6
`
`“Diffusible-PEG-Lipid Stabilized
`Plasmid Lipid Particles”
`
`Ian MacLachlan* and Pieter Cullis* ,{
`*Protiva Biotherapeutics Incorporated, Burnaby, BC, Canada V5G 4Y1
`{Inex Pharmaceuticals Inc., Burnaby, BC, Canada V5J 5J8
`
`I. Introduction
`II. Properties of a Plasmid Delivery System for the
`Treatment of Systemic Disease
`A. Definition of an appropriate vector
`B. Overcoming the barriers to transfection
`C. Proposed mechanism of stabilized plasmid lipid particle
`mediated transfection
`III. Methods of Encapsulating Plasmid DNA
`IV. Pharmacology of Encapsulated Plasmid DNA
`A. Biodistribution following systemic administration of SPLP
`B. Biodistribution of protein expression following systemic
`administration of SPLP
`V. Conclusion
`References
`
`ABSTRACT
`
`Many viral and non-viral gene transfer systems suffer from common phar-
`macological issues that limit their utility in a systemic context. By application
`of the liposomal drug delivery paradigm, many of the limitations of the first
`generation non-viral delivery systems can be overcome. Encapsulation in small,
`long-circulating particles called stabilized plasmid lipid particles (SPLP) results
`in enhanced accumulation at disease sites and selective protein expression. This
`work compares the detergent dialysis method of SPLP manufacture with an
`alternative method, spontaneous vesicle formation by ethanol dilution. The
`
`Advances in Genetics, Vol. 53
`Copyright 2005, Elsevier Inc. All rights reserved.
`
`0065-2660/05 $35.00
`DOI: 10.1016/S0065-2660(05)53006-2
`
`ARBUTUS - EXHIBIT 2035
`Moderna Therapeutics, Inc. v. Arbutus Biopharma Corporation
`IPR2019-00554
`
`

`

`158
`
`MacLachlan and Cullis
`
`pharmacology of SPLP, as determined by monitoring lipid label and quantitative
`real time PCR, is also presented. ß 2005, Elsevier Inc.
`
`I. INTRODUCTION
`
`Current efforts in gene transfer research focus on the development of genetic
`drugs capable of treating acquired diseases such as cancer, inflammation, viral
`infection or cardiovascular disease. The disseminated nature of these diseases
`requires the development of vector systems capable of accessing distal sites
`following systemic or intravenous administration. Unfortunately, most vectors
`have limited utility for systemic applications. Viral vectors, for example, are
`rapidly cleared from the circulation, limiting transfection to “first-pass” organs
`such as the lungs, liver and spleen. In addition, many viruses induce immune
`responses that compromise potency upon subsequent administration. In the case
`of most non-viral vectors such as plasmid DNA-cationic lipid complexes (lipo-
`plexes), the large size and positively charged nature of these systems also results
`in rapid clearance upon systemic administration with the highest expression
`levels observed in first-pass organs, particularly the lungs (Huang and Li, 1997;
`Hofland et al., 1997; Templeton et al., 1997; Thierry et al., 1995). In addition,
`lipoplexes often give rise to significant toxicities both in vitro and in vivo
`(Harrison et al., 1995; Li and Huang, 1997; Tousignant et al., 2000, 2003). In
`spite of these limitations, non-viral gene transfer systems offer specific clinical
`and commercial advantages as therapeutics. Because non-viral systems use
`synthetic or highly purified components, they are chemically defined and free
`of adventitious agents. Non-viral systems can be manufactured under controlled
`conditions, relatively unconstrained by the biological considerations that define
`the scale-up of viral production in mammalian cell culture. These advantages
`have encouraged a number of investigators to focus on the development of non-
`viral gene transfer systems that have utility in a systemic context (Dzau et al.,
`1996; Li and Huang, 1997; Templeton et al., 1997; Wheeler et al., 1999; Zhu
`et al., 1993). Here we will describe one system that specifically attempts to
`address the inability of current vector systems to overcome the first barrier to
`systemic gene delivery, delivery to the disease site and the target cell.
`
`II. PROPERTIES OF A PLASMID DELIVERY SYSTEM FOR THE
`TREATMENT OF SYSTEMIC DISEASE
`
`A. Definition of an appropriate vector
`
`We propose the following definition of an ideal carrier for systemic gene transfer:
`The ideal vector will (i) be safe and well tolerated upon systemic administration;
`
`

`

`6. Diffusible-PEG-Lipid Stabilized Plasmid Lipid Particles
`
`159
`
`(ii) have the appropriate pharmacokinetic attributes to ensure delivery to
`disseminated disease sites; (iii) deliver intact DNA to target tissue and mediate
`transfection of that tissue; (iv) be non-immunogenic; and (v) be stable
`upon manufacture to facilitate production at commercial scale with uniform,
`reproducible performance specifications.
`Gene-based drugs must maximize the benefit to patient health while
`minimizing the risks associated with treatment. Accordingly, gene transfer
`systems must be safe and well tolerated. Attempts to bypass the inherent
`pharmacology of a given vector by invoking elaborate or invasive treatment
`methodologies are likely to result in an increased, potentially unacceptable, risk
`to the patient. Methods such as ‘hydrodynamic injection’ or direct portal vein
`infusion may continue to generate exciting preclinical results, but translation
`of these methods to a clinical setting will be limited. Gene-based drugs will
`be adopted more readily if they can be delivered in a manner analogous to
`conventional medicines, for example by intravenous injection or in oral form.
`The toxicity associated with systemic administration of poorly tolerated
`compounds is exacerbated by accumulation in non-target tissue and can be
`reduced by optimizing delivery to the target site. In the case of gene-based drugs,
`‘delivery’ is determined by physical and biochemical properties including stabil-
`ity, size, charge, hydrophobicity, interaction with serum proteins and non-target
`cell surfaces, as well as the mechanism of action of the nucleic acid payload. In
`the context of a disease site, effective delivery requires that a vector overcome
`obstacles associated with heterogeneous cell populations that are often prolifer-
`ating rapidly, at different stages of the cell cycle and not conforming to the
`patterns of organization established during the development of normal tissue. As
`demonstrated in this work, these challenges, and other potential barriers to
`transfection, can represent opportunities for conferring a degree of selectivity
`greater than that associated with the use of conventional therapeutics.
`
`B. Overcoming the barriers to transfection
`
`The barriers to transfection include the pharmacological barriers inhibiting
`delivery to the target cell, and the intracellular barriers that inhibit nuclear
`delivery and expression of the plasmid DNA construct. An effective delivery
`system must be able to confer stability to the nucleic acid payload in the blood
`despite the presence of serum nucleases and membrane lipases. Systemic delivery
`requires the use of a ‘stealthy’ delivery system, since indiscriminate interaction
`with blood components, lipoproteins or serum opsonins, can cause aggregation
`before the carrier reaches the disease site. This is especially important in the case
`of systemic delivery systems containing large polyanionic molecules such as
`plasmid DNA, which have a greater potential for inducing toxicity through
`interaction with complement and coagulation pathways (Chonn et al., 1991).
`
`

`

`160
`
`MacLachlan and Cullis
`
`Other barriers to gene delivery may include the microcapillary beds of the “first
`pass” organs, the lung and the liver, and the phagocytic cells of the reticuloen-
`dothelial system. Accessing target cell populations requires extravasation from
`the blood compartment to the disease site. Carriers of appropriate size can pass
`through the fenestrated epithelium of tumor neovasculature and accumulate at
`the tumor site via the “enhanced permeation and retention” (EPR) effect
`(Mayer et al., 1990), also referred to as “passive” targeting or “disease site”
`targeting. In order to take advantage of the EPR effect, which can result in
`accumulation of up to 10% of the injected dose per gram of tumor tissue, the
`gene carriers must be small (diameter on the order of 100 nm) and long-
`circulating (circulation lifetimes of 5 h or more following intravenous injection
`in mice). Clearly, nucleic acids require pharmaceutical enablement in the form
`of appropriate carriers that confer: protection from degradation, an extended
`circulation lifetime, appropriate biodistribution and delivery facilitation of the
`nucleic acid payload to the disease site.
`While delivery of intact plasmid DNA to a target cell is a prerequisite, it
`in no way guarantees transfection. Once at the cell surface, vectors are con-
`fronted with a number of physical and biochemical barriers, each of which must
`be overcome in order to effect transfection and transgene expression. The first
`physical barrier to transfection is the plasma membrane, protected by the
`carbohydrate coating, or glycocalyx, formed by the post-translational glycosyla-
`tion of transmembrane proteins. Although early models of lipid-mediated trans-
`fection invoked a putative fusion event between the plasma membrane and the
`membrane of the lipid vesicle, it is now generally agreed that the majority of
`intracellular delivery occurs through endocytosis.
`Endocytosis is a complex process by which cells take up extracellular
`material. This occurs through a number of discrete pathways, reviewed else-
`where in this volume. While there is some evidence that non-viral vectors may
`be taken up by caveolae, syndecan-mediated endocytosis or other clathrin-
`independent pathways, the classical endocytic pathway involves the activity of
`cell surface clathrin-coated pits, invaginations in the plasma membrane that are
`subsequently pinched off into the cytoplasm (Goldstein et al., 1985). When this
`occurs, internalized material remains trapped on the exoplasmic side of the
`internalized vesicle, without direct access to the cytoplasm or the nucleus.
`Endocytic vesicles undergo a series of biochemical changes that represent escape
`opportunities for a non-viral vector. The first such change occurs within 5 min of
`uptake as internalized vesicles form the early endosome containing the “Com-
`partment of Uncoupling of Receptor and Ligand” (CURL) (Geuze et al., 1983).
`Early endosomes are transiently fusogenic (Dunn and Maxfield, 1992) with a pH
`close to that of the exoplasm, while late endosomes have a significantly lower
`luminal pH (Murphy et al., 1993). As endosomes mature to form lysosomes they
`experience a further decrease in internal pH and an increase in fusogenicity.
`
`

`

`6. Diffusible-PEG-Lipid Stabilized Plasmid Lipid Particles
`
`161
`
`Although the process of clatharin-dependent endocytosis has been well char-
`acterized, the processing and release of internalized non-viral vectors or their
`DNA payload is not well understood. Even less clear is the relative import of
`clathrin-independent uptake through mechanisms that share some, but not all of
`the features of the classical pathway. Improvements in our understanding
`of these alternative pathways, and their role in non-viral gene transfer, will
`be important for the rational design of more effective intracellular delivery
`strategies for non-viral vectors.
`Following uptake, plasmid DNA spends some indeterminate residency
`time in the cytoplasm prior to gaining entry to the nucleus. Unlike viral systems
`that have evolved specific mechanisms to traverse this barrier, untargeted non-
`viral vectors rely on diffusion to facilitate interaction with the nuclear envelope
`(Kopatz et al., 2004). However the cytoplasm, rather than an empty space, is a
`highly organized compartment containing networks of cytoskeletal elements and
`membrane-bound organelles that have the potential to interact with and accu-
`mulate vector systems that arrive at the cytosol intact. When plasmid DNA is
`delivered by direct microinjection into the cytosol of mammalian cells it is
`rapidly degraded by divalent-cation-dependent cytosolic nucleases (Howell
`et al., 2003; Lechardeur et al., 1999). This has implications for vector design.
`Vector systems that either protect the DNA payload from degradation following
`endosome release or effectively minimize the cytoplasmic residency time are to
`be expected to yield improved transfection efficiencies.
`The final physical barrier to transfection is delivery to the nucleus. The
`nucleus has evolved as a means of organizing, isolating and protecting the
`genome of eukaryotic cells from adventitious agents such as viruses or transpo-
`sons. The nuclear uptake of DNA is limited by the presence of an intact nuclear
`envelope and as such non-viral transfection is considerably more efficient in
`highly mitotic cells (Mortimer et al., 1999; Wilke et al., 1996). Strategies to
`overcome this barrier to transfection take one of two forms: either targeting
`transfection reagents to cell populations with a high degree of mitotic activity,
`such as tumor tissue; or enhancing the low level of transfection that occurs in
`quiescent cells by using either nuclear targeting technologies or condensing
`agents that compact plasmid DNA to a size more amenable to uptake through
`the nucleopore complex (Blessing et al., 1998; Sebestyen et al., 1998).
`
`C. Proposed mechanism of stabilized plasmid lipid particle
`mediated transfection
`
`1. Delivery to the target cell
`
`systemic applications are
`imposed upon vectors used for
`The demands
`conflicting. First, the carrier must be stable and long-circulating, circulating
`long enough to facilitate accumulation at disease sites via the EPR effect.
`
`

`

`162
`
`MacLachlan and Cullis
`
`Figure 6.1. Structure of stabilized plasmid lipid particles.
`
`Second, the carrier must interact with—and be taken up by—target cells
`following arrival at the target site in order to facilitate gene expression. The
`“stabilized plasmid-lipid particle” (SPLP) attempts to satisfy both of these
`requirements.
`SPLPs consist of a single plasmid encapsulated in a lipid bilayer con-
`taining a diffusible polyethylene glycol (PEG)-lipid conjugate (Fig. 6.1). The
`PEG-lipid conjugates in the SPLP play an essential role during the formulation
`process, stabilizing the nascent particle and preventing aggregation of the
`particles in the vial. In the blood, the PEG-lipid shields the positive surface
`charge, preventing rapid clearance following intravenous injection. Following
`administration, at 37C and in the presence of sufficient lipid sink, the PEG
`conjugate dissociates from the SPLP, revealing the positive charge and an
`increasingly fusogenic lipid bilayer, transforming the particle into a transfec-
`tion-competent entity. The residency of the PEG conjugate in the SPLP bilayer
`is determined by the length of the lipid anchor. PEGs with shorter lipid anchors,
`such as ceramide-C8 or dimyristoyl-glycerol, dissociate more quickly from the
`bilayer, quickly ‘activating’ the SPLP into which they are incorporated. As a
`result, particles incorporating PEG-lipids with shorter lipid anchors show higher
`transfection potency in vitro than those containing longer lipid anchors (e.g.,
`ceramide-C20 or distearoyl-glycerol) (Mok et al., 1999). When injected systemi-
`cally, PEG conjugates with a larger, more securely fastened anchor and will
`confer greater stability and extended circulation lifetimes, leading to greater
`levels of accumulation at disseminated disease sites (Monck et al., 2000; Tam
`et al., 2000).
`
`

`

`6. Diffusible-PEG-Lipid Stabilized Plasmid Lipid Particles
`
`163
`
`2. The role of cationic lipids in promoting intracellular delivery
`
`While the factors facilitating intracellular delivery of non-viral vectors are poorly
`understood, it is believed that both polycation and cationic lipid-containing
`systems function, at least in part, by coating plasmid DNA with a positive
`charge that enables binding of the DNA complex to anionic cell surface
`molecules, such as cell surface proteoglycans that appear to facilitate transfec-
`tion both in vitro (Mislick and Baldeschwieler, 1996) and in vivo (Mounkes et al.,
`1998). Inhibition of the interaction between the positively charged lipoplex and
`negatively charged cell surface molecules by pretreatment with polyanionic
`compounds greatly inhibits lipoplex-mediated transfection while having no
`effect on electroporation or adenoviral transfection. Intravenous administration
`of heparinase I, an enzyme specific for the cleavage of heparan sulfate proteo-
`glycans, also inhibits cationic lipoplex-mediated transfection. Given that the
`basis of the interaction between proteoglycans and cationic vectors appears to be
`electrostatic, differences in charge and charge density between vector systems
`could yield differences in transfection efficiency. This has certain implications
`for the design of vector systems for systemic gene therapy. SPLP are transiently
`charge shielded due to the incorporation of diffusible PEG-lipids. As the PEG-
`lipid leaves the particle in the blood compartment, the positive charge conferred
`by the cationic lipid component is revealed. Although systems with a lower
`surface charge might be expected to benefit from increased circulation time,
`incorporation of additional cationic lipid in the SPLP lipid bilayer yields an
`appreciable gain in potency in vitro (Zhang et al., 1999).
`It is also believed that cationic lipids play a direct role to facilitate
`intracellular delivery following internalization in the endosome. The proposed
`mechanism involves the ability of cationic lipids to promote formation of the
`HII phase in combination with anionic lipids (Hafez et al., 2001), thereby
`destabilizing the bilayer structure of the endosomal membrane, encouraging
`fusion with the SPLP bilayer and facilitating the cytoplasmic translocation of
`the associated plasmid DNA. Clearly, the cationic lipid content of systemic
`carrier systems must be optimized with a view towards achieving both an
`extended circulation lifetime and effective intracellular delivery. In particular,
`a compromise must be made between incorporating high amounts of cationic
`lipid, which facilitates transfection and the fact that high amounts of catio-
`nic lipid result in shorter circulation lifetimes, reducing the amount of the
`material that arrives at the disease site.
`
`3. The role of helper lipids in promoting intracellular delivery
`
`The majority of cationic lipids require the addition of a fusogenic ‘helper’ lipid
`for efficient in vitro gene transfer (Farhood et al., 1995; Felgner et al., 1994; Gao
`and Huang, 1995; Hui et al., 1996). Inclusion of lipids, such as unsaturated
`
`

`

`164
`
`MacLachlan and Cullis
`
`like dioleoylphosphatidylethanolamine (DOPE),
`phosphatidylethanolamines
`promote destabilization of the lipid bilayer and fusion (Farhood et al., 1995;
`Hui et al., 1981; Litzinger and Huang, 1992). The fusogenicity of DOPE-con-
`taining bilayers is thought to be due to their polymorphic nature. Upon formu-
`lation, most lipids adopt the bilayer-forming Lamellar Phase (L), while DOPE
`has a tendency to form the inverse hexagonal (HII) phase (Cullis PR and B.,
`1978; Koltover et al., 1998). Several researchers have noted that increasing
`the degree of unsaturation of the lipid hydrophobic domain increases the affinity
`for the HII phase (Cullis PR and B., 1979; Dekker et al., 1983; Epand et al.,
`1991; Sankaram et al., 1989; Szule et al., 2002). As a result, the fusogenicity of an
`SPLP bilayer can be increased by increasing the degree of unsaturation in the
`hydrophobic domain of either the helper lipid or cationic lipid components
`(Heyes et al., 2004). Furthermore, certain cationic lipids can function in
`the absence of fusogenic helper lipids, either alone (Felgner et al., 1994; Gao
`and Huang, 1995) or in the presence of the non-fusogenic lipid cholesterol
`(Liu et al., 1995).
`The specific role of fusogenic helper lipids in the transfection process,
`and whether this role is conserved between lipoplex and systems such as SPLP
`which fully encapsulate plasmid DNA, is not clear. Membrane fusion events
`could theoretically occur at a number of different stages in the gene delivery
`process, either at the plasma membrane, endosome or nuclear envelope. In order
`for fusion with the plasma membrane to occur, positively charged lipid particles
`must first bypass the negatively charged glycocalyx. Fusion of lipoplex systems
`with the plasma membrane would be expected to be a particularly inefficient
`method of introducing DNA into the cytosol since lipoplex fusion events may
`resolve with plasmid DNA, formerly attached to the cationic liposome surface,
`deposited on the outside surface of the plasma membrane. Encapsulated systems
`differ from lipoplex in this respect. Fusion with the plasma membrane could
`result in an encapsulated carrier delivering its contents into the cytosol. How-
`ever, the bulk of both lipoplex- and SPLP-mediated transfection is thought to be
`by fusion with the endosomal membrane of particles that are taken up intact by
`endocytosis (Wrobel and Collins, 1995). There is considerable biochemical
`evidence to support an endosomal route for internalized plasmid DNA. One
`example is the transient inhibition of endocytosis and concomitant transfection
`upon treatment of cells with cytochalasin-B, an inhibitor of actin polymeriza-
`tion required in the endocytic process (Hui et al., 1996). Another example
`utilizes fluorescently labeled lipids to track the fate of lipoplex or SPLP upon
`delivery to the cell. Fusion of labeled liposomes with the plasma membrane
`would result in the transfer of lipid label to the membrane. Cells exposed
`to lipoplex or SPLP containing rhodamine-phosphatidyl-ethanolamine accumu-
`late fluorescent label in endocytic granules, well before plasma membranes
`become fluorescent (Hui et al., 1996; Palmer et al., 2003). In the absence of a
`
`

`

`6. Diffusible-PEG-Lipid Stabilized Plasmid Lipid Particles
`
`165
`
`fusion-induced translocation event, fusion of lipoplex systems with endosomal
`compartments results in a gradual destabilization and disruption of the endoso-
`mal membrane. Encapsulated systems have an advantage over lipoplex in that a
`single fusion event within an endosomal compartment would be expected to
`result in efficient delivery of the DNA payload to the cytosol.
`The role of fusogenic lipids in vivo remains unclear. A number of
`investigators have reported that replacement of fusogenic DOPE with the less
`fusogenic lipid cholesterol yields higher levels of gene expression upon systemic
`administration of either lipoplex or encapsulated systems (Sakurai et al., 2001;
`Templeton et al., 1997). However, it is important to distinguish the effect of
`helper lipids on biodistribution from the effect on intracellular delivery. The
`enhanced gene expression observed upon incorporation of cholesterol in lipo-
`plex or SPLP formulations may be a result of either an increase in transfection
`efficiency or improved pharmacokinetics and delivery to the target cell. Fuso-
`genic formulations are more likely to interact with the vascular endothelium,
`blood cells, lipoproteins and the fixed and free macrophages of the mononuclear
`phagocyte system while in the blood compartment, leading to rapid clearance
`and decreasing the proportion of carriers that reach target tissue. Incorporation
`of cholesterol may simply render vectors less promiscuous and thereby improve
`delivery to the target cell. The implication is that there is further rationale for
`transiently shielding the fusogenic potential of systemic carriers through the use
`of diffusible PEG-lipids.
`A variety of approaches can be considered for enhancing the endoso-
`mal release of internalized liposomes. In addition to the use of fusogenic lipids
`that are thought to facilitate endosome release, another strategy involves the
`incorporation of specific lipids that render the liposome pH-sensitive such that it
`becomes more fusogenic in low pH compartments such as the late endosome and
`lysosome (Wang and Huang, 1987, 1989; Lee and Huang, 1996). One example
`of this approach utilizes titratable cationic lipids that become positively charged
`at the reduced pH values that may be encountered in endosomes. Cationic lipids
`such as 1,2-dioleoyl-3-(N,N-dimethylamino)propane (AL1) that exhibit pK of
`approximately 6.6 (Bailey and Cullis, 1994) confer no significant positive charge
`to carriers at neutral pH yet are fully positively charged at the pH values
`commonly encountered in endosomes.
`
`4. Nuclear delivery
`
`Attempts to improve the nuclear uptake of plasmid DNA must take into
`consideration the physical constraints of the nucleopore complex that mediates
`the uptake of plasmid DNA into the intact nucleus. When fully condensed by
`monovalent detergent counterions, a 5.5 kb supercoiled plasmid DNA molecule
`becomes a sphere of about 25 nm in diameter (Blessing et al., 1998) while the
`
`

`

`166
`
`MacLachlan and Cullis
`
`passive diffusion channel of the nuclear pore complex has an internal diameter
`of 9 nm (Ohno et al., 1998). The diameter of the activated nuclear pore
`complex through which active transport occurs, and therefore the size limit for
`signal mediated nuclear import, is 25 nm. Although there does appear to be
`considerable potential for improving the nuclear uptake of supercoiled plasmids
`through attachment of nuclear localization peptides or other nuclear import
`signals, it remains to be seen if this can be accomplished in a manner that is
`compatible with large-scale formulation and systemic gene delivery (Sebestyen
`et al., 1998).
`
`III. METHODS OF ENCAPSULATING PLASMID DNA
`
`In order to capitalize on the pharmacology and disease site targeting demon-
`strated by liposomal drug carriers it is necessary to completely entrap plasmid
`DNA within the contents of a liposome. Unlike small molecule drugs, plasmid
`DNA cannot easily be “loaded” into preformed liposomes using pH gradients or
`other similar strategies. Lipid encapsulation of high molecular weight DNA was
`first demonstrated in the late 1970s, prior to the development of cationic lipid-
`containing lipoplex (Hoffman et al., 1978; Mannino et al., 1979; Mukherjee
`et al., 1978). Plasmid DNA has subsequently been encapsulated by reverse-
`phase evaporation (Cudd and Nicolau, 1985; Fraley et al., 1980; Nakanishi
`et al., 1985; Soriano et al., 1983), ether injection (Fraley et al., 1979; Nicolau
`and Rottem, 1982), lipid hydration-dehydration techniques (Alino et al., 1993;
`Baru et al., 1995; Lurquin, 1979) sonication (Jay and Gilbert, 1987; Ibanez et al.,
`1997; Puyal et al., 1995), spontaneous internalization into pre-formed liposomes
`(Templeton et al., 1997) and others methods (Monnard et al., 1997; Szelei
`and Duda, 1989) (summarized in Table 6.1). Early attempts to encapsulate
`plasmid DNA yielded mostly large multilamellar vesicles with poor transfection
`efficiency (Baru et al., 1995; Nicolau et al., 1983; Scaefer-Ridder et al., 1982),
`while more recently, improvements in formulation technology have resulted in
`the production of cationic lipid-containing particles with a much greater trans-
`fection potential. SPLP initially utilized detergent dialysis, a process in which
`unilamellar vesicles are formed upon removal of detergent from a DNA:lipid
`solution. While early efforts to encapsulate plasmid DNA using detergent
`dialysis yielded low encapsulation efficiencies (Fraley et al., 1979; Nakanishi
`et al., 1985), these results were significantly improved upon through the use of
`PEG lipids to stabilize the vesicles during the formulation process (Wheeler
`et al., 1999). In this way plasmid-containing cationic liposomes are stabilized in
`a manner analogous to PEGylated liposomal drug formulations that exhibit
`extended circulation lifetimes (Allen and Chong, 1987; Klibanov et al., 1990;
`Needham et al., 1992; Papahadjopoulos et al., 1991; Wu et al., 1993). PEG
`
`

`

`6. Diffusible-PEG-Lipid Stabilized Plasmid Lipid Particles
`
`167
`
`conjugates sterically stabilize liposomes by forming a protective hydrophilic
`layer that shields the hydrophobic lipid layer, preventing the association of
`serum proteins and resulting uptake by the reticuloendothelial system (Gabizon
`and Papahadjopoulos, 1988; Senior et al., 1991). Although this approach has
`been investigated with a view towards improving the stability and pharmacoki-
`netics of lipoplex (Hong et al., 1997), lipoplex incorporating PEG-lipids systems
`suffer from the heterogeneity common to most complexes of plasmid DNA and
`cationic lipid.
`The detergent dialysis method of plasmid encapsulation involves the
`simultaneous solubilization of hydrophobic (cationic and helper lipid) and
`hydrophilic (PEG lipid and plasmid DNA) components in a single detergent-
`containing phase (Fenske et al., 2002;Wheeler et al ., 1999). Particle formation
`occurs spontaneously upon removal of the detergent by dialysis. This technique
`can result in the formation of small (approximately 100 nm diameter) ‘stabilized
`plasmid lipid particles’ (SPLP) containing one plasmid per vesicle in combina-
`tion with optimized plasmid trapping efficiencies approaching 70%. The SPLP
`protocol results in stable particles with low levels of cationic lipids, high levels of
`fusogenic lipids and high DNA-to-lipid ratios. SPLP can be concentrated to
`achieve plasmid DNA concentrations of >5 mg/ml. These attributes compare
`favorably with the previously reported plasmid encapsulation processes (Table
`6.1). The SPLP method yields the highest plasmid DNA-to-lipid ratio of any
`method and SPLP are remarkably stable when compared to other encapsulated
`systems. Although the detergent dialysis process results in 30–50% unencapsu-
`lated DNA,
`free plasmid DNA can be removed by simple ion-exchange
`chromatography.
`Although SPLP shows considerable potential as systemic gene transfer
`agents, the detergent dialysis method suffers from a number of limitations.
`Detergent dialysis is exquisitely sensitive to minor changes in the ionic strength
`of the formulation buffer. Changes as small as 10 mM result in dramatic decrease
`in encapsulation efficiency (Fenske et al., 2002). Even when SPLP are formed
`under ideal conditions, the detergent dialysis method results in the formation of
`large numbers of empty vesicles that are usually separated from SPLP by gradient
`ultracentrifugation (Fenske et al., 2002). The detergent dialysis method is also
`difficult to scale to the size required to support preclinical and clinical develop-
`ment of the technology. For these reasons, alternative methods of preparing
`stable plasmid lipid particles have been explored.
`One such method uses ethanol-destabilized cationic liposomes (Maurer
`et al., 2001). Though this method does not require gradual detergent removal or
`ultracentrifugation steps, it does require the formation of cationic vesicles prior
`to the encapsulation of pDNA. Once cationic liposomes of the desired size have
`been prepared, they are destabilized by ethanol addition to 40% v/v. Destabili-
`zation of vesicles with ethanol requires very slow addition of ethanol to a rapidly
`
`

`

`trifugation)
`filter,ultracen-
`54.6nm(400nm
`
`nmfilter)
`
`(400nm)
`
`1.97g/mol
`
`(200nm)
`
`142.5nm(200
`
`0.83 g/ mol
`
`0.5 to 7.5 m
`
`ND
`
`ND
`
`ND
`
`290nm(DOPE)
`180nm(DOPC)
`
`2.25g/mol
`
`14to17%
`
`4.6kbplasmid
`
`50nm
`
`0.26g/mol
`
`ND
`
`15  g/ mol
`
`11%
`
`15%
`
`Aug¼230nm
`
`250,000MW)
`(approximately
`genomicDNA
`
`sonicated
`
`3.9 kb plasmid
`
`(43.5:43.5:13)
`
`EPC:Chol:stearylamine
`PC:PG:Chol(40:10:50)
`PC:PS:Chol (40:10:50)
`
`0.1to1.5m;
`
`<1g/mol
`
`2to6%
`
`3.9kbplasmid
`
`EPC:EPG(91:9)
`
`400nm
`
`0.38g/mol
`
`ND
`
`0.97  g/ mol
`
`12%
`
`10%
`
`3.9kbplasmid
`
`EPC:PS:Chol(40:10:50)
`
`plasmid
`
`8.3kb,14.2kbp
`
`PC:PS:Chol(50:10:40)
`
`etal.,1986)
`
`Detergentdialysis(Stavridis
`
`andRottem,1982)
`
`Etherinjection(Nicolau
`
`(Fraleyetal.,1979)
`
`Etherinjection
`
`(CuddandNicolau,1985)
`
`Reverse-phaseevaporation
`(Nakanishietal.,1985)
`Reverse-phaseevaporation
`
`(Sorianoetal.,1983)
`
`100nmto1m
`
`0.23g/mol
`
`13to16%
`
`11.9kbplasmid
`
`PC:PS:Chol(40:10:50)
`
`Reverse-phaseevaporation
`
`400nm
`
`g/mol
`
`<4.2
`
`30to50%
`
`SV40DNA
`
`PSorPS:Chol(50:50)
`
`Reverse-phaseevaporation
`
`(Fraleyetal.,1980)
`
`Diameter
`
`ratioa
`
`DNA-to-lipid
`
`efficiencya
`Trapping
`
`LengthofDNA
`
`Lipidcomposition
`
`Procedure
`
`Table6.1.ProceduresforEncapsulatingPlasmidinLipid-BasedSystems
`
`ND
`
`Chol:EPC:PS (50:40:10)
`
`Dehydration-rehydration,
`
`filters)(Alinoetal.,1993)
`extrusion(400or200nm
`
`plasmid
`
`3.9kb,13kb
`
`orEPC
`
`EPC:Chol(65:35)
`
`(40:40:20)
`DOPE:Chol:oleicacid
`DOPC:Chol:oleicacidor
`
`(Lurquin,1979)
`
`Lipidhydration
`
`(WangandHuang,1987)
`Detergentdialysis,extrusion
`
`

`

`(QELS)
`
`100–150nm
`
`70g/mol
`
`80to95%
`
`fracture)
`65nm(freeze-
`
`75nm(QELS);
`
`62.5g/mol
`
`60to70%
`
`80 to 120 nm
`
`ND
`
`17 to 50%
`
`plasmid
`
`4.4to15kb
`
`plasmid
`
`4.4to15kb
`
`plasmid
`
`3.4kblinear
`
`bacteriophage
`
`ND¼Notdetermined.
`aSomevaluescalculatedbasedonpresenteddata.
`
`Various