The AAPS Journal 2007; 9 (1) Article 9 (http://www.aapsj.org).
` Themed Issue: Non-viral Gene Delivery: Opportunities and Challenges
`Guest Editors - Dexi Liu and Craig K. Svensson
` Nonviral Gene Delivery: What We Know and What Is Next
` Submitted: January 6 , 2007 ; Accepted: January 26 , 2007 ; Published: March 23 , 2007
` Xiang Gao , 1 Keun-Sik Kim , 1 and Dexi Liu 1
` 1 Department of Pharmaceutical Sciences, School of Pharmacy, University of Pittsburgh, Pittsburgh, PA 15261
`
` A BSTRACT
` Gene delivery using nonviral approaches has been exten-
`sively studied as a basic tool for intracellular gene transfer
`and gene therapy. In the past, the primary focus has been on
`application of physical, chemical, and biological principles to
`development of a safe and effi cient method that delivers a
`transgene into target cells for appropriate expression. This
`review summarizes the current status of the most commonly
`used nonviral methods, with an emphasis on their mechanism
`of action for gene delivery, and their advantages and limita-
`tions for gene therapy applications. The technical aspects of
`each delivery system are also reviewed, with a focus on how
`to achieve optimal delivery effi ciency. A brief discussion of
`future development and further improvement of the current
`systems is intended to stimulate new ideas and encourage
`rapid advancement in this new and promising fi eld.
`
` K EYWORDS: Gene delivery , gene therapy , nonviral vectors ,
` transfection
`
` INTRODUCTION
` The primary challenge for gene therapy is to develop a
`method that delivers a therapeutic gene (transgene) to
`selected cells where proper gene expression can be achieved.
`An ideal gene delivery method needs to meet 3 major crite-
`ria: (1) it should protect the transgene against degradation
`by nucleases in intercellular matrices, (2) it should bring
`the transgene across the plasma membrane and into the
`nucleus of target cells, and (3) it should have no detrimental
`effects.
` Viral vectors are able to mediate gene transfer with high
`effi ciency and the possibility of long-term gene expression,
`and satisfy 2 out of 3 criteria. The acute immune response,
`immunogenicity, and insertion mutagenesis uncovered in
`gene therapy clinical trials have raised serious safety con-
`cerns about some commonly used viral vectors. The limita-
`tion in the size of the transgene that recombinant viruses can
`carry and issues related to the production of viral vectors
`
` Corresponding Author: Dexi Liu, Department of
`Pharmaceutical Sciences, School of Pharmacy, University
`of Pittsburgh, Pittsburgh, PA 15261 . Tel: (412) 648-8553 ;
` Fax: (412) 383-7436 ; E-mail: dliu@pitt.edu
`
`present additional practical challenges. Methods of nonviral
`gene delivery have also been explored using physical
` (carrier-free gene delivery) and chemical approaches (syn-
`thetic vector-based gene delivery). Physical approaches,
`including needle injection, 1 electroporation, 2 , 3 gene gun, 4 , 5
`ultrasound, 6 and hydrodynamic delivery, 7 , 8 employ a physi-
`cal force that permeates the cell membrane and facilitates in-
`tracellular gene transfer. The chemical approaches 9-12 use
` synthetic or naturally occurring compounds as carriers to
`deliver the transgene into cells. Although signifi cant prog-
`ress has been made in the basic science and applications of
`various nonviral gene delivery systems, the majority of
`nonviral approaches are still much less effi cient than viral
`vectors, especially for in vivo gene delivery. In this review,
`we will briefl y discuss the advantages and limitations of the
`nonviral gene delivery systems that are shown to be active
`for in vivo gene delivery. We will also highlight approaches
`toward development of improved nonviral systems for
`human gene therapy. We hope that our discussion here will
`stimulate new thoughts and efforts toward advancement of
`this diverse and promising new fi eld.
`
` GENE TRANSFER BY NEEDLE INJECTION
`OF NAKED DNA
` Simple injection of plasmid DNA directly into a tissue with-
`out additional help from either a chemical agent or a physi-
`cal force is able to transfect cells. Local injection of plasmid
`DNA into the muscle, 1 liver, 13-15 or skin, 16 or airway instil-
`lation into the lungs, 17 leads to low-level gene expression.
`Specifi c or nonspecifi c receptors on the cell surface that
`bind and internalize DNA have been implicated as a mecha-
`nism, though the details are sketchy at this point. Neverthe-
`less, gene transfer with naked DNA is attractive to many
`researchers because of its simplicity and lack of toxicity.
`Practically, airway gene delivery and intramuscular injec-
`tion of naked DNA for the treatment of acute diseases and
`DNA-based immunization, respectively, are 2 areas that are
`likely to benefi t from naked DNA-mediated gene transfer,
`provided that further improvements are made in delivery
`effi ciency and duration of transgene expression. A broad
`application of naked DNA – mediated gene transfer to gene
`therapy may not be conceivable because DNA, being large
`in size and highly hydrophilic, is effi ciently kept out of the
`cells in a whole animal by several physical barriers. These
`
`E92
`
`Moderna Ex 1008-p. 1
`Moderna v Protiva
`
`

`

`The AAPS Journal 2007; 9 (1) Article 9 (http://www.aapsj.org).
`include the blood endothelium, the interstitial matrices, the
`increased transgene expression in isolated lung in an ex
`vivo organ culture setting, 28 and Dean et al showed that
`mucus lining and specialized ciliate/tight junction of epi-
`such a strategy also worked in live animals when a pair of
`thelial cells, and the plasma membrane of all cells. In addi-
`electrodes was placed on the chest. 29 The level of reporter
`tion, DNA degradation by intra- and extracellular nuclease
`gene expression obtained was 2 to 3 orders of magnitude
`activities further reduces the chance that DNA entering
`higher than that with plasmid DNA alone. DNA as large as
`nuclei will be intact and functional. The current strategy for
`100 kb has been effectively delivered into muscle cells. 30
`improving naked DNA – based gene transfer is to include in
`Long-term expression over 1 year after a single electropora-
`DNA solution substances capable of enhancing the effi ciency
`tion treatment was seen. 31 Gene transfer by electroporation
`of DNA internalization by target cells. For example, trans-
`ferrin has been shown to enhance transfection in vitro. 18
`showed less variation in effi ciency across species than did
`The addition of water-immiscible solvents, 19 , 20 non-ionic
`direct DNA injection. The amount of DNA and how well the
`polymers, 21 or surfactants, 22 or the use of hypotonic solu-
`injected plasmid DNA distributes within the treated tissue
`tion, 23 has also been shown to elevate gene transfer across
`prior to electroporation appear to have an important impact
`cell membranes. Also, several nuclease inhibitors have been
`on transfection effi ciency. It was also reported that age
`shown to enhance naked DNA – mediated gene transfer in
`of the recipient animals affects the transfection effi ciency
`cultured cells, 24 muscle, 25 and lungs. 26
`in mice. 32 Treatment of muscle with hyaluronidase prior
`to injection of plasmid DNA to loosen up the surrounding
`extracellular matrix signifi cantly enhanced transfection,
`possibly because of improved distribution of plasmid DNA
`in the tissue. 32 , 33 Alternatively, plasmid DNA administra-
`tion through the portal vein followed by localized electro-
`poration on rat liver resulted in widespread transfection in
`hepatocytes in the treated lobe but not in the surrounding
`lobes. 34 This result raises the possibility that one can supply
`cells with plasmid DNA via blood circulation and then apply
`electroporation to a selected area to achieve localized gene
`transfer. A short time interval between DNA administration
`and electroporation is critical to minimize DNA degradation
`by extracellular nucleases.
` Several major drawbacks exist for in vivo application of
`electroporation. First, it has a limited effective range of
`~1 cm between the electrodes, which makes it diffi cult to
`transfect cells in a large area of tissues. Second, a surgical
`procedure is required to place the electrodes deep into the
`internal organs. Third, high voltage applied to tissues can
`result in irreversible tissue damage as a result of thermal
`heating. 35 Ca 2+ infl ux due to disruption of cell membranes
`may induce tissue damage because of Ca 2+ -mediated prote-
`ase activation. 36 The possibility that the high voltage applied
`to cells could affect the stability of genomic DNA is an
`additional safety concern. However, some of these concerns
`may be resolvable by optimizing the design of electrodes,
`their spatial arrangement, the fi eld strength, and the dura-
`tion and frequency of electric pulses.
`
` Transfer by Gene Gun
` Particle bombardment through a gene gun is an ideal method
`for gene transfer to skin, mucosa, or surgically exposed
` tissues within a confi ned area. 4 DNA is deposited on the
` surface of gold particles, which are then accelerated by pres-
`surized gas and expelled onto cells or a tissue. The momen-
`tum allows the gold particles to penetrate a few millimeters
`deep into a tissue and release DNA into cells on the path. Such
`a simple and effective method of gene delivery is expected
`to have important applications as an effective tool for DNA-
`based immunization. Further improvements could include
`chemical modifi cation of the surface of the gold particles to
`allow higher capacity and better consistency for DNA coat-
`ing, and fi ne-tuning of the expelling force for precise control
`of DNA deposition into cells in various tissues. 27
`
` GENE TRANSFER BY PHYSICAL METHODS
` Physical approaches have been explored for gene transfer
`into cells in vitro and in vivo. Physical approaches induce
`transient injuries or defects on cell membranes, so that DNA
`can enter the cells by diffusion. Gene delivery employing
`mechanical (particle bombardment or gene gun), electric
`(electroporation), ultrasonic, hydrodynamic (hydrodynamic
`gene transfer), or laser-based energy has been explored in
`recent years.
`
` Gene Transfer by Electroporation
` Electroporation is a versatile method that has been exten-
`sively tested in many types of tissues in vivo, 2 , 3 among
`which skin and muscles are the most extensively investi-
`gated, although the system should work in any tissues into
`which a pair of electrodes can be inserted. For example,
`Hasson et al demonstrated that electroporation substantially
`E93
`
` Ultrasound-Facilitated Gene Transfer
` The discovery that ultrasound can facilitate gene transfer at
`cellular 37 and tissue levels 38 expands the methodology of
`gene transfer by physical methods. A 10- to 20-fold enhance-
`ment of reporter gene expression over that of naked DNA
`has been achieved. The transfection effi ciency of this sys-
`tem is determined by several factors, including the frequency,
`
`Moderna Ex 1008-p. 2
`Moderna v Protiva
`
`

`

`The AAPS Journal 2007; 9 (1) Article 9 (http://www.aapsj.org).
`the output strength of the ultrasound applied, the duration of
`structure — including small dye molecules, proteins, oligo-
`ultrasound treatment, 39 and the amount of plasmid DNA
`nucleotides, small interfering RNA, and linear or circular
`used. The effi ciency can be enhanced by the use of contrast
`DNA fragments as large as 175 kb — have been delivered by
`agents or conditions that make membranes more fl uidic. 40 , 41
`this method. 30 , 46 The nonspecifi c nature of hydrodynamic
`The contrast agents are air-fi lled microbubbles that rapidly
`delivery suggests that this method can be applied to intra-
`expand and shrink under ultrasound irritation, generating
`cellular delivery of any water-soluble compounds, small
`local shock waves that transiently permeate the nearby
`colloidal particles (molecular assembly), or viral particles.
`cell membranes. Unlike electroporation, which moves DNA
`Hydro dynamic delivery allows direct transfer of substances
`along the electric fi eld, ultrasound creates membrane pores
`into cytoplasm without endocytosis.
`and facilitates intracellular gene transfer through passive
` Such a simple, reproducible, and highly effi cient method for
`diffusion of DNA across the membrane pores. 37 , 42 Conse-
`gene delivery has been used to express proteins of therapeu-
`quently, the size and local concentration of plasmid DNA
`tic value such as hemophilia factors, 47 , 48 alpha-1 antitrypsin,
`play an important role in determining the transfection effi -
` 49-51 cytokines, 52 hepatic growth factors, 53 and erythropoie-
`ciency. Efforts to reduce DNA size for gene transfer by
`tin 54 in mouse and rat models. Depending on the plasmid
`methods of standard molecular biology or through proper
`construct and the regulatory elements driving expression of
`formulation could result in further improvement. Interestingly,
`the transgene, the level of gene expression in some cases
`signifi cant enhancement has been reported in cell culture
`has reached or exceeded the physiological level. 49-51 The
`and in vivo when complexes of DNA and cationic lipids
`fact that a bacterial artifi cial chromosome that contains an
`have been used. 42 , 43 Since ultrasound can penetrate soft
`entire chromosomal transcription unit and replication origin
` tissue and be applied to a specifi c area, it could become an
`(>150 kb) can be delivered successfully to the liver using
`ideal method for noninvasive gene transfer into cells of the
`this method 30 opens up many possibilities for gene therapy
`internal organs. Evidence supporting this possibility has
`applications in liver-associated genetic diseases.
`been presented: in one study, plasmid DNA was coadminis-
` The real challenge for gene transfer by the hydrodynamic
`tered with a contrast agent to blood circulation, and this was
`method is how to translate this simple and effective proce-
`followed by ultrasound treatment of a selected tissue. 44 So
`dure to one that is applicable to humans. Rat liver can be
`far, the major problem for ultrasound-facilitated gene deliv-
`transfected similarly through tail vein injection using an
`ery is low gene delivery effi ciency.
`injection volume equivalent to 8% to 9% of body weight
`(T. Suda and D. Liu, unpublished data, 2006). If the same
`ratio is extrapolated to humans, one would have to inject up
`to 7.5 L of saline at a high rate, which is obviously many
`times over the maximal volume that a person can tolerate.
`However, successful liver transfection has been achieved
`using balloon catheter – based and occlusion-assisted infu-
`sion to specifi c lobes in rabbit 55 and swine models, 56 , 57 indi-
`cating that with modifi cation, hydrodynamic gene delivery
`can become a clinically relevant procedure.
`
` Hydrodynamic Gene Delivery
` Hydrodynamic gene delivery is a simple method that intro-
`duces naked plasmid DNA into cells in highly perfused
`internal organs (eg, the liver) with an impressive effi -
`ciency. 7 , 8 The gene delivery effi ciency is determined by the
`anatomic structure of the organ, the injection volume, and
`the speed of injection. In a mouse model, the optimal condi-
`tion involves 1.6 to 1.8 mL of DNA solution in saline for a
`20 g mouse (8%-9% of the body weight) and an injection
`time of ~5 seconds via the tail vein. Mechanistically, the
`rapid tail vein injection of a large volume of DNA solution
`causes a transient overfl ow of injected solution at the infe-
`rior vena cava that exceeds the cardiac output. As a result,
`the injection induces a fl ow of DNA solution in retrograde
`into the liver, a rapid rise of intrahepatic pressure, liver
`expansion, and reversible disruption of the liver fenestrae. 45
`Electron microscopy shows the existence of transient mem-
`brane defects in hepatocytes shortly after the hydrodynamic
`treatment, which could be the mechanism for plasmid DNA
`to enter the hepatocytes. 45 The gene transfer effi ciency of
`this simple procedure is the highest so far achieved in vivo
`using nonviral approaches. Approximately 30% to 40% of
`the hepatocytes are transfected by a single hydrodynamic
`injection of less than 50 μg of plasmid DNA. 7 Various
` substances of different molecular weight and chemical
`E94
`
` GENE DELIVERY BY CHEMICAL METHODS
` By far the most frequently studied strategy for nonviral gene
`delivery is the formulation of DNA into condensed particles
`by using cationic lipids or cationic polymers. The DNA-
`containing particles are subsequently taken up by cells via
`endocytosis, macropinocytosis, or phagocytosis in the form
`of intracellular vesicles, from which a small fraction of the
`DNA is released into the cytoplasm and migrates into the
`nucleus, where transgene expression takes place.
`
` Cationic Lipid-Mediated Gene Delivery
` Since 1987, when Felgner et al fi rst reported that a double-
`chain monovalent quaternary ammonium lipid, N-[1-(2,3-
`dioleyloxy)propyl]-N,N,N-trimethylammonium chloride,
`
`Moderna Ex 1008-p. 3
`Moderna v Protiva
`
`

`

`The AAPS Journal 2007; 9 (1) Article 9 (http://www.aapsj.org).
`have protonable amine groups that apparently intercept the
`effectively binds and delivers DNA to cultured cells, 58
`endosome maturation by absorbing protons to slow down
` hundreds of new cationic lipids have been developed (for
`review, see Liu et al 10 ). These lipids differ by the number of
`the acidifi cation process inside the endosomes, preventing
`charges in their hydrophilic head group and by the detailed
`the endosome-lysosome transition. It has also been sug-
`structure of their hydrophobic moiety. Although some cat-
`gested that a local destabilization effect of some micelle-
`ionic lipids alone exhibit good transfection activity, they are
`forming cationic lipids on the endosomal membrane ’ s
`often formulated with a noncharged phospholipid or choles-
`integrity is part of the underlying mechanism of lipid-based
`gene delivery. 63
`terol as a helper lipid to form liposomes. Upon mixing with
`cationic liposomes, plasmid DNA is condensed into small
` Lipoplexes form spontaneously when cationic liposomes
`quasi-stable particles called lipoplexes. DNA in lipoplexes
`are mixed with DNA. The process involves an initial
`is well protected from nuclease degradation. Lipoplexes are
`rapid association of polycationic liposomes and poly-
`able to trigger cellular uptake and facilitate the release of
`anionic DNA through electrostatic interaction, followed
`DNA from the intracellular vesicles before reaching destruc-
`by a slower lipid rearrangement process. 65 The structure
`tive lysosomal compartments.
`of lipoplexes is infl uenced by multiple factors, including
`the charge ratio, the concentration of individual lipids and
`DNA, the structure of the cationic lipid and the helper lipid,
`the physical aggregation state of the lipids (multilamellar
`or unilamellar liposomes, or micelles), the salt concentra-
`tion, and the method of preparation. Lipoplexes come in
`various forms, including fully condensed lipid/DNA com-
`plexes, partially condensed lipid/DNA complexes, DNA
`sandwiched between cationic lipid bilayers, lipid-coated
`DNA arranged in a hexagonal lattice, or partially condensed
`DNA surrounded by a lipid bilayer. 66 , 67 With the same lipid
`composition and charge ratio, lipoplexes that are prepared
`from multilamellar liposomes with a size of ~500 nm and
`those that are intrinsically less stable exhibit better activity
`in transfection. 68 , 69
` The simplest way to prepare lipoplexes is to mix diluted
`solutions of plasmid DNA and preformed liposomes. The
`resulting lipoplexes are generally heterogeneous in size
`and morphology. The heterogeneity is primarily due to the
`relatively large sizes of DNA and liposomes, and the multi-
`variant nature of the interaction between the DNA and
` liposomes. Alternative methods involving forms of lipid
`assembly other than liposomes have been designed to over-
`come these problems. For example, direct addition of DNA
`solution to a dried fi lm of cationic lipid and DOPE promotes
`entrapment of DNA within multilamellar liposomes, rather
`than sandwiching of DNA between liposomes. 70 A method
`of lipoplex preparation by slow dialysis has also been devel-
`oped. This procedure involves DNA condensation in mixed
`micelles consisting of cationic lipid and non-ionic detergent,
`and removal of the detergent by dialysis. 70 At a concentra-
`tion below the critical micelle concentration of single-chain
`cationic lipids, DNA collapses into unimolecular lipid-DNA
`nanoparticles that are much smaller (20-30 nm). Small parti-
`cles are preferred for in vivo gene delivery because of their
`slower clearance rate in the blood and, therefore, their high
`probability of reaching the target cells. Conjugation to these
`small-sized complexes with polyethylene glycol (PEG) and
`targeting ligands on their surface makes it possible to con-
`struct target-specifi c gene carriers. 71
`
` Most of our understanding of lipid-mediated gene deliv-
`ery derives from characterization work on lipoplexes pre-
`pared in low-salt solution and transfection tests on cells in
`the absence of interfering substances such as serum. Under
`these conditions, the transfection effi ciency of lipoplexes is
`affected by (1) the chemical structure of the cationic lipid,
`(2) the charge ratio between the cationic lipid and the DNA,
`(3) the structure and proportion of the helper lipid in the
`complexes, (4) the size and structure of the liposomes, (5)
`the total amount of the lipoplexes applied, and (6) the cell
`type. The fi rst 4 factors determine the structure, charge prop-
`erty, and transfection activity of the lipoplexes. The remain-
`ing 2 defi ne the overall toxicity to the treated cells, and the
`susceptibility of the cells to a particular lipid-based trans-
`fection reagent. The chemical structure of the cationic lipid
`has a major impact on the transfection effi ciency. In general,
`multivalent lipids with long and unsaturated hydrocarbon
`chains tend to be more effi cient than monovalent cationic
`lipids with the same hydrophobic chains. Transfection typi-
`cally requires that the cationic lipid be in slight excess over
`DNA such that the lipoplexes have net positive charges on
`the surface. Spontaneous mixing between cationic lipids and
`cellular lipids in the membrane of the endocytic vesicles is
`crucial to the endosome-releasing process. 59 Spontaneous
`lipid mixing in endosomes becomes more profound when
`a non-bilayer-forming lipid such as dioleoylphosphatidyl-
`ethanolamine (DOPE) is used as the helper lipid, rather than
`a bilayer-forming lipid, dioleoylphosphatidylcholine. 60 , 61
`Inclusion of DOPE is believed to increase membrane fl u-
`idity and facilitate lipid exchange and membrane fusion
`between lipoplexes and the endosomal membrane. A high
`local concentration of DOPE, which has a strong tendency
`to form an inverse hexagonal phase, may lead to a nonbi-
`layer lipid structure and cause membrane perturbation and
`endosome destruction. 62 However, some multivalent lipids
`have intrinsic transfection activity, and a helper lipid does
`not have a major impact on overall transfection activity,
`indicating that multivalent cationic lipids work on a differ-
`ent mechanism. 63 , 64 Often, these cationic lipopolyamines
`E95
`
`Moderna Ex 1008-p. 4
`Moderna v Protiva
`
`

`

`The AAPS Journal 2007; 9 (1) Article 9 (http://www.aapsj.org).
`membrane-active peptide seem to be logical and incremen-
` Many cationic lipids show excellent transfection activity in
`tal steps to solving some of these problems.
`cell culture, but most do not perform well in the presence of
`serum, and only a few are active in vivo. 10 A dramatic
` In vivo gene transfer by systemic administration of lipo-
`change in size, surface charge, and lipid composition occurs
`plexes mainly transfects endothelial cells in the pulmonary
`when lipoplexes are exposed to the overwhelming amount
`vasculature. 74 A large excess of cationic lipids was needed
`of negatively charged and often amphipathic proteins and
`to mediate optimal gene transfer. 74 Although using DOPE
`polysaccharides that are present in blood, mucus, epithelial
`as the helper lipid makes the formulation more effi cient for
`lining fl uid, or tissue matrix. Once administered in vivo,
`airway gene delivery, it has an adverse effect on intravenous
`lipoplexes tend to interact with negatively charged blood
`transfection. Cholesterol was found to be a better helper
`components and form large aggregates that could be
`lipid for systemic transfection. 74 In systemic gene delivery,
`absorbed onto the surface of circulating red blood cells,
`cholesterol seems to stabilize the lipoplex structure in blood,
`trapped in a thick mucus layer, or embolized in microvascu-
`while formulations containing DOPE tend to fall apart more
`latures, preventing them from reaching the intended target
`easily in the presence of blood components. 72 , 74 Moreover,
`cells in the distal location. Some even undergo dissolution
`it has been shown that one can effi ciently transfect pulmo-
`after they are introduced to the blood circulation. 72 , 73
`nary endothelial cells by injecting free cationic liposomes
`and following shortly thereafter with a second injection of
` Despite these undesirable characteristics, lipoplexes have
`naked DNA solution, 84 which suggests that various forms of
`been used for in vivo gene delivery to the lungs by intra-
`lipoplex structures that seem to be important for transfec-
`venous 74-76 and airway 77-79 administration. Transgene
`tion of cells in tissue culture are not critical for transfection
`expression was clearly detectable but in most cases was
`by intravenous injection. The expression of a reporter gene
`insuffi cient for a meaningful therapeutic outcome. For air-
`in transfected endothelial cells in the lung follows a quick
`way gene delivery to the lungs, animal studies using lipo-
`onset, which reaches its peak level ~8 to 16 hours posttrans-
`plexes prepared from 3 b -[N-(N ′ ,N ′ -dimethylaminoethane)
`fection, then declines rapidly. The rapid decline is not solely
`carbamoyl]cholesterol (DC-Chol) and DOPE have shown
`due to DNA degradation, as a second transfection 1 week
`that this procedure was mild to the host and partially effec-
`later did not result in signifi cant expression, 74 suggesting
`tive in correcting genetic defects in a cystic fi brosis trans-
`that it is likely that initially transfected cells become resis-
`membrane regulator mutant model. 80 By screening a large
`tant to the same type of transfection. The lack of cellular
`cationic lipid library in earlier studies, the Genzyme group
`response by the transfected cells to subsequent transfection
`revealed some structure-activity relationships important
`suggests a negative transcription regulatory mechanism.
`to the transfection activity of cationic lipids in a mouse
`model. 81 Several cholesterol derivatives with polyamine
` Toxicity related to gene transfer by lipoplexes has been
`groups linked to cholesterol through a carbamoyl bond
`observed. Acute infl ammation reactions have been reported
`exhibited signifi cantly higher activity in the lung compared
`in animals treated with airway instillation or intravenous
`with that of naked DNA or DC-Chol/DOPE lipoplexes. 81
`injection of lipoplexes. 74 Detailed toxicological studies on
`Most transfected cells were found in the lower airways in
`one of the Genzyme Lipid formulations, GL-67/DOPE,
`the alveoli region, not the intended bronchial epithelial
`revealed that the cationic lipid contributes signifi cantly to
`cells.
`the toxicity observed. 85 , 86 Similar toxic effects are also
`noticeable in systemic gene delivery via the tail vein with
` Several inhibitory factors for lipoplex-based gene delivery
`other types of cationic lipids. Symptoms include acute pul-
`have been identifi ed for airway gene transfer. 82 , 83 A critical
`monary hypotension, induction of infl ammatory cytokines,
`factor is that upper-airway epithelial cells are covered by a
`tissue infi ltration of neutrophils in lungs, decrease in white
`negatively charged and viscous mucus layer that often traps
`cell counts, and in some cases tissue injury in liver and
`and neutralizes the surface charges of the lipoplexes. In
`spleen. 87 In humans, various degrees of adverse infl ammatory
`patients with cystic fi brosis, the epithelial cells are further
`reactions, including fl ulike symptoms with fever and airway
`covered with a thick layer of sputum that contains genomic
`infl ammation, were noted among subjects who received
`DNA released from dead cells and bacteria. In lower air-
`aerosolized GL67 liposomes alone or lipoplexes. 85 These
`ways, the surfactant layer enriched with several phospho-
`early clinical data suggest that these lipoplex formulations
`lipids and surfactant proteins is also believed to inhibit
`are inadequate for use in humans.
`the transfection activity of lipoplexes. In addition, well-
` differentiated upper airway epithelial cells are known to be
` Part of the infl ammation response seen in treated lungs is
`less active in taking up lipoplexes than are those in the lower
`related to the unmethylated CpG (umCpG) sequences found
`airways. Proper shielding of surface charges of lipoplexes
`in the plasmid of bacterial origin. A potent immune stimulant,
`to reduce nonspecifi c protein/mucin association, inclusion
`umCpG triggers release of proinfl ammatory cytokines. 88 , 89
`of a target ligand to enhance specifi c binding, and substitu-
`Cationic lipids in lipoplexes are capable of enhancing the
`tion of a portion of the cationic lipids with less toxic and
`umCpG effect. 90 Another factor related to the severity of
`E96
`
`Moderna Ex 1008-p. 5
`Moderna v Protiva
`
`

`

`The AAPS Journal 2007; 9 (1) Article 9 (http://www.aapsj.org).
`transfection-related side effects is complement activation 91
` PEI is perhaps the most active and most studied polymer for
`and adsorption of serum proteins onto their surface, which
`gene delivery. PEI is made either from an acid-catalyzed
`in turn act as opsonins to trigger the uptake of opsonized
`aziridine ring opening reaction that leads to branched poly-
`particles by macrophages and other immune cells. Various
`mers, or by hydrolysis of poly(2-ethyl-2-oxazolium) that
`strategies have been considered to deal with the toxic
`results in linear polymers. Jean Paul Behr ’ s group fi rst intro-
`responses. For example, covering the lipoplex surface with
`duced PEI as an effi cient and economic synthetic polymeric
`gene transfer agent. 101 Chemically, PEI is one of the most
`inert polymers could, in principle, reduce protein adsorption
`densely charged polymers: one third of the atoms are nitro-
`and their affi nity to immune cells and thereby minimize the
`gen, and one sixth of the nitrogen atoms carry a positive
`toxic responses. Toward this end, PEG-lipid conjugates
`charge at physiological pH. Branched PEI has a ratio of
`have been incorporated into the lipoplexes to minimize the
` primary:secondary:tertiary amine groups close to 1:1:1,
`nonspecifi c interaction of lipoplexes with blood compo-
`according to a recently revised estimate. 116 For PEI- mediated
`nents. 92-94 It is believed that PEG, being hydrophilic and
`transfection, DNA-to-PEI ratios, the molecular weight and
`unable to interact with either DNA or cationic lipids, pro-
`confi guration of PEI, the concentration of DNA and poly-
`vides longer circulation times of liposomes in blood circula-
`mer, and the ionic strength of the solvent for preparation
`tion by minimizing the binding of blood components and
`are all important factors that determine the physical prop-
`lipoplexes. Unfortunately, inclusion of such bulky PEG lip-
`erties of the DNA/PEI complexes (polyplexes) and their
`ids into lipoplexes causes dose-dependent inhibition in
`transfection activity. Most of the amines in linear PEI are
`transfection activity. For this reason a different length of
`secondary amines except the terminal groups. Both linear
`hydrocarbons in PEG-lipid derivatives was used to adjust
`PEI (LPEI) and branched PEI (BPEI) have excellent trans-
`the time of PEG-lipid association with lipoplexes. The
`fection activities in vitro and exhibit moderate transfection
`objective of this strategy is to use the PEG as a cover for
`activity in vivo. LPEI is reportedly less toxic to