Linear and branched chitosan oligomers as delivery systems for pDNA and siRNA in vitro and in vivo

Targeted gene delivery with trisaccharide-substituted chitosan oligomers in vitro and after lung administration in vivo.. Introduction...11 Nucleic acids and oligonucleotides as potentia

145 &9((9%889.

#  

1445 %(9%&# 145 &9((9%889

“You are not given aught of knowledge but a little”

Holy Qur’an: 17:85

To my family

Papers discussed

The thesis is based on the following papers, which will be referred to

by their Roman numerals:

I Köping-Höggård M, Vårum KM, Issa M, Danielsen S,

Christen-sen BE, Stokke BT, Artursson P Improved chitosan-mediated gene delivery based on easily dissociated chitosan polyplexes of

highly defined chitosan oligomers Gene Ther (2004)11(19):

1441-52

II Issa M, Köping-Höggård M, Tømmeraas K, Vårum KM,

Chris-tensen BE, Strand SP, Artursson P Targeted gene delivery with trisaccharide-substituted chitosan oligomers in vitro and after

lung administration in vivo J Control Rel (2006) 115 (1):

103-112

III Issa M, Strand SP, Vårum KM, Artursson P Chitosan oligomers

as siRNA delivery systems in vitro In Manuscript.

IV Köping-Höggård M*, Issa MM * , Köhler T, Tronde A, Vårum

KM, Artursson P A miniaturized nebulization catheter for proved gene delivery to the mouse lung, J Gene Med 7(9) (2005): 1215-22 *shared first authorship

im-Also published:

Mohamed M Issa, Magnus Köping-Höggård and Per Artursson Chitosan

and the mucosal delivery of biotechnology drugs Drug Discovery Today:

Technologies (2005), 2 (1) 1-6.

Reprints were made with permission from the publishers

Introduction 11

Nucleic acids and oligonucleotides as potential pharmaceutical products 11

Gene delivery systems 14

Viral gene delivery systems 14

Non-viral gene delivery systems 14

Naked (unformulated) nucleic acids 15

Barriers to non-viral gene delivery 15

Pharmaceutical barriers 15

Extracellular barriers 16

Cell-surface and intracellular barriers 17

Bacterial gene delivery (bactofection) and alternative gene therapy (AGT) 19

Physical methods of non-viral gene delivery 20

Chemical methods of non-viral gene delivery 20

Lipids 21

Polymers 22

Strategies to improve the in vitro/in vivo efficiency of non-viral gene delivery: Structure-activity relationship 25

Molecular weight reduction 25

Positive charge shielding 26

Active targeting 26

In vivo toxicity of polyplexes 26

Chitosan chemical structure 27

Properties of chitosan 28

Physicochemical properties 28

Biological properties 29

General applications of chitosan in drug delivery 29

Chitosan as a delivery system for proteins and peptide drugs 29

Chitosan as non-viral gene delivery system 30

Aims 32

Materials and methods 33

Nucleic acids 33

Polycations 33

Formulation of complexes 34

Size and morphology of the complexes 35

Physical and enzyme stability 35

In vitro studies 35

Transfection experiments 35

Cellular uptake of chitosan complexes 36

Cellular toxicity (Intracellular dehydrogenase activity) 37

In vivo studies 37

Luciferase gene expression 37

Distribution pattern of gene expression in the mouse lung 37

Toxicological evaluations 37

Statistics 38

Results and discussion 39

Optimised linear chitosan oligomers as non-viral gene delivery systems (Paper I) 39

Characterisation of DPn18 polyplexes in vitro and in vivo 39

Characterisation of polyplexes based on oligomer fractions isolated from DPn18 in vitro and in vivo 41

Intracellular release of pDNA from chitosan oligomer-based polyplexes 43

Trisaccharide-substituted chitosan oligomers as non-viral gene delivery systems (Paper II) 44

Structure-activity relationship of polyplexes based on trisaccharide-substituted chitosan oligomers (TCO) .45

Impact of trisaccharide substitution on chitosan oligomer-based polyplexes 45

Chitosan oligomers as siRNA delivery systems in vitro (Paper III) 50

In vitro siRNA delivery under conditions previously optimised for pDNA 50

Structure-activity relationship of chitosan oligomer-based siRNA complexes 52

Improved aerosol gene delivery to the mouse lung in vivo (Paper IV) 56

Physical stability and in vitro transfection efficiency following nebulisation with the NCD 57

In vivo nebulisation with the NCD VS intratracheal instillation 58

Summary and conclusions 60

Acknowledgements 62

References 64

A-A-M 2-acetamido-2-deoxy-D

-glucopyranosyl-E-(1-4)-2-acetamido-2-deoxy-Danhydro-D-mannofuranose

-glucopyranosyl-E-(1-4)-2,5-AGT Alternative gene therapy

AUC Area under the curve

FA Fraction of acetylated units

GFP Green fluorescent protein

GlcNAc N-acetylglucosamine

GMP Good manufacturing practice

HSPG Heparan sulphate proteoglycans

mRNA Messenger ribonucleic acid

NCD Nebulisation catheter device

NLS Nuclear localisation signal

PBS Phosphate-buffered saline

PEI Polyethyleneimine

PLL Poly-L-lysine

siRNA Small interfering RNA

TCO Trisaccharide-substituted chitosan oligomers

of the natural polysaccharide chitosan as a delivery vehicle for nucleic acids

is described as one of these newly developed approaches

Nucleic acids and oligonucleotides as potential

pharmaceutical products

The concept of gene medicine is based on the use of nucleic acids as drugs for gene therapy with the aim of restoring or shutting down a specific cellu-lar function [2] In contrast to conventional medicines that often focus on the treatment of clinical symptoms, gene medicines provide treatment at the molecular level of deoxyribonucleic acids (DNA) or ribonucleic acids (RNA), i.e at the intracellular gene expression level The first successful gene therapy-based clinical trial was reported in the early 1990s In this clinical trial, immune competence was restored to children suffering from X-linked severe combined immunodeficiency (SCID), so that they were able to live outside the sterile environment “bubbles” to which they have been con-fined [3,4] By September 2006, 1192 clinical gene therapy trials had been approved and conducted worldwide; most of those involved devastating, acquired or inherited diseases such as cancer, monogenic disorders (e.g cystic fibrosis) and infectious diseases (e.g AIDS) (http://www.wiley.co.uk/genetherapy/clinical/) In addition to gene therapy, gene vaccination is another application of gene medicine [5] The introduc-tion of genes encoding for various pathogenic antigens into target cells can result in the production of cellular and humoral (antibody) immune re-sponses [6]

Two major approaches are used for the transfer of therapeutic nucleic acids

(genetic material) to the target area for gene therapy (Figure 1) 1) Ex vivo:

the genetic material is first inserted into cells grown in vitro (cell cultures)

The transfected cells are then selected, expanded, and introduced into the patient To avoid rejection by the host immune system, cells (e.g bone mar-row cells) are usually obtained from the same patient (autologous cells) prior

to the procedure 2) In vivo: the genetic material is transferred directly into

the target cells in the patient This process involves less manipulation than

the ex vivo approach It may also be the only option in tissues where

individ-ual cells cannot be obtained or cultured in sufficient quantities or the tured cells cannot be re-implanted

cul-Figure 1. In vivo and ex vivo strategies for the transfer of genetic materials in

human gene therapy [7]

Nucleic acids that can be used as gene medicines can take a variety of forms One such is plasmid DNA (pDNA), a macromolecule that carries a specific gene sequence (transgene) encoding the desired protein(s) Upon delivery to the target cells (the transfection process), the gene sequences are eventually translated as the desired therapeutic functional or structural protein(s) (Fig-ure 2) In contrast, gene-silencing techniques (antigenes) such as antisense

oligodeoxynucleotides (ODN), ribozymes, DNAzymes and the more cently introduced small interfering, double-stranded RNA (siRNA) are nu-cleic acids that can bind to specific target sequences of intracellular mRNA and subsequently block its translation (Figure 2) [8-10] This process can lead to specific knock down (silencing) of the targeted cellular proteins or functions More specifically, the process of RNA interference (RNAi) by siRNA involves incorporation of the short double-stranded RNA into a pro-tein structure to produce an RNA-induced silencing complex (RISC), which recognizes and binds to the target mRNA sequence, resulting in cleavage [11] In this thesis, pDNA and, to a lesser extent, siRNA are used

re-Figure 2.The central dogma presented by Crick, showing the flow of genetic information from DNA to messenger RNA (mRNA) to protein [12] In addition, a general model of gene silencing is presented ODN and siRNA recognize mRNA sequences and subsequently cleave the mRNA or block its translation

One antisense drug, for the treatment of cytomegalovirus (CMV) rhinitis in AIDS patients, has been approved by the FDA for clinical use [9] Several more drug products that are based on gene-silencing techniques, especially siRNA are being investigated in clinical trials at various phases, which sug-gests that more marketed nucleic acid-based drug products may be available

in the coming few years [13,14]

The achievement of significant clinical or therapeutic benefits with nucleic acid-based drugs has, however, been challenged by several obstacles The low efficiency and toxicity of various systems (viral or non-viral) or methods (physical or chemical) used for the cellular delivery of nucleic acids have caused various problems for successful clinical gene therapy [15,16]

Gene delivery systems

Viral gene delivery systems

Viral gene delivery systems are based on recombinant viruses that are

ge-netically modified to be replication-defective and unable to cause diseases when introduced into the human body Viruses that have been used in gene therapy protocols include retroviruses, adenoviruses, adeno-associated vi-

ruses, vaccinia viruses and herpes simplex viruses [17-22] The first clinical gene therapy protocol approved for the treatment of SCID, in 1990, involved

a retroviral vector Since then, almost 70% of clinical protocols for gene therapy have used viral delivery systems (http://www.wiley.co.uk/genetherapy/clinical/) The ability of viruses to condense nucleic acids and provide protection against enzymatic degrada-

tion, together with their highly specialised mechanisms for cell infection (cell binding and penetration, escape from the intracellular compartments, active transport of the genetic material into the nucleus, etc.) has resulted in their current position as the most effective gene delivery systems [23] How-

ever, significant limitations are inherent to their use in humans Viral vectors (e.g retroviruses) may provoke insertional mutagenesis, in which the ran-

dom integration of the viral DNA with the host cell genome may result in disruption of the expression of a tumor-supressor gene or activation of an oncogene leading to the malignant transformation Recent reports on the development of leukaemia in children treated with viral vectors for SCID provided evidence for the toxicity problems associated with viral delivery systems [24] Repeated administration of viral vectors (e.g adenoviruses) may induce an immune response, which can abolish transgene expression [18] Other problems associated with viral delivery systems include their limited gene-carrying capacity, restricted cell-targeting and high large-scale production cost

Non-viral gene delivery systems

The increased risk of using viral delivery systems for gene therapy of genetic

or acquired human diseases has motivated the search for, and utilisation of, safer carrier molecules Non-viral delivery systems such as naked or formu-

lated nucleic acids have important advantages over viral approaches,

includ-ing their reduced propensity for insertional mutagenesis and pathogenicity,

as well as their relatively low cost and ease of production [25]

Naked (unformulated) nucleic acids

In non-viral gene delivery, the desired therapeutic gene(s) is usually rated into pDNA, which is a circular double-stranded DNA molecule of bac-terial origin Besides the therapeutic gene(s), pDNA contains other important gene sequences such as promoter/enhancer elements, which are responsible for controlling the transcription and expression levels of the encoded protein once it is introduced into the target cells [26] As a result of the inclusion of these elements, a typical plasmid comprises 5,000-10,000 base (nucleotide) pairs and has a molecular mass of more than 1 million Daltons (Da) [23] pDNA is a strong polyanion because of the phosphate groups in its nucleo-tide backbone It can exist in three different conformations (supercoiled, open circular and linear), of which the supercoiled form is preferred because

incorpo-of its more compact structure In physiological salt solutions, the phosphate groups of the pDNA bind counter ions, which may further contribute to the compactness of the plasmid structure [27] pDNA can elicit immune re-sponses as a result of the unmethylated CpG motifs [28] While this may be

an advantage in development gene vaccines, this effect can be masked by the methylation or exclusion of these motifs from the pDNA sequence [29]

Barriers to non-viral gene delivery

Although non-viral gene delivery systems largely lack the adverse effects of viral delivery systems, their clinical application has been limited as a result

of their poor gene transfection efficiency A better understanding of the ous barriers encountered by non-viral gene delivery will significantly con-tribute to the development of nucleic acid-based formulations into therapeu-tic products for human gene therapy These barriers can generally be classi-fied into three major categories: pharmaceutical, extracellular and cell-surface and intercellular barriers (Figure 3) [25,30,31]

Figure 3.General schematic representation of cell-surface and intracellular ers to the expression of a transgene delivered by non-viral plasmid-based formula-tions These include (1) complexation of pDNA with the delivery system; interaction

barri-of DNA-based complexes with the cell membrane; cellular internalisation via (2) nonspecific or (3) receptor-mediated endocytic pathways; (4) endosomal entrap-ment followed by (5) the development of lysosomes or (6) the rupture and cyto-plasmic release of the complexes or pDNA alone [cytoplasm is the site of action for antisense oligonucleotides (ODN and siRNA), ribozymes, DNAzymes and aptam-ers]; (7) nuclear translocation [the nucleus is the site of action for transgenes in plasmids for gene therapy and siRNA-generating plasmids]; and (8) transcription followed by (9) the expression of the transgene into the desired therapeutic product

Extracellular barriers

On introduction into physiological fluids, either cell culture media in vitro

or, the body fluids such as blood in vivo, naked or formulated nucleic acids will encounter a hostile environment In vitro, particle size stability is one of

the most critical issues that can affect the transfection process [33] stabilised nucleic acid formulations tend to aggregate in physiological media and sediment onto the surface of cells as a result of the reduced colloidal stability While this could be an advantage in transfecting adherent cells, uncontrolled aggregations could impair cellular uptake because of size con-straints associated with cellular uptake mechanisms, mainly endocytosis

Non-(vesicular uptake of extracellular molecules) [34,35] Various uptake ways involved in non-viral gene delivery are shown in Figure 4

path-Following in vivo systemic administration of formulated nucleic acids,

inter-action with the plasma proteins and blood cells can result in aggregation, fast clearance and a reduced circulation time in the human body [36] Further, the abundance of nucleic acid-degrading enzymes (nucleases) in the blood or

in the mucosal sites can lead to a substantial loss of the expected therapeutic effect [37] However, besides the poor serum stability and unfavourable

pharmacokinetics in vivo, one of the most critical toxic effects preventing

repeated administration of these formulations is stimulation of the immune response [38]

Figure 4.Portals of entry into the

mammalian cell The endocytic

pathways differ with regard to the

size of the endocytic vesicle, the

nature of the cargo and the

mecha-nism of vesicle formation [34,35]

Cell-surface and intracellular barriers

Binding and uptake

Once the naked nucleic acids are at the cell surface, the negatively charged pDNA and oligonucleic acids, in general, exhibit little interaction with the negatively charged components of the cell membrane (glycoproteins, glycol-ipids and proteoglycans) Further, size-restrictions can hamper cellular up-take via endocytosis [34,35] Therefore, an optimal delivery system should physically condense nucleic acids into small particles (e.g through ionic complexation) and should facilitate their cellular binding and uptake This can be achieved, either non-specifically by providing excess positive surface charges on the particles or specifically via receptor-mediated endocytosis by actively targeting specific receptors on the cell surface [39-42]

Release from endosomes

Following uptake and internalisation, the formulated nucleic acids are trapped in intracellular compartments called endosomes The endosomes mature from early to late stages with a simultaneous drop in pH from 6.0 to 5.0 They then fuse with lysosomes, and a significant fraction of the trapped nucleic acids is eventually degraded by the lysosomal hydrolytic enzymes, leading to a substantial reduction in the transfection efficiency [43] It would thus be preferable if the delivery system could facilitate early escape of the nucleic acids from the endosomes This can be achieved either by disrupting the endosomal membrane or by inducing endosomal swelling and rupturing [44-46] Membrane active peptides, either synthetic or derived from viruses have been shown to enhance the endosomal release of non-viral gene deliv-ery systems [47-49]

Unpacking

In general, the disassembly of nucleic acids from the delivery system is sential for the transcription and expression of the desired function [50] This process can take place during endosomal release, where anionic cell mole-cules such as lipids can displace the nucleic acids [44] However, unpacking

es-of the formulated nucleic acids can occur after endosomal release in the tosol or even later in the nucleus [51-53] The chemical nature of the deliv-ery system can influence the release by controlling the strength (tightness) of the interaction with the nucleic acids In other words, the physical stability of the nucleic acid formulation can be a limiting factor in the disassembly of the nucleic acids [54]

cy-Diffusion in the cytoplasm

For oligonucleic acids that exert their mechanism of action in the cytoplasm, such as siRNA, the critical intracellular steps include efficient cellular bind-ing and uptake, early escape from the endocytic vesicles and stability against intracellular enzymatic degradation [55] However, larger pDNA has to overcome additional hurdles to access the nucleus for the transcription of the encoded therapeutic genes into mRNA After endosomal escape, pDNA has

to diffuse through the dense microfilaments and microtubule (cytoskeleton) network, as well as through a variety of subcellular organelles bathing in the cytosol [56] In the highly packed cytosol, the movement of pDNA is size-dependent, and DNA larger than 3,000 base pairs appears to be practically immobile [57] The restricted mobility of pDNA in the cytosol necessitates its active transport to the nucleus [58]

Nuclear translocation and expression

The nuclear pore complex (NPC), which is located on the nuclear brane, is the ultimate obstacle to the entry of pDNA into the nucleus [59] As

mem-with diffusion in the cytoplasm, passage through the NPC is size-dependent [60,61] The diameter of the NPC is 9 nm, but this can be extended to 20-25

nm during active translocation [62,63] This pore size range implies that it is almost impossible for pDNA to passively diffuse to the nucleus through the NPC [58,64] However, coupling of a nuclear localising signal (NLS) to the pDNA or to the delivery system can provoke the enlargement of the NPC through conformational changes [65-67] Alternatively, pDNA can access the nucleus during cell division (mitosis) when the nuclear envelope is disas-sembled [68] However, this method does not significantly contribute to the

efficiency of nuclear translocation in vivo since most human cells are

non-dividing [69]

Nuclear entry does not guarantee long-term expression of the encoded gene In most cases, only transient gene expression can be obtained as a re-sult of transcriptional silencing of the episomal pDNA (extrachromosomal; non-integrating pDNA into the cellular genome) in the nucleus [70,71] Recent advances suggest that efficient, long-term expression could be attain-able through site-specific integration of pDNA into “safe sites” on the cellu-lar genome, i.e those that are not associated with cell proliferation or tumour suppression [72] In addition, the introduced transgene may be maintained in the nucleus in a stable manner by the design of DNA molecules that can be replicated and transferred to daughter cells in an extrachromosomal form [73,74]

trans-Bacterial gene delivery (bactofection) and alternative gene therapy (AGT)

Bacteria-mediated transfer of pDNA (bactofection) is a technique, in which genetically modified (nonpathogenic) bacteria are used to transfer genes directly into the target organ or tissue [75] The basic idea of bactofection was introduced in 1980; transformed bacteria were used to deliver genes located on plasmids into cultured mammalian cells [76] The main advan-tages of bactofection are the simplicity of application and the possibility of selective gene transfer Another interesting approach; alternative gene ther-apy (AGT) was also proposed to be useful for gene therapy application In AGT, instead of using the transferred bacteria for pDNA delivery, they are exploited as a factory for the production of therapeutic peptides or proteins

in situ [77] AGT allows the levels of the expressed therapeutic product to be

controlled or even stopped by using specific antibiotics or activation of cidal genes in the bacteria Several studies have reported successful applica-tion of bactofection and AGT in genetic vaccination and gene therapy of

sui-several tumors, and monogenic and infectious diseases in vivo [78-80]

How-ever, both bactofection and AGT share several serious adverse effects that

are mainly related to host-bacteria interactions Stimulation of the host mune system by the introduced bacteria or the products of their lysis can lead to rapid clearance of the bacteria or even autoimmune reactions

im-Physical methods of non-viral gene delivery

Several strategies have been developed with the aim of enhancing the

effi-ciency of gene transfection using pDNA in vitro and in vivo One strategy

involves the use of physical methods of delivery such as nuclear tion, particle bombardment (ballistic delivery via gene gun), and electropora-tion (short, controlled electric pulses) [81-84] These methods avoid the problems associated with endocytosis and facilitate the introduction of pDNA into the intracellular environment either directly (injection) or through the disrupted cell membrane Examples of physical methods other

microinjec-than nuclear microinjection were shown to be efficient in the local in vivo

delivery of naked pDNA to tissues such as the skin and skeletal muscles [85,86] One important application of such methods is gene vaccination [86,87]

An alternative physical method for in vivo targeting involves hydrodynamic

injection, where the naked nucleic acids are injected in large volume under high pressure into the tail vein of animal models This technique has proved

to be efficient for pDNA and siRNA targeted delivery (passive targeting) mainly to the liver and, to a lesser extent to the skeletal muscles [88,89] However, as for nuclear microinjection, hydrodynamic injection is of minor application in clinical practice [90]

Chemical methods of non-viral gene delivery

Chemical methods of gene delivery involve the formulation of negatively charged nucleic acids with various polycations e.g cationic lipids and cati-onic polymers, into micro- or nanoparticulate structures (complexes) through electrostatic, ionic interactions (Figure 5) [7,91] The driving force behind the complex formation is the release of the counter ions associated with the polycations and the accompanying substantial gain in entropy [92] The physicochemical (size, morphology, surface charge and stability) and bio-logical properties of the complexes are dependent on factors such as the chemical structure of the polycation, the polycation/nucleic acid stoichiome-try (+/- charge ratio), the pH and the order of mixing To distinguish the origins of the complexing molecule, lipid-based formulations are referred to

as lipoplexes and polymer-based formulations as polyplexes [93]

Figure 5.Schematic illustration of polyplex (pDNA/cationic polymer) formation and pDNA compaction as visualised by atomic force microscopy

Lipids

The early attempts to formulate pDNA with neutral or anionic lipids were not successful because the resultant liposomes (spherical lipid-bilayer vesi-cles that can encapsulate nucleic acids) had poor physical properties (large particle size) as well as poor transfection efficiency [94] In 1987, the suc-cessful use of a cationic lipid (DOTMA; N-(1-(2,3-diolyloxy)-propyl)-N,N-trimethylammonium bromide) as a non-viral gene delivery system was re-ported for the first time [95] Following this report, a variety of cationic lip-ids were introduced to the field of gene delivery These lipids can be either monovalent (DOTMA, DOTAP) or multivalent (DOSPA: diolyloxy sper-minecarboxamidoethyl diethylpropanaminium trifluoroacetate; DOGS: dioc-tadecylamido-glycylspermine, Transfectam™) [96,97] Other neutral, helper lipids (co-lipids) were also introduced to improve the transfection efficiency

of cationic lipids especially those of the monovalent type Such helper lipids include DOPE and cholesterol [98-100] Commercially available lipid com-binations for gene delivery include Lipofectin™ (DOTMA/DOPE) and Li-pofectamin™ (DOSPA/DOPE)

Upon lipoplex formation, multilamellar structures of a size range of 0.2-1.0

µm are produced with the nucleic acid monolayers sandwiched between the lipid bilayers In general, lipoplexes formulated with monovalent lipids are physically unstable; they aggregate to give rise to heterogeneous shapes that have been described as “spaghetti and meatballs” [101,102] In order to pre-pare more defined lipid-based complexes, cationic thiol-detergents have been used to compact individual pDNA molecules into particles of around

32 nm The lipoplexes packed with pDNA molecules are physically lised by an oxidation-induced dimerisation of the detergent into a disulphide

stabi-lipid on the template pDNA [103] Recently, these lipoplexes were also

shown to be stable following intravenous injection in vivo [104]

The improved transfection efficiency of lipoplexes over naked pDNA is attributed to their resistance to enzymatic degradation, improved cellular uptake and efficient endosomal release Besides their efficiency as pDNA delivery systems, lipid-based formulations were reported as efficient deliv-

ery systems for siRNA in vitro and in vivo [105-107]

Although by 2005 8.3% of the gene therapy clinical trials were based on lipoplexes (http://www.wiley.co.uk/genetherapy/clinical/), several problems continue to limit their application as drug products for human use For in-stance, the formation of lipoplexes involves complex interactions between the lipid molecules, in addition to those with nucleic acids Additionally, the ability to control the size and morphology (colloidal satiability) of lipoplexes

is rather limited, with resultant instability problems over time [108]

Furthermore, toxicity associated with the use of lipoplexes in terms of

im-mune stimulation of the host can also limit their in vivo application The

toxicity of lipoplexes is closely associated with the administered dose and may, in part, result from their large size and the high positive surface charge [109,110] However, the resulting inflammatory responses of lipoplexes could be advantageous in specific applications such as vaccination and anti-tumour immunotherapies [111]

Polymers

In contrast to the large number of clinical trials investigating lipoplex-based gene therapy, gene therapy based on cationic polymers (polyplexes) is still in its infancy This seems “paradoxical” since polycations were being used for the insertion of DNA into cells long before lipid formulations [112] How-ever, progress was marginal until the introduction of polyethyleneimine (PEI) in 1995 This was primarily due to the low transfection efficiency of the used systems such as poly-L-lysine and protamine sulphate [46] Cati-onic polymers are interesting alternatives to cationic lipids in many respects The self assembly of polyplexes does not involve interactions of the polyca-tion molecules with each other, which results in better control of their physi-cal properties compared with lipoplexes In addition, the chemical structure

of various polycations comprises repeated units that can be easily lated by chemical modification to improve the physical and biological prop-erties of the resultant polyplexes, and consequently enhance their transfec-tion efficiency [108,113]

manipu-Several naturally occurring proteins, such as histones, cationised human serum albumin and gelatin, have been employed as non-viral gene delivery systems [114-116] However, the low transfection efficiency of these pro-

teins compared with that of the recently introduced synthetic cationic mers has limited their further application Examples of commonly used syn-thetic cationic polymers as non-viral delivery systems for nucleic acids in-clude PEI, polyamidoamine (PAMAM) dendrimers, poly-L-lysine (PLL) and chitosan (Figure 6)

Figure 6.Cationic polymers most commonly used for nucleic acid delivery [117]

Of the various cationic polymers, PEI has displayed several properties that placed it as the gold standard and one of the most efficient non-viral systems for nucleic acid delivery Therefore, PEI has been selected as a reference delivery system in this thesis, and the properties of PEI-based polyplexes will be discussed below in more detail

Polyethyleneimine (PEI)

Since the introduction of branched and linear PEI, the properties of based polyplexes have been extensively studied with the aim of devising safer, target-specific PEI derivatives [118-122] Various PEI derivatives

PEI-have been used to deliver oligonucleotides, ribozymes, RNA and pDNA in

vitro and in vivo The transfection efficiency and the toxicity of PEI depend

to a great extent on material characteristics such as the Mw, degree of

branching and cationic charge density While high Mw PEI (800 kDa) was associated with increased cellular toxicity in some studies, lower Mw coun-terparts (5-48 kDa) formulated at higher charge ratios were better tolerated

in cell cultures [123,124]

PEI is able to physically condense nucleic acids, especially pDNA, into small nanoparticles (less than 100 nm) that are suitable for cellular uptake The particle size and shape of PEI-based polyplexes depend on the structure and Mw of PEI, the pH and ionic strength of the surrounding environment, the method of preparation and the charge ratio used for polyplex formula-tion An excess of PEI (i.e higher charge ratios than 1:1 +/-) is required to produce smaller, positively surface-charged polyplexes that display im-proved colloidal stability [125]

In addition to enhanced pDNA packing and physical stability, the presence

of excess positive charges on the surface of the polyplexes facilitates tion with the negatively charged cell membrane components and conse-

interac-quently enhances the internalisation of the polyplexes in vitro [126] As for

cationic lipids, it has been proposed that extensive damage of the cell brane due to high positive charge density contributes to the cytotoxicity of PEI polyplexes [109,127] Fluid-phase and adsorptive endocytosis have been

mem-reported as possible cellular uptake mechanisms of PEI polyplexes in vitro

[128,129]

It has been postulated that the high transfection efficiency of PEI may be due

to the ability of the polymer to capture protons in the slightly acidic dosomal compartments This buffering capacity makes PEI act as a “proton sponge”, allowing the cell to pump more protons together with chloride ions and water into the endosomes, leading to the swelling and eventual rupture

en-of the endosomal membrane [46] Although this mechanism has been cised, several recent reports support the role of the buffering capacity of PEI

criti-in its superior transfection efficiency criti-in vitro [130,131]

Upon cytoplasmic release, PEI can protect the complexed nucleic acids from enzymatic degradation Moreover, PEI polyplexes undergo intracellular traf-ficking and nuclear translocation more efficiently than lipoplexes or naked pDNA [61] Interestingly, PEI polyplexes were shown to be transported in the cytoplasm by more than one mechanism: diffusive transport, restricted transport (by altering the structure of the cytoskeleton) and active transport through the microtubules [132] Unlike lipids, intact polyplexes formulated with branched PEI were found inside the nucleus [52,128] Accordingly, cell division is not a prerequisite for the nuclear translocation and expression of

nucleic acids formulated with PEI in vitro [133] However, the presence of

free or complexed PEI in the nucleus may, in part, explain its increased tracellular toxicity [134,135]

in-In contrast to studies of cationic lipids, initial investigations indicated that cationic polymers (e.g PEI) were unsuitable for the delivery of oligonucleic acids such as ODN and siRNA [136] However, several recent studies have reported contradictory findings [137,138] One possible reason for this dis-crepancy is poor characterisation of the critical formulation parameters re-quired for the efficient delivery of oligonucleic acids Therefore, the struc-ture-activity relationship of siRNA delivery will be addressed in this thesis

Polyamidoamine (PAMAM) dendrimers

PAMAM dendrimers (e.g Superfect™) are highly branched cationic mers that are of comparable efficiency to PEI [139]

poly-Poly-L-lysine (PLL)

PLL is a linear biodegradable polymer composed of repeated lysine units Polyplexes based on PLL are generally less efficient than those of PEI and PAMAM dendrimers The most important reason for the poor transfection efficiency of PLL-based polyplexes is their minimal ability to escape from the endosomes [140] Thus, PLL-mediated gene delivery requires co-administration of endosomolytic agents such as chloroquine for an efficient escape from the endosomal compartments [141] However, this approach is

not practical for in vivo application because of the dose-dependent toxicity of

chloroquine Further, it is hard to target chloroquine to the specific cells transfected by PLL polyplexes [142] The introduction of chemical residues such as histidine, which can be protonated at the slightly acidic pH of the endosomal compartments, to PLL has significantly improved the transfection efficiency by facilitating endosomal escape of the polyplexes [142] The greater efficiency of the histidylated PLL compared to unmodified PLL pro-vides further evidence for the “proton sponge” theory of the mechanism of action of PEI

Strategies to improve the in vitro/in vivo efficiency of

non-viral gene delivery: Structure-activity relationship

Molecular weight reduction

Several recent studies have reported higher transfection efficiency and lower cellular toxicity of low Mw PEI (less than 25 kDa) than for the higher Mwcounterpart [123,137,143-145] In one of these studies, the biocompatibility

of low Mw modified PEI was superior to that of non-degradable PEI (25

kDa) and the modified version mediated gene transfection in cultured

neu-rons and in the central nervous system (rat spinal cord) in vivo [143]

Positive charge shielding

The introduction of a hydrophilic polymer layer to the surface of the plexes via covalent or non-covalent linking to the polymer backbone can significantly enhance the transfection efficiency of the polyplexes The steri-cally-stabilised polyplexes have enhanced colloidal stability in terms of a reduced susceptibility towards aggregation in physiological surroundings

poly-Moreover, in vivo, the reduced positive surface charge on the polyplexes will

suppress non-specific interactions, leading to increased circulation time and reduced toxicity Examples of hydrophilic polymers that have been used for such purpose include poly(ethyleneglycol) PEG, pluronics, polyacrylic acid (PAA) and methacrylate derivatives [89,146-149] However, uncontrolled substitution of the polymer backbone with such hydrophilic moieties can result in a reduced charge interaction between the cationic polymer and the negatively charged nucleic acids i.e impaired complexation [147]

Active targeting

The in vivo targeting of non-viral gene delivery systems to cells or tissues

can be accomplished by taking advantage of special physiological tions, e.g irregular fenestration in the liver, spleen, bone marrow and lungs

condi-or increased vascularisation (blood supply) to certain tumcondi-ors, to facilitate passive accumulation (passive targeting) However, receptor targeting seems

to be a more reliable means of achieving site-directed gene delivery The concept of receptor-mediated endocytosis has been extensively exploited to

achieve targeted gene delivery in vitro and in vivo Various ligands that

tar-get cell-surface receptors (transferrin, folate, lectin, growth factors and tegrin receptors) have been coupled to cationic polymers [147,150-156] For example, coupling of sugar residues such as galactose, lactose and N-acetyl glucosamine (GlcNAc) to cationic polymers has been reported to signifi-cantly enhance the transfection efficiency of PEI and PLL polyplexes in various cell lines expressing lectin receptors such as liver and airway epithe-

in-lial cells in vitro [151-153,157] However, studies on the use of targeted polyplexes to the lungs in vivo have not yet been reported at the

lectin-onset of this thesis

In vivo toxicity of polyplexes

As described above, cationic polymers can complex nucleic acids into small, positively charged nanoparticles that are internalised via non-specific ad-

sorptive endocytosis in cell cultures in vitro However, the in vivo situation

is different, and the highly charged particles may fail to reach target cells due to non-specific interactions with blood components (e.g plasma proteins including the complement components, blood cells), extracellular matrix and non-target cells or tissues While interaction with the complement compo-nents can trigger the activation of the complement system and removal of the polyplexes, interaction with plasma proteins such as albumin leads to aggre-gation of the complexes and subsequent accelerated clearance from the cir-culation [38,109] Moreover, aggregated polyplexes can cause obstruction of blood vessels or lung capillaries, which can be fatal Further, recent studies

in mice have shown that polyplexes formulated with the highly efficient cationic polymer; PEI resulted in a dose-dependent toxicity and up-

regulation of genes involved in inflammatory processes following in vivo lung administration [134,135] The unfavourable in vitro/in vivo safety pro-

file of most available non-viral gene delivery systems motivated a search for alternative biocompatible cationic polymers that can form efficient nontoxic polyplexes In this thesis, the natural polymer chitosan is investigated as such an alternative

Chitosan chemical structure

Chitosan is a linear, binary hetero-polysaccharide consisting of randomly distributed (1-4)-linked 2-acetamido-2-deoxy-E-D-glucose and 2-amino-2-deoxy-E-D-glucose molecules (Figure 7) It is mainly obtained by partial

alkaline de-N-acetylation of the second most abundant polysaccharide in

nature after cellulose: chitin, primarily sourced from crustacean and insect shells [158] Chitin also exists naturally in some microorganisms and fungi such as yeast The word chitosan refers to a family of aminopolysaccharides, which differ according to their degree of acetylation (DA, 2%-60%) and Mw(50-2000 kDa) [159] These two characteristic features are of utmost impor-tance because of their heavy impact on the physicochemical, biological and the formulation properties of chitosan [160]

Figure 7.Chemical structure of chitin (R = COCH3) and chitosan (R = H or COCH3)

Estimation of the number of papers published on the use of chitosan in drug delivery over the last 10 years suggests that this molecule is attracting in-creasing interest in the drug delivery community (Figure 8)

On closer examination of the published articles on the pharmaceutical and biomedical applications of chitosan in the PubMed database, it is evident that more than half are related to drug delivery, and a significant fraction of these are related to delivery of macromolecules such as peptide and protein drugs, vaccines or genes [161]

applica-tions of chitosan This

survey did not include

on factors such as the Mw, DA, pH, ionic strength and temperature [164] Protonated chitosan can form gels with polyanions, a property that was ex-ploited in the development of controlled-release formulations of various drugs [159] The possibility of chemical modification of chitosan, for in-stance, by introducing chemical groups to its backbone structure through covalent binding has resulted in a series of versatile chitosan derivatives with improved physicochemical properties and transfection efficiency [165-168]

as the probable mechanism of action of chitosan as a transmucosal tion enhancer [161]

absorp-General applications of chitosan in drug delivery

Examples of potential applications of chitosan in conventional cal devices include its use as binder, disintegrant and coating material [171] Because of its unique polymeric character, gel-forming properties and pro-pensity for degradation, chitosan-based controlled and targeted delivery sys-tems have also been developed and examined [159] Several reports have confirmed the absorption enhancing characteristics of chitosan formulations following nasal administration of vaccines, peptides, proteins and low Mwdrugs such as morphine [172-174] In addition to the nasal route, chitosan has been investigated as a vehicle for ocular and peroral drug delivery in order to prolong contact time and improve drug absorption [175,176]

pharmaceuti-Chitosan as a delivery system for proteins and peptide drugs

Since Illum et al [177] reported in 1994 that high Mw chitosan solutions significantly increased the transport of insulin across the nasal mucosa of rats and sheep, several human studies have confirmed the potential of chito-san to improve the mucosal absorption of peptides [178-180] Furthermore, studies on the nasal administration of protein-antigens showed that chitosan solutions, and in particular chitosan powders, enhanced the immune response

to antigens such as influenza, pertussis and diphtheria [181-183] For ple, a chitosan-based nasal vaccine system for influenza was evaluated in

exam-human subjects [182] The results were encouraging; over 70% of the teers had protective levels for some strains after intranasal administration, a rate which was comparable to that obtained after intramuscular administra-tion In a study from 2004, a chitosan powder formulation of a diphtheria vaccine administered intranasally to humans resulted in high T-helper cells

volun-of type 2 responses, which correlated with protective levels volun-of neutralising antibodies [174] Collectively, these studies provided convinc-ing evidence for the potential of chitosan to improve the transmucosal deliv-ery of macromolecular drugs The positive effect of chitosan solu-tions/powders on the transmucosal delivery of peptides and proteins has been explained by its mucoadhesive properties and its capacity to open the tight junctions between epithelial cells, thus facilitating the transport of mac-romolecular drugs through well organised epithelia [184]

toxin-Chitosan as non-viral gene delivery system

The use of chitosan in gene delivery was described for the first time in 1995 [185] Since then, slow but steady progress has been made, resulting in chi-tosan now being placed among the most effective non-viral gene delivery systems for mucosal application As for chitosan-mediated protein delivery,

the most promising in vivo results have been obtained in vaccine research

The first significant result was obtained in mice after oral administration of chitosan nanoparticles carrying plasmid pDNA encoding a peanut allergen [186] The treatment resulted in induced tolerance to peanut allergy, which was related to allergen-specific secretory Immunoglobulin A and serum Im-munoglobulin G (IgA and IgG, respectively)

Several studies have suggested that chitosan would be a suitable delivery vehicle for administration of pDNA to the airway mucosa Recently, intrana-sal vaccination with chitosan nanoparticles carrying pDNA encoding anti-gens of Respiratory Syncytial Virus (RSV) protected mice against challenge with the virus [187] The utility of chitosan-pDNA vaccines against RSV infections was further corroborated by Iqbal et al [188] In 2005, nanoparti-cles of chitosan complexed to pDNA coding for siRNA against the RSV NS1 protein were intranasally delivered to mice, resulting in a substantial decrease in virus titres in the lungs and decreased inflammation and airway reactivity [189] The significance of these results is related to the current lack of effective prophylaxis against RSV infection in the lower respiratory tract

It has been proposed that the mechanism of chitosan-mediated gene delivery,

as with PEI, involves initial interaction between the positively charged san-pDNA polyplexes and the negatively charged cell membrane compo-nents such as heparan sulphate proteoglycans (HSPG) on the cell membrane

chito-After binding, the polyplexes are taken up into endocytic vesicles by cytosis Subsequently, early escape from the endosomes followed by nuclear translocation and gene transcription are required for successful gene expres-sion [190] However, further studies are needed to elucidate the transfection mechanism for chitosan-mediated gene delivery It remains to be seen if chitosan nanoparticles facilitate transcytosis across mucosal epithelia and in particular across the follicle-associated epithelium

endo-At the onset of this thesis, chitosan-based gene formulations were generally based on high Mw chitosans that were poorly characterised in terms of Mwdistribution and DA The use of such chitosans resulted in poor transfection

efficiency in vitro and in vivo [191,192] Therefore, in order to improve the

various properties of chitosan-based gene formulations, this thesis oured to establish the structure-property relationships of well-defined, low

endeav-Mw chitosans (chitosan oligomers) as delivery systems for nucleic acids

(pDNA and siRNA) in vitro and after lung administration in vivo In

addi-tion, the concept of receptor targeting was exploited to further enhance the gene delivery efficiency of chitosan oligomer-based polyplexes Finally, improved methodology for the administration of gene formulations to the

lung in vivo was explored

The overall aim of the thesis was to develop efficient chitosan based gene formulations to deliver nucleic acids (pDNA or siRNA) into mammalian cells The specific aims were to:

oligomer-x Investigate the effect of well-defined low Mw chitosan oligomers on the physicochemical properties and the transfection efficiency of the formulated pDNA complexes compared to high Mw conventional

chitosans in vitro and in vivo (Paper I)

x Enhance the transfection efficiency of chitosan oligomer-based

pDNA formulations in vitro and after lung administration in vivo by

coupling a targeting moiety (a trisaccharide branch) to the chitosan backbone (Paper II)

x Investigate the potential of chitosan oligomers as delivery systems

for siRNA in vitro, and to study the effect of controlling chitosan

structural variables and formulation parameters on the gene ing activity of siRNA complexes formulated with chitosan oli-gomers (Paper III)

silenc-x Optimise a new catheter device for improved aerosol delivery of gene formulations to the mouse lung (Paper IV)

Materials and methods

Nucleic acids

GMP-grade plasmids (gWizTM) containing a cytomegalovirus (CMV) moter and a firefly luciferase (pLuc) or green fluorescent protein (pGFP) were purchased from Aldevron, Fargo, ND, USA The unmodified siRNA duplex, siGL3 (sense, 5'-CUUACGCUGAGUACUUCGAdTdT-3'; an-tisense, 5'-UCGAAGUACUCAGCGUAAGdTdT-3') that targets the luciferase gene (siRNA-Luc) was ordered from MedProbe (Lund, Sweden)

pro-A mismatching siRNpro-A: siCONTROL non-targeting siRNpro-A #1 (siCON1; sense, 5'-UAGCGACUAAACACAUCAAUU-3'; antisense, 5'-UUGAUGUGUUUAGUCGCUAUU-3'), was ordered from Dharmacon Research, Inc (Lafayette, CO)

Polycations

In Paper I, Ultrapure chitosan (UPC), Protosan UPG 210, with a degree of polymerisation (DP) of around 1000 and a Mw of 162 kDa (endotoxin con-tent < 1500 EU/g, total viable count < 100 cfu/g and heavy metal content <

100 ppm) was obtained from Pronova Biopolymer, Oslo, Norway In Papers

I-IV, fully de-N-acetylated chitosan oligomers (fraction of acetylated units;

FA< 0.001) with number-average degree of polymerisation (DPn) of 18, 34,

50 and 85 monomer units were prepared by alkaline de-N-acetylation with

subsequent nitrous acid depolymerisation and reduction as described ously [193] The chain length distributions were analysed by size exclusion chromatography with a multiangle laser light scattering detector (SEC-

previ-MALLS) The chitosans were defined according to their DP as DPn18,

DPn34, DPn50 and DPn85 In Paper I, DPn18 chitosan oligomers were further fractionated by Superdex 30 gel filtration as previously described and frac-tions corresponding to DP intervals of 10-14, 15-21, 22-35 and 36-50 were pooled [193] In Paper II, DPn18 and DPn34 chitosan oligomers were substi-tuted by reductive N-alkylation with the trimer 2-acetamido-2-deoxy-D-

2,5-anhydro-D-mannofuranose (A-A-M) to obtain oligomers with 7, 23 and

glucopyranosyl-E-(1-4)-2-acetamido-2-deoxy-D-glucopyranosyl-E-(1-4)-40 % substituted amino groups [194] 1H NMR spectroscopy was used to determine the degree of substitution (DS) Branched PEI (Mw 25 kDa) was purchased from Aldrich Sweden, Stockholm, Sweden Linear PEI (ExGen 500;Mw 22 kDa) was purchased from Ferementas, Germany Lipofectamine

2000 (LF2000) was purchased from Invitrogen

Cells

The human embryonic kidney cell line HEK 293 and the human cervix epithelial cell line HeLa were obtained from ATCC, Rockville, MD, USA The human airway epithelial cell line Calu-3 was a kind gift from Dr Ursula Hultkvist-Bengtsson, AstraZeneca R&D Lund, Sweden The human bron-chial cell line 16HBE14o- was a gift from Professor Godfried Roomans, Department of Medical Cell Biology, Uppsala University, Sweden The hu-man liver hepatocyte cell line HepG2 was a gift from Professor Edvard Smith, Unit for Molecular Cell Biology and Gene Therapy Science, Karolin-ska Institute, Sweden The HEK 293 (293-Luc) cell line that stably expresses firefly luciferase was a gift from Dr Paavo Honkakoski, Department of Pharmaceutics, University of Kuopio, Finland [195] A stably luciferase-expressing ovarian carcinoma cell line (SKOV-3-Luc) was a gift from Dr Achim Aigner, Department of Pharmacology and Toxicology, Philipps-University Marburg, Germany [138].The cells were maintained according to the supplier's recommendations

Formulation of complexes

Chitosan stock solutions (0.2-2 mg/ml) were prepared by dissolving chitosan

in sterile MilliQ water at pH 6.2 followed by sterile filtration Chitosan plexes were formulated by adding chitosan and then pDNA or siRNA stock solutions to sterile MilliQ water during intense stirring on a vortex mixer (Heidolph REAX 2000, level 4, Kebo Lab, Spånga, Sweden) as previously

com-described [54] For in vitro studies (i.e gel retardation and physical stability

assays, morphology, and transfection experiments), different charge ratios of chitosan complexes were produced at a constant pDNA or siRNA concentra-tion (13.3 µg/ml) The charge ratio was defined as the ratio of the maximum number of protonable primary amines in chitosan to the number of negative

phosphates on pDNA or siRNA [93] For in vivo studies, chitosan complexes

were prepared at pDNA concentrations of 5, 10 and 25 µg/ml in 1 ml sterile MilliQ water The complexes were concentrated in a SpeedVac Plus centri-fuge (Savant Instruments, Holbrook, NY) at 1,400 rpm for approximately 90 minutes to obtain DNA concentrations of around 50, 100 and 250 µg/ml, respectively PEI 25 kDa stock solutions (10 mM) and PEI complexes were

prepared as described previously [54] In Papers I, II and IV, an optimal charge ratio of 5:1 (+/-) was used for PEI complexes [196] In Paper III, an optimal charge ratio of 15:1 and a weight ratio of 2:1 (+/-) were used to for-mulate PEI and LF2000 complexes with siRNA

Size and morphology of the complexes

The size (hydrodynamic diameter) of the complexes was determined by ton correlation spectroscopy (Zetasizer 4000, Malvern Instruments, Malvern,

pho-UK for pDNA complexes and Nanosizer ZS, Malvern Instruments, Malvern,

UK for siRNA complexes) as previously described [54] In Paper I, san complexes were imaged by atomic force microscopy (AFM)

Chito-Physical and enzyme stability

The physical stability of the formulated complexes was studied using the agarose gel retardation assay Concentrations of 0.8 and 4 % agarose in 40

mM TAE buffer were used as previously described for pDNA and siRNA complexes, respectively [54] Protection of complexed pDNA and siRNA against enzyme degradation was studied after incubating the complexes with DNase I and RNase A, respectively as described previously The integrity of pDNA and siRNA was investigated by incubating the complexes with the polyanion heparin (5 mg/ml) followed by gel electrophoresis

In vitro studies

Transfection experiments

Most of the in vitro transfection experiments were carried out in HEK 293

cells, since previous studies had demonstrated a qualitative correlation tween transfection efficacy in this cell line and that in the mouse lung epithe-

be-lium after intratracheal instillation in vivo [54] Twenty-four hours before the

transfection experiments, HEK 293 (normal and 293-Luc), HepG2, 16HBE14o- and SKOV-3-Luc cells were seeded in 96-well tissue culture plates (Costar, Cambridge, UK) to obtain a cell confluency of 70-80% on the day of transfection Transfections were carried out at pH 7.4 in serum-free medium (OptiMEM I Reduced Serum Media, Gibco/BRL Life Technologies

AB, Täby, Sweden), at pH 5.0 in acetate buffer or in phosphate-buffered saline (PBS) that was free from magnesium and calcium ions (PBS, Sigma) Isotonicity (300 mOsm/kg) was obtained by the addition of mannitol In papers I, II and IV, the cells were washed in pre-heated OptiMEM and 50 µl

of the formulation (corresponding to 0.33 µg pDNA) was added to each well In paper III, various concentrations of siRNA complexed with the se-lected polycations were used with or without the co-transfection of 0.3 µg pDNA (pLuc or pGFP) per well To investigate the endosomolytic properties

of the chitosan oligomer formulations (Paper II), the transfection medium was supplemented with 200 nM Bafilomycin A1 (Sigma-Aldrich Sweden

AB, Stockholm, Sweden) as previously described [142] For the competition studies (Paper II), various amounts of the free A-A-M branch (5 and 10 mM) were added to the branched chitosan complexes After 5 h incubation, the formulations were removed and 0.2 ml of fresh culture medium was added The medium was changed every second day for experiments that exceeded two days At pre-specified time points, ranging from 24 to 144 hours after transfection, the cells were washed with PBS (pH 7.4), and lysed with luciferase lysis buffer (Promega, Madison, WI) The luciferase gene expres-sion was then measured with a luminometer (Mediators PhL, Vienna, Aus-tria) The amount of luciferase expressed was determined from a standard curve prepared with firefly luciferase (Sigma, St Louis, MO) and total cell protein was measured using the bichinchoninic acid test (Pierce, Rockford, IL.)

Cellular uptake of chitosan complexes

In Paper II, HEK 293 cells were plated one day before transfection on well plates or 4-well Lab-Tek chambered coverglasses (Nalge Nunc Interna-tional, Napperville, IL, USA) FITC-labelled pCMV-Luc complexed with linear or branched chitosan oligomers was added to the cells After 5 h incu-bation, the cells were washed with the polyanion heparin (2 mg/ml in PBS)

96-to dissociate and remove the membrane-bound complexes and fresh culture medium was added [126] The fluorescence emitted from the internalised particles was measured directly with a plate reader (TECAN Safire2, Tecan Austria GmbH, Grödig, Austria) Fluorescent images of transfected HEK 293 cells were obtained with a confocal laser scanning microscope (LSM 510 META confocal microscope, Carl Zeiss) A live-cell compatible dye (Cell tracker orange CMTMR, Molecular Probes, The Netherlands) was used to stain the transfected HEK 293 cells Measurements were performed directly

on the Lab-Tek dishes, on living cells at 37o C In order to quantitatively assess the data obtained from the confocal images, stacks of images (com-posed of 30 optical sections) for each formulation were analysed Statistical analysis of the number of internalised particles was performed using the non-parametric Mann-Whitney test

Cellular toxicity (Intracellular dehydrogenase activity)

In Paper III, the effect of various siRNA formulations on the intracellular dehydrogenase activity (a measure of cellular toxicity) in 293-Luc cells was evaluated by the MTT method as described previously [197]

In vivo studies

The animal experiments were approved by The Swedish National Board for Laboratory Animals (the local ethical committee in Uppsala, Sweden) Balb/c mice aged 6-8 weeks were anaesthetised with ketamin/xylazine (5/20 vol%, 0.1 ml/10 g of body weight), and the trachea was surgically exposed

by a 0.5 cm skin incision in the neck The formulations were administered into the trachea through a 28 G needle (2 X 50 µl portions, Papers I and II)

or via nebulisation with the NCD (10 µl, Paper IV)

Luciferase gene expression

At the pre-specified time points, the mice were sacrificed (CO2), and the lungs were removed, washed in ice-cold PBS, homogenised in a Beadbeater (Biospec Products, Bartlesville, OK) for 1 min in ice-cold luciferase lysis buffer (Promega) with a protease inhibitor cocktail (Complete, Boehringer Mannheim Scandinavia AB, Bromma, Sweden), centrifuged at 15,000 rpm

at 4° C, and mixed with luciferase reagent (Promega); luciferase gene

ex-pression was determined as described for in vitro transfection above

Distribution pattern of gene expression in the mouse lung

In Paper I, in order to determine the distribution of luciferase gene sion in mouse lungs, lungs were removed, placed on ice and dissected into parenchyma, bronchial tree and trachea A blunt scalpel was used to care-fully scrape off the lung parenchyma, leaving a network of bronchi and bronchioles Samples were then processed and assayed for luciferase gene expression In Paper IV, the lung distribution pattern of Evans blue dye solu-tion and fluorescein-labeled pCMV-Luc (5 Pg) complexed with DP 36-50 chitosan oligomers were investigated in the mouse lung after administration via intratracheal nebulisation through the NCD or conventional intratracheal instillation

expres-Toxicological evaluations

The lung histology was evaluated blind for the appearance of inflammatory cells and structural damage as previously described [198] Mice were killed

24 h after intratracheal administration of the polyplexes and, after cardiac perfusion with PBS and 3% paraformaldehyde in PBS, the lungs were re-moved, rinsed in PBS and stored overnight in 3% paraformaldehyde The lungs were frozen the next day in OCT Embedding Medium (Sakura Finetek Europe, Zoeterwoude, The Netherlands) Cryosections (5 µm) were cut in a Leica Jung CM 3000 cryostat (Leica Instruments GmbH, Nussloch, Ger-many), stained briefly with hematoxylin and eosin, mounted and examined under a light microscope

Statistics

The experiments were performed on a minimum of two occasions using quadruplicate samples Gene expression is presented as the amount of the

expressed transgene per µg (in vitro) or mg (in vivo) of tissue protein All

data are expressed as mean values ± one standard deviation Statistical ferences between mean values were investigated using ANOVA Statistical differences between the physical shapes of the complexes were investigated using contingency table analysis Statistical differences are denoted as *(p < 0.05), ** (p < 0.01) and *** (p < 0.001)

dif-Results and discussion

Optimised linear chitosan oligomers as non-viral gene delivery systems (Paper I)

pDNA polyplexes formulated with low Mw chitosans were previously

char-acterised in our laboratory [199] The in vitro and in vivo transfection

effi-ciencies of these polyplexes were shown to be dependent on their physical properties While polyplexes formulated with chitosan oligomers with a DP less than 14 monomer units (DP 14) were physically unstable, the use of chitosan oligomers with DP > 24 resulted in polyplexes that were signifi-cantly stable Therefore, the aim of Paper I was to examine the effect of chi-tosan oligomers of intermediate chain length (14-24 monomer units) on the physical properties and transfection efficiency of the polyplexes It was hy-pothesised that such chitosans would provide the advantage of easier release

of the complexed pDNA while maintaining the required protection against enzymatic degradation Therefore, a fully deacetylated chitosan oligomer fraction with a number-average degree of polymerisation (DPn) of 18 monomer units (DPn18) was produced DPn18 was a polydispersed fraction that contained oligomers between 6 and 50 mers (Figure 9A)

Characterisation of DPn18 polyplexes in vitro and in vivo

High charge ratios were required to obtain DPn18 polyplexes that were physically stable in the gel retardation assay and to obtain high luciferase

gene expression in HEK 293 cells in vitro (Figure 9B, 9C) The excess of

low Mw chitosan in the polyplex formulations suggested that the toxicity of the formulations could have been increased However, in agreement with the reduced toxicity of low Mw PEI, no acute toxicity of chitosan oligomers was

detected in the MTT assay in vitro or in the histological sections of the mouse lung after intratracheal administration in vivo [123] Further, initial in

vivo investigations showed that expression of the luciferase gene mediated

by DPn18 polyplexes in the mouse lung was significantly higher than that mediated by both shorter and longer chitosan oligomers (Figure 9D)