05) 6

tosa

Shirui Maoa,b, Xintao Shuaia, Florian Ungera, Matthias Wittmara,

the cytotoxicity results, PEG 5kDa is superior for PEGylation when compared to PEG 550Da at similar graft ratios. Complexation

with insulin further increased cell viability. In addition, Lactate dehydrogenase (LDH) assays were performed to quantify the

cationic polysaccharide produced by partial deacetyla-tion of chitin derived from naturally occurring crusta-

its poor solubility in physiological media. Most chit-osans are only soluble in aqueous acidic solutions below

ARTICLE IN PRESScean shells. Due to its specic structure and properties,chitosan has found a number of applications in drugdelivery including that of as an absorption enhancer for

pH 6.5, where primary amino groups of chitosan areprotonated. To improve water solubility of chitosan,several derivatives have been studied. For example, themodication of chitosan by quaternization of the aminogroups [3,4], N-carboxymethylation [5] and PEGylation[6,7] have been reported.

Corresponding author. Tel.: +496421 282 5881;fax: +496421 282 7016.

E-mail address: kissel@staff.uni-marburg.de (T. Kissel).0142-9612/$ - se

doi:10.1016/j.bimembrane-damaging effects of the copolymers, which is in line with the conclusion drawn from MTT assay. Moreover, the safety of

the copolymers was corroborated by observing the morphological change of the cells with inverted phase contrast microscopy.

Based upon these results PEG-g-TMC merits further investigations as a drug delivery vehicle.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Trimethyl chitosan; PEGylation; Copolymer; In vitro cytotoxicity; MTT assay; LDH assay

1. Introduction

Chitosan (poly[b-(1-4)-2-amino-2-deoxy-D-glucopyra-nose]) is a relatively less toxic and biocompatible

hydrophilic macromolecular drugs, as promising poly-meric excipients for mucosal drug and vaccine delivery[1], and as a gene delivery system [2]. Applications ofchitosan in the biomedical eld are limited, however, byXiulan Xie , Thomas KisselaDepartment of Pharmaceutics and Biopharmacy, Philipps-University of Marburg, Ketzerbach 63, D-35032 Marburg, Germany

bDepartment of Pharmacy, Shenyang Pharmaceutical University, Wenhua Road 103, 110016 Shenyang, ChinacDepartment of Chemistry, Hans-Meerwein-Street, Philipps-University of Marburg, D-35032 Marburg, Germany

Received 8 December 2004; accepted 29 March 2005

Available online 23 May 2005

Abstract

PEGylated trimethyl chitosan (TMC) copolymers were synthesized in an attempt to both increase the solubility of chitosan in

water, and improve the biocompatibility of TMC. A series of copolymers with different degrees of substitution were obtained by

grafting activated poly(ethylene glycol)s (PEG) of different MW onto TMC via primary amino groups. Structure of the copolymers

was characterized using 1H, 13C NMR spectroscopy and GPC. Solubility experiments demonstrated that PEG-g-TMC copolymers

were completely water-soluble over the entire pH range of 114 regardless of the PEG MW, even when the graft density was as low

as 10%. Using the methyl tetrazolium (MTT) assay, the effect of TMC molecular weight, PEGylation ratio, PEG and TMC

molecular weight in the copolymers, and complexation with insulin on the cytotoxicity of TMC was examined, and IC50 values were

calculated with L929 cell line. All polymers exhibited a time- and dose-dependent cytotoxic response that increased with molecular

weight. PEGylation can decrease the cytotoxicity of TMC to a great extent in the case of low molecular weight TMCs. According toc a,Biomaterials 26 (20

Synthesis, characterization and cygraft-trimethyl chitoe front matter r 2005 Elsevier Ltd. All rights reserved.

omaterials.2005.03.0363436356

toxicity of poly(ethylene glycol)-n block copolymers

www.elsevier.com/locate/biomaterials

84.7%) was purchased from Fluka (Steinheim, Ger-many) and depolymerized according to the method

INerialsHowever, these efforts were only partially successfulin the case of PEG(5k)-g-chitosan, which was preparedvia reacting chitosan (degree of deacetylation 70%, MW70kDa) with methoxy-PEG(5k)-g-nitrophenyl carbo-nate [7]. With a graft density of 78.5% (calculated as aweight ratio of PEG 5kDa in the graft copolymer, thesame below), PEG(5k)-g-chitosan copolymer solutionsbecame cloudy at pH 6.8, which is suggestive of polymerprecipitation. Even at graft densities as high as 92.7%,the copolymer solution became cloudy [6] and formedaggregates [7] when the pH was 47.0. It is noteworthythat phase separation appeared at pH 6.5 or higher forunmodied chitosan, depending on the molecularweight [8]. This indicates that PEGylation can onlymarginally improve the solubility of chitosan in a verylimited range (pH 6.57.0), even with high graftdensities. Accordingly, the objective of the present studywas to design chitosan derivatives with superiorsolubility and biocompatibility.Another possible approach involves trimethyl chit-

osan (TMC). TMC is a permanently quaternizedchitosan derivative, which has been proven to be highlysoluble over a wide pH range (pH 19) up to 10% (w/v)concentration. Moreover, TMC is capable of openingtight junctions of intestinal epithelial cells at physiolo-gical pH values, thus increasing paracellular perme-ability of intestinal epithelia [9,10] whereas chitosanitself is insoluble and thus ineffective in this role. TMChas been demonstrated to be a promising excipient inthe development of solid dosage forms for the peroraldelivery and intestinal absorption of peptides and otherdrug substances [11]. The degree of quaternization ofTMC plays an important role on its absorptionenhancing properties and the effects of TMC onintestinal epithelia reached an optimum value atapproximately 40% degree of quaternization [12,13].Similar results were obtained on absorption enhance-ment across rabbit corneal epithelia [14] and nasalepithelia of rats [15]. It was also shown that the degreeof quaternization of the TMC had a pronounced effecton the mucoadhesive properties of this polymer [16].Moreover, TMC or modied TMC seems to be anefcient gene delivery system [17,18]. Using intestinalCaco-2 cell monolayers and ciliated chicken embryotrachea, the cytotoxicity and ciliotoxicity of TMCpolymers with different degree of quaternization werestudied. No substantial cell membrane damage could bedetected on the Caco-2 cells, while the effect on the CBFin vitro was found to be marginal at a concentration of1.0% (w/v) [19]. However, TMC was shown to becytotoxic in L929 mouse broblast cells as indicated by

ARTICLES. Mao et al. / Biomat6344MTT assay in our laboratory (data shown below).Similarly, it was found that reversibility of transepithe-lial resistance at 0.5% concentrations of TMC withdifferent degree of quaternization could not be demon-strated at pH 6.2 and 7.4 in Caco-2 cells [20]. Therefore,described previously [8] to obtain chitosans of differentmolecular weights. Methoxypoly(ethylene glycol) (Mn550 and 5000Da), water free toluene (99.9%), dichlor-omethane (DCM) (99.9%), diethyl ether (99.9%),maleic anhydrides, N-hydroxysuccinimide (NHS) anddicyclohexylcarbodiimide (DCC) were obtained fromAldrich (Steinheim, Germany) and used as received.Human recombinant insulin powder (26.2 IU/mg) was agift from Aventis Pharma AG (Germany). Dulbeccosmodied Eagles medium (DMEM) was obtained fromPAA (Colbe, Germany). MTT (3-(4,5-dimethyl-thiazol-2-yl)-2, 5-diphenyl tetrazolium bromide) was purchasedfrom Sigma (Deisenhofen, Germany). Lactate dehydro-genase (LDH) assay kit (Product No. LK 100) wasobtained from Sigma (Taufkirchen, Germany). All otherchemicals and solvents were of analytical grade.

2.2. Synthesis of TMC with different MW

Different MW TMCs were prepared according to atwo-step method described previously using depolymer-further improvement of the biocompatibility of TMC isdesirable.As water soluble, biocompatible, non-toxic and non-

immunogenic material, not only can PEG enhancebiocompatibility but also favorably affect pharmacoki-netics and tissue distribution. Conjugation of PEG toproteins is well known to enhance the in vivo half-life ofthe conjugated drugs, assist penetration into the cellmembrane, alter pharmacological properties and in-crease biocompatibility [21,22]. Therefore, PEGylatedTMC copolymers could possibly provide further biolo-gical functionality, in addition to improving solubilityand decreasing cytotoxicity. Here we described thesynthesis and physicochemical characterization of PE-Gylated TMC copolymers in an effort to improve bothsolubility and biocompatibility. Additionally, the effectof TMC molecular weight, PEGylation ratio, PEG andTMC molecular weight in the copolymers and com-plexation with insulin on the cytotoxicity of TMC wasexamined, providing experimental support for thedevelopment of drug carriers from such materials.

2. Materials and methods

2.1. Materials

Chitosan (MW 400 kDa, degree of deacetylation

PRESS26 (2005) 63436356ized chitosans of appropriate MW as starting materialsand were subsequently characterized by 1H NMR [23].Throughout, we used the abbreviation TMC 400 kDa todenote the polymer prepared from chitosan 400 kDa,and the same for the other polymers. Since the degree of

INerialsquaternization of TMC plays an important role inopening the tight junctions and a higher degree ofsubstitution had improved permeation enhancement[13,14], TMCs with a 40% degree of substitution wereprepared.

2.3. Activation of mPEG

The monohydroxy-terminated PEG was converted toa carboxyl-terminated intermediate by estericationwith cyclic aliphatic anhydride according to the litera-ture report [24]. Briey, in a 50ml round-bottomed askequipped with a reux condenser and an oil bubbler, 5 g(0.0091mol) of pre-dried mPEG 550Da was dissolved in20ml of water free toluene, 4.457 g maleic anhydride(0.04545mol, 1:5molar ratio vs. PEG) was added underthe protection of argon. The reaction mixture wasstirred at 70 1C for 48 h under an argon atmosphere.After the esterication process, toluene and the excessmaleic anhydride were eliminated by distillation andsublimation at 40 1C under vacuum. Subsequently, theintermediate mentioned above (0.54mmol) togetherwith NHS (2.7mmol) (1:5 in molar ratio) were dissolvedin 20ml DCM in a ask equipped with a magneticstirring bar. The ask was then cooled in an ice-waterbath and DCC (0.54mmol) was added under argon. Thereaction mixture was sealed under argon and stirred for1 h at 0 1C, and further 24 h at room temperature. Theprecipitated 1,3-dicyclohexylurea (DCU) was removedby ltration. The ltrate was added to diethyl ether(50ml) and cooled at 4 1C for 2 h. The precipitatedproduct was then redissolved in DCM and reprecipi-tated with diethyl ether. This procedure was repeated atleast three times to completely remove excess NHS.Finally, the product was dried under vacuum and storedat room temperature under argon. mPEG 5kDa wasactivated in the same manner.

2.4. Copolymers preparation

A predetermined quantity of TMC was dissolved inpuried water at a concentration of 10mg/ml. NHS-mPEG was dissolved in water free DMSO (50mg/ml)solution. Subsequently, the TMC solution was added tothe NHS-mPEG solution and the mixture was stirred atroom temperature for 24 h. Weight ratios of TMC toNHS-mPEG were varied in order to achieve the optimalPEGylation level. After 24 h of stirring, the solution waspuried by ultra ltration with an Amicon systemequipped with a 10 000MW cutoff membrane. Theconcentrated solution was diluted and ultra ltrated

ARTICLES. Mao et al. / Biomatagain. This procedure was repeated at least three timesand the dialyzed solution was nally freeze-dried. Thegraft ratio (wt%) of PEG was calculated from integralvalues of the characteristic peaks of PEG block at3.35 ppm (OCH3) and TMC blocks at 3.0 ppm(N (CH3)2) and 3.3 ppm (N+ (CH3)3) obtained inthe 1H NMR spectra, using the known molecular weightof mPEG. This method of comparing integration of 1HNMR peaks has been reported in the literature [25].PEGylated chitosans were prepared in the same

manner for the purpose of comparison. However, dueto its higher viscosity, concentration of chitosan400 kDa was 0.2% instead of 1% for TMC 400 kDa.The following nomenclature for the copolymers was

adopted: PEG(X)n-g-TMC/chitosan(Y), where X de-notes the MW of PEG in Da, Y denotes the MW ofTMC/chitosan in kDa, and the subscript n represents theaverage number of PEG chains per TMC/chitosanmacromolecule of Y kDa.

2.5. Characterization of polymers

Nuclear magnetic resonance spectroscopy (NMR): 1HNMR and 13C NMR spectra were recorded on a JEOLGX 400 D spectrometer operating at 400MHz at roomtemperature. Samples were dissolved in D2O or CDCl3.

Gel permeation chromatography (GPC) was carriedout using a Supremamax 3000 column (PolymerStandard Service, Mainz, Germany) with 0.3 M aceticacid and 0.2M sodium acetate as an eluent (1ml/min).Forty microliters samples were injected with a MerckHitachi autosampler AS-2000A. The samples wereanalyzed with a differential refractive index (RI)detector RI-71 from Merck.

Differential scanning calorimetry (DSC) measure-ments were performed on a PerkinElmer DSC (DSC-7) at a heating rate of 20 1C/min. The sample(approximately 5mg) was rst heated from 100 1C to170 1C. It was then quenched to 120 1C with liquidnitrogen, and heated again to 170 1C. The meltingtemperature (Tm) was determined from the endothermicpeak of the DSC curve recorded in the rst heating scan.

2.6. Water solubility testing

Solubility of the various copolymers was measured atdifferent pH values at room temperature. Briey, thecopolymers were dissolved in a 0.25% acetic acidsolution (2mg/ml), the pH value of the solution wasadjusted using 1 N NaOH and transmittance of thesolution at 600 nm as a function of pH was recorded onan UV/Vis spectrophotometer (UV-160, Shimadzu,Japan). The copolymers were considered insoluble whenthe transmittance was less than 90%, compared to thatof distilled water.

PRESS26 (2005) 63436356 63452.7. Preparation and characterization of insulin

complexes

The complexes were prepared via self-assembly usingthe electrostatic interactions between polymers and

For data analysis, the viscosity (0.88mPa s) and the

test, as recommended by USP 26. The experiment was

provided by the supplier, are listed in Table 1. Theaverage number of free amino groups in the copolymer

INerialscarried out according to the method described pre-viously [27,28]. The relative cell viability compared tocontrol cells containing cell culture medium withoutpolymer was calculated by [A]test/[A]control. Polyethylene600 (1mg/ml) and polyethylenimine (10mg/ml) wereused as positive and negative controls of the method,respectively.

2.9. LDH assay

A total of 50 000 L929 cells/well were seeded in 12-well cell culture plates. After 24 h incubation, cells werewashed with phosphate buffered solution (PBS) andwere then incubated with selected polymers (1mg/ml) in2ml PBS. One hundred microliters per well sampleswere collected at predetermined points. The LDHcontent in these samples was assayed utilizing acommercial kit according to the manufacturers proto-col, which spectrophotometrically determined theamount of reduced nicotinamide adenine dinucleotide(NAD) at 492 nm in the presence of lactate and LDH.Control experiments were performed with 0.1% (w/v)refractive index (1.33) of distilled water at 25 1C wereused. Measurements were analyzed using the CONTINalgorithm. The zeta potential measurements of thecomplexes were carried out in the standard capillaryelectrophoresis cell of the Zetasizer 3000 HS fromMalvern Instruments at 25 1C in 0.01M Tris buffer pH7.4. Values given are mean7SD (n 10).

2.8. MTT assay

A mouse connective tissue broblast cell line, L929,was selected to evaluate cytotoxicity as a direct contactinsulin. The process parameters were optimised asdescribed elsewhere in detail [26]. Briey, appropriatequantity of copolymer and insulin (1mg/ml) weredissolved in 0.1 N Tris buffer pH 7.4, respectively.Complexes were prepared by mixing equal volume ofinsulin and polymer solution under gentle magneticstirring, and incubating for a further 20min at roomtemperature. The complexes were characterized bydynamic light scattering and laser Doppler anemometry(LDA) measurements. Complex size measurements werecarried out with a Zetasizer 3000 HS from MalvernInstruments, Herrenberg, Germany (10mW HeNe laser,633 nm). Scattering light was detected at 901 anglethrough a 400 mm pin hole at a temperature of 25 1C.

ARTICLES. Mao et al. / Biomat6346Triton X-100 and set as 100% cytotoxicity. LDH releasewas calculated by the following equation:

LDH% Asample AmediumA100% Amedium 100,was measured using uorescamine assay [29].Obviously, the graft ratios increased with feed ratio

and a linear correlation between NHS-mPEG/polymerratio and graft density was found (Table 1). It isgenerally assumed that the activity of PEG is molecularweight-dependent, for example, mPEG 550Da shouldbe more reactive than mPEG 5kDa. This is the case forPEG-g-chitosan(400) copolymers, as shown in Table 1.However, the opposite is true for TMC 400 kDa, asindicated by the number of PEG chains per TMCmolecule. For instance, when the NHS-mPEG/TMC400 kDa feed ratio was 3:1, only 83 PEG 550Dawhere [A]sample, [A]medium, [A]100% denote the absor-bance of the sample, medium control and Triton X-100control, respectively. All experiments were run intriplicate.

2.10. Calculations and statistics

Results are depicted as mean7SD from at least threeseparate measurements. Signicance between the meanvalues was calculated using ANOVA one-way analysis(Origin 7.0 SRO, Northampton, MA, USA). Probabilityvalues Po0:05 were considered signicant.

3. Results

3.1. Copolymer preparation and characterization

Hydroxyl-terminated PEGs were converted to car-boxyl-terminated intermediates by esterication withcyclic aliphatic anhydride. Maleic anhydride was chosenin the present study to activate PEG in preference tosuccinic anhydride because it has been demonstrated tobe more reactive to PEG hydroxyl group, and isstraightforward to eliminate via sublimation after theesterication process [24]. After activation, PEGs werecoupled to the amino groups of TMC, forming a graftcopolymer, i.e., PEG-g-TMC. The synthetic route isshown in Scheme 1.We investigated the inuence of polymer solution pH

on the nal graft ratios, with no signicant differencebeing observed at polymer pH 5.0, 7.0, 8.5, respectively.Therefore all copolymers were prepared at a constantpolymer solution pH of 7.0. Taking TMC 400 kDa as anexample, the effect of feed ratios was investigated andthe graft ratios of the copolymers, calculated from 1HNMR spectra based on the MW of the polymer

PRESS26 (2005) 63436356molecules were coupled to TMC molecule comparedto 298mPEG 5kDa molecules at the same feedratio and prepared under the same conditions. Theresults were reproducible with different batches ofactivated PEGs, as annotated in Table 1. Furthermore,

INerialsARTICLES. Mao et al. / Biomatcopolymers were prepared with TMC 100 kDa (with thesame substitution ratio as that of TMC 400 kDa) in thesame manner and were found that the graft ratios ofPEG(550)-g-TMC(100) copolymers were similar to thatof PEG(550)-g-chitosan in the same feed ratio (data notshown). Steric effects could not be a cause for the lowgraft ratio of PEG(550)-g-TMC(400), because at the

(a)

(b)

Scheme 1. Synthetic route of PEG

Table 1

Properties of PEG-g-TMC(400) and PEG-g-chitosan(400) copolymers

Feeding

ratio

(PEG/

polymer)

(w/w)

PEG(550)-g-TMC(400) PEG(5k)-g-TMC(400)

Graft% PEG550a Free

NH2b

Graft% PEG5ka Free

NH2b

3:1 10.3e 83 553 75.4f 298 250

4:1 16.1 139 524 88.9 641 233

5:1 17.1 148 341 89.4 679 168

aNumber of PEG chains per TMC 400 kDa molecule, calculation based obNumber of free amino groups per TMC 400kDa molecule, calculation bcCalculated as a weight ratio of PEG in the graft copolymer.dNumber of free amino groups per chitosan 400 kDa molecule, calculation

PEG. Fluorescamine assay is not feasible due to the solubility limitation ofeCalculated as a weight ratio of PEG in the graft copolymer based on NM

different batches of activated PEG 550Da were investigated, the values werfReproducibility of the copolymer preparation method with different batch

77.72%, 75.44%, respectively, with SD of 1.73.PRESS26 (2005) 63436356 6347same feed ratio even higher graft ratio was obtainedwith PEG(5k)-g-TMC(400) (75.44%) compared toPEG(5k)-g-TMC(100) (68.25%) due to relative highamount of primary amino groups. Possible explanationsare competitive reactions of activated PEG 550Da withwater. PEG-NHS is moisture-sensitive and can lose itsreactivity by interacting with water. The shorter PEG

ylated chitosan derivatives.

PEG(550)-g-chitosan(400) PEG(5k)-g-chitosan(400)

Graft%c PEG550a Free

NH2d

Graft%c PEG5ka Free

NH2d

43.7 564 1547 75.2 243 1868

49.4 710 1401 82.8 386 1725

55.1 892 1219 86.6 515 1596

n 1H NMR data.

ased on uorescamine assay.

based on the substitution of chitosan 400 kDa and the graft density of

the copolymers.

R analysis. Reproducibility of the copolymer preparation method with

e 8.93%, 12.6%, 10.3%, respectively, with SD of 1.85.

es of activated PEG 5kDa were investigated, the values were 78.83%,

INerialsARTICLES. Mao et al. / Biomat6348chains were more sensitive. Although short PEG-NHSpossesses the highest reactivity to amino groups, theyalso show the highest reactivity to water. Principally,reactivity with NH2 is higher than with water.However, some substituted groups may shield part ofNH2 groups of TMC and thus grafting will becomeslow, requiring the PEG-NHS to maintain reactive overa longer period.

1H NMR spectra of a graft copolymer and corre-sponding intermediate prepolymers are shown in Fig. 1.1H NMR resonance signals are assigned accordingly.NHS-terminated mPEG 550Da (Fig. 1A(b)) showed

Fig. 1. 1H (A) and 13C NMR (B) spectra of HO-terminated mPEG 550Da

TMC(400) copolymer in D2O (c).PRESS26 (2005) 63436356resonance signals at 4.3(CH2O C (QO)), 6.2 and6.6 ppm (doublet, C (QO) CHQCHC (QO) ),which are not present in the spectrum of HO-terminatedmPEG containing no maleic moiety (Fig. 1A(a)). Onlyone signal at 6.0 (CQC) was observed in the graftcopolymer (Fig. 1A(c)), implying the environment of thedouble bond has changed, and likely the protonshielding effect of local chemical environment to thetwo methine groups become closer. Additionally, thegraft copolymer showed strong absorption at 3.0 (N(CH3)2) and 3.3 (N+(CH3)3) ppm, which are from theTMC. The structure of the copolymers was further

(a), NHS-terminated mPEG 550Da in CDCl3 (b), and PEG(5k)42-g-

between chitosan and PEG 550Da, leading to improvedmiscibility.

3.3. Water solubility of copolymers

The primary aim of our present study is to improvethe solubility of chitosan. As such, water solubility ofthe copolymers was evaluated under different pHconditions. Solutions of PEG-g-TMC copolymers re-mained transparent over the entire pH range irrespectiveof PEG MW even when the graft density was as low as10% and the concentration was as high as 50mg/ml.The PEG moieties presented on the TMC chainincreased the hydrophilicity of the copolymer, thusleading to improved solubility. In contrast, no clearsolutions were observed with PEG grafted chitosancopolymers at pH 7, despite of high graft densities. Thisis due to the strong interaction between chitosan andPEG, as demonstrated by the signicantly decreased Tmvalue in DSC experiments.

3.4. Effect of TMC MW on cytotoxicity

INerialsinvestigated with 13C NMR (Fig. 1B). These results areconsistent with the expected chemical structure of thecopolymers.GPC measurements were also carried out to verify the

successful synthesis of the graft copolymers. Copoly-mers showed unimodal molecular weight distribution inthe GPC eluograms with increased molecular weight(data not shown), indicating that negligible PEGhomopolymer was present in the copolymer products.

3.2. Thermal properties of the copolymers

The thermal properties of PEG-g-TMC(400) copoly-mers were studied using DSC. DSC thermogramsshowed an endothermic peak at its melting temperaturefor PEG and a MW dependence was observed, Tm 14.6and 63 1C for PEG 550Da and PEG 5kDa, respectively.However, no glass transition temperature was detected,probably because PEG crystallized too rapidly. Addi-tionally, no Tg and Tm were observed for both chitosan400 kDa and TMC 400 kDa.The melting temperature of PEG 5kDa in the

copolymers decreased with increasing TMC content(Table 2). For instance, Tm of homo-PEG 5kDa is65.9 1C, and only 55.5 1C in the copolymer with a graftratio of 75.44%, a decrease about 10 1C. This implies thecrystallization of PEG 5kDa was greatly affected byTMC 400 kDa.It is also noted that the decrease of Tm in PEG(5k)-g-

chitosan(400) copolymers was larger than that inPEG(5k)-g-TMC(400) copolymers with similar PEG5kDa content (Table 2). The depression of Tm can beexplained with the following equation [30]:

Tm T0m 12geDH fL

,

where ge is the free energy of chain folding at the surfaceof the lamellae, DH f is the heat of fusion of the lamellae,L is the thickness of the lamellae, and T0m is the meltingpoint of a 100% perfect polymer crystal. As we can see,the reduction of the lamellar thickness can lead to amelting point depression. The interactions betweendifferent blocks are expected to have a dominant effecton reducing L in the diblock copolymer, leading to adecreased Tm. Therefore, it is reasonable to assume thatthe interaction between PEG and chitosan is strongerthan that of PEG and TMC. This result can be used toexplain the solubility difference of the copolymers. Onthe other hand, decreased Tm indicated that those twocomponents are miscible and compatible. As to thesmall MW mPEG 550Da, the Tm change was not so

ARTICLES. Mao et al. / Biomatobvious, as summarized in Table 2. Tm of the puremPEG 550Da was observed at 14.63 1C and no Tm wasobserved in the PEG(550)-g-chitosan(400) copolymerswhen the content of mPEG 550Da was less than55.09%, implying that the crystallization of mPEG550Da was fully suppressed by chitosan 400 kDa, and adecreased Tm was identied only when PEG 550Dacontent was above 60%. In contrast, a Tm decrease in allmPEG(550)-g-TMC(400) copolymers was found. Thiscan probably be attributed to the stronger interaction

PRESS

Table 2

Melting temperature of the copolymers

Polymers Substitution

(%)aGraft ratio

(%)bTm

mPEG 550Da 14.6

mPEG 5000Da 65.9

Chitosan 400kDa c

TMC 400 kDa c

mPEG(550)-g-Chitosan(400) 26.8 43.7 c

33.7 49.4 c

42.4 55.1 12.2

mPEG(550)-g-TMC(400) 3.9 10.3 13.6

6.6 16.1 12.1

7.0 17.1 13.2

mPEG(5k)-g-Chitosan (400) 11.6 75.2 54.1

18.4 82.8 53.1

24.5 86.6 54.2

mPEG(5k)-g-TMC(400) 14.2 75.4 55.5

30.5 88.9 60.0

32.3 89.4 61.5

aCalculated based on the primary amino group content in chitosan

400 kDa, 84.7%.bCalculated as a weight ratio of PEG in the graft copolymer.cNot detected.

26 (2005) 63436356 6349MTT assays were performed to test the effects ofpolymer structure on the metabolic activity of cells. Allthe polymers showed a dose and MW dependent effecton cytotoxicity. Cell viabilities of different MW TMCsvs. concentration were investigated, and the IC50 values,

which represent concentration of the copolymers result-ing in 50% inhibition of cell growth, were calculated.With similar charge ratios, the cytotoxicity of TMCs

increased with increasing MW. TMC 400 kDa wasespecially toxic with an IC50 of 15 mg/ml. However,TMC 5kDa was shown to be completely non-toxic withan IC5041mg/ml. An exponential relationship betweenTMC MW and IC50 after 3 h incubation was estab-lished, as presented in Fig. 2, which can be used topredict the cytotoxicity of different MW TMC.

3.5. Effect of PEGylation degree on the cytotoxicity of

PEG(5k)-g-TMC(400) copolymers

Taking the extremely toxic TMC 400 kDa as a model,effect of PEGylation degree on cytotoxicity of PEG(5k)-g-TMC(400) was assessed with MTT assay. Thetheoretical molecular weights and compositions werecalculated from the degree of substitution based on 1HNMR spectra and listed in Table 3. Compared to theunmodied TMC, PEGylation decreased cytotoxicitysignicantly (Po0:05) despite of the increased MW ofthe copolymers and was a function of PEG substitutiondegree. A linear correlation between substitution degreeand IC50 value after 3 h incubation was found (Fig. 3).When the substitution degree reached 25%, the resultingdecrease in toxicity was reduced.

3.6. Effect of TMC MW in PEG(5k)-g-TMC

copolymers

As demonstrated, cytotoxicity of TMC was MW

ARTICLE IN PRESS

00

50

100

150

200

250

300

100 200 300 400

IC50

IC50 = 2195.2M-0.7349

TMC MW (KDa)

R2 = 0.9225

Fig. 2. Relationship between TMC MW and IC50 values.

Table 3

Properties and IC50 values of TMC and its derivatives (Mean7SD)

Polymer (kDa) Degree of

substitution

(%)a

TMC content

[%(w/w)]

Theoretical

MW (g/mol)

Dalton/

chargeb particle size (mv)

TMC 400 100 400000 189.4

TMC 100 100 100000 189.4

TMC 50 100 50000 189.4

TMC 25 100 25000 189.4

S. Mao et al. / Biomaterials 26 (2005) 634363566350TMC 5 100 5000 189.4

PEG(5k)298-g-

TMC(400)

12.0 22.771.7 1890000 1042

PEG(5k)640-g-

TMC(400)

25.7 11.172.1 3600000 2446

PEG(5k)680-g-

TMC(400)

27.4 10.672.5 3800000 2655

PEG(5k)40-g-

TMC(100)

6.44 32.871.0 300000 640

PEG(5k)19-g-

TMC(50)

6.13 34.270.9 145000 590

PEG(550)228-

g-TMC(100)

36.7 41.571.5 225000 750PEG(550)116-

g-TMC(50)

37.4 40.873.4 120000 930

Cell viability was quantied by MTT assay n 7.aCalculated by 1H NMR measurement.bDenoted as Dalton per charge.(nm)

3 h 24 h

30 15 256.670.8 19.672.370 22 273.374.6 22.772.590 37 236.671.3 21.771.9270 125

41000 41000220 40 171.771.9 11.270.9

370 4500 181.872.1 2.370.3

380 4500 245.672.9 1.771.0

4500 4500 233.377.2 18.870.3

4500 4500 203.977.2 13.770.3

4500 270 224.672.3 13.972.9dependent. However, whether PEGylation has the sameinuence with different MW TMCs remains unclear.Therefore, PEG(5k)-g-TMC copolymers were preparedwith similar graft ratio but different TMC MW (TMC400, 100 and 50 kDa, respectively), and the cytotoxicitywas evaluated. Copolymer properties are listed in Table3 and cytotoxicity results are shown in Fig. 4. The

IC50 of pure polymers (mg/ml) Complexes Zeta potential4500 460 220.375.5 16.471.9

inuence of TMC MW on cytotoxicity was morepronounced in copolymers. The cytotoxicity of TMC100 kDa and TMC 50kDa was decreased more than 10-fold after PEGylation with a substitution degree ofapproximately 6%, with more than 80% of the cells stillviable after 24 h incubation with a 500 mg/ml polymersolution. In contrast, PEG(5k)298-g-TMC(400) wasmore toxic compared to PEG(5k)40-g-TMC(100) andPEG(5k)19-g-TMC(50) copolymers, despite its highersubstitution degree (12.0% vs. 6%) and lower TMCcontent. Additionally, for the copolymers PEG(5k)40-g-TMC(100) and PEG(5k)19-g-TMC(50), no apparenttime and dose dependent cytotoxicity was observed,implying that they were non-toxic in the concentrationrange used in this study.

3.7. Effect of PEG MW in the copolymers

It is believed that PEG MW in the copolymers willinuence the cytotoxicity to some extent. Therefore,taking TMC100 kDa as an example, PEG(5k)40-g-

TMC(100) and PEG(550)228-g-TMC(100) copolymers,which have similar graft ratios, were synthesized and thecell viabilities were investigated with MTT assay.Results are shown in Fig. 5.After 3 h of incubation, no signicant difference in cell

viability was observed between the two copolymersaccording to two sample-paired t-test P40:05. How-ever, the difference was signicant in the concentrationrange of 10500 mg/ml after 24 h of incubation(Po0:05). Higher cell viability, especially at a concen-

decrease the cytotoxicity of the polymers, the relatively

released upon cell lysis. This assay permits the investiga-

ARTICLE IN PRESS

R2 = 0.9889

200

300

400

IC50

S. Mao et al. / Biomaterials 26 (2005) 63436356 63510

100

0 5 10 15 20 25 30Degree of Substitution (%)

Fig. 3. Correlation between degree of substitution of PEG in

PEG(5k)-g-TMC(400) copolymers and IC50 values after 3 h incubation

with the L929 cells measured by MTT assay. Each point represents the

mean7SD of seven experiments.(a) (b

Fig. 4. Effect of TMC MW on the cytotoxicity of PEG(5k)-g-TMC copolym

L929 cells. Each point represents the mean7SD of seven experiments.tion of chemicals that may induce alternations in cellintegrity. It was performed to measure the membrane-damaging effects of the copolymers via the quantity ofLDH in the culture media at different time points. Basedtoxic polymers TMC 400, 100, 50 kDa were used asexamples. Polymer concentration was kept constant at0.1mg/ml. Complexes were prepared at optimizedpolymer/insulin mass ratio 0.3:1. All of the complexesare comparable in size and carry positive charge (Table3). Effects of complexation on cell viability are shown inFig. 6. After 3 h of incubation, the effect of complexa-tion was only apparent with TMC 400 kDa. Nosignicant differences were found with TMC 100 and50 kDa P40:05. However, after 24 h incubation cellviabilities increased approximately two fold with thecomplexes for all TMCs investigated.

3.9. LDH assay

LDH is a stable enzyme present in the cytosol that istration of 500 mg/ml, was observed for PEG(5k)40-g-TMC(100) compared to that of PEG(550)228-g-TMC(100), despite the higher substitution degree ofPEG(550)228-g-TMC(100) (36.7% vs. 6.4%).

3.8. Effect of complexation with insulin

Based on the assumption that complexation may)

ers measured by MTT assay after (a) 3 h and (b) 24 h incubation with

IN

(b)

ers measured by MTT assay after (a) 3 h and (b) 24 h incubation with L929

20

30

40

50

60Ce

ll V

iabi

lity

(%)

polymer complex

erialsARTICLE

(a)

Fig. 5. Effect of PEG MW on the cytotoxicity of PEG-g-TMC copolym

cells. Each point represents the mean7SD of seven experiments.

20

30

40

50

60

Cell

Via

bilit

y (%

)

polymer complex

S. Mao et al. / Biomat6352on MTT assay, PEG(5k)40-g-TMC(100), PEG(5k)19-g-TMC(50), and PEG(550)130-g-TMC(50) copolymerswere practically non-toxic, with IC504500 mg/ml after24 h of incubation. In order to elucidate any membrane-damaging effect caused by the copolymers, theirinuence on LDH release was investigated using TMC100 kDa as a positive control, as shown in Fig. 7. After3 h of incubation, the LDH released was less than 6%for the three copolymers investigated, compared to50.573.1% for TMC 100 kDa, which is in agreementwith the results of the MTT assay.

3.10. Microscopic observations

After performing the LDH assay, changes in cellmorphology were observed using a Nikon inverse phasecontrast microscope (Nikon TMS, Nikon, Japan)equipped with an objective (Plan 10/0.30Dl/Ph1, Nikon,Japan) of 100 magnication. Fig. 8 shows a selectionof phase contrast microscopy images obtained after 3 hof incubation with 1mg/ml polymer solutions, andcompared with medium control (PBS solution). Ingeneral, L929 mouse broblasts are large, spindle-shaped, adherent cells that grow as a conuent mono-layer (Fig. 8a). Complete cell death was observed using0.1% Triton X-100 in PBS as a positive control(Fig. 8b). The cells were grainy and lacked normal

0

10

TMC400kDa TMC100kDa TMC50kDa(a) (b)

Fig. 6. Effect of complexation with insulin on the cytotoxicity of TMC (100mL929 cells. Each point represents the mean7SD of seven experiments. ThePRESS26 (2005) 63436356cytoplasmatic space, the open area between cellsindicated cell lysis had occurred. In contrast, thebroblast L929 cells incubated with copolymersPEG(5k)40-g-TMC(100), PEG(5k)19-g-TMC(50) andPEG(550)130-g-TMC(50) maintained a polygonal shapewith stretched lapodia (Fig. 8df), no cell debris, no

0

10

TMC400kDa TMC100kDa TMC50kDa

g/ml) measured by MTT assay after (a) 3 h and (b) 24 h incubation with

complexes were prepared at polymer/insulin mass ratio 0.3:1.

-20 0 20 40 60 80 100 120 140 160 180 200

0

10

20

30

40

50

60

LD

H R

elea

se (

%)

Time (min)

Fig. 7. Cytotoxicity of the TMC and its derivatives by LDH assay.

Each point represents the mean7SD of three experiments. ()TMC100 kDa, () PEG(5k)40-g-TMC(100), (m)PEG(550)228-g-TMC(100),(.)PEG(5k)19-g-TMC(50).

IN

(b

erialsARTICLE

(a)

S. Mao et al. / Biomatdetachment from dish bottom was observed, which wascomparable to that of the medium control. As acontrast, the boundary of the cells became blurry afterincubation with TMC 100 kDa and spindle shape waslost (Fig. 8c). These morphological observations wereconsistent with the results obtained from both the MTTand LDH assays.

4. Discussion

TMC and PEGylated TMC copolymers were synthe-sized and their in vitro cytotoxicity were studied with theMTT and LDH assays in the current work. All thepolymers exhibited a time- and dose-dependent cyto-

(d(c)

(e) (f

Fig. 8. Phase contrast microscopy images of L929 cells after incubation wit

cytotoxic control, the arrows showing the open area between cells indicated c

boundary of the cells became blurry and spindle shape was lost, (d) PE

TMC(100). All gures are of the same magnication (100 ).PRESS

)

26 (2005) 63436356 6353toxic response that increased with MW. PEGylationdecreased the cytotoxicity and was substitution degreedependent. Complexation with insulin decreased thecytotoxicity after 24 h incubation.In our study, the degree of quaternization of different

MW TMCs was similar (40%), implying that theactivity of primary amino groups was chitosan MWindependent. Our result is consistent with Florystheory, which suggests the intrinsic activity of allfunctional groups on a polymer remains the same [31].Generally, the determination of cell viability is an

assay to evaluate the in vitro cytotoxicity of biomater-ials. The predictive value of in vitro cytotoxicity tests isbased on the concept that toxic chemicals affect thebasic functions of cells. Such functions are common to

)

)

h different polymers (1mg/ml) for 3 h. (a) Medium control, (b) 100%

ell lysis had occurred, (c) TMC 100 kDa, the arrows indicated that the

G(5k)40-g-TMC(100), (e) PEG(5k)19-g-TMC(50), (f) PEG(550)228-g-

increasing PEG substitution degree. This was particularrelevant in the case of small MW TMC. At a similar

INerialsall cells, and hence the toxicity can be measured byassessing cellular damage. MTT and LDH assays aretwo methods commonly used for this purpose. Normallyan early indication of cellular damage is a reduction inmetabolic activity and this is the principle of MTT assay[27]. By contrast, LDH reects the damage/leakage ofplasma membranes. It has been shown that changes inmetabolic activity are superior indicators of early cellinjury, and effects on membrane integrity are indicativeof more serious damage, leading to cell death [32].Therefore, in this study, MTT assay was employed rstto evaluate the correlation between polymer structureand toxicity, and LDH assay was used for corrobora-tion. As indicated, MTT and LDH assays gave similarresults. PEG(5k)40-g-TMC(100), PEG(5k)19-g-TMC(50), and PEG(550)130-g-TMC(50) copolymersdid not induce a signicant decrease in metabolicactivity after 3 h incubation. Additionally, no signicantLDH release was measured. In contrast, a remarkableLDH release was observed with TMC 100 kDa, a toxicpolymer indicated by MTT assay using L929 mousebroblast cells, a cell line recommended by USP 26 andseveral other pharmacopoeias as a standard method forcytotoxicity testing [28]. However, using trypan blueexclusion assay, a direct measurement of cell number,Kotze et al. suggested that TMC was almost non-toxic[20,33]. This discrepancy could probably be attributed tothe different methods employed. Generally, trypan blueexclusion assay is a direct measurement of cell number,since dead cells normally detach from a culture plate,and are washed away in the medium; therefore, it cannotdifferentiate between dead cells that may have beendamaged. In general, polycations are considered to becytotoxic [34]. TMC, like most cationic macromoleculessuch as protamine and polylysine, probably interactswith anionic components (sialic acid) of the glycopro-teins on the surface of epithelial cells, causing cytotoxiceffects. Changes in cell morphology have already beenobserved with 1.0 mg/ml polylysine [35].In general, biocompatibility is inuenced by different

properties of the polymers such as MW, charge densityand type of the cationic functionalities, structure andsequence (block, linear, branched), and conformationalexibility [28]. TMC 400 kDa was found to display thehighest cytotoxicity, whilst TMC 25 kDa and TMC5kDa were almost non-toxic. An increase in cytotoxicityas a function of MW, which was observed for TMC inthis study, was also reported for other polycations, suchas DEAE-dextran [28] and PEI (polyethylenimine) [36].PEGylation decreased the cytotoxicity of TMC con-siderably, the extent of which was substitution degree,

ARTICLES. Mao et al. / Biomat6354TMCMW and PEG MW dependent. The effect of PEGcan be explained by steric effects, which acts to shield aproportion of the positive charges present on TMC, asshown in Table 3, the zeta potential of PEG(5k)-g-TMC(400) copolymer insulin complex decreased withdegree of substitution (6%), the zeta potential ofPEG(5k)19-g-TMC(50) insulin complex was 13.770.3,compared to 18.870.3 for that of PEG(5k)40-g-TMC(100) insulin complex (Table 3). However, sinceparts of the primary amino groups in TMCs weresubstituted by PEG, positive charge density decreasedoverall. When considering the cytotoxicity, PEG 5kDais preferable to PEG 550Da for PEGylation as itscomparatively long chain structure probably shields thepositive charges of TMC more efciently. This pointwas supported by the data in Table 3. The zetapotentials of PEG(550)116-g-TMC(50) copolymer insu-lin complex did not decrease signicantly even thesubstitution degree of PEG was 6 times higher (Table 3)than that of PEG(5k)19-g-TMC(50).High cytotoxicity of the PEG(5k)298-g-TMC(400)

copolymer was observed, despite the high degree ofsubstitution. This effect was probably related to thepolymers high MW. Hence, a higher substitution degreeis essential to decrease the cytotoxicity of TMC 400 kDaand 25% substitution was demonstrated to be sufcient.It should be noted here that a degree of substitution of440% is impossible to attain, since only primary aminofunction groups can participate in the reaction (40%).Complexation with insulin decreased the cytotoxicity

of TMC after 24 h of incubation with cells. Thisphenomenon can be related to the electrostatic interac-tion between TMC and insulin, which decreased theinteraction of the positively charged amino groups ofTMC with the anionic components of the glycoproteinson the cell membrane, leading to higher cell viability.Similar results were reported with PEI and galactosy-lated PEI [37].Based on studies with modied PLL (poly-L-lysine), it

was noted that macromolecules with tertiary aminegroups exhibit a lower toxicity than those with primaryand secondary residues [38]. Dekie et al. found that thepresence of primary amines had a signicant toxic effecton red blood cells with poly(L-glutamic acid) derivatives[39]. However, TMC is toxic despite of its high tertiaryamino group content (approximately 40%) and relativelow primary amine groups (approximately 40%) com-pared to chitosan (85%). In contrast, chitosan, despiteits high primary amine group content, was biocompa-tible (8). Therefore, it is clear that both the type of aminegroups and nature of the polymer inuence thecytotoxicity.PRESS26 (2005) 634363565. Conclusions

PEGylated trimethyl chitosan copolymers with vary-ing PEG molecular weight and graft ratios weresuccessfully synthesized and characterized by 1H

Eur J Pharm Sci 2001;14(3):2017.

[2] Singla AK, Chawla M. Chitosan: some pharmaceutical and

exclusively to diblock copolymers with enhanced DNA condensa-

INerialsbiological aspects-an update. J Pharm Pharmacol

2001;53:104767.

[3] Sieval AB, Thanou M, Kotze AF, Verhoef JC, Brussee J,

Junginger HE. Preparation and NMR characterization of highly

substituted N-trimethyl chitosan chloride. Carbohydrate Polym

1998;36:15765.

[4] Le Dung P, Milas M, Rinando M, Desbrieres J. Water soluble

derivatives obtained by controlled chemical modication of

chitosans. Carbohydrate Polym 1994;24:20914.

[5] Muzzarelli RAA, Tanfani F, Emanuelli M. N-(carboxymethyli-

dene) chitosans and N-(carboxymethyl) chitosans: novel chelating

polyampholytes obtained from chitosan glyoxylate. Carbohydrate

Res 1982;107:199214.

[6] Saito H, Wu X, Harris J, Hoffman A. Graft copolymers of poly

(ethylene glycol)(PEG) and chitosan. Macromol Rapid Commun

1997;18:54750.

[7] Ohya Y, Cai R, Nishizawa H, Hara K, Ouchi T. Preparation of

PEG-grafted chitosan nanoparticles as peptide drug carriers. STP

Pharma Sci 2000;10:7782.

[8] Mao S, Shuai X, Unger F, Simon M, Bi D, Kissel T. The

depolymerization of chitosan: Effects on physicochemical and

biological properties. Int J Pharm 2004;281:4554.

[9] Thanou MM, Kotze AF, Scharringhausen T, Luessen HL, deReferences

[1] Van der Lubben IM, Verhoef JC, Borchard G, Junginger HE.

Chitosan and its derivatives in mucosal drug and vaccine delivery.NMR, 13C NMR, GPC and DSC measurements.Decreased melting temperature of PEG in the copoly-mers indicated that the two blocks are miscible andcompatible. PEG-g-TMC copolymers were completelywater-soluble over the entire pH range, irrespective ofPEG MW. Cytotoxicity of TMC and the copolymerswas investigated with the MTT and LDH assays, whichallow the quantication of the cell metabolic activityand membrane integrity, respectively. Similar resultswere obtained with the two methods, which were in turnin agreement with observations from inverted phasecontrast microscope images. MW dependent cytotoxi-city was observed for TMC, and PEGylation led toincreased biocompatibility. PEG 5kDa is preferable toPEG 550Da for efcacious PEGylation. Complexationwith insulin decreased the toxicity of TMCs after 24 hincubation. These insights could be particular helpfuldue to the promising application of TMC as drugdelivery vehicles.

Acknowledgements

Shirui Mao cordially thanks DAAD (DeutscheAkademische Austauschdienst) for the nancial support.

ARTICLES. Mao et al. / BiomatBoer AG, Verhoef JC, Junginger HE. Effect of degree of

quaterization of N-trimethyl chitosan chloride for enhanced

transport of hydrophilic compounds across intestinal Caco-2 cell

monolayers. J Controlled Rel 2000;64:1525.

[10] Van der Merwe SM, Verhoef JC, Verheijden JH, Kotze AF,

Junginger HE. Trimethylated chitosan as polymeric absorptionenhancer for improved peroral delivery of peptide drugs. Eur J

Pharm Biopharm 2004;58(2):22535.

[11] Van der Merwe SM, Verhoef JC, Kotze AF, Junginger HE. N-

trimethyl chitosan chloride as absorption enhancer in oral peptide

drug delivery. Development and characterization of minitablet

and granule formulations. Eur J Pharm Biopharm

2004;57(1):8591.

[12] Hamman JH, Schultz CM, Kotze AF. N-trimethyl chitosan

chloride: optimum degree of quaternization for drug absorption

enhancement across epithelial cells. Drug Dev Ind Pharm

2003;29(2):16172.

[13] Jonker C, Hamman JH, Kotze AF. Intestinal paracellular

permeation enhancement with quaternised chitosan: in situ and

in vitro evaluation. Int J Pharm 2002;238(12):20513.

[14] Di Colo G, Burgalassi S, Zambito Y, Monti D, Chetoni P. Effect

of different N-trimethyl chitosans on in vitro/in vivo ooxacin

transcorneal permeations. J Pharm Sci 2004;93(11):285162.

[15] Hamman JH, Stander M, Kotze AF. Effect of the degree of

quaternization of N-trimethyl chitosan chloride on absorption

enhancement: in vivo evaluation in rat nasal epithelia. Int J

Pharm 2002;232(12):23542.

[16] Snyman D, Hamman JH, Kotze AF. Evaluation of the

mucoadhesive properties of N-trimethyl chitosan chloride. Drug

Dev Ind Pharm 2003;29(1):619.

[17] Thanou M, Florea BI, Geldof M, Junginger HE, Borchard G.

Quaternized chitosan oligomers as novel gene delivery vectors in

epithelial cell lines. Biomaterials 2002;23:1539.

[18] Jansma CA, Thanou M, Junginger HE, Borchard G. Preparation

and characterization of 6-O-carboxymethyl-N-trimethyl chitosan

derivative as a potential carrier for targeted polymeric gene and

drug delivery. STP Pharma Sci 2003;13(1):637.

[19] Thanou MM, Verhoef JC, Romeijn SG, Nagelkerke JF, Merkus

FW, Junginger HE. Effects of N-trimethyl chitosan chloride, a

novel absorption enhancer, on caco-2 intestinal epithelia and the

ciliary beat frequency of chicken embryo trachea. Int J Pharm

1999;185(1):7382.

[20] Kotze AF, Thanou MM, Lueben HL, De Boer AG, Verhoef JC,

Junginger HE. Enhancement of paracellular drug transport with

highly quaternized N-trimethyl chitosan chloride in neutral

environments: In vitro evaluation in intestinal epithelial cells

(Caco-2). J Pharm Sci 1999;88:2537.

[21] Ogris M, Brunner S, Schuller S, Kircheis R, Wagner E.

PEGylated DNA/transferring-PEI complexes: reduced interaction

with blood components, extended circulation in blood and

potential for systemic gene delivery. Gene Therapy

1999;6:595605.

[22] Veronese FM. Peptide and protein PEGylation: a review of

problems and solutions. Biomaterials 2001;22:40517.

[23] Hamman JH, Schultz CM, Kotze AF. N-trimethyl chitosan

chloride: optimum degree of quaternization for drug absorption

enhancement across epithelial cells. Drug Dev Ind Pharm

2003;29:16172.

[24] Shuai X, Jedlinski Z, Luo Q, Farhod N. Synthesis of novel block

copolymers of poly(3-hydroxybutyric acid) with poly(ethylene

glycol) through anionic polymerisation. Chinese J Polym Sci

2000;18:1923.

[25] Petersen H, Martin AL, Stolnik S, Roberts CJ, Davies MC, Kissel

T. The macrostopper route: A new synthesis concept leading

PRESS26 (2005) 63436356 6355tion potential. Macromolecules 2002;35:98546.

[26] Mao S, Kissel T. Nanocomplex formation between chitosan

derivatives and insulin: Effect of pH, polymer structure and

molecular weight. CRS Meeting in Heidelberg, Germany; April

2004.

[27] Mosmann T. Rapid colorimetric assays for cellular growth and

survival: application to proliferation and cytotoxicity assays. J

Immunol Methods 1983;65:5563.

[28] Fischer D, Li Y, Ahlemeyer B, Krieglstein J, Kissel T. In vitro

cytotoxicity testing of polycations: inuence of polymer structure

on cell viability and hemolysis. Biomaterials 2003;24:112131.

[29] Read ML, Etrych T, Ulbrich K, Seymour LW. Characterisation

of the binding interaction between poly(L-lysine) and DNA using

the uorescamine assay in the preparation of non-viral gene

delivery vectors. FEBS Lett 1999;461:96100.

[30] Wei M, Shuai X, Tonelli AE. Melting and crystallization

behaviors of biodegradable polymers enzymatically coalesced

from their cyclodextrin inclusion complexes. Biomacromolecules

2003;4:78392.

[31] Flory PJ. Principles of polymer chemistry. New York: Cornell

University Press; 1953. p. 69.

[32] Davila JC, Reddy CG, Davis PJ, Acosta D. Toxicity assessment

of papaverine hydrochloride and papaverine-derived metabolites

in primary cultures of rat hepatocytes. In Vitro Cell Dev Biol

1990;26:51524.

[33] Kotze AF, Luessen HL, de Leeuw BJ, de Boer BG, Coos Verhoef

J, Junginger HE. Comparison of the effect of different chitosan

salts and N-trimethyl chitosan chloride on the permeability of

intestinal epithelial cells (Caco-2). J Controlled Rel

1998;51:3546.

[34] Morgan DM, Larvin VL, Pearson JL. Biochemical characteriza-

tion of polycation-induced cytotoxicity to human vascular

endothelial cells. J Cell Sci 1989;94:5539.

[35] Quinton PM, Phillpot CW. A role for anionic sites in epithelial

architecture. Effect of cationic polymers on cell membrane

structure. J Cell Biol 1973;56:78796.

[36] Fischer D, Bieber T, Li Y, Elsasser HP, Kissel T. A novel non-

viral vector for DNA delivery based on low molecular weight,

branched polyethylenimine: effect of molecular weight on

transfection efciency and cytotoxicity. Pharm Res

1999;19:12739.

[37] Kunath K, von Harpe A, Fischer D, Kissel T. Galactose-

PEIDNA complexes for targeted gene delivery: degree of

substitution affects complex size and transfection efciency. J

Controlled Rel 2003;88:15972.

[38] Ferruti P, Knobloch S, Ranucci E, Gianasi E, Duncan R. A novel

chemical modication of poly-L-lysine reducing toxicity while

preserving cationic properties. Proc Int Symp Control Rel Bioact

Mater 1997;24:456.

[39] Dekie L, Toncheva V, Dubruel P, Schacht EH, Barrett L,

Seymour LW. Polyl-glutamic acid derivatives as vectors for gene

therapy. J Controlled Rel 2000;65:187202.

ARTICLE IN PRESSS. Mao et al. / Biomaterials 26 (2005) 634363566356

Synthesis, characterization and cytotoxicity of poly(ethylene glycol)-graft-trimethyl chitosan block copolymersIntroductionMaterials and methodsMaterialsSynthesis of TMC with different MWActivation of mPEGCopolymers preparationCharacterization of polymersWater solubility testingPreparation and characterization of insulin complexesMTT assayLDH assayCalculations and statistics

ResultsCopolymer preparation and characterizationThermal properties of the copolymersWater solubility of copolymersEffect of TMC MW on cytotoxicityEffect of PEGylation degree on the cytotoxicity of PEG(5k)-g-TMC(400) copolymersEffect of TMC MW in PEG(5k)-g-TMC copolymersEffect of PEG MW in the copolymersEffect of complexation with insulinLDH assayMicroscopic observations

DiscussionConclusionsAcknowledgementsReferences