A scalable, extrusion-free method for efficient liposomal encapsulation of plasmid DNA

Article abstract of DOI:10.1007/s11095-004-1873-z

Purpose. A fully scalable and extrusion-free method was developed to prepare rapidly and reproducibly stabilized plasmid lipid particles (SPLP) for nonviral, systemic gene therapy. Methods. Liposomes encapsulating plasmid DNA were formed instantaneously by mixing lipids dissolved in ethanol with an aqueous solution of DNA in a controlled, stepwise manner. Combining DNA-buffer and lipid-ethanol flow streams in a T-shaped mixing chamber resulted in instantaneous dilution of ethanol below the concentration required to support lipid solubility. The resulting DNA-containing liposomes were further stabilized by a second stepwise dilution. Results. Using this method, monodisperse vesicles were prepared with particle sizes less than 200 nm and DNA encapsulation efficiencies greater than 80%. In mice possessing Neuro 2a tumors, SPLP demonstrated a 13 h circulation half-life in vivo, good tumor accumulation and gene expression profiles similar to SPLP previously prepared by detergent dialysis. Cryo transmission electron microscopy analysis showed that SPLP prepared by stepwise ethanol dilution were a mixed population of unilamellar, bilamellar, and oligolamellar vesicles. Vesicles of similar lipid composition, prepared without DNA, were also <200 nm but were predominantly bilamellar with unusual elongate d morphologies, suggesting that the plasmid particle affects the morphology of the encapsulating liposome. A similar approach was used to prepare neutral egg phosphatidylcholine:cholesterol (EPC:Chol) liposomes possessing a pH gradient, which was confirmed by the uptake of the lipophilic cation safranin O. Conclusions. This new method will enable the scale-up and manufacture of SPLP required for preclinical and clinical studies. Additionally, this method now allows for the acceleration of SPLP formulation development, enabling the rapid development and evaluation of novel carrier systems.

Full text of DOI:10.1007/s11095-004-1873-z

Pharmaceutical Research, Vol. 22, No. 3, March 2005 (© 2005)
DOI: 10.1007/s11095-004-1873-z
Research Paper
A Scalable, Extrusion-Free Method for Efficient Liposomal Encapsulation of
Plasmid DNA
Lloyd B. Jeffs,1 Lorne R. Palmer,1 Ellen G. Ambegia,1 Cory Giesbrecht,1 Shannon Ewanick,1 and
Ian MacLachlan1,2
Received September 17, 2004; accepted December 6, 2004
Purpose. A fully scalable and extrusion-free method was developed to prepare rapidly and reproducibly
stabilized plasmid lipid particles (SPLP) for nonviral, systemic gene therapy.
Methods. Liposomes encapsulating plasmid DNA were formed instantaneously by mixing lipids dis-
solved in ethanol with an aqueous solution of DNA in a controlled, stepwise manner. Combining
DNA-buffer and lipid-ethanol flow streams in a T-shaped mixing chamber resulted in instantaneous
dilution of ethanol below the concentration required to support lipid solubility. The resulting DNA-
containing liposomes were further stabilized by a second stepwise dilution.
Results. Using this method, monodisperse vesicles were prepared with particle sizes less than 200 nm
and DNA encapsulation efficiencies greater than 80%. In mice possessing Neuro 2a tumors, SPLP
demonstrated a 13 h circulation half-life in vivo, good tumor accumulation and gene expression profiles
similar to SPLP previously prepared by detergent dialysis. Cryo transmission electron microscopy
analysis showed that SPLP prepared by stepwise ethanol dilution were a mixed population of unila-
mellar, bilamellar, and oligolamellar vesicles. Vesicles of similar lipid composition, prepared without
DNA, were also <200 nm but were predominantly bilamellar with unusual elongated morphologies,
suggesting that the plasmid particle affects the morphology of the encapsulating liposome. A similar
approach was used to prepare neutral egg phosphatidylcholine:cholesterol (EPC:Chol) liposomes pos-
sessing a pH gradient, which was confirmed by the uptake of the lipophilic cation safranin O.
Conclusions. This new method will enable the scale-up and manufacture of SPLP required for pre-
clinical and clinical studies. Additionally, this method now allows for the acceleration of SPLP formu-
lation development, enabling the rapid development and evaluation of novel carrier systems.
KEY WORDS: DNA; liposome; plasmid; systemic gene delivery.
Following delivery to the target tissue, the vector must then
be taken up by the tissue and express its DNA payload at
sufficient levels to give a therapeutic effect. Importantly, a
systemic vector must also be safe, well tolerated upon admin-
istration, and be non-immunogenic. The utility of viral vec-
tors, cationic lipid–containing DNA complexes (lipoplexes),
and polycationic polymer nucleic acid complexes (polyplexes)
is limited in a systemic context (1).
Liposomes encapsulating small-molecule drugs such as
doxorubicin (2) and vincristine (3) have been shown to be
effective in humans. Liposomes can increase drug accumula-
tion at the tumor site by 50 to 100 times compared to the
administration of free drug (4–6). This is the result of the
INTRODUCTION
Systemic delivery of therapeutic genes to disease sites
requires a vector with suitable properties. An ideal systemic
vector should have the appropriate physical attributes to en-
sure favorable pharmacokinetics and delivery to disease sites.
1 Protiva Biotherapeutics Inc., Burnaby, British Columbia, Canada
V5G 4Y1.
2
To whom correspondence should be addressed. (e-mail:
ian@protivabio.com)
ABBREVIATIONS: Chol, cholesterol; DODAC, dioleyldimethyl-
ammonium chloride; DODAP, 1,2-dioleoyl-N,N-dimethyl-3-
aminopropane; DODMA, 1,2-dioleyloxy-N,N-dimethylaminopro- phenomenon referred to as “disease site targeting”—a pas-
pane; DOPE, dioleoylphosphatidylethanolamine; DOPG, dio-
leoylphosphatidylglycerol; DSPC, distearoylphosphatidylcholine;
EPC, egg phosphatidylcholine; HBS, Hepes buffered saline; 3H-
CHE, tritium-labeled cholesteryl hexadecyl ether; OGP, octylgluco-
pyranoside; PBS, phosphate buffered saline; PEG-CerC20, 1-O-(2’-
(␻-methoxypolyethyleneglycol)2000)-2-N-arachidoylsphingosine;
PEG-S-DSG, 3-O-(2’(␻-methoxypolyethyleneglycol)2000)-1,2-
distearoyl-sn-glycerol; QELS, quasi-elastic light scattering; SPLP, sta-
bilized plasmid lipid particles; SVF, spontaneous vesicle formation;
TE, Tris EDTA buffer; TEM, transmission electron microscopy;
TNS, potassium 2-(p-toluidino)-6-naphthalenesulfonic acid.
sive targeting process that requires liposomes of around 100
nm in diameter and a circulation half-life >5 h. Although
encapsulation of small-molecule drugs is relatively straight-
forward, encapsulating macromolecules, such as plasmid
DNA, within a liposomal bilayer poses a greater challenge.
Efficient encapsulation of plasmid DNA has been achieved
using a detergent dialysis approach where lipid bilayers are
formed around plasmid DNA as the detergent is removed (7).
The resulting stabilized plasmid lipid particles (SPLP) con-
tained DNA fully encapsulated within a lipid bilayer that incor-
0724-8741/05/0300-0362/0 © 2005 Springer Science+Business Media, Inc.
362

Efficient Liposomal Encapsulation of Plasmid DNA
363
porated a cationic lipid (DODAC), a fusogenic lipid (DOPE), bromide (10.0 g, 30 mmol) was added and the reaction re-
and a PEG-lipid (PEG-CerC20). These particles possessed an fluxed under argon for 18 h. The reaction mixture was
average diameter of less than 100 nm, and unlike DNA- quenched with ethanol, washed (3 × 250 ml distilled water),
cationic lipid complexes formed by mixing preformed cationic dried (magnesium sulfate), and evaporated. The organic
lipid–containing vesicles with plasmid DNA, fully protect phase was dried over magnesium sulfate and evaporated. The
DNA following incubation with Escherichia coli DNase I. crude product was purified on a silica gel column (1% metha-
SPLP prepared using this approach have a circulation half- nol in chloroform). It was subsequently decolorized by stir-
life of >6 h, accumulate in distal tumors following intravenous ring for 30 min in a suspension of activated charcoal (1 g) in
administration in tumor-bearing mice, and lead to reporter ethanol (75 ml) at 60°C. The charcoal was removed by filtra-
gene expression in murine tumors (8). SPLP have also been tion through Celite and the ethanol solution concentrated to
prepared by destabilizing preformed empty vesicles in etha- yield 4.8 g (65%) of pure product.
nol at 40% v/v and incubating in the presence of plasmid
DNA (9). This ethanol destabilization method yielded DNA distearoyl-sn-glycerol (PEG-S-DSG; Fig. 1B) was prepared
encapsulation efficiencies of approximately 65%. by treating a solution of monomethoxypolyethylene glycol
3-O-[2Ј(␻-methoxypolyethyleneglycol)2000]-1,2-
Although SPLP show considerable potential for sys- (average MW ס 2000; PEG2000) (30 g, 15 mmol) in pyridine
temic gene transfer, the detergent dialysis method of prepa- (150 ml) with succinic anhydride (10 g, 100 mmol) overnight.
ration suffers from a number of limitations. The method is The solvent was removed by rotovap and the solution diluted
exquisitely sensitive to minor changes in the ionic strength of with water (200 ml). The solution was acidified with concen-
the formulation buffer; changes as small as 10 mM result in a trated HCl and extracted with choloroform (3 × 150 ml). The
dramatic decrease in encapsulation efficiency (10). Even organic fraction was dried over magnesium sulfate and con-
when SPLP are formed under ideal conditions, the detergent centrated to yield ∼30 g of crude product as a white wax.
dialysis method results in the formation of large numbers of Purification was carried out by flash column chromatography
empty vesicles that are typically separated from SPLP by with chloroform containing 0–7% methanol. The pure prod-
gradient ultracentrifugation (10). The detergent dialysis uct was taken up in water (300 ml) and lyophilized to give 20
method is also difficult to scale to the size required to support g of PEG2000-succinate (PEG2000-S) as a white powder.
preclinical and clinical development. For this reason, we PEG2000-S (9 g, 4.3 mmol) was dissolved in benzene (100 ml)
sought to develop a more simple, robust, and fully scalable and treated with oxalyl chloride (4.44 g, 35 mmol). The solu-
method for the encapsulation of plasmid DNA in SPLP. The tion was stirred for 2 h at room temperature and the solvent
method reported here, a type of spontaneous vesicle forma- removed under vacuum. The residue was dissolved in etha-
tion, produces SPLP with the same desirable properties as nol-free chloroform (100 ml) and 1,2-distearoyl-sn-glycerol
those prepared by detergent dialysis. We discuss the use of (3.0 g, 4.8 mmol) and triethylamine (3 ml, 20 mmol) added.
this method for the preparation of other types of drug deliv- The solution was stirred for 48 h, added to water (100 ml),
ery systems formerly prepared using extrusion methods.
acidified with HCl, and the organic phase collected. The
aqueous phase was extracted with chloroform (2 × 100 ml).
The organic phases were combined, dried over MgSO4, fil-
tered, and the solvent removed. The crude product was pu-
rified by column chromatography, eluting with chloroform
containing 0–7% methanol. The product was decolorized by
stirring for 30 min in a suspension of activated charcoal (1 g)
in ethanol (100 ml) at 60°C. The suspension was filtered, the
solvent removed, and the product taken up in water (300 ml)
and lyophilized, yielding PEG2000DSG (6.0 g) as a fluffy white
powder.
Dioleoylphosphatidylethanolamine (DOPE) and egg
phosphatidylcholine (EPC) were purchased from Northern
Lipids Inc. (Vancouver, BC, Canada). Distearoylphosphati-
dylcholine (DSPC) was purchased from Avanti Polar Lipids,
Inc. (Alabaster, AL, USA). Synthetic cholesterol (Chol) and
the detergent octylglucopyranoside (OGP) were obtained
from Sigma-Aldrich Co. (Oakville, ON, Canada). Tritium-
labeled cholesteryl hexadecyl ether (3H-CHE) was obtained
from Mandel NEN Products (Guelph, ON, Canada).
The pCMVluc plasmid, encoding the luciferase reporter
gene under the control of the cytomegalovirus promoter, was
propagated in Escherichia coli strain DH5␣ and purified by
standard alkaline lysis/cesium chloride density gradient cen-
trifugation (14,15).
MATERIALS AND METHODS
Lipids and Plasmid DNA
The cationic lipids dioleyldimethylammonium chloride
(DODAC) and 1-O-[2Ј-(␻-methoxypolyethylenegly-
col)2000]-2-N-arachidoylsphingosine (PEG-CerC20) were
synthesized as described previously (11,12). The cationic lipid
1,2-dioleyloxy-N,N-dimethylaminopropane (DODMA; Fig.
1A) was synthesized using methodology derived from that
used previously for the synthesis of a DOTMA precursor
(13). 3-(Dimethylamino)-1,2-propanediol (1.43 g, 12 mmol)
and 95% sodium hydride (NaH, 2.52 g, 100 mmol) were
stirred in benzene (60 ml) under argon for 30 min. Oleyl
SPLP Prepared by Detergent Dialysis
SPLP were prepared by detergent dialysis as described
elsewhere (7,10). Briefly, DOPE, DODAC, and PEG-CerC20
Fig. 1. Chemical structures of the lipids (A) DODMA and (B) PEG-
S-DSG.

364
Jeffs et al.
at a molar ratio of 82.5:7.5:10 were dissolved in aqueous so-
The plasmid and lipid solutions were combined using a
lutions of OGP. A pCMVluc plasmid solution was then added single peristaltic pump with dual pump heads (Watson Mar-
to give a solution of 0.4 mg/ml DNA, 10 mg/ml lipid, and 200 low), joined in opposition using a polypropylene T-connector
mM detergent. This solution was dialyzed for 48 h and unen- (Cole-Parmer) and platinum-cured silicone tubing (Nalgene)
capsulated DNA removed by anion exchange chromatogra- both with an internal diameter of 1.6 mm. This configuration
phy (DEAE Sepharose CL6B; Sigma). Empty vesicles were is conceptually similar to one that has been used to prepare
then removed by a one-step sucrose density ultracentrifuga- lipoplex by the controlled mixing of preformed cationic lipo-
tion. Fractions containing SPLP were consolidated and dia- somes with plasmid DNA (16). The two peristaltic pump
lyzed with Hepes buffered saline (HBS) to remove sucrose. heads were calibrated and adjusted to deliver 1 ml/s of mixed
Isolated SPLP were concentrated by Amicon ultrafiltration to solution. The apparatus was then used to prepare SPLP by
a final DNA concentration of 0.5 mg/ml. All samples were mixing the lipid and plasmid solutions, preheated to 37°C,
passed through sterile 0.2-␮m filters and then stored at 5
3°C until use.
resulting in the instantaneous formation of a liposomal sus-
pension in 45% ethanol with approximately 60% DNA en-
capsulation. The volumes of lipid and plasmid solutions are
not critical for the preparation of SPLP using this method.
Solution volumes as low as 1 ml (0.9 mg DNA) and as large
as 5000 ml (4.5 g DNA) have been used to prepare SPLP
with similar results. Upon formation, the SPLP were then
immediately diluted with 300 mM NaCl with 20 mM citrate
pH 6.0 at 37°C using a similar apparatus. This second ethanol
dilution step was found to stabilize further the metastable
vesicles and resulted in a significant increase in the DNA
encapsulation from 60% to 80–90%. The diluted vesicles
were then incubated at 37°C for 30 min prior to removal of
unencapsulated DNA by charged membrane filtration
through a Mustang Q coin filter (Pall Corporation, Ann Ar-
bor, MI, USA). After removal of unencapsulated DNA, en-
capsulated DNA was typically greater than 98% of that re-
maining. SPLP were concentrated using a tangential flow
ultrafiltration system (Amersham Biosciences, Piscataway,
NJ, USA) and diafiltered against 15 volumes of phosphate-
buffered saline (PBS), pH 7.4, reducing the ethanol concen-
tration to less than 0.5% [determined using a standard color-
imetric alcohol dehydrogenase assay (Sigma)]. After ultrafil-
tration, samples were filtered through sterile 0.2-␮m filters
(Pall) and stored at 5 3°C until use.
Preparation of SPLP and Vesicles by Spontaneous
Vesicle Formation
A flow diagram summarizing the process for preparing
SPLP by spontaneous vesicle formation (SVF) is shown in
Fig. 2. First, a plasmid solution was prepared by combining
DNA in 10 mM Tris EDTA (TE) buffer with 100 mM citrate
buffer (pH 5.0) and distilled deionized water to achieve a
plasmid DNA concentration of 0.9 mg/ml in 20 mM citrate. A
lipid solution in ethanol was prepared by dissolving Chol,
DSPC, DODMA, and PEG-S-DSG at molar ratios of 55:20:
15:10 in absolute, anhydrous ethanol and then adding distilled
water to achieve an ethanol concentration of 90 vol%. This
lipid composition was similar to the lipid composition of pre-
formed vesicles used previously to encapsulate plasmid (9),
except the more stable DODMA was selected instead of the
titratable cationic lipid 1,2-dioleyloxy-N,N-dimethylamino-
propane (DODAP), and PEG-S-DSG replaced PEG-CerC20.
The total concentration of lipid in solution was 20 mM. Equal
volumes of both lipid and plasmid solutions were heated to
37°C prior to vesicle formation.
Empty vesicles of similar lipid composition, but lacking
DNA, were prepared in a similar manner as SPLP, with the
exception that an equivalent volume of 10 mM Tris-EDTA
(TE) buffer was added to the initial citrate buffer solution
instead of DNA.
Vesicles comprising EPC:Chol at a 55:45 molar lipid ratio
were also prepared using this apparatus. Lipids were dis-
solved at 20 mM in 80% ethanol, and vesicles were formed by
mixing these lipids with either PBS (pH 7.4) or 150 mM cit-
rate (pH 4.0). The second dilution step was performed using
the same buffer as the first mixing step. Following sample
concentration by ultrafiltration, EPC:Chol vesicles were first
diafiltered with 15 volumes of the original mixing buffer to
ensure maintenance of the internal pH of the vesicles. Once
the ethanol was removed, the vesicles were diafiltered against
PBS, pH 7.4. These vesicles were then adjusted to the desired
lipid concentration and filtered using sterile 0.2-␮m filters.
Radiolabeled SPLP were prepared by adding 3H-CHE
at 0.5 ␮Ci/mg lipid to the lipid-ethanol solution prior to mix-
ing with the DNA solution.
Lipid Analysis
The total lipid content of SPLP and liposomes was cal-
culated from the phospholipid content determined using the
colorimetric method of Fiske and Subbarow (17). For SPLP,
Fig. 2. Process flowchart for SPLP prepared by spontaneous vesicle
formation. Please refer to Materials and Methods for a full descrip-
tion of the process.

Efficient Liposomal Encapsulation of Plasmid DNA
365
lipid was first isolated from DNA using chloroform extraction sample was incubated with 500 U of DNase 1 (Life Technolo-
(18). The apparent pKa for the tertiary amino lipid DODMA gies, Mississauga, ON, Canada) in a total volume of 100 ␮l
in the lipid bilayer component of SPLP was determined as React 4 buffer at pH 7.4 (Life Technologies) for 60 min at
previously described (19) by measuring the fluorescence of 37°C. Following incubation, DNA was isolated by sequen-
potassium 2-(p-toluidino)-6-naphthalenesulfonic acid (TNS) tially adding 500 ␮l DNAzol (Life Technologies) and 1000 ␮l
over a pH range using a spectrofluorometer (SLM-Aminco) absolute ethanol. After centrifuging for 30 min at 13,000 × g,
using excitation and emission wavelengths of 321 and 445 nm, the supernatant was decanted and the DNA washed twice
respectively.
with 70% ethanol and dried. The resultant DNA pellets were
dissolved in 30 ␮l Tris-EDTA buffer, pH 8.0, and analyzed by
gel electrophoresis using a 0.8% agarose gel in Tris-acetate-
EDTA, with a high DNA mass ladder (Invitrogen).
Particle Size Analysis
Particle size was determined using a Nicomp model 370
Submicron Particle Sizer (Particle Sizing Systems, Santa Bar-
bara, CA, USA). The instrument employs quasi-elastic light
scattering (QELS) to determine the hydrodynamic equivalent
spherical diameter of the particles. Volume-weighted, Gaus-
sian distribution analysis was used to determine mean vesicle
diameters and population standard deviations. Particle poly-
dispersity was evaluated by the goodness-of-fit parameter ␹2,
where a value of <3 indicates a monodisperse suspension (ac-
cording to the instrument manufacturer).
Cryo Transmission Electron Microscopy
Cryo transmission electron microscopy (Cryo-TEM)
analysis of SPLP and liposomes was performed at Uppsala
University, Sweden, using a Zeiss EM 902A with cryo equip-
ment.
Dye Loading of EPC-Chol Liposomes
The feasibility of preparing EPC:Chol vesicles with an
acidic aqueous interior by stepwise ethanol dilution was in-
vestigated by formulating with either citrate buffer (pH 4.0)
or PBS (pH 7.4). Vesicles at 5 mM lipid in PBS (pH 7.4) were
incubated with 0.2 mM of the lipophilic dye safranin O
(Sigma) at 37°C for 30 min. After 30 min, 500 ␮l aliquots were
passed through a 2-ml Sepharose CL4B gel filtration column
to separate free dye from the vesicles. Fractions were ana-
lyzed for lipid content and safranin O content, which was
determined using a spectrofluorometer (SLM-Aminco) at an
excitation wavelength of 516 nm and emission wavelength of
585 nm.
DNA Content
Concentration and percent encapsulation of plasmid
DNA in SPLP were determined using the membrane imper-
meable dye PicoGreen (Molecular Probes, Eugene, OR,
USA) that binds specifically to double-stranded DNA and
fluoresces only when bound. Picogreen fluorescence was
measured with a spectrofluorometer (SLM-Aminco) using
excitation and emission wavelengths of 495 and 525 nm, re-
spectively. Plasmid encapsulation was determined by measur-
ing fluorescence upon addition of PicoGreen to SPLP (Fi)
and comparing this value to that obtained upon lysis of the
SPLP lipid bilayers with 0.1% Triton X-100 (Ft):
In Vivo Transfection, Plasma Clearance, and Tumor
Accumulation Studies
Ft − Fi͒
Ft
͑
SPLP Plasmid Encapsulation %͒ =
× 100%
͑
All animal studies were performed in accordance with
the guidelines of the Canadian Council on Animal Care. Ten
to 14 days prior to SPLP treatment, 8-week-old male A/J mice
(Harlan, IN, USA) were inoculated subcutaneously in the
hind flank with 1.5 × 106 Neuro 2a cells (ATCC, VA, USA).
For in vivo transfection studies, SPLP (200 ␮l total volume
containing 2 mg total lipid and 100 ␮g total DNA) were ad-
ministered by lateral tail vein injection. For plasma clearance
and tumor accumulation studies, SPLP incorporating 3H-
CHE were used at the same lipid and DNA concentrations.
At appropriate time points, mice were anesthetized and blood
was collected by cardiac puncture into microtainer tubes.
Plasma was separated from red blood cells via centrifugation
The total concentration of plasmid was then determined
by comparing the fluorescence after addition of detergent
(Ft) to a DNA standard curve prepared using a plasmid DNA
standard:
Ft − b͒ × Dilution
͑
Total DNA Concentration =
m
where b is the y intercept of the DNA standard curve, and m
is the slope of the standard curve.
The concentration of encapsulated plasmid DNA in
SPLP was then determined by multiplying the total DNA
concentration value by the value for percent plasmid encap-
sulation. In vivo experiments were performed using the con-
centration of encapsulated DNA to calculate DNA dose.
3
and analyzed for H-CHE by liquid scintillation counting us-
ing Picofluor 20 (Perkin Elmer) and a Beckman LS6500
(Beckman Instruments). For clearance and accumulation
studies, tumors were harvested at the specified time points
and homogenized in lysing matrix tubes containing 500 ␮l of
distilled water. Liver lysate (100 ␮l) and 200 ␮l of all other
DNase Protection Assay
The ability of SPLP to protect encapsulated plasmid tissue lysates were assayed for radioactivity by liquid scintil-
from exogenous DNase was assessed using a method modi- lation counting with Picofluor 40 (Perkin Elmer). For gene
fied slightly from that reported previously (7). The following expression studies, tumors were lysed with 750 ␮l of cell cul-
samples were tested: naked plasmid DNA, SPLP prepared ture lysis reagent (CCLR, Promega, Madison, WI, USA), and
by SVF, and plasmid DNA added to empty vesicles with the 20 ␮l of lysate was assayed for luciferase activity in a 96-well
same lipid composition as SPLP to achieve a similar microplate luminometer (Bernholdt Technologies, Oakridge,
DNA:lipid ratio as the SPLP. A 1-␮g DNA aliquot of each TN, USA) using the Promega Luciferase Assay reagent kit

366
Jeffs et al.
(Promega) according to the manufacturer’s instructions. Re-
combinant luciferase from the American firefly, Photinus
pyralis (Roche, PQ, Canada) was used to calibrate lumines-
cence readings.
RESULTS
Characterization of the Spontaneous Vesicle
Formation Method
Lipid encapsulated plasmid DNA can be prepared by the
dropwise addition of lipid dissolved in ethanol to a rapidly
mixing aqueous buffer containing DNA. Small-scale SPLP
samples were prepared using either the stepwise dilution
method or the ethanol drop approach diluted in each case to
a final ethanol concentration of 22.5% and incubated for 60
min before DNA encapsulation efficiency was determined
(Fig. 3). SPLP prepared by the stepwise approach had a
DNA encapsulation efficiency of 81% and were monodis-
perse with a mean vesicle diameter of 132 55 nm (␹2 ס 0.6).
SPLP prepared by ethanol drop to an initial target ethanol
concentration of 45% and then diluted dropwise to 22.5%
ethanol had a DNA encapsulation efficiency of 74% and
2
mean vesicle size of 116 54 nm (␹ ס 0.4). SPLP prepared
by a single-step ethanol drop procedure to 22.5% ethanol in
Fig. 4. Effects of (A) initial ethanol concentration and (B) pH on
pH 6.0 citrate buffer had a mean vesicle size of 107 53 nm
but was a polydisperse formulation (␹ ס 14.1), and DNA ethanol concentration in the lipid solution was varied, while the pH of
DNA encapsulation at the vesicle formation step. (A) The initial
2
the plasmid solution was kept constant at pH 5.0. (B) The pH of the
plasmid solution was varied while the ethanol concentration was kept
constant at 90%. DNA trapping efficiencies for the SPLP samples
were determined after the dilution step with 300 mM NaCl, 20 mM
citrate, pH 6.0. Data represent mean SD.
encapsulation was only 17%. These results suggest an ideal
range of ethanol concentration, referred to here as the “DNA
encapsulation zone” (Fig. 2). In this zone, lipid bilayers effi-
ciently form around the DNA particles. This same SPLP for-
mulation can be prepared with similar DNA encapsulation
efficiencies and vesicle size at a starting concentration of 80–
100% ethanol. When encapsulation was attempted in 75%
ethanol, DNA encapsulation decreased dramatically and
tempting to prepare large-scale SPLP preparations. In this
case, it also becomes difficult to achieve rapid mixing of etha-
nol with larger volumes of DNA solution. This limitation may
contribute to low encapsulation efficiencies observed when
we attempted to scale this method (data not shown).
2
mean particle size increased (166 101 nm, ␹ ס 4.5), pre-
sumably as a result of reduced solubility of the lipid compo-
nents prior to mixing (Fig. 4A). Clearly, vesicles formed by
ethanol drop are exposed to variable ethanol concentrations
with the first vesicles formed at lower ethanol concentrations,
concentrations that preclude DNA encapsulation. This limi-
tation of the ethanol drop method is more evident when at-
An important feature of the SPLP prepared here is the
titratable cationic lipid DODMA. This lipid has an apparent
pKa of 6.8 in a lipid bilayer (as determined by TNS assay) and
while predominantly uncharged at physiologic pH, possesses
a positive charge under acidic conditions that is thought to
promote the association of DODMA-containing bilayer frag-
ments with negatively charged DNA. Because DNA hydro-
lyzes when incubated under acidic conditions (20), we varied
the pH of the initial DNA buffer to determine the ideal pH
for maximizing DNA encapsulation while minimizing DNA
degradation. Predictably, DNA encapsulation efficiency
dropped dramatically at mixing buffer pH values above 5
(Fig. 4B). The vesicles sizes changed little over the pH range
3 to 7 (110 37 nm to 122 55 nm), thus the pH of 5.0 was
selected for the initial step of this process.
While developing the encapsulation approach described
here, it was determined that the addition of a second dilution
step increased the in-process stability of SPLP and at the
same time increased the extent of DNA encapsulation from
60% to 70% seen after the initial mixing step to 80–90%. The
pH and salt concentration of the dilution buffer were varied
to determine the optimal conditions for promoting the in-
crease in DNA encapsulation (Fig. 5). Dilution buffer con-
taining 300 mM NaCl with pH adjusted from pH 4 to 10 was
Fig. 3. Comparison of spontaneous vesicle formation by stepwise di-
lution to the ethanol drop method for encapsulating DNA. Lipids
were dissolved in (A) 90% ethanol and diluted either stepwise to (B)
45% and (C) 22.5% ethanol, represented by the solid line, or (C)
added dropwise into a stirred plasmid solution to a final ethanol
concentration of 22.5%, represented by the dotted line.

Efficient Liposomal Encapsulation of Plasmid DNA
367
Fig. 6. SPLP prepared by spontaneous vesicle formation protect
DNA from digestion by DNase 1. Lanes: Stability of free plasmid
DNA, free plasmid in the presence of liposomes, and plasmid encap-
sulated in SPLP in the presence of DNase I. Each sample consisted of
1 ␮g of DNA and was incubated with either 500 U of DNase 1 (+) or
without DNase 1 (−) at 37°C for 1 h prior to analysis by gel electro-
phoresis.
Fig. 5. Effect of varying (A) pH and (B) ionic strength on DNA
encapsulation. Following the formation of SPLP by SVF. (A) The
pH of the dilution buffer was varied while keeping the ionic strength
of the diluted sample constant at 150 mM NaCl. (B) The ionic
strength of the dilution buffed was varied while keeping the pH of the
dilution buffer constant at pH 6.0. The dilution buffers possessed
twice the concentration of the target NaCl concentration for the di-
luted sample (e.g., the dilution buffer was 300 mM NaCl, to achieve
a salt concentration of 150 mM after dilution). Data represent mean
SD.
DNA in the SPLP sample showed no sign of degradation,
indicating that the DNA is completely protected. A similar
result was seen previously for SPLP prepared by detergent
dialysis (7).
evaluated. Optimal DNA encapsulation was observed at
buffer pH values of 4 to 6 (Fig. 5A). Above pH 6, DNA
encapsulation diminished until there was no enhancement at
pH 8 or above. This effect is thought to be a result of the net
charge on the DODMA in the SPLP. The dilution buffer was
Liposome Formulations Prepared by Spontaneous
Vesicle Formation
SPLP or empty vesicles similar in lipid composition to
set at pH 6.0 with citrate, and the optimal ionic strength of the SPLP were prepared by SVF and analyzed for particle size,
dilution buffer was determined (Fig. 5B). Interestingly, dilu- DNA and lipid content. Empty liposomes prepared by SVF
2
tion in 20 mM citrate buffer at pH 6.0 without NaCl yielded had mean vesicle diameter of 120 11 nm, ␹ ס 0.3, and a
2
no additional DNA encapsulation compared to the mixed narrower size distribution than SPLP (101
37 nm, ␹
ס
SPLP sample, suggesting that a simple reduction of ethanol 0.7). The efficiency of DNA encapsulation for SPLP pre-
concentration and shift in pH from 5.0 to 6.0 is insufficient to pared by SVF was 88% prior to free DNA removal and 96%
increase DNA encapsulation. Increasing the NaCl concentra- in the final sterile filtered sample and following the free DNA
tion in the dilution buffer to 300 mM resulted in an increase removal step. For this formulation, the starting DNA:lipid
in DNA encapsulation to 90%. As the NaCl concentration in ratio was 45 ␮g DNA/␮mol lipid while the DNA:lipid ratio in
the dilution buffer was increased beyond 300 mM, the DNA the final filtered sample was 40 ␮g DNA/␮mol, indicating that
encapsulation efficiency was reduced. In the presence of the ratio of encapsulated DNA to lipid remains constant
NaCl, supercoiled DNA is known to adopt a more compact throughout the formulation process. For small-scale SPLP
structure (21). Therefore, the dilution step may be increasing formulation (<10 mg starting DNA), 50–60% recoveries of
DNA encapsulation by causing the DNA to adopt a more both encapsulated plasmid and lipid are typical in the final
compact structure, increasing the charge density and provid- sterile filtered sample. Apart from the loss of unencapsulated
ing more efficient association with lipid bilayer fragments that DNA (8–15%), the major sources of product loss are holdup
assemble to become liposomes containing DNA. Based on in the ultrafiltration cartridges and sterile filters. When en-
these observations, a dilution buffer of 300 mM NaCl with 20 capsulating plasmid DNA at larger scale (>100 mg plasmid
mM citrate pH 6.0 was chosen for SPLP formation.
DNA), we routinely achieve DNA yields of 75%, as the in-
The ability of SPLP prepared by SVF to protect encap- process holdup losses represent a smaller proportion of the
sulated DNA was tested by incubating with DNase I (Fig. 6). starting material. SPLP samples have been prepared with
Naked DNA was completely degraded by DNase treatment. plasmid DNA concentrations as high as 9 mg/ml DNA. A
DNA mixed with empty vesicles with the same lipid compo- variety of lipid compositions have been used to prepare
sition as SPLP was also completely degraded, indicating that SPLP using the SVF method, and plasmids varying in size
lipid alone did not interfere with the activity of the DNase. from 4.4 to 15 kb have been encapsulated using a lipid com-

368
Jeffs et al.
position containing a range of tertiary and quaternary cationic vesicles; and small dense particles (Fig. 7B). These dense par-
lipids, including DODAP and DODAC. ticles have a narrow size range and appear to possess mem-
It was of considerable interest to determine whether SVF brane structures. The 15% DODMA vesicles prepared by
could be used to prepare liposomes suitable for the encapsu- SVF have a very uniform bilamellar morphology (Fig. 7C).
lation of small-molecule drugs. Vesicles consisting of EPC: Many of these vesicles also exhibit protrusions of the outer
Chol (55:45 mol%) were also prepared by SVF. These bilayers that were not observed when the vesicles were pre-
vesicles possessed a mean particle diameter of 92 41 nm, pared in the presence of DNA. DNA appears to play an
2
␹
ס 1.7, well within the typical size range of vesicles pre- important role in determining the morphology of SPLP. It
pared by hydration from a lipid film followed by repeated may be that, in SPLP, the titratable cationic lipid DODMA
extrusion through 100-nm polycarbonate membranes (22,23). in the inner bilayer is in contact with the DNA, effecting
The morphologies of liposomes prepared by SVF were particle morphology in a manner that is not evident in the
analyzed by Cryo-TEM (Fig. 7). Cryo-TEM has been used empty vesicles. Furthermore, higher concentrations of
extensively to study liposome morphology, including DNA- DODMA in the outer bilayer may make the empty vesicles
lipid complexes (24). As reported previously, SPLP prepared more fusogenic, a property that has been described for analo-
by detergent dialysis are predominantly unilamellar (Fig. 7B) gous vesicles containing DODAP (19). Vesicles with tubular
(7). SPLP and vesicles prepared by SVF are morphologically protrusions have previously been observed by cryo-TEM and
distinct from the unilamellar vesicles prepared by detergent attributed to an asymmetric distribution of the anionic lipid
dialysis or by extrusion (23). SPLP prepared using SVF ap- dioleoylphosphatidylglycerol (DOPG) between the outer and
pear to have three distinct particle morphologies: unilamel- inner faces of the lipid bilayer in response to a transmem-
lar vesicles similar in appearance to SPLP prepared by deter- brane pH gradient (25), a mechanism that may be responsible
gent dialysis; as well as bilamellar or oligolamellar vesicles for the structures observed in Fig. 7C.
that are more abundant and smaller than the unilamellar
Bilamellar structures were also very common for the
EPC:Chol vesicles (data not shown). The lipophilic cation
safranin O was next used to determine whether SVF could be
used to prepare EPC:Chol vesicles possessing a transmem-
brane pH gradient. Liposomes possessing transmembrane pH
gradients (acidic inside) have been used to encapsulate doxo-
rubicin, vincristine, and other lipophilic amino-containing
drugs within liposomes (26). Safranin O is taken up by model
membrane systems in response to a membrane potential and
is a useful tool to measure pH gradients across liposome
membranes (27). When EPC:Chol vesicles were formulated
with citrate buffer at pH 4, they demonstrated a significant
increase in safranin O uptake of the cationic dye safranin
(55% encapsulation efficiency) as compared to similar
vesicles formulated with PBS pH 7.4 (9% encapsulation). Al-
though the conditions for uptake were not optimized, this
experiment clearly demonstrates that liposomal formulations
possessing a pH gradient can be prepared by SVF.
Animal Studies with SPLP
Extended circulation lifetime is a prerequisite for disease
site targeting and gene expression at distal tumor sites. To
determine the clearance properties of SPLP, we prepared
3
SPLP labeled with H-CHE by SVF and evaluated their be-
havior in mice. At 24 h after intravenous administration, 40%
of the injected dose was still present in the plasma, yielding a
SPLP plasma half-life of 13 h (Fig. 8A). This plasma clearance
profile was comparable to DOPE:DODAC:PEG-CerC20
SPLP that demonstrated a 10 h circulation half-life in mice
(12). The accumulation of SPLP at distal tumor sites was
significant, with 8% of the injected dose per gram of tumor
accumulating 24 h after administration (Fig. 8B). Other tis-
sues exhibited high levels of SPLP accumulation including the
spleen, liver, adrenal glands, and small intestine each accu-
mulating 44%, 26%, 14%, and 12% of the injected SPLP dose
per gram of tissue, respectively (data not shown). Although,
SPLP accumulation in these nontarget tissues was quite high,
gene expression in these tissues was consistently two to three
orders of magnitude lower than the tumor tissue, consistent
with previous observations using SPLP prepared by detergent
Fig. 7. Cryo transmission electron microscopy of SPLP. (A) SPLP
prepared by detergent dialysis. (B) SPLP prepared by stepwise etha-
nol dilution. (C) Vesicles as in (B) but prepared without DNA. Scale
bars ס 200 nm.

Efficient Liposomal Encapsulation of Plasmid DNA
369
sensitive to concentration effects. A significant benefit of the
SVF method reported here is that it facilitates the rapid
preparation of SPLP, allowing formulation development to
proceed at a much faster pace than previous methods. Where
it previously took 5 days to prepare a preclinical batch of
SPLP by detergent dialysis (10), the same SPLP formulation
can be prepared in a single day using this method. Spontane-
ous vesicle formation makes it possible to routinely prepare a
large number of small-scale formulations (1 to 5 mg DNA) in
a day, facilitating the accelerated optimization of formulation
composition (e.g., lipid ratios, plasmid content) and process
parameters (e.g., pH, ionic strength).
Spontaneous vesicle formation represents an effective
solution to the issues confounding SPLP preparation by de-
tergent dialysis (7) or using ethanol-destabilized cationic li-
posomes (9). For the detergent dialysis approach, liposomes
encapsulating DNA are formed using a 48-h dialysis step.
Vesicle formation is very sensitive to the rate of detergent
removal, and attempts to convert this dialysis step to a more
rapid and scalable tangential flow process have failed (data
not shown). The DNA trapping efficiency of the detergent
dialysis process is quite low. Unencapsulated DNA, as well as
empty liposomes, need to be removed from the formulation
to give an acceptable DNA/lipid ratio. Although the removal
of free DNA is straightforward, empty liposomes are re-
moved using sucrose density ultracentrifugation (10), a step
that is difficult to scale. Unlike the detergent dialysis ap-
proach, preparing SPLP using ethanol destabilized cationic
liposomes does not require gradual detergent removal or ul-
tracentrifugation steps (9). However, this method requires the
formation of cationic vesicles prior to the encapsulation of
DNA. The formation of vesicles by ethanol drop with an
optional vesicle size reduction by extrusion through polycar-
bonate membrane filters is considerably more complicated
and time consuming than the method described here. Once
cationic liposomes of the desired size have been prepared,
they need to be destabilized by ethanol addition to 40% v/v.
Destabilization of vesicles with ethanol requires very slow
addition of ethanol to a rapidly mixing aqueous suspension of
cationic lipids to avoid localized areas of high ethanol con-
centration (>50% v/v) that would promote fusion and con-
version of liposomes into large lipid structures (9). The addi-
tion of nucleic acid (oligonucleotide or DNA) must then be
accomplished, again slowly, in a dropwise manner. The un-
controlled nature of both the vesicle destabilization and
nucleic acid addition steps certainly poses challenges for re-
producibly preparing SPLP at a scale suitable for preclinical
evaluation.
Fig. 8. In vivo (A) plasma clearance and (B) tumor accumulation of
SPLP prepared by spontaneous vesicle formation. Each mouse re-
ceived a single intravenous injection of 100 ␮g SPLP DNA. Error
bars given as standard errors; n ס 4 to 5.
dialysis. Figure 9 illustrates a direct comparison of tumor gene
expression following intravenous administration of SPLP
prepared by detergent dialysis and SVF, demonstrating that
both methods yield particles with very similar patterns of lu-
ciferase gene expression over a 7-day period. Gene expres-
sion was highest 72 h after treatment with either preparation.
DISCUSSION
Controlled stepwise mixing has the advantage of main-
taining an accurately and consistently mixed product through-
out a process, facilitating scale-up of those processes that are
Neutral lipid complexes with 80% DNA encapsulation
have previously been prepared by mixing DOPC/DOPE
vesicles (1:1 molar ratio) with ethanol and calcium, followed
by dialysis (28). Although these particles were under 200 nm
and had longer in vivo circulation times than cationic lipid-
DNA complexes, they accumulated predominantly in the
liver and showed very little tumor accumulation. Numerous
other approaches have been used to encapsulate plasmid in
lipid particles, including reverse phase evaporation, ether in-
jection, and the dehydration of DNA lipid complexes and
subsequent rehydration (Table I). This dehydration and re-
hydration of plasmid lipid vesicles (DRV) produces particles
greater than 500 nm that are not suitable for systemic delivery
of genes but may show promise as DNA-based vaccines (29).
Fig. 9. In vivo luciferase gene expression following intravenous ad-
ministration of SPLP prepared by spontaneous vesicle formation
and detergent dialysis. Each mouse received a single intravenous
injection of 100 ␮g SPLP DNA. Data for SPLP prepared with deter-
gent dialysis and stepwise dilution are represented by open and solid
circles, respectively. Error bars given as standard errors, n ס 4.

370
Jeffs et al.
Table I. Procedures for Encapsulating Plasmid in Lipid-Based Systems
Lipid composition
and molar proportions
Trapping DNA-to-lipid
Procedure
Length of DNA
SV40 DNA
efficiencya
ratioa
Diameter
Reverse-phase
evaporation (30)
Reverse-phase
evaporation (31)
Reverse-phase
evaporation (32)
Reverse-phase
PS or PS:Chol (50:50)
PC:PS:Chol (40:10:50)
PC:PS:Chol (50:10:40)
EPC:PS:Chol (40:10:50)
30–50%
<4.2 ␮g/␮mol 400 nm
11.9 kb plasmid
13–16%
10%
0.23 ␮g/␮mol 100 nm to 1 ␮m
0.97 ␮g/␮mol ND
8.3 kb, 14.2 kbp
plasmid
3.9 kb plasmid
12%
0.38 ␮g/␮mol 400 nm
evaporation (33)
Ether injection (34)
Ether injection (35)
EPC:EPG (91:9)
3.9 kb plasmid
3.9 kb plasmid
2–6%
15%
<1 ␮g/␮mol
15 ␮g/␮mol
0.1 to 1.5 ␮m
ND
PC:PS:Chol (40:10:50)
PC:PG:Chol (40:10:50)
EPC:Chol:stearylamine
(43.5:43.5:13)
Detergent dialysis (36)
Sonicated genomic 11%
DNA (approx.
0.26 ␮g/␮mol 50 nm
250,000 MW)
Detergent dialysis, extrusion DOPC:Chol:oleic acid or 4.6 kb plasmid
14–17%
2.25 ␮g/␮mol 180 nm (DOPC) 290 nm
(37)
DOPE:Chol:oleic acid
(40:40:20)
(DOPE)
Lipid hydration (38)
EPC:Chol (65:35) or EPC 3.9 kb, 13 kb
plasmid
ND
ND
ND
0.5 to 7.5 ␮m
Dehydration-rehydration,
extrusion (400 or 200 nm
filters) (39)
Chol:EPC:PS (50:40:10)
ND
0.83 ␮g/␮mol 142.5 nm (200 nm filter) 54.6
(200 nm)
1.97 ␮g/␮mol
(400 nm)
2.65–3.0
␮g/␮mol
2.5–4.2
nm (400 nm filter,
ultracentrifugation)
Dehydration-rehydration
(40)
Dehydration-rehydration
(DRV) for DNA vaccines
(29)
EPC
2.96 kb, 7.25 kb
plsmid
35–40%
94–97%
1–2 ␮m
580–700 ␮m
PC:DOPE:DOTAP
various DOTAP molar
ratios
pRc/CMV HBS
5.6 kb
␮g/␮mol
Sonication (in the presence
of lysozyme) (41)
Sonication (42)
Asolectin (soybean
phospholipids)
EPC:Chol:lysine-DPPE
(55:30:15)
1.0 kb linear DNA 50%
0.08 ␮g/␮mol 100–200 nm
6.3 kb ssDNA
1.0 kb dsRNA
60–95%
13 ␮g/␮mol
ssDNA; 14
␮g/␮mol
dsRNA
100–150 nm
ssDNA
80–90%
dsRNA
46–52%
Spermidine-condensed
DNA, sonication,
extrusion (43)
EPC:Chol:PS (40:50:10)
EPC:Chol:EPA (40:50:10)
or EPC:Chol:CL
4.4 kb, 7.2 kb
plasmid
2.53–2.87
␮g/␮mol
400–500 nm
(50:40:10)
PS:Chol (50:50)
Ca2+-EDTA entrapment of
DNA-protein complexes
(44)
42.1 kbp
52–59%
22 ␮g/␮mol
ND
bacteriophage
Freeze-thaw extrusion (45)
POPC:DDAB (99:1)
3.4 kb linear
plasmid
17–50%
60–70%
ND
80–120 nm
SPLP-detergent dialysis (7)
DOPE:PEG-Cer:DODAC 4.4 to 10 kb
62.5 ␮g/␮mol 75 nm (QELS); 65 nm
(84:10:6)
plasmid
(after
(freeze-fracture)
removal of
empty
vesicles)
5 ␮g/␮mol
Ethanol-Ca2+ destabilization DOPC:DOPE (50:50)
pCMV-CAT 4.4
kb
pCMV-Luc 5.7 kb 70%
70–80%
160 nm (QELS)
(28)
(20:45:10:25)
DSPC:Chol:PEG-
CerC14:DODAP
(20:45:10:25)
Ethanol destabilization (9)
22.6 ␮g/␮mol ND
SPLP detergent dialysis (46) (PEG-DSG or
POD):DOTAP:DOPE
(20:11:69)
Various
pCMV-Gal,
pCMV-Luc,
pCMV-GFP
4.4 to 15 kb
plasmid
Up to 46% 20 ␮g/␮mol
53–60 nm (QELS)
SPLP-ethanol dilution (this
work)
80–95%
40–45 ␮g/
␮mol
100–150 nm (QELS)
ND, not determined.
a Some values calculated based on presented data.

Efficient Liposomal Encapsulation of Plasmid DNA
371
Stark, P. Scherrer, R. W. Graham, and P. R. Cullis. Stabilized
plasmid lipid particles for systemic gene therapy. Gene Ther. 7:
1867–1874 (2000).
9. N. Maurer, K. F. Wong, H. Stark, L. Louie, D. McIntosh, T.
Wong, P. Scherrer, S. C. Semple, and P. R. Cullis. Spontaneous
entrapment of polynucleotides upon electrostatic interaction with
ethanol-destabilized cationic liposomes. Biophys. J. 80:2310–2326
(2001).
10. D. B. Fenske, I. MacLachlan, and P. R. Cullis. Stabilized plasmid-
lipid particles: a systemic gene therapy vector. Methods Enzymol.
346:36–71 (2002).
11. I. M. Hafez, S. Ansell, and P. R. Cullis. Tunable pH-sensitive
liposomes composed of mixtures of cationic and anionic lipids.
Biophys. J. 79:1438–1446 (2000).
12. M. A. Monck, A. Mori, D. Lee, P. Tam, J. J. Wheeler, P. R.
Cullis, and P. Scherrer. Stabilized plasmid-lipid particles: phar-
macokinetics and plasmid delivery to distal tumors following in-
travenous injection. J. Drug Target. 7:439–452 (2000).
13. P. L. Felgner, T. R. Gadek, M. Holm, R. Roman, H. W. Chan, M.
Wenz, J. P. Northrop, G. M. Ringold, and M. Danielsen. Lipo-
fection—a highly efficient, lipid-mediated DNA-transfection pro-
cedure. Proc. Natl. Acad. Sci. U. S. A. 84:7413–7417 (1987).
14. H. C. Birnboim and J. Doly. A rapid alkaline extraction proce-
dure for screening recombinant plasmid DNA. Nucleic Acids Res.
7:1513–1523 (1979).
15. J. Sambrook, E. F. Fritsch, and T. Maniatis. In N. Ford, C. Nolan,
and M. Ferguson (eds.), Molecular Cloning, Cold Spring Harbor
Laboratory Press, New York, 1989, pp. 1.38–31.39.
16. S. Hirota, C. T. de Ilarduya, L. G. Barron, and F. C. J. Szoka.
Simple mixing device to reproducibly prepare cationic lipid-DNA
complexes (lipoplexes). Biotechniques 27:286–290 (1999).
17. C. H. Fiske and Y. Subbarow. The colorimetric determination of
phosphorus. J. Biol. Chem. 66:375–400 (1925).
18. E. G. Bligh and W. J. Dyer. A rapid method of total lipid extrac-
tion and purification. Can. J. Biochem. Physiol. 37:911–917
(1959).
19. A. L. Bailey and P. R. Cullis. Modulation of membrane fusion by
asymmetric transbilayer distributions of amino lipids. Biochem-
istry 33:12573–12580 (1994).
20. R. Thomas. The denaturation of DNA. Gene. 135:77–79 (1993).
21. T. Schlick, B. Li, and W. K. Olson. The influence of salt on the
structure and energetics of supercoiled DNA. Biophys. J. 67:
2146–2166 (1994).
22. L. D. Mayer, M. J. Hope, and P. R. Cullis. Vesicles of variable
sizes produced by a rapid extrusion procedure. Biochim. Biophys.
Acta 858:161–168 (1986).
23. B. L. Mui, P. R. Cullis, E. A. Evans, and T. D. Madden. Osmotic
properties of large unilamellar vesicles prepared by extrusion.
Biophys. J. 64:443–453 (1993).
24. M. Almgren, K. Edwards, and G. Karlsson. Cryo transmission
electron microscopy of liposomes and related structures. Colloids
and Surfaces A: Physiochemical and Engineering Aspects 174:3–
21 (2000).
25. B. L. Mui, H. G. Dobereiner, T. D. Madden, and P. R. Cullis.
Influence of transbilayer area asymmetry on the morphology of
large unilamellar vesicles. Biophys. J. 69:930–941 (1995).
26. E. Maurer-Spurej, K. F. Wong, N. Maurer, D. B. Fenske, and
P. R. Cullis. Factors influencing uptake and retention of amino-
containing drugs in large unilamellar vesicles exhibiting trans-
membrane pH gradients. Biochim. Biophys. Acta 1416:1–10
(1999).
27. M. B. Bally, M. J. Hope, C. J. A. Van Echteld, and P. R. Cullis.
Uptake of safranine and other lipophilic cations into model mem-
brane systems in response to a membrane potential. Biochim.
Biophys. Acta 812:66–76 (1985).
The simplicity of the stepwise dilution method to prepare
rapidly liposomes of desirable size and encapsulate DNA with
high efficiency is thought to result from the precise control of
the conditions that lipids enter the aqueous environment and
self arrange into lipid bilayer fragments and then liposomes.
A common feature of liposomes formed by SVF is the abun-
dance of bilamellar structures. Although we offer no expla-
nation for why this occurs, it appears that this morphology has
no effect on the stability or activity of SPLP in vivo. This
method may also be an attractive alternative to the prepara-
tion of liposomes by extrusion (22). The fact that liposomes
possessing transmembrane pH gradients can be prepared by
SVF, as demonstrated by safranin O uptake, indicates that
similarly prepared vesicles could be loaded with lipophilic
cationic drugs such as doxorubicin and vincristine (26,27).
Using this SVF method, we have prepared a 225 mg
DNA batch of SPLP from start (dissolution of lipids) to finish
(sterile filtration) in one working day. The properties of SPLP
manufactured at this scale were equivalent to those prepared
at 1/10 or 1/100 of this scale using smaller process steps. To
date, an appreciable hurdle preventing the development of
systemic nonviral gene delivery has been achieving predict-
able and controlled manufacture of stable systems. This
hurdle has been overcome: SPLP batches containing up to 5
g of DNA have been prepared using this approach and are
now under clinical evaluation.
ACKNOWLEDGMENTS
We dedicate this work to Cory Giesbrecht, 1972–2003,
our valued colleague and friend. We thank James Heyes for
preparing DODMA, DODAC, PEG-CerC20, and PEG-S-
DSG; Kevin McClintock for animal studies; Jay Petkau for
preparing the pCMVluc plasmid; and Catherine Simmons for
assisting with the preparation of the manuscript.
REFERENCES
1. I. MacLachlan, P. Cullis, and R. W. Graham. Progress towards a
synthetic virus for systemic gene therapy. Curr. Opin. Mol. Ther.
1:252–259 (1999).
2. A. Gabizon, H. Shmeeda, and Y. Barenholz. Pharmacokinetics of
pegylated liposomal doxorubicin: review of animal and human
studies. Clin. Pharmacokinet. 42:419–436 (2003).
3. A. H. Sarris, F. Hagemeister, J. Romaguera, M. A. Rodriguez, P.
McLaughlin, A. M. Tsimberidou, L. J. Medeiros, B. Samuels, O.
Pate, M. Oholendt, H. Kantarjian, C. Burge, and F. Cabanillas.
Liposomal vincristine in relapsed non-Hodgkin’s lymphomas:
early results of an ongoing phase II trial. Ann. Oncol. 11:69–72
(2000).
4. L. D. Mayer, R. Nayar, R. L. Thies, N. L. Boman, P. R. Cullis,
and M. B. Bally. Identification of vesicle properties that enhance
the antitumour activity of liposomal vincristine against murine
L1210 leukemia. Cancer Chemother. Pharmacol. 33:17–24 (1993).
5. N. L. Boman, M. B. Bally, P. R. Cullis, L. D. Mayer, and M. S.
Webb. Encapsulation of vincristine in liposomes reduces its tox-
icity and improves its anti-tumor efficacy. J. Liposome Res. 5:
523–541 (1995).
6. M. S. Webb, T. O. Harasym, D. Masin, M. B. Bally, and L. D.
Mayer. Sphingomyelin-cholesterol liposomes significantly en-
hance the pharmacokinetic and therapeutic properties of vincris-
tine in murine and human tumour models. Br. J. Cancer 72:896–
904 (1995).
7. J. J. Wheeler, L. Palmer, M. Ossanlou, I. MacLachlan, R. W.
Graham, Y. P. Zhang, M. J. Hope, P. Scherrer, and P. R. Cullis.
Stabilized plasmid-lipid particles: construction and characteriza-
tion. Gene Ther. 6:271–281 (1999).
28. A. L. Bailey and S. M. Sullivan. Efficient encapsulation of DNA
plasmids in small neutral liposomes induced by ethanol and cal-
cium. Biochim. Biophys. Acta 1468:239–252 (2000).
29. Y. Perrie and G. Gregoriadis. Liposome-entrapped plasmid
DNA: characterisation studies. Biochim. Biophys. Acta 1475:125–
132 (2000).
30. R. Fraley, S. Subramani, P. Berg, and D. Papahadjopoulos. In-
troduction of liposome-encapsulated SV-40 DNA into cells. J.
Biol. Chem. 255:10431–10435 (1980).
8. P. Tam, M. Monck, D. Lee, O. Ludkovski, E. Leng, K. Clow, H.

372
Jeffs et al.
31. P. Soriano, J. Dijkstra, A. Legrand, H. Spanjer, D. Londos-
Gagliardi, F. Roerdink, G. Scherphof, and C. Nicolau. Targeted
their interactions with plant protoplasts. Nucleic Acids Res. 6:
3773–3784 (1979).
and non-targeted liposomes for in vivo transfer to rat liver cells of 39. S. F. Alino, M. Bobadilla, M. Garcia-Sanz, M. Lejarreta, F. Unda,
plasmid containing the preproinsulin I gene. Proc. Natl. Acad.
Sci. USA 80:7128–7131 (1983).
and E. Hilario. In vivo delivery of human alpha 1-antitrypsin
gene to mouse hepatocytes by liposomes. Biochim. Biophys. Res.
Comm. 192:174–181 (1993).
32. M. Nakanishi, T. Uchida, H. Sugawa, M. Ishiura, and Y. Okada.
Efficient introduction of contents of liposomes into cells using 40. M. Baru, J. H. Axelrod, and I. Nur. Liposome-encapsulated
HVJ (sendai virus). Exp. Cell Res. 159:399–409 (1985).
DNA-mediated gene transfer and synthesis of human factor IX in
mice. Gene. 161:143–150 (1995).
33. A. Cudd and C. Nicolau. Intracellular fate of liposome encapsu-
lated DNA in mouse liver: analysis using electron microscope 41. D. G. Jay and W. Gilbert. Basic protein enhances the incorpora-
autoradiography and subcellular fractionation. Biochim. Biophys.
Acta 845:477–491 (1985).
tion of DNA into lipid vesicles: model for the formation of pri-
mordial cells. Proc. Natl. Acad. Sci. USA 84:1978–1980 (1987).
34. R. T. Fraley, C. S. Fornari, and S. Kaplan. Entrapment of a 42. C. Puyal, P. Milhaud, A. Bienvenue, and J. R. Philippot. A new
bacterial plasmid in phospholipid vesicles: potential for gene
therapy. Proc. Natl. Acad. Sci. USA 76:3348–3352 (1979).
35. C. Nicolau and S. Rottem. Expression of beta-lactamase activity
in Mycoplasma carpicolum transfected with the liposome-
encapsulated E. coli pBR32 plasmid. Biochim. Biophys. Res.
Comm. 108:982–986 (1982).
cationic liposome encapsulating genetic material. A potential de-
livery system for polynucleotides. Eur. J. Biochem. 228:697–703
(1995).
43. M. Ibanez, P. Gariglio, P. Chavez, R. Santiago, C. Wong, and I.
Baeza. Spermidine-condensed DNA and cone-shaped lipids im-
prove delivery and expression of exogenous DNA transfer by
liposomes. Biochem. Cell Biol. 74:633–643 (1996).
36. J. C. Stavridis, G. Deliconstantinos, M. C. Psallidopoulos, N. A.
Armenakas, D. J. Hadjiminas, and J. Hadjiminas. Construc- 44. J. Szelei and E. Duda. Entrapment of high molecular mass DNA
tion of transferrin-coated liposomes for in vivo transport of ex-
ogenous DNA to bone marrow erythroblasts in rabbits. Exp. Cell
Res. 164:568–572 (1986).
molecules in liposomes for the genetic transformation of animal
cells. Biochem. J. 259:549–553 (1989).
45. P. A. Monnard, T. Oberholzer, and P. L. Luisi. Entrapment of
nucleic acids in liposomes. Biochim. Biophys. Acta-Biomembr.
1329:39–50 (1997).
46. J. S. Choi, J. A. MacKay, and F. C. Szoka. Low-pH-sensitive
PEG-stabilized plasmid-lipid nanoparticles: Preparation and
characterization. Bioconjugate Chem. 14:420–429 (2003).
37. C. Y. Wang and L. Huang. pH-sensitive immunoliposomes me-
diate target cell-specific delivery and controlled expression of a
foreign gene in mouse. Proc. Natl. Acad. Sci. USA. 84:7851–7855
(1987).
38. P. F. Lurquin. Entrapment of plasmid DNA by liposomes and