymers for genedelivery. Biomacromolecules 2002; 3: 1197–1207.

41. Thomas M, Klibanov AM. Enhancing polyethylenimine’s deliveryof plasmid DNA into mammalian cells. Proc Natl Acad Sci USA2002; 99: 14640–14645.

42. Dubruel P, Dekie L, Christiaens B, et al. Poly l-glutamic acidderivatives as multifunctional vectors for gene delivery. Part B.Biological evaluation. Biomacromolecules 2003; 4: 1177–1183.

43. Ogris M, Brunner S, Schuller S, et al. PEGylated DNA/transferrin-PEI complexes: reduced interaction with bloodcomponents, extended circulation in blood and potential forsystemic gene delivery. Gene Ther 1999; 6: 595–605.

44. Malik N, Wiwattanapatapee R, Klopsch R, et al. Dendrimers:relationship between structure and biocompatibility in vitro,and preliminary studies on the biodistribution of 125I-labelledpolyamidoamine dendrimers in vivo. J Control Release 2000; 65:133–148.


Alkylcarboxylate grafting to polyethylenimine: a simple approach to producing a DNA nanocarrier with...

Documents

Transcript of Alkylcarboxylate grafting to polyethylenimine: a simple approach to producing a DNA nanocarrier with...