Chitosan-g-PEG/DNA complexes deliver gene to the rat liver via intrabiliary and intraportal...

THE JOURNAL OF GENE MEDICINE R E S E A R C H A R T I C L EJ Gene Med 2006; 8: 477–487.Published online 3 January 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jgm.868

Chitosan-g-PEG/DNA complexes deliver gene to therat liver via intrabiliary and intraportal infusions

Xuan Jiang1,2

Hui Dai3

Kam W. Leong3,4

Suat-Hong Goh2

Hai-Quan Mao3,5*Yi-Yan Yang1**

1Institute of Bioengineering andNanotechnology, Singapore 138669,Singapore2Department of Chemistry, NationalUniversity of Singapore, Singapore119260, Singapore3Division of Johns Hopkins inSingapore, Singapore 138669,Singapore4Department of BiomedicalEngineering, School of Medicine,Johns Hopkins University, Baltimore,MD 21205, USA5Department of Materials Science andEngineering, Johns HopkinsUniversity, Baltimore, MD 21218,USA

*Correspondence to: Hai-Quan Mao,102 Maryland Hall, 3400 N. CharlesStreet, Baltimore, MD 21218, USA.E-mail: [email protected]

**Correspondence to: Yi-Yan Yang,IBN, 31 Biopolis Way, The Nanos,#04-01, Singapore138669, Singapore. E-mail:[email protected]

Received: 17 January 2005Revised: 19 October 2005Accepted: 19 October 2005

Abstract

Background Chitosan has been shown to be a non-toxic and efficient vectorfor in vitro gene transfection and in vivo gene delivery through pulmonaryand oral administrations. Recently, we have shown that chitosan/DNAnanoparticles could mediate high levels of gene expression followingintrabiliary infusion [1]. In this study, we have examined the possibility ofusing polyethylene glycol (PEG)-grafted chitosan/DNA complexes to delivergenes to the liver through bile duct and portal vein infusions.

Methods PEG (Mw: 5 kDa) was grafted onto chitosan (Mw: 47 kDa,deacetylation degree: 94%) with grafting degrees of 3.6% and 9.6% (molarpercentage of chitosan monosaccharide units grafted with PEG). The stabilityof chitosan-g-PEG/DNA complexes was studied by measuring the changein particle size and by agarose gel electrophoresis against bile or serumchallenge. The influence of PEG grafting on gene transfection efficiencywas evaluated in HepG2 cells using luciferase reporter gene. Chitosan andchitosan-g-PEG/DNA complexes were delivered to the liver through bile ductand portal vein infusions with a syringe pump. Gene expression in the liverand the distribution of gene expression in other organs were evaluated.The acute liver toxicity of chitosan and chitosan-g-PEG/DNA complexes wasexamined by measuring serum alanine aminotranferase (ALT) and aspartateaminotransferase (AST) activities as a function of time.

Results Both chitosan and chitosan-g-PEG displayed comparable genetransfection efficiency in HepG2 cells. After challenge with serum and bile,chitosan-g-PEG/DNA complexes, especially those prepared with chitosan-g-PEG (GD = 9.6%), did not form large aggregates like chitosan/DNAcomplexes but remained stable for up to 30 min. In addition, chitosan-g-PEG prevented the degradation of DNA in the presence of serum and bile. Onday 3 after bile duct infusion, chitosan-g-PEG (GD = 9.6%)/DNA complexesmediated three times higher gene expression in the liver than chitosan/DNAcomplexes and yielded background levels of gene expression in other organs.On day 1 following portal vein infusion, gene expression level induced bychitosan/DNA complexes was hardly detectable but chitosan-g-PEG (GD =9.6%) mediated significant transgene expression. Interestingly, transgeneexpression by chitosan-g-PEG/DNA complexes in other organs after portalvein infusion increased with increasing grafting degree of PEG. The ALT andAST assays indicated that grafting of PEG to chitosan reduced the acute livertoxicity towards the complexes.

Conclusion This study demonstrated the potential of chitosan-g-PEG as asafe and more stable gene carrier to the liver. Copyright 2006 John Wiley& Sons, Ltd.

Keywords gene delivery; PEGylation; chitosan; bile duct; portal vein; livertargeting

Copyright 2006 John Wiley & Sons, Ltd.

478 X. Jiang et al.

Introduction

Liver-targeted gene delivery has great potential in treatingmany hepatic diseases. Over the past decade, genetransfer with viral and non-viral vectors has beenextensively attempted for liver-targeted gene therapy.Viral-based vectors have been shown to be moreefficient in mediating gene transfer to hepatocytesboth in vitro and in vivo; however, the safety issuesassociated with viral vectors have seriously hamperedtheir clinical applications. Non-viral vectors such ascationic polymers have been increasingly proposed assafer alternatives to viral vectors because of theirlow immune response, targetability, and the ease ofsynthesis and scaling up [2,3]. A number of deliverysystems have been developed to deliver genes to theliver. Galactosylated cationic liposomes, galactosylatedpolylysine (PLL) and galactosylated polyethyleneimine(PEI) [4–6] have been synthesized and evaluated asliver-targeted gene carriers, since the galactose moietycan be specifically recognized by asialoglycoproteinreceptors on hepatocytes. After intravenous injection,these vectors were preferentially taken up by liverparenchymal cells. On the other hand, the optimization ofadministration routes such as infusion through portalvein and bile duct has also been shown to enhancegene transfection in the liver [7–10]. In particular, thebile duct route provides many advantages over systemicadministrations [1]. It directly delivers complexes to theliver parenchymal cells, bypassing the Kupffer cells. Thisnot only improves the delivery efficiency to hepatocytes,but also avoids the opsonization of the particles inserum. It will lead to a directed gene expression inthe hepatocytes. Furthermore, bile duct injection can beeasily adopted in clinical settings through endoscopicretrograde cholangiopancreatography (ERCP), a routinebile duct canulation procedure [9,11]. Therefore, liver-targeted gene delivery through the bile duct will havewide implications for treating inherited and metabolicliver diseases, liver cancer, viral hepatitis (e.g. interferonα and γ delivery), and systemic diseases, like hemophiliaA and B (factor VIII and IX delivery).

Chitosan is a natural biodegradable polymer extractedfrom crustacean shells. It has been investigated as a non-toxic gene carrier for in vitro gene transfection [12–17].Successful in vivo gene transfection has been demon-strated following oral feeding [18–20] and pulmonaryadministration [21]. More recently, we have shown thathigh levels of gene expression in rat liver can be achievedby infusion of chitosan/DNA nanoparticles through thebile duct [1]. The same nanoparticles failed to elicitany gene transfer in the liver by portal vein infusion.This is primarily due to the aggregation of chitosan/DNAnanoparticles in the serum [17]. The size of aggregatescould reach up to microns, preventing the penetrationof chitosan/DNA nanoparticles through the fenestrae(150–170 nm in diameter) of the sinusoidal endothelium[22] to reach hepatocytes in the case of portal vein admin-istration. Bile contains mainly cholesterol, phospholipids

and bile salts at a concentration of 1–1.5% (w/w). Pro-teins are also rich in the bile, comprising 5% (w/w) of thetotal bile constituents. Chitosan was reported to be ableto bind bile acids and cholesterols in the bile [23]. Theadsorption of bile acids, cholesterol and bile proteins mayalso result in aggregation of chitosan/DNA complexes,hampering the entry of chitosan/DNA complexes intohepatocytes through bile canaliculi (0.5–1.2 µm) [24] inthe case of bile duct infusion. This has been confirmedby our recent report showing that gene expression in theliver is localized to the region in the proximity to theinjection point through intrabiliary infusion [1]. PEGyla-tion of proteins, drugs, and liposomes has been proven tobe an effective approach in extending circulation in theblood stream [25]. Similarly, cationic polymers for genedelivery like PLL and PEI were grafted with polyethy-lene glycol (PEG) to form block copolymers [26–30]. Thesteric shield provided by highly hydrophilic PEG layerseffectively prevented aggregation of the polymer/DNAcomplexes and reduced interaction with plasma proteins.Moreover, PEG-grafted PEI/oligodeoxynucleotide com-plexes demonstrated prolonged blood circulation time andslower uptake by the reticuloendothelial system [29,30].In this study, we grafted PEG onto the chitosan polymerchain to reduce the aggregation of complexes. PEG wasconjugated to chitosan at two different grafting degrees(GD). The effect of PEG grafting and GD on the stabil-ity of the complexes, and in vitro transfection efficiency,were investigated. The in vivo gene transfection efficiencyof PEG-grafted chitosan/DNA complexes was analyzed inrats through bile duct and portal vein infusions. Transgeneexpression by chitosan-g-PEG/DNA complexes in varioustissues and its acute liver toxicity were evaluated in ratsin comparison with chitosan/DNA complexes.

Materials and methods

Materials

Chitosan (Mw: 47 kDa, deacetylation degree: 94%) wasa gift from Prof. W. J. Li, Department of MolecularSciences, Fudan University (Shanghai, China). It waspurified prior to modification. Briefly, chitosan was firstdissolved in 2% acetic acid solution and filtered through a0.45 µm membrane, and then precipitated in 1 M sodiumhydroxide solution. The precipitate was filtered, washedwith distilled water, and dried under vacuum at 60 ◦C.Methoxy PEG-nitrophenol carbonate (Mw: 5 kDa) waspurchased from Sunbio, Korea. PEI (branched, 25 kDa),sodium acetate, acetic acid, dimethyl sulfoxide (DMSO),ethidium bromide (EtBr) and other chemicals if notspecified were obtained from Sigma-Aldrich Singapore.Fetal calf serum (FCS) was purchased from BiologicalIndustries, Israel. Plasmid DNA encoding the 6.4 kbfirefly luciferase (pCMV-luciferase VR1255C) driven bythe cytomegalovirus (CMV) promoter was a gift from CarlWheeler, Vical (San Diego, CA, USA). The plasmid wastransformed in Escherichia coli DH5α and amplified in

Copyright 2006 John Wiley & Sons, Ltd. J Gene Med 2006; 8: 477–487.

Chitosan-g-PEG/DNA Complexes for Gene Delivery 479

Terrific broth media at 37 ◦C with stirring overnight. Theamplified plasmid DNA was purified by an endotoxin-free Giga plasmid purification kit (Qiagen, Valencia,CA, USA). Purified plasmid DNA was dissolved in Tris-EDTA (TE) buffer, and its purity and concentration weredetermined by ultraviolet (UV) absorbance at 260 and280 nm. p43-clz1 plasmid DNA (7.8 kb) encoding theβ-galactosidase gene driven by the CMV promoter waspurchased from Aldevron LLC (Fargo, ND, USA). Thehuman hepatocellular carcinoma cell line (HepG2 cells)was purchased from ATCC (Manassas, VA, USA).

Preparation of chitosan-g-PEG

Synthesis of chitosan-g-PEG is shown in Scheme 1.A chitosan solution (5 mg/mL) was prepared in 2%acetic acid with gentle heating and the pH of thesolution was adjusted to 6.0 with sodium hydroxide.The chitosan solution and methoxy PEG-nitrophenolcarbonate solution (15 mg/mL in DMSO) were purgedwith nitrogen to remove oxygen. A volume of 15 or30 mL of methoxy PEG-nitrophenol carbonate solutionwas added to 20 mL of chitosan solution. The mixturewas incubated at 4 ◦C for 2 days. The chitosan solutionwas dialyzed (MWCO = 12 000) against deionized waterat room temperature for 2 days to remove unreactedmethoxy PEG-nitrophenol carbonate and nitrophenol.The dialyzed solution was freeze-dried. The productwas further washed with acetone to remove residues ofnitrophenol and then dried in an oven overnight at 60 ◦C.PEG-grafted chitosan (chitosan-g-PEG) was characterizedby 1H nuclear magnetic resonance (NMR) spectroscopy(model AV400, Bruker 400 MHz) at 70 ◦C using 2% (w/w)of CD3COOD/D2O as solvent.

Preparation of chitosan/DNA andchitosan-g-PEG/DNA complexes

Chitosan or chitosan-g-PEG was dissolved in 25 mMsodium acetate buffer (pH 5.5) at a concentration of0.5 to 3.0 mg/mL. Plasmid DNA solution of 100 µg/mLwas prepared in the buffer separately. Equal volumes of

the two solutions were mixed and vortexed for 10 s. Theresulting complexes were incubated at room temperature(22 ◦C) for 30 min prior to further analyses.

Competition binding assay

Plasmid DNA (20 µg/mL) was stained with EtBr(400 ng/mL) in 25 mM sodium acetate buffer (pH 5.5).An increasing amount of chitosan or chitosan-g-PEG solu-tion was added gradually to the stained DNA solution.The fluorescence intensity of the solution was mea-sured using a microplate reader (SpectraMax, MolecularDevices, Sunnyvale, CA, USA) with excitation and emis-sion wavelengths of 366 and 590 nm, respectively.

Size and zeta potential of thecomplexes and their stability in serumand bile

The size and zeta potential of chitosan/DNA andPEG-g-chitosan/DNA complexes were analyzed usinga Zetasizer 3000 HAS (Malvern Instruments, South-borough, MA, USA). The complexes were preparedat different N/P ratios, and the DNA concentrationswere fixed at 50 and 10 µg/mL for size and zetapotential measurements, respectively. At each N/Pratio, the complexes were prepared in triplicate andmeasured. The average of these measurements wasused.

The physical stability of the complexes in serum andbile was examined by measuring size change and usingthe agarose gel retardation assay after challenge withserum and bile. Bile was produced from the bile ductof Wistar rats and stored at −80 ◦C. For size analysis,0.1 mL of FCS or rat bile was added to 0.9 mL ofthe complexes containing 50 µg of plasmid DNA. Theaverage size of the complexes was measured as afunction of time at room temperature. For the agarosegel retardation assay, equal volumes of the complexescontaining 50 µg/mL of DNA and FCS or rat bilewere mixed and incubated at 37 ◦C. At predeterminedtime intervals, 10 µL of the complexes were collected

Scheme 1. Sythesis of chitosan-g-PEG polymers


480 X. Jiang et al.

and loaded into individual wells of a 0.8% agarosegel, and electrophoresed at 80 V for 40 min. Thegel was stained with EtBr and visualized on a UVilluminator (Eagle Eye II, Stratagene, La Jolla, CA,USA).

In vitro gene transfection

In vitro gene transfection studies were performed inHepG2 cells. The cells were maintained in Dulbecco’smodified Eagle’s medium (DMEM) supplemented with10% FCS at 37 ◦C, 5% CO2. The cells were seeded onto24-well plates at a density of 6 × 104 cells per well andincubated for 1 day. The culture media were replacedwith fresh media, and the complexes containing 2 µgof plasmid DNA were added to each well. After 4 h ofincubation, the culture media were replaced with DMEMcontaining 10% FCS. After 44 h, the culture media wereremoved, and the cells were washed with 0.5 mL ofphosphate-buffered saline (PBS, pH 7.4). Then, 0.2 mLof reporter lysis buffer (Promega, Madison, WI, USA) wasadded to each well to lyse the cells. The cell suspensionwas next subjected to two cycles of freezing and thawing,and was centrifuged at 14 000 rpm for 5 min. A volumeof 20 µL of the supernatant of the cell lysate wasmixed with 100 µL of luciferase substrate (Promega).The relative light units (RLU) were measured using aluminometer (Lumat LB9507, Berthold, Germany). Theluciferase activity was converted into the amount ofluciferase using pure, recombinant luciferase (Promega)as the standard, and normalized to protein content usingthe BCA protein assay (Bio-Rad Laboratories, Hercules,CA, USA).

In vivo gene transfection

The animal protocols were approved by the AnimalCare and Use Committee, Faculty of Medicine, NationalUniversity of Singapore. Male Wistar rats aged 6–8 weeks(200–300 g, three per group) were anesthetized. Thelivers were separated from the surrounding tissue. A33-gauge needle was inserted into the common bile ductand a tie was used to secure the needle. PEI/plasmidDNA complexes (200 µg) were prepared in PBS solutionwith an N/P ratio of 10 : 1. Chitosan, chitosan-g-PEG(GD = 3.6%) and chitosan-g-PEG (GD = 9.6%)/plasmidDNA (200 µg) complexes were prepared in 25 mM sodiumacetate buffer (pH 5.5) with an N/P ratio of 3 : 1 or5 : 1. Then, 4 mL of the complexes were administeredthrough the needle over 20 min (0.2 mL/min) using asyringe pump. A tie was then placed around the bile ductbetween the liver and the point of infusion to preventback flow, and the needle was withdrawn. After 30 min,all ties were removed. A single 10-O nylon (Ethicon,Somerville, NJ) stitch was used to repair the needlehole in the bile duct to prevent bile leakage whennecessary.

Portal vein infusion was performed as reported byUehara et al. [31]. Briefly, the portal vein was tied and4 mL of PEI, chitosan or chitosan-g-PEG/DNA complexeswere infused over 4 min using a syringe pump and a33-gauge needle into the portal vein. Fifteen minutesafter the portal vein had been tied, the tie was removedto restore blood flow.

One or three days after administration, three ratsfrom each group were sacrificed, and major organs(liver, heart, lung, spleen and kidney) were harvested,weighed, and homogenized with a tissue homoge-nizer (Heidolph, Schwabach, Germany). The homogenatewas subjected to two cycles of freezing and thaw-ing, and centrifuged at 14 000 rpm for 5 min at4 ◦C. Then, 20 µL of the supernatant were mixedwith 100 µL of luciferase substrate (Promega). RLUwere measured with a luminometer (Lumat LB9507,Berthold, Oak Ridge, TN, USA). The luciferase activ-ity was converted into the amount of luciferase usingpure, recombinant luciferase (Promega) as the stan-dard.

Detection of β-galactosidase geneexpression

A volume of 4 mL of PBS or chitosan-g-PEG (GD =9.6%)/p43-clz1 plasmid DNA (200 µg) complexes at anN/P ratio of 5 : 1 was administrated through the bileduct as described above. Two days after infusion, ratlivers were harvested and frozen using liquid nitrogen.Cryostat sections of 5–8 mm were cut from differentlobes of the liver. The sections were then fixed in PBScontaining 0.5% glutaraldehyde at 4 ◦C for 15 min andwashed twice with PBS. The sections were stained witha LacZ reporter assay kit (tissue staining) (Invivogen,USA) at 37 ◦C for 3 h. The stained sections werewashed twice with PBS, counterstained with 0.1%(w/v) Nuclear Fast Red for 3 min, then dehydrated andmounted.

Liver function tests

Blood samples were collected from tail veins of the ratson days 1, 2 and 3 after bile duct or portal vein infusionand the rats in the control group. Blood samples wereincubated at 4 ◦C for 4 h and centrifuged at 2000 rpm for15 min. Serum was isolated and stored at −80 ◦C. Liverfunctions including ALT (alanine aminotranferase) andAST (aspartate aminotransferase) activities were testedby the Clinical Chemistry Laboratory at the NationalUniversity Hospital, Singapore.

Statistical analysis

In vivo gene expression data were analyzed by analysisof variance and Student’s t test. Data were consideredstatistically significant if p values were <0.05.



0

0.2

0.4

0.6

0.8

1

1.2

0 1.0 2.0 3.0 4.0 5.0

Chitosan

Chitosan-g-PEG(3.6%)


Rel

ativ

e F

luo

resc

ence

Inte

nsi

ty

N/P ratio

Figure 1. Ethidium bromide (EtBr) exclusion from chitosan/DNAor chitosan-g-PEG/DNA complexes. Various amounts of chitosanor chitosan-g-PEG solution were added to 25 mM sodium acetatebuffer (pH 5.5) containing 20 µg/mL of DNA and 400 ng/mL ofEtBr. Fluorescence intensity (λex = 366 nm, λem = 590 nm) wasmeasured after each addition. Relative fluorescence intensitywas calculated as the ratio of the fluorescence intensity ofthe solution after the addition of polymer to the fluorescenceintensity of DNA/EtBr solution without polymer

Results

Effect of PEG on DNA-binding ability ofchitosan-g-PEG

Competition assays were conducted to investigate theeffect of PEG conjugation on the DNA-binding ability ofchitosan-g-PEG. Formation of well-condensed complexesof DNA with chitosan will block access of EtBr toits binding sites on DNA. Intercalated EtBr is stronglyfluorescent but the fluorescent intensity of free EtBris weak, allowing for observation of the degree ofDNA condensation by chitosan. As shown in Figure 1,EtBr fluorescence decreased with increasing chitosanor chitosan-g-PEG content in the complexes. The DNA-binding capacity of chitosan-g-PEG was similar to that ofchitosan, yielding the maximum reduction in fluorescenceintensity at an N/P ratio of about 2.

Particle size and surface charge ofchitosan and chitosan-g-PEG/DNAcomplexes

The particle size of chitosan-g-PEG/DNA complexes isan important factor that influences the access andpassage of the complexes through both bile canaliculiand discontinuous fenestraes after bile and portal veininfusions, respectively. Figure 2a shows that the averagesize of chitosan/DNA complexes decreased rapidly from235 to 147 nm with increasing N/P ratio from 1 to3, whereas the average size of chitosan-g-PEG/DNAcomplexes did not change significantly as the N/P ratioincreased from 1 to 8.

Surface charge of polymer/DNA complexes plays animportant role in non-specific uptake of the complexes

0

100

200

300

400

500

0 4 6 8 10

N/P ratio

Mea

n P

arti

cle

Siz

e (n

m) Chitosan



-50-40-30-20-10

01020304050

2 6 10

N/P ratio

Zet

a P

ote

nti

al (

mV

)Chitosan



(a)

(b)

2

4 8

Figure 2. Effect of N/P ratio on the average particle size(a) and zeta potential (b) of chitosan and chitosan-g-PEG/DNAcomplexes. Chitosan and chitosan-g-PEG/DNA complexes wereprepared in 25 mM sodium acetate buffer (pH 5.5) and theconcentrations of DNA were 50 and 10 µg/mL for particle sizeand zeta potential analyses, respectively

by cells and their physical stability in blood. Serumprotein tends to adsorb strongly onto positively chargedparticle surfaces. This opsonization will lead to severeaggregation of the complexes, thereby reducing theirability to transfer through fenestrae or bile canaliculi. Inaddition, complexes opsonized with plasma proteins aremore susceptible to clearance by the reticuloendothelialsystem [25]. Figure 2b shows the zeta potentials ofchitosan and chitosan-g-PEG/DNA complexes at variousN/P ratios. At an N/P ratio of 1, chitosan and chitosan-g-PEG/DNA complexes had a negatively charged surface,suggesting that the amount of chitosan or chitosan-g-PEG was not enough to condense the DNA moleculescompletely. When the N/P ratio was increased to 3,the zeta potential of all the complexes turned positive.Grafted PEG molecules partially shielded the surfacecharge of the complexes and thus yielded lower surfacecharge as compared to chitosan. Both chitosan-g-PEG(GD = 3.6%) and chitosan-g-PEG (GD = 9.6%) yieldedthe same surface charges at N/P ratios higher than 3.

Gene transfection in HepG2 cells

Both chitosan and chitosan-g-PEG (GD = 3.6%) mediatedthe highest transfection efficiency at an N/P ratioof 3 and the transfection efficiency of chitosan-g-PEG/DNA complexes at this N/P ratio was comparable to


482 X. Jiang et al.

10-2

10-1

1

101-

102

103

104

N/P=3 N/P=8

N/P ratio

Luci

fera

se E

xpre

ssio

n(p

g/ m

g pr

otei

n)

N/P=2 N/P=5

Chitosan



PEI

Figure 3. Gene transfection efficiency of chitosan and chitosan-g-PEG/DNA complexes in HepG2 cells. Chitosan andchitosan-g-PEG/DNA complexes were prepared in 25 mM sodium acetate buffer (pH 5.5) and the concentration of DNA was50 µg/mL. Complexes were incubated with HepG2 cells for 4 h and luciferase activity was measured 48 h after transfection

that of unmodified chitosan/DNA complexes (Figure 3).However, their highest transfection efficiency was still1000-fold lower than the PEI control. Chitosan-g-PEG(GD = 9.6%) yielded much lower gene transfectionefficiency and the highest gene transfection efficiencywas achieved at an N/P ratio of 5. Probably, it isbecause chitosan-g-PEG (GD = 9.6%)/DNA complexeswith denser PEG layer were more resistant to celluptake compared to chitosan-g-PEG (GD = 3.6%)/DNAcomplexes, resulting in lower gene expression.

Stability of the nanoparticles in thepresence of serum and bile

The purpose of PEG grafting on chitosan was toreduce the aggregation and increase the stability ofchitosan/DNA complexes in bile and serum. Effects ofserum and bile on the stability of chitosan-g-PEG/DNAcomplexes were examined by particle size analysis. Uponthe challenge with 10% FCS, chitosan/DNA complexesstarted to aggregate rapidly (Figure 4a). After 10 min,complexes with a size below 1000 nm were no longerdetectable (Figure 4b). However, for chitosan-g-PEG(GD = 3.6%)/DNA complexes, the average particle sizedropped slightly to about 100 nm after being challengedwith serum and remained the same after 30 min.Chitosan-g-PEG (GD = 9.6%)/DNA complexes were themost stable and the change in particle size in 10% FCSwas negligible for the examining period.

The addition of rat bile also resulted in rapidaggregation of chitosan/DNA complexes (Figure 4c).After 5 min, more than 70% of complexes had a sizeabove 1000 nm and the percentage increased to 90% after

30 min (Figure 4d). As in the serum test, chitosan-g-PEG(GD = 3.6% or 9.6%)/DNA complexes were stable afterthe addition of rat bile. They did not show any significantchange in size. Particles with a size above 1000 nm werenot detectable.

It has been demonstrated that serum nucleases in bloodand biliverdin in bile cause rapid degradation of plasmidDNA [32,33]. The stability of chitosan and chitosan-g-PEG/DNA complexes in the presence of serum or bile wasfurther examined by agarose gel electrophoresis (datanot shown). Plasmid DNA incubated with 10% (v/v) ratserum and bile rapidly degraded into small molecularweight fractions within 30 and 10 min, respectively.However, with the protection of chitosan or chitosan-g-PEG, the degradation of DNA in both serum and bile wassignificantly delayed. No DNA fragments were releasedand detected within 2 h.

In vivo gene transfection following bileduct infusion

Figure 5a shows gene expression in the liver followingbile duct infusion of chitosan/DNA complexes (N/P = 3),chitosan-g-PEG (GD = 3.6%)/DNA complexes (N/P = 3)or chitosan-g-PEG (GD = 9.6%)/DNA complexes (N/P =5). These different N/P ratios corresponded to thepeak transfection efficiencies of the three different com-plexes in vitro. The chitosan-g-PEG (GD = 3.6%)/DNAcomplexes yielded slightly higher gene expression com-pared to unmodified chitosan. Chitosan-g-PEG (GD =9.6%)/DNA complexes achieved the highest luciferaseexpression, which was about 3 times higher than thatmediated by chitosan/DNA complexes. Interestingly, after



0

100

200

300

400

500

0 10 20 30 40Time (minute)

Chitosan



0

20

40

60

80

100

120

0 10 20 30 40Time (minute)

Per

cen

tag

e (%

)

Chitosan



Time (minute)

0200

400600800

1000

1200140016001800

0 10 20 30 40

Mea

n P

arti

cle

Siz

e (n

m)

Chitosan



0

20

40

60

80

100

120

0 10 20 30 40Time (minute)

Per

cen

tag

e (%

)

Chitosan



(a) (b)

(c) (d)

Mea

n P

arti

cle

Siz

e (n

m)

Figure 4. Average particle size of chitosan and chitosan-g-PEG/DNA complexes in the presence of 10% (v/v) FCS (a) or 10% (v/v)rat bile (c) as a function of time. Chitosan and chitosan-g-PEG/DNA complexes were prepared in 25 mM sodium acetate buffer (pH5.5) and the concentration of DNA was 50 µg/mL. FCS or rat bile (100 µL) was added to 900 µL of complex solution. The particlesize of the complexes was measured at 5, 10, 15, 20 and 30 minutes after the addition of FCS or rat bile. Percentage of chitosanand chitosan-g-PEG/DNA complexes with a size above 1000 nm after the addition of FCS (b) or rat bile (d) was also plotted as afunction of time. The percentage was calculated from the intensity distribution given by the Zetasizer 3000HS

0

10

20

30

40

50

60

70

80

90

100

Lu

cife

rase

Exp

ress

ion

(pg

/ g t

issu

e)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Spleen Heart Lung Kidney

Lu

cife

rase

Exp

ress

ion

(pg

/g t

issu

e)

*

PEIChitosanChitosan-g-PEG(3.6%)Chitosan-g-PEG(9.6%)

(a) (b)PEIChitosanChitosan-g-PEG(3.6%)Chitosan-g-PEG(9.6%)

Figure 5. Luciferase expression in rat liver (a) and other major organs (b) after infusion of the complexes from common bile duct.The liver and other organs were harvested 3 days after administration (∗p < 0.05). The detection limit of luciferase expression is0.05 pg/g tissue

bile duct infusion, PEI, which was much more efficientthan chitosan and chitosan-g-PEGs in the HepG2 cell line,mediated almost 10-, 15- and 40-fold lower gene expres-sion in liver than chitosan, chitosan-g-PEG (GD = 3.6%)and chitosan-g-PEG (GD = 9.6%), respectively.

The β-galactosidase staining results show that thepositive cells scattered around the central vein andhad the morphology of hepatocytes (Figure 6). Thegene expression levels in other major organs were alsoexamined, as shown in Figure 5b. Gene expression level

was close to background level in the spleen, heart, lungand kidney. Comparing Figure 5a with Figure 5b, it isclear that this delivery method yielded liver-specific geneexpression for all three carriers.

Gene transfection following portal veininfusion

Chitosan and chitosan-g-PEG showed low levels of genetransfection following portal vein infusion. Figure 7a


484 X. Jiang et al.

Figure 6. Localization of gene expression after infusion of the complexes through the common bile duct. The livers were harvested2 days after administration. Rat liver injected with PBS (A) and chitosan-g-PEG (GD = 9.6%)/p43-clz1 plasmid DNA complexes (B,C and D). Magnification for (A), (B) and (C): ×100; magnification for (D): ×200

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Lu

cife

rase

Exp

ress

ion

(pg

/g t

issu

e)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Spleen Lung Kidney

Lu

cife

rase

Exp

ress

ion

(pg

/g t

issu

e)

*

*


(a) (b)


Heart

Figure 7. Luciferase expression in rat liver (a) and other major organs (b) after infusion of the complexes from the portal vein. Theliver and other organs were harvested 1 day after administration (∗p < 0.05). The detection limit of luciferase expression is 0.05pg/g tissue

shows gene expression in the liver on day 1, mediatedby PEI/DNA complexes, chitosan/DNA complexes (N/P =3), chitosan-g-PEG (GD = 3.6%)/DNA complexes (N/P =3) and chitosan-g-PEG (GD = 9.6%)/DNA complexes(N/P = 5). The gene expression level mediated bychitosan/DNA complexes was about 0.1 pg luciferase/gtissue. Both chitosan-g-PEG (GD = 3.6%) and chitosan-g-PEG (GD = 9.6%) mediated 3 and 7 times higher geneexpression levels than chitosan, respectively. The geneexpression level induced by PEI/DNA complexes wasalso much lower than those of chitosan-g-PEG (GD =3.6%) and chitosan-g-PEG (GD = 9.6%). In other organs(Figure 7b), gene expression was not detected in the ratstreated with chitosan and PEI/DNA complexes. However,

a low level of expression was observed in the rats injectedwith chitosan-g-PEG/DNA complexes. Gene expressionlevel was also low in other organs (Figure 7b). In general,luciferase levels in all organs tested, mediated by PEG-grafted chitosan carriers, were significantly higher thanthat of chitosan. Gene expression was undetectable onday 3 for all three carriers in all organs tested.

From Figures 5 and 7, it is noted that the luciferaseexpression level in the liver was higher than other organsin the case of bile duct infusion. However, it was notsignificantly higher in the liver when compared withother organs in portal vein infusion. After the chitosan-g-PEG/DNA complexes have been administrated throughthe bile duct and incubated for 30 min, the residual



chitosan-g-PEG/DNA complexes that have not been takenup by hepatocytes will flow with the secreted bile andeventually reach the intestine. In the digestive tract,chitosan-g-PEG/DNA complexes have very little chanceof penetrating intestinal epithelial cells and entering theblood stream. Therefore, chitosan-g-PEG/DNA complexesadministrated through the bile duct could hardly betransported to other organs, resulting in low genetransfection. However, in the portal vein infusion, onlythe portal vein was tied up. Therefore, a significantamount of the complexes was probably washed awayby the blood flow and transported to other organs. Inaddition, the incubation time for the portal vein injectionwas much shorter than that in the bile duct infusion(10 vs. 30 min), leading to a reduced uptake of chitosan-g-PEG/DNA complexes by hepatocytes. Thus, the luciferaselevel was not significantly higher in the liver as comparedwith other organs.

Liver toxicity

To evaluate the acute toxicity of chitosan/DNA com-plexes, chitosan-g-PEG/DNA complexes and PEI/DNAcomplexes to rat liver, blood samples were takenat different time points and analyzed for ALT andAST activities after administration of these complexes.On day 1, higher levels of ALT (400–600 U/L)and AST (480–600 U/L) were detected in the ratstreated with chitosan and chitosan-g-PEG/DNA complexes(Figure 8). However, PEI/DNA complexes resulted inmuch higher ALT (1010U/L) and AST (900U/L) lev-els on day 1. ALT and AST levels in the rats injectedwith chitosan-g-PEG (GD = 3.6%) and chitosan-g-PEG(GD = 9.6%)/DNA complexes appeared to be signifi-cantly lower on the second day when compared tothose of the rats injected with PEI and chitosan/DNAcomplexes. On day 3, ALT and AST levels droppedrapidly to close to the normal level (ALT: 39 ± 13 U/Land AST: 119 ± 25 U/L) in all rats injected with chi-tosan and chitosan-g-PEG/DNA complexes but the ALTand AST levels of the rats treated with PEI/DNA com-plexes were still high, being 130 and 150, respec-tively.

Discussion

Modification of chitosan for liver-targeted gene deliveryhas been attempted in several studies [34–37]. Aftercoupling the galactose moiety onto chitosan, specificinteraction of the asialoglycoprotein receptor withgalactose ligands has been observed, but there is verylimited enhancement in both cell uptake and transfectionefficiency of galactosylated chitosan in hepatocyte celllines. Moreover, these modified chitosan carriers have notbeen tested for their in vivo gene transfer efficiencies.Issues related to the stability and opsonization inbiological media and access to target cells or passagethrough biological barriers, in this case, sinusoidalfenestrae and bile canaliculi, can only be addressedin in vivo studies. In this study, we chose bile ductand portal vein infusions as the administration routesto deliver complexes directly to the liver. We studiedthe effect of PEG modification on the stability ofpolymer/DNA complexes in the presence of bile andserum, and gene transfection efficiency in vitro andin vivo.

Although their average size was slightly smaller thanthat of chitosan-g-PEG/DNA complexes in the preparationmedia, chitosan/DNA complexes were highly susceptibleto the challenge of serum and bile. The presence ofserum or bile led to serious aggregation of chitosan/DNAcomplexes. Immunohistochemical results also showedthat large aggregates of chitosan/DNA complexes werepresent in bile canaliculi after infusion (data not shown).However, PEG-grafted chitosan effectively shielded thepositively charged surface when compared to unmodifiedchitosan. Due to the reduced surface charge and the sterichindrance of PEG chains, chitosan-g-PEG/DNA complexeswere able to maintain their size upon challenge by bile andserum, and offered protection to DNA in the complexesagainst enzymatic degradation. More importantly, thesizes of chitosan-g-PEG/DNA complexes are small enoughfor complexes to pass through the fenestrae or bilecanaliculi.

The highest gene expression level mediated bychitosan-g-PEG (GD = 3.6%)/DNA complexes (N/P = 3)in HepG2 cells was comparable to chitosan/DNA

0

200

400

600

800

1000

1200

1400

1600

0 3 4

Time (day)

AL

T (

U/L

) Chitosan-g-PEG(3.6%)


Chitosan(47kDa)

PEI(a)

0

200

400

600

800

1000

1200

1400

1600

Time (day)

AS

T (

U/L

) Chitosan-g-PEG(3.6%)


Chitosan(47kDa)

PEI(b)

1 2 0 3 41 2

Figure 8. (a) ALT and (b) AST levels of the rats after bile duct infusion of the complexes


486 X. Jiang et al.

complexes (N/P = 3). Increasing the PEG graftingdegree decreased transfection efficiency in HepG2cells. Among chitosan, chitosan-g-PEG (GD = 3.6%)and chitosan-g-PEG (GD = 9.6%), chitosan-g-PEG (GD =9.6%) mediated the highest level of transgene expressionin the rat liver after bile duct infusion. Although PEIexcelled chitosan and chitosan-g-PEGs in the transfectionof the HepG2 cell line, the transfection efficiency ofPEI/DNA complexes was the lowest after bile ductinfusion, which may be related to its highest liver toxicitydemonstrated in ALT and AST tests.

Following portal vein infusion, chitosan-g-PEG carriersmediated much lower luciferase expression in theliver, even though it was more efficient than chitosanand PEI. The expression was more transient, onlydetectable on day 1. With increasing grafting degreeof PEG, the gene expression level was not onlyenhanced in the liver, but also in other major organs,suggesting that the PEG grafting to chitosan reduced theopsonization and prolonged the circulation time of thecomplexes.

When the two administration routes were compared,bile duct infusion was obviously superior to portalvein infusion. Intrabiliary infusion directly delivers thecomplexes to the hepatocytes through bile canaliculi[38], avoiding clearance by Kupffer cells. In portal veininfusion, significant amounts of complexes are probablywashed away by blood flow, as the blood flow isnot fully stopped in the liver during the process ofportal vein infusion. This is to reduce liver damagedue to ischemia re-perfusion, whereas, in bile ductinfusion, the complexes are incubated in a relatively morestatic condition. Therefore, the complexes by intrabiliaryinfusion stand a better chance of being taken up byhepatocytes.

The hepatic toxicity of polycation/DNA complexes isone of the major concerns. As demonstrated in ALT andAST assays, PEG grafting slightly reduced the acute livertoxicity of the carrier. Chitosan-g-PEG/DNA complexesinduced less hepatic injury than chitosan/DNA complexesat early time points, but the mechanism involved is notclear.

In conclusion, PEG grafting to chitosan efficientlyshields the positive charge on the surface of chitosan/DNAcomplexes, improving the particle stability in bile andserum, and thus the transfection efficiency in vivo.Chitosan-g-PEG (GD = 9.6%) mediated 3 times higherluciferase expression in the liver than unmodifiedchitosan following intrabiliary infusion. Chitosan-g-PEGalso exhibited slightly lower acute toxicity to the liverthan chitosan.

Acknowledgements

This work was funded by the Institute of Bioengineering andNanotechnology (Agency for Science, Technology and Research,Singapore) and the Division of Biomedical Sciences, JohnsHopkins in Singapore.

References1. Dai H, Zhang PC, Lim WS, Chen Y, Leong KW, Mao HQ. Liver-

targeted gene transfer mediated by chitosan-DNA nanoparticlesthrough bile duct, portal vein and tail vein injections. Mol Ther2003; 7: S322–S323 (Part 2).

2. Herweijer H, Wolff JA. Progress and prospects: naked DNA genetransfer and therapy. Gene Ther 2003; 10: 453–458.

3. Niidome T, Huang L. Gene therapy progress and prospects:nonviral vectors. Gene Ther 2002; 9: 1647–1652.

4. Kawakami S, Fumoto S, Nishikawa S, Yamashita M, Hashida F.In vivo gene delivery to the liver using novel galactosylatedcationic liposomes. Pharm Res 2000; 17: 306–313.

5. Wu GY, Wu CH. Receptor-mediated gene delivery andexpression in vivo. J Biol Chem 1988; 263: 14 621–14 624.

6. Morimoto K, Nishikawa M, Kawakami S, et al. Molecularweight-dependent gene transfection activity of unmodified andgalactosylated polyethyleneimine on hepatoma cells and mouseLiver Mol Ther 2003; 7: 254–261.

7. De Godoy JL, Malafosse R, Fabre M, Mehtali M, Houssin D,Soubrane O. In vivo hepatocyte retrovirus-mediated genetransfer through the rat biliary tract. Hum Gene Ther 1999;10: 249–257.

8. Otsuka M, Baru M, Delriviere L, Talpe S, Nur I, Gianello P.In vivo liver-directed gene transfer in rats and pigs with largeanionic multilamellar liposomes: routes of administration andeffects of surgical manipulations on transfection efficiency. JDrug Target 2000; 8: 267–279.

9. Zhang X, Collins L, Sawyer GJ, Dong X, Qiu Y, Fabre JW. In vivogene delivery via portal vein and bile duct to individual lobesof the rat liver using a polylysine-based nonviral DNA vectorin combination with chloroquine. Hum Gene Ther 2001; 12:2179–2190.

10. Zhang X, Sawyer GJ, Dong X, Qiu Y, Collins L, Fabre JW. Thein vivo use of chloroquine to promote non-viral gene delivery tothe liver via the portal vein and bile duct. J Gene Med 2003; 5:209–218.

11. Xie X, Forsmark CE, Lau JY. Effect of bile and pancreatic juiceon adenoviral-mediated gene delivery: implications on thefeasibility of gene delivery through ERCP. Dig Dis Sci 2000;45: 230–236.

12. MacLaughlin FC, Mumper RJ, Wang JJ, et al. Chitosan anddepolymerized chitosan oligomers as condensing carriers forin vivo plasmid delivery. J Control Rel 1998; 56: 259–272.

13. Richardson SCW, Kolbe HJV, Duncan R. Potential of lowmolecular mass chitosan as a DNA delivery system:biocompatibility, body distribution and ability to complex andprotect DNA. Int J Pharm 1999; 178: 231–243.

14. Thanou M, Florea BI, Geldof M, Junginger HE, Borchard G.Quaternized chitosan oligomers as novel gene delivery vectorsin epithelial cell lines. Biomaterials 2002; 23: 153–159.

15. Lee KY, Kwon IC, Kim YH, Jo WH, Jeong SY. Preparation ofchitosan self-aggregates as a gene delivery system. J Control Rel1998; 51: 213–220.

16. Kim YH, Gihm SH, Park CR, et al. Structural characteristicsof size-controlled self-aggregates of deoxycholic acid-modifiedchitosan and their application as a DNA delivery Carrier.Bioconjugate Chem 2001; 12: 932–938.

17. Mao HQ, Roy K, Troung-Le VL, et al. Chitosan-DNA nanoparti-cles as gene carriers: synthesis, characterization and transfectionefficiency. J Control Rel 2001; 70: 399–421.

18. Roy K, Mao HQ, Huang SK, Leong KW. Oral gene delivery withchitosan–DNA nanoparticles generates immunologic protectionin a murine model of peanut allergy. Nat Med 1999; 5: 387–391.

19. Chew JL, Wolfowicz CB, Mao HQ, Leong KW, Chua KY.Chitosan nanoparticles containing plasmid DNA encoding housedust mite allergen, Der p 1 for oral vaccination in mice. Vaccine2003; 21: 2720–2729.

20. Kai E, Ochiya T. A method for oral DNA delivery with N-acetylated chitosan. Pharm Res 2004; 21: 838–842.

21. Okamoto H, Nishida S, Todo H, Sakakura Y, Iida K, Danjo K.Pulmonary gene delivery by chitosan–pDNA complex powderprepared by a supercritical carbon dioxide process. J Pharm Sci2003; 92: 371–380.

22. Wisse E, De Zanger RB, Charels K, Van Der Smissen P,McCuskey RS. The liver sieve: considerations concerning thestructure and function of endothelial fenestrae, the sinusoidalwall and the space of Disse. Hepatology 1985; 5: 683–692.



23. Nauss JL, Thompson JL, Nagyvary JJ. The binding of micellarlipids to chitosan. Lipids 1983; 18: 714–719.

24. Bioulac-Sage P, Le Bail B, Balabaud C. Liver and biliary-tracthistology. In Oxford Textbook of Clinical Hepatology, Bircher J,Benhamon JP, McIntyre N, Rizzetto M, Rodes J (eds). OxfordUniversity Press: Oxford, 1999; 13–21.

25. Patel HM. Serum opsonins and liposomes: their interaction andopsonophagocytosis. Crit Rev Ther Drug Carrier Syst 1992; 9:39–90.

26. Choi YH, Liu F, Kim JS, Choi YK, Park JS, Kim SW. Polyethyleneglycol-grafted poly-L-lysine as polymeric gene carrier. J ControlRel 1998; 54: 39–48.

27. Katayose S, Kataoka K. Remarkable increase in nucleaseresistance of plasmid DNA through supra-molecular assemblywith poly(ethylene glycol)-poly(L-lysine) block copolymer. JPharm Sci 1997; 87: 160–163.

28. Tang GP, Zeng JM, Gao SJ, et al. Polyethylene glycolmodified polyethylenimine for improved CNS gene transfer:effects of PEGylation extent. Biomaterial 2003; 24:2351–2362.

29. 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-kappa B decoy inmice. Pharm Res 2002; 19: 810–817.

30. Fischer D, Osburg B, Petersen H, Kissel T, Bickel U. Effect ofpoly(ethylene imine) molecular weight and pegylation onorgan distribution and pharmacokinetics of polyplexes with

oligodeoxynucleotides in mice. Drug Metab Dispos 2004; 32:983–992.

31. Uehara T, Honda K, Hatano E, et al. Gene transfer to the ratbiliary tract with the HVJ-cationic liposome method. J Hepatol1999; 30: 836–842.

32. Houk BE, Hochhaus G, Hughes JA. Kinetic modeling of plasmidDNA degradation in rat Plasma. AAPS Pharm Sci 1999; 1:U1–U9.

33. Asad SF, Singh S, Ahmad A, et al. Bilirubin/biliverdin-Cu(II)induced DNA breakage; reaction mechanism and biologicalsignificance. Toxicol Lett 2002; 131: 181–189.

34. Gao SY, Chen JN, Xu XR, et al. Galactosylated low molecularweight chitosan as DNA carrier for hepatocyte-targeting. Int JPharm 2003; 255: 57–68.

35. Park YK, Park YH, Shin BA, et al. Galactosylated chitosan–graft-dextran as hepatocyte-targeting DNA carrier. J Control Rel 2000;69: 97–108.

36. Park IK, Kim TH, Park YH, et al. Galactosylated chitosan-graft-poly(ethylene glycol) as hepatocyte-targeting DNA carrier. JControl Rel 2001; 76: 349–362.

37. Park IK, Kim TH, Park YH, Kim WJ, Akaide T, Cho CS.Visualization of transfection of hepatocytes by galactocylatedchitosan-graft-poly (ethylene glycol)/DNA complexes byconfocal laser scanning microscopy. Int J Pharm 2003; 257:103–110.

38. Junqueira LC, Carneiro J, Kelly RO (eds). Basic Histology (9thedn). Appleton & Lange: Norwalk, CT, 1989; 315.


Chitosan-g-PEG/DNA complexes deliver gene to the rat liver via intrabiliary and intraportal...

Documents

Transcript of Chitosan-g-PEG/DNA complexes deliver gene to the rat liver via intrabiliary and intraportal...