MacKay et al. 2008 supplemental
1
Supplemental materials and methods
1.1
Materials
1-palmitoyl,
2-oleoyl-sn-glycero-3-phosphocholine
(POPC),
1,2-dioleoyl-sn-glycero-3-
phosphoethanolamine (DOPE), 1,2-dipalmitoyl-3-trimethylammonium-propane (DPTAP), 1,2dioleoyl-3-trimethylammonium-propane
phosphoglycerol-N-(polyethyleneglycol)2
(DOTAP),
kD
1,2-distearoyl-sn-glycero-3-
(PEG-DSPE),
1,2-dioleoyl-sn-Glycero-3-
phosphoethanolamine-N-(carboxyfluorescein) (CF-PE), and cholesterol (Chol) were obtained
from
Avanti
Polar
Lipids
dioctadecyloxacarbocyanine
tetramethylindocarbocyanine
(Alabaster,
perchlorate
perchlorate
AL).
Fluorescent
(DiO),
(DiI),
and
lipid
tracers
3,3’-
1,1’-dioctadecyl-3,3,3',3'1,1’-dioctadecyl-3,3,3',3'-
tetramethylindodicarbocyanine perchlorate (DiD) were purchased from Molecular Probes
(Eugene, OR). The pH-sensitive PEG-lipid (POD) was synthesized in our lab as described
elsewhere1(Fig. S1a). A reduction-sensitive cationic thiocholesterol lipid (TCL-6) was also
synthesized in our lab as described elsewhere2(Fig. S1b). Green fluorescent protein (GFP) and
luciferase encoding plasmids were generous gifts from Valentis (Burlingame, CA).
1.2
Synthesis and purification of TATp and TATp lipids
The HIV TAT peptide (TATp), GRKKRRQRRRGYG, was synthesized by solid-phase FMOC
chemistry as described by Wyman et al.3. ~300 mg of resin containing either the peptide or
peptide conjugates were cleaved from the resin by stirring 4 hrs at room temperature in a 10 mL
solution of 95 % trifluoroacetic acid (TFA), 2.5 % distilled deionized water, and 2.5 %
triisopropylsilane. Resin was removed by filtration through a bed of glass fibers, and deprotected
material was precipitated into 80 mL of cold ether on ice. The precipitate was centrifuged and
30
MacKay et al. 2008 supplemental
the supernatant was discarded. The pellet was dried and hydrated in 4 mL of distilled deionized
water. The solution was frozen and water was removed under high vacuum. Samples were
purified by high-pressure liquid chromatography (HPLC) (Dionex Sunnyvale, CA). The
unmodified peptide was purified using a semi-preparative scale C18 column at 10 mL /min with
a 100 min gradient from 0 to 100 % acetonitrile with 0.1 vol % TFA. The unmodified peptide
eluted at 13 % acetonitrile (ACN) and the mass was confirmed by matrix assisted laser
desorption ionization (MALDI) mass spectroscopy ([M+H]+expected: 1,673.0 g/mol; [M+H]+observed:
1,673.0 g/mol). After the peptide was synthesized, lipids anchors were conjugated to the
deprotected amino terminus. Lipid conjugation was performed while the peptide remained on
solid support, prior to deprotecting the peptide sidechains. This prevented the modification of
primary amines on the lysine residues. After reaction, the peptide-lipid conjugates were sidechain deprotected, cleaved from the solid support, purified using HPLC, and confirmed using
mass spectrometry.
The TATp-lipid (Fig. 1a) conjugate was prepared using 1,2 dipalmitoyl-sn-glycero-3 succinate
(Avanti; Alabaster, AL). 50 moles of lipid were dissolved in 4 mL dichloromethane (DCM):
7.5 mL n-methyl-2-pyrrolidone (NMP) in the presence of 400 moles of n-methyl-morpholine
and
120
moles
of
hexafluorophosphate(1-),3-oxide
1-H-Benzotriazolium,
1-[bis(dimethylamino)methylene]-
O-(Benzotriazol-1-yl)-N,N,N',N'
tetramethyluronium
hexafluorophosphate (HBTU). 60 moles of protected TATp on the resin were added to the
reaction cocktail and stirred overnight at room temperature. The reaction was stopped by
washing the resin with 20 mL of DCM over a bed of glass fibers. The resin was recovered, dried,
and TATp-lipid was deprotected as described above. TATp-lipid was dissolved in
31
MacKay et al. 2008 supplemental
water:ACN:TFA [33:66:0.1] and purified on a semipreparative C4 column at 4 mL/min using a
20 min gradient from 45 to 65 % ACN in the presence of 0.1 vol % TFA. The TATp-lipid peak
eluted at about 50 % ACN and was identified using MALDI ([M+H]+expected: 2,323.5 g/mol;
[M+H]+observed: 2,323.1 g/mol).
The TATp-PEG-lipid (Fig. 1b) conjugate was prepared using 1,2, dipalmitoyl-sn-glycero-3phosphoethanolamine -N- amino(polyethylene glycol)2 kD (Avanti; Alabaster, AL) in a 2 step
reaction. For reaction 1, in 4 mL of NMP, the unprotected amino-terminal end of the peptide (60
moles) was reacted for 4 hrs at room temperature under stirring with excess 180 moles of
disuccinimidyl suberate (Pierce, Rockford, IL, USA) in the presence of 160 moles of n-methyl
morpholine. The reaction was stopped by washing with 20 mL of NMP over a bed of glass fibers.
For reaction 2, the resin was stirred with 72 moles of lipid dissolved in 4 mL NMP in the
presence of 360 moles of n-methyl morpholine. The reaction proceeded for 4 hrs at room
temperature and was stopped by washing the resin with 20 mL of DCM over a bed of glass fibers.
The resin was recovered, dried, and TATp-PEG-lipid was deprotected as described above.
TATp-PEG-lipid was dissolved in water:ACN:TFA [50:50:0.1] and then purified on a
semipreparative C4 column at 4 mL/min using a 10 min gradient from 50 to 70 % ACN in the
presence of 0.1 vol % TFA. The TATp-PEG-lipid peak eluted at about 60 % ACN. Characterized
using MALDI, peaks with an average spacing of 44.2 g/mol were observed. For the species in
Fig. 1b with (CH2CH2O)n=34 to 48 , the peaks with greater than 50 % of the maximal intensity
corresponded to the molecular mass expected ([M+H]+expected: 4,046, 4,090, 4,134, 4,178, 4,222,
4,266, 4,310, 4,354, 4,398, 4,442, 4,486, 4,530, 4,574, 4,618, 4,662 g/mol; [M+H]+observed: 4,046,
4,090, 4,133, 4,178, 4,222, 4,267, 4,310, 4,354, 4,400, 4,444, 4,487, 4,531, 4,576, 4,620, 4,664
32
MacKay et al. 2008 supplemental
g/mol). Obtaining a MALDI spectra from the PEG lipid series was enabled by first dissolving the
TATp-PEG-lipid in water:ACN:TFA [50:50:0.1] and then mixing this 1:1 with ACN saturated
with sinapinic acid.
1.3
Preparation and characterization of liposomes
For the preparation of liposomes, 5-10 moles of total lipid was mixed in chloroform and dried
in a glass test tube by rotary evaporation under reduced pressure. The lipid film was placed under
high vacuum for 4 hrs to remove residual solvent. Liposomes were formed either in pH 7.5
Hepes buffered saline (145 mM NaCl, 5 mM Hepes) for the radiological liposome uptake studies
or in pH 8.5 Tris buffered saline (150 mM NaCl, 5 mM Tris/Acetate) for the CED and NLP
studies. The lipid (10 mM) was hydrated at 20ºC with intermittent vortexing. Liposomes were
formed by extrusion (11x) through 200, 100, or 80 nm pores in polycarbonate membranes.
Particle size and zeta potential were determined using a Malvern (Southborough, MA) Zeta3000
Dynamic Light Scattering Instrument and are presented in Table S1.
1.4
In vitro uptake of TATp modified radiolabeled liposomes
To quantitatively determine the kinetic uptake of TATp modified liposomes in cell culture, we
labeled liposomes with a headgroup iodinated benzamidine phospholipid (125I-BPE), as
described previously 4. Radiolabeled lipid (less than 0.01% mole percent of the total lipid) were
deposited as a film with other lipid components from a chloroform solution before hydration.
Gamma radiation activity was determined using a Perkin Elmer Wizard Gamma Counter (Boston,
MA). These experiments were performed with B16F10 cells in a 96 well assay as follows. Cells
were trypsinized and seeded into plates at 0.5 – 1 * 104 cells per well per 100 L Modified Eagle
33
MacKay et al. 2008 supplemental
Media (UCSF Cell Culture Facility) supplemented with 10% Fetal Bovine Serum (FBS) (UCSF
Cell Culture Facility). Cells were allowed to grow for 24 hrs prior to the assay. At the time of
assay, 3 wells were trypsized into 50 L and the cell number per well was counted, typically
containing 2 * 104 cells per well at the time of assay.
During the assay, liquid manipulations were performed gently using a multichannel pipettor. At
the start of each assay (t=0), the cells were washed once with 100 L of media. During the
course of the assay, this wash was removed and replaced with 100 L of media containing
radiolabeled liposomes (60, 90, 105, 120 min). The media used for the washing and binding
assay was FBS free MEM supplemented with 10 mM Hepes buffer, pH 7.4. All cells were
incubated with liposomes in a 37C incubator for 2 hrs under gentle orbital shaking. At 120 min,
all cell incubations (15, 30, 60, 120 min) were washed simultaneously to recover radioactive
counts. For every well, 3 fractions were obtained. Fraction 1 included the supernatant and
unbound liposomes that washed off with 2 changes of PBS. Fraction 2 included surface bound
liposomes that were washed off by 3 changes of 1.6 M NaCl PBS. Fraction 3 included cellinternalized liposomes obtained by solubilizing the cells in 0.5M NaOH. All 3 fractions were
measured on a gamma counter. Each data point was performed in triplicate and the average
recovery of radioactivity was typically 85%.
1.5
Formation of bioresponsive NLP
Bioresponsive NLP (Fig. 1c) were prepared as described by a modification of the method of
Huang et al.2. All NLP in these studies were formed using TCL-6 cationic lipid. Compound
TCL-6 is a thiocholesterol cationic lipid that can release DNA following exposure to the
34
MacKay et al. 2008 supplemental
cytosolic reducing environment. Briefly, the lipid components were dissolved in chloroform,
mixed, and dried in a glass vial. pH-insensitive NLP were prepared from a mixture of PEGDSPE:TCL-6:DOPE:DiD [10:50:40:0.1] and pH-sensitive NLP were prepared from POD:TCL6:DOPE:DiD [20:40:40:0.1]. 20 moles of total lipid were dissolved into 1.5 mL ethanol:buffer
[1:1] (5 mM Tris/acetate, pH 8.5) and mixed with 1 mg DNA in 1.5 mL ethanol:buffer [1:1] (5
mM Tris/acetate, pH 8.5). About half of the lipid components were positively charged (TCL-6),
and these associate with the negatively charged phosphate DNA backbone at a charge ratio of
3:1 TCL-6 nitrogen to DNA phosphate. Ethanol was dialyzed away at 20 ºC through a 10,000
MWCO membrane, allowing formation of lipid bilayers around the DNA. Particle formation was
confirmed by dynamic light scattering. For Fig. 3c-h, NLP were incubated with 10 fold molar
excess glutathione after formation to reduce non-specific adsorption to cells2. Immediately prior
to CED, the salt concentration was adjusted to 150 mM NaCl. DNA concentration and protection
was confirmed using the picogreen assay as described elsewhere5.
1.6
Incorporation of TATp into NLP
TATp linked to a lipid anchor via a PEG (Mw=2 kD) can easily be incorporated into liposomes
or NLP after particle formation using a micelle transfer method 6. Before micelle transfer,
unencapsulated DNA was removed from the particle solution using anion exchange
chromatography with DEAE-SepharoseTM CL6B (Sigma; St. Louis, MO) as described
previously5. This is necessary to prevent electrostatic binding between positively charged TATp
and surface bound DNA. TATp-PEG-lipid was dried in a glass tube under high vacuum and then
rehydrated into micelles at 100 M in 5 mM Tris/acetate pH 8.5. Next, an appropriate volume of
TATp micelle solution was added such that 20 nanomoles of TATp were added per mol of
35
MacKay et al. 2008 supplemental
DNA measured by picogreen, giving approximately 0.5 mol % TATp by mol lipid. After mixing,
the micelles were incubated for 5 hours at 20C. TATp incorporation was monitored by cation
exchange chromatography. For this assay, 100 L NLP at a DNA concentration of about 100
ng/L were washed through a CM-25 SephadexTM (Sigma; St. Louis, MO) bed (1 x 5 cm)
equilibrated in 150 mM NaCl Tris buffer pH 8.5. Unmodified NLP flow through the column in
the void volume. TATp modified NLP bind to the column and can be eluted with the addition of
1.6 M NaCl Tris buffer. DNA concentration and liposome turbidity were observed using
absorbance at 260 nm. This cation exchange method provides quantitative information regarding
what fraction of NLP/DNA becomes significantly modified with TATp-PEG-lipid. This has
helped to overcome two problems: 1) dry PEG-lipids do not mix quickly enough to
modify >95% of the NLP and 2) excess unencapsulated DNA binds to the TATp, preventing
interaction with the column and enhancement of in vitro transfection. This simple
chromatography step enabled us to find and overcome both of these problems, and now after our
TATp micelle transfer >95% of the A260 binds to the column and elutes in the high salt wash.
Improved quantification can be achieved by elution with a salt gradient
1.7
Intracranial implantation of U87-MG tumors in athymic rats
Athymic rats were purchased from the National Cancer Institute and cared for at the UCSF
Laboratory Animal Resource Center. We adhered to protocols recommended by the National
Institute of Health Guide for the Care and Use of Laboratory Animals and as approved by the
UCSF Institutional Animal Care and Use Committee. To minimize pain and discomfort, animals
were sedated with an intra peritoneal injection of Ketamine (60 mg / kg body weight) and
Xylazine (7.5 mg / kg body weight). After anesthesia was attained, an incision was made in the
36
MacKay et al. 2008 supplemental
scalp above the bregma and the animal was mounted in a stereotactic frame (Stoelting, Wood
Dale IL). A hole was drilled in the skull 3 mm right of the bregma and a guide screw was
installed. Subconfluent U87-MG cells were trypsinized on the day of implantation, washed once
in Hanks Balanced Salt Solution (UCSF Cell Culture Facility), stored on ice, and centrifuged to
obtain a cell pellet, which was then injected into the brain. Through the guide screw, at a depth
of 5 mm, a Hamilton syringe was used to inject 10 L of cells of a U87-MG cell pellet over a 1
min period. The stereotactic coordinates of injection were centered within the caudate putamen 7.
To relieve pain following surgery, animals were given a subcutaneous injection of
Buprenorphine (0.05 mg / kg body weight).
1.8
Convection enhanced delivery
Convection enhanced delivery (peak flow rate of 0.8 L / min) of liposomes and NLP were
performed into both normal rat brains and U87-MG intracranial tumors respectively. NLP
infusions were performed at day 19-21 post implantation when the tumor was approximately
100-300 mg. Prior to infusion, liposomes were diluted to a concentration of 1 mM lipid and NLP
were diluted to a concentration of 4 mM lipid in Tris buffered saline, pH 8.5. Animals were
sedated and mounted on a stereotactic frame. To prevent backflow at this flow rate,
8
a narrow
cannula was prepared from fused silica tubing with an outer diameter of 0.16 mm (Polymicro
Technologies, Phoenix, AZ), extending 3 mm from the tip of a 24 gauge needle used for support
9
. This cannula was inserted into the same stereotactic coordinates used for tumor implantation
except the cannula was inserted only 4 mm rather than 5 mm. This was done to insure that the
cannula was positioned within the tumor mass rather than at the edge of or below the tumor. A
syringe pump was used to infuse solutions into the brain at an increasing flow rate as follows: 0.1
37
MacKay et al. 2008 supplemental
L /min for 5 min, 0.2 L /min for 5 min, 0.5 L /min for 5 min, and 0.8 L /min for 30 min for
a total volume of 28 L infused over 45 min 10,11. The infusion cannula was withdrawn over a 5
min period following cessation of infusion. Post infusion, animals were sacrificed at 1 or 24 hrs.
After sedation, animals were prepared by perfusion fixation with 200 mL of phosphate buffered
saline PBS (100 mM phosphate, 150 mM NaCl, pH 7.4) and 200 mL of 4 % paraformaldehyde
in PBS. Brains were removed and post-fixed overnight in 4 % paraformaldehyde at 4ºC.
1.9
Cell culture luciferase detection
The in vitro luciferase assay was performed using U87-MG cells. 50,000 cells were seeded per 1
mL per well in 24-well plates and grown in modified eagle media supplemented with 10 % fetal
bovine serum, 1x non-essential amino acids, and 0.1 mg/mL sodium pyruvate, 100 units/ml
penicillin, and 0.1 mg/ml streptomycin. (UCSF Cell Culture Facility; San Francisco, CA). After
24 hours, the media was replaced with 1 mL of the same media containing 1 g luciferase
plasmid / well. As a positive control, PEI and DNA were complexed with a charge ratio of 7:1
PEI nitrogen to DNA phosphate, 30 minutes prior to addition to the cells. 4 hours later the media
was removed and replaced with fresh media. After 20 hours, the cells were washed with 1 mL of
PBS per well and then lysed in 0.2 mL of Reporter Lysis Buffer (Promega; Madison, WI) for 30
minutes. The lysate was obtained and centrifuged at 14,000 rpm for 5 minutes and luciferase was
immediately assayed by injecting 10 L of the lysate into 50 L rehydrated Luciferase Substrate
(Promega E1501; Madison, WI). The sample was immediately read using an Optocomp I
luminometer (MGM Inc.; Hamden, CT) for 10 seconds. The Micro BCA protein assay kit
(Pierce; Rockford, IL) was used to determine total protein.
38
MacKay et al. 2008 supplemental
1.10
Intracranial tumor transfection
Luciferase assay was performed on fresh tissue obtained 1 day post-infusion of NLP and 19 days
post-implantation of U87-MG cells. To prevent degradation, samples were processed
immediately after sacrifice and kept on ice. The right and left lobes of the brain were placed
separately in vials containing about 0.5 mL of 2.4 mm zirconia beads (Biospec; Bartlesville, OK )
and their tissue mass was recorded. The average mass of tissue obtained per sample was 0.67 
0.08 grams (n=16). The tumor is contained in the right lobe of the brain. To minimize bubble
formation, the vial was filled completely with Reporter Lysis Buffer (Promega; Madison, WI)
and the sealed vials were homogenized on a Mini Bead Beater (Biospec; Bartlesville, OK) at
5,000 beats for 1 minute. Samples were cooled on ice before and after homogenization. Samples
were centrifuged for 5 minutes at 10,000 rpm to remove debris. 10 L of the supernatant was
injected into 50 L of the rehydrated luciferase assay substrate and the emitted light was read
immediately. To determine the total amount of luciferase in the tumor a standard curve was
prepared using dilutions of commercially available luciferase (Promega; Madison, WI). Standard
curves were prepared in a tissue homogenate from an untreated animal. Additionally, to confirm
that this homogenization procedure does not destroy luciferase activity, luciferase was spiked
into a tissue sample prior to homogenization. After homogenization greater than 75 % of the
activity remained. Also infused as a control, a conventional cationic lipoplex was prepared by
mixing DOTAP:Chol [6:4] with DNA at a charge ratio of 6:1 DOTAP nitrogen to DNA
phosphate to form stable aggregates and used on the same day of formation.
1.11
Luciferase immunohistochemistry
39
MacKay et al. 2008 supplemental
Luciferase immunohistochemistry was performed on paraformaldehyde fixed rat brains isolated
1-day post-infusion of NLP and 19 days post-implantation of U87-MG tumors. Staining
conditions were worked out using a ‘positive control’ fixed cell pellet of U87-MG transfected
with 50 g of luciferase DNA per 20 mL per T-150 flask complexed with DOTAP:DOPE [1:1]
mixed at charge ratio of 4:1 DOTAP nitrogen to DNA phosphate. The pellet was mixed with
untransfected cells, fixed, and suspended in HystogelTM (Richard-Allan Sci; Kalamazoo, MI).
Both pellet and brain were similarly embedded in paraffin and cut into 4 um sections. Heat
induced epitope retrieval was performed in citrate pH 6.0 for 10 minutes, 3 % hydrogen peroxide
for 20 min. Sections were blocked with casein for 20 minutes and primed with the FireflyLuciferase Goat Polyclonal Antibody Horseradish Peroxidase Conjugate (Abcam #AB635-100;
Cambridge, MA) for 2 hours at a dilution of 1:1500. DAKO DD+ diluent was used to reduce
background. Sections were stained with HRP Chromagen for 5 minutes (Abcam #AB8183;
Cambridge, MA). Hematoxylin staining and a two minute acid wash were performed before
mounting with Faramount aqueous mounting media (Dako; Carpinteria, CA).
1.12
Macroscopic imaging
The entire brain was sliced into 1 mm axial sections and imaged macroscopically at a resolution
of 10 pixels per mm using a macroscopic fluorescence scanner, the Molecular Dynamics
StormTM (Amersham Biosciences; Piscataway, NJ) for DiD far red fluorescence (excitation=640
nm; emission > 650 nm). 16 bit images were saved as TIFF files for analysis. For illustrative
purposes, the images have been processed to show 3 edges. Edges were created using the
threshold function and edge-detect function in Image J version 1.34i (http://rsb.info.nih.gov/ij/).
To define the brain background, a 1 mm by 1 mm control box was drawn on the caudate
40
MacKay et al. 2008 supplemental
putamen of the un-infused left side of the brain. The average and standard deviation of the
control box were calculated, and this average was defined as the background intensity. The
outermost edge (black) was thresholded to show the outer edge of the brain slice. The middle
edge (red) was thresholded at a pixel intensity equal to the brain background plus 5 times the
standard deviation, defined as the outer limit of detection. The inner edge (yellow) was
thresholded at the half-maximal intensity between the background and the maximum pixel
intensity for the section. Thus, the red edge gives the farthest detectable signal from particles,
and the yellow edge illustrates the area where half of the fluorescence intensity had dissipated.
The sections with the highest fluorescence intensity occur near the tip of the infusion cannula,
and these areas were imaged using confocal scanning laser microscopy.
1.13
Fluorescence microscopy
Confocal images were obtained to observe the in vitro cellular distribution of liposome uptake
and the in vivo distribution of liposomes in brain sections. All images were taken using
LaserSharpe Software on a Biorad 1024 Confocal Scanning Laser Microscope (Hercules, CA)
mounted on a Nikon Diaphot 200 microscope. The DiI signal was acquired in photomultiplier 1
with excitation from the 568 nm line of a krypton-argon laser. The DiO signal was acquired in
photomultiplier 2 with excitation from the 488 nm line of a krypton-argon laser. DiD signals
were acquired in photomultiplier 3 with excitation from the 633 nm line of a helium-neon laser.
Brain sections were imaged using a 4x objective lens at a resolution of 2.5 m per pixel. Within
the figures, GFP or DiO are depicted in green color, DiI are depicted in red color, and DiD is
depicted in blue color.
41
MacKay et al. 2008 supplemental
In Fig. 3, cells were viewed with epifluorescence microscopy performed on a Nikon Eclipse
TS1000 fluorescence microscope with a 20x objective, and images were recorded using SPOT
version 3.5 (Diagnostic Instruments Inc.; Kanagawa, Japan). DiD signal was imaged using a
532-587 nm excitation filter, a 608-683 nm emission filter, and a 10 sec exposure. GFP signal
was imaged using a 460-500 nm excitation filter, a 510-560 nm emission filter, and a 10 sec
exposure.
1.14
Data analysis for co-infused liposomes
Comparisons of radii of penetration were made as described previously
12
by quantifying the
distribution radius for 3 co-infused liposomes. Briefly, three liposomes, one of which serves as a
standard control, were labeled with either DiO, DiI, or DiD. Images were obtained using
sequential excitation/detection of each fluorophore to prevent crossover between dyes. A radially
directed profile in a region extending from the tip of the cannula outward was extracted from the
confocal microscopy using Image J. Profiles were obtained in regions of convection through gray
matter where staining was uninterrupted by anatomical features. In all cases, the profiles were
obtained by taking a box of 0.5 mm in height and extending 2-6 mm from the center of the
infusion. This RGB image was separated into both red, green, and blue 8 bit TIFFs. For each
color box, a plot profile was obtained and the raw data were saved to an Excel TM spreadsheet for
further analysis. To characterize the extent of particle penetration, the radius of 90% area under
the curve (AUC), r90% AUC, was calculated as described previously
12
. This radius estimates the
distance where 90% of the fluorescent intensity has been deposited. A radius was calculated for
each profile by first subtracting the background signal in unstained gray matter, and then
integrating from r=0 outward using the trapezoidal rule to obtain the AUC as a function of radius.
42
MacKay et al. 2008 supplemental
The maximum AUC near the end of the plot was used to normalize the profile. The radius where
the normalized AUC was 90% was then picked from the curve.
2.
Supplemental results
2.1.
In vitro uptake of TATp modified fluorescent liposomes
To characterize these two ligands, both lipid conjugates (~ 0.3 mol % TATp) were dried with
POPC:Chol [6:4], hydrated in aqueous buffer, and extruded through 200 nm pores to form
liposomes. TATp modified liposomes have a zeta potential similar to that for liposomes with a
2.4 mol % cationic lipid (Table S1). Both lipid conjugates were then characterized for extent of
binding and internalization. Using confocal scanning laser microscopy, we observed that TATp
modified liposomes bind strongly to the cell surface and internalize into vesicular compartments
within the cell within 1 hour (Fig. S2a). The strong degree of surface binding made
determination of internalization difficult; however, much of the cell surface fluorescence was
removable by high salt washes (1.0 M NaCl) (Fig. S2b).
2.2
In vitro uptake of TATp modified radiolabeled liposomes
Liposomes were prepared with and without the TATp ligands by extrusion to a size of about 200
nm (Table S1). Because TATp has positive charge (8 per peptide), it was important to include a
control liposome with a conventional cationic lipid, DPTAP. TATp liposomes were prepared at a
mole percent ligand of 0.3% and have about 2.4% by mole positive charge, so 2.4% by mole
DPTAP liposome were prepared as a control for charge induced liposome uptake. Both
liposomes with 2.4% DPTAP and 0.3% TATp ligand were confirmed to have a similar amount
of positive surface charge using zeta potential measurements (Table S1). Both the DPTAP
43
MacKay et al. 2008 supplemental
control (2.2  0.4 mV) and the TATp-lipid liposomes (1.9  0.8 mV) had detectably higher
surface charge than did neutral POPC liposomes (-5.7  0.6 mV), but neither formulation was as
high as a 10% DOTAP liposome (31.9  0.8 mV). We concluded that the 2.4% DPTAP
liposomes have a similar surface charge to the 0.3% TATp-lipid liposomes and proceeded to use
these as controls for the kinetic uptake of radiolabeled liposomes.
2.3
Validation of high salt washing
TATp-lipid successfully induced cell surface binding and internalization in cell culture (Fig. S2).
High salt washes were used to displace the electrostatically bound TATp from cell surface
proteoglycans (Fig. S2,S3)13. TATp ligands appear to bind to a variety of cell types via cell
surface proteoglycans, and binding is not thought to be cell type dependent14. B16F10 cells are
of mouse melanoma origin; however, due to the non-specificity of TATp ligands, trends in
binding kinetics are assumed to be representative of many cell lines, including human brain
tumor cells. High salt PBS was sufficient to displace much, but not the all, of the cell associated
radiolabel; this removable fraction is defined as “surface bound” liposomes. There was a
significant amount of internalized radiolabel that could not be removed by high salt washes.
Before proceeding with the assay (Fig. S3), the number of washes needed to remove all surface
bound liposomes was validated (Fig. S3a). B16F10 cells were incubated for 1 hr with 0.3%
TATp-lipid or 2.4% DPTAP liposomes. At the end of the incubation period, radioactive
liposomes and media were removed. Wells were washed 0 to 4 times with 1.6 M NaCl PBS,
dissolved, and counted for radioactivity. The percent of counts per well that remained in the cells
is plotted (Fig. S3a), showing that 10% of the TATp modified liposomes bound to cells, but only
44
MacKay et al. 2008 supplemental
1% of the dose was actually internalized. It was observed that after 3 high salt washes, no
additional radiolabel from either formulation was removed from the cell bound material. Thus, 3
high salt washes are sufficient to distinguish between surface bound and internalized liposomes.
We developed this stringent salt (NaCl) washing technique to determine the extent of
internalization of electrostatically mediated liposome binding, such as the TATp interaction with
negatively charged cell surface proteoglycans. High salt washes destabilize cationic/anionic
polymer interactions and enable separation of surface associated liposomes from strongly
associated liposomes, which appear internal to the cell (Fig. S2c). We also performed the
experiments at 4 C (data not shown); however, the 2.4% DPTAP positive control liposomes
bound and internalized significantly more at 4C (15% bound, 1 hr) than at 37C (0.8 % bound,
1 hr) and the neutral controls did not have nearly the same effect 4C (2 % bound, 1 hr) versus
37C (0.5 % bound, 1 hr). For conventional ligands, the amount internalized is defined as the
cell-associated percentage at 37 C minus the percentage at 4C; however, this approach gives a
negative number for uptake for our 2.4% DPTAP liposome. We attribute this artifact to the
properties of the fluid lipid bilayer, which may form temperature dependent rafts that
significantly alter binding. Therefore, we did not to pursue temperature shift methods to study
TATp liposome uptake.
2.3
Liposome binding is dependent upon method of chemical attachment
Two TATp lipid conjugates were prepared. One ligand consisted of a TATp directly linked to a
lipid anchor on the surface of the liposome, TATp-lipid (Fig. 1a). The second ligand was
prepared by linking the TATp peptide to the end of a flexible 2 kD PEG linked to a phospholipid
45
MacKay et al. 2008 supplemental
anchor (Fig. 1b). A range of lipid concentrations was observed from 3 to 100 M. At every
concentration, all four liposome compositions maintain the same relationship for internalization
(Fig. S3b). This range of concentration does not saturate the binding sites on the cell surface or
the internalization process into the cell for the 0.3% TATp-PEG-lipid. All subsequent kinetic
binding studies were performed at a 30 M lipid concentration.
At 37 C, both ligands had rapid binding kinetics followed by quick internalization (Fig. S3c)
and the amount of cell surface lipid quickly reached a steady state (Fig. S3c). At a density of
0.3% by mole, surface density 0.005 ligands / nm2, the TATp-PEG-lipid performed the best, with
24  1% of liposomes internalized after a 120 min incubation. This uptake did not saturate under
the period of time observed. The 0.3% TATp-lipid liposomes internalized only 6  1% of the
liposomes in the cells. Additionally, this internalization reached a plateau at 60 min, suggesting
that TATp-lipid uptake saturated the cellular capacity for uptake, which did not regenerate in this
period. Both the neutral control liposome and the 2.4% DPTAP control liposome showed less
than 0.5% uptake over the course of this assay. After 2 hours, all four liposome formulations
were significantly different when compared by ANOVA of the Log10(percent internalized)
(F3,8=508.6, R2=0.995, p=1.8 x 10-9) and Tukey HSD (p<0.01 for each individual comparison).
We conclude that the TATp-PEG-lipid is a suitable ligand for inducing sustainable uptake.
2.4
Effect of TATp surface density on liposome binding
Liposomes were prepared with surface densities of the TATp-lipid ranging from 0.03 to 1.0%
(Table S1). As expected, the extent and rate of internalization (Fig. S3e) was dependent upon the
surface density of ligand; however, there was very little difference in liposome binding to the cell
46
MacKay et al. 2008 supplemental
surface (Fig. S3f). The 1% condition gave the most favorable binding kinetics seen for the
TATp-lipid, with 8.3  0.9% of the liposomes bound after 2 hrs of incubation. The fact that
surface binding was not substantially increased for the 1% TATp-lipid liposomes suggests that
the increase in surface density has an effect only upon the transfer rate of liposomes from the
surface to the cellular interior.
2.5
2 kD PEG can shield binding and uptake of TATp liposomes
In contrast to the increased internalization rate of the 1% TATp-lipid, the presence of 5% PEGDSPE caused a dramatic reduction in the rate of cell surface binding (Fig. S3g,h). This reduction
in the surface binding of liposomes translated to less liposome internalization (Fig. S3g). We
conclude that the presence of this particular PEG lipid can hinder the binding of TATp.
47
MacKay et al. 2008 supplemental
Supplemental References
1.
Li W, Huang Z, MacKay JA, Grube S, Szoka FC, Jr. (2005). Low-pH-sensitive
poly(ethylene glycol) (PEG)-stabilized plasmid nanolipoparticles: effects of PEG chain
length, lipid composition and assembly conditions on gene delivery. J Gene Med 7: 67-79.
2.
Huang Z, Li W, MacKay JA, Szoka FC, Jr. (2005). Thiocholesterol-based lipids for
ordered assembly of bioresponsive gene carriers. Mol Ther 11: 409-417.
3.
Wyman TB, Nicol F, Zelphati O, Scaria PV, Plank C, Szoka FC, Jr. (1997). Design,
synthesis, and characterization of a cationic peptide that binds to nucleic acids and
permeabilizes bilayers. Biochemistry 36: 3008-3017.
4.
Abra RM, Schreier H, Szoka FC (1982). The use of a new radioactive-iodine labeled
lipid marker to follow in vivo disposition of liposomes: comparison with an encapsulated
aqueous space marker. Res Commun Chem Pathol Pharmacol 37: 199-213.
5.
Choi JS, MacKay JA, Szoka FC, Jr. (2003). Low-pH-sensitive PEG-stabilized plasmidlipid nanoparticles: preparation and characterization. Bioconjug Chem 14: 420-429.
6.
Tseng YL, Liu JJ, Hong RL (2002). Translocation of liposomes into cancer cells by cellpenetrating peptides penetratin and tat: a kinetic and efficacy study. Mol Pharmacol 62:
864-872.
7.
Paxinos G, Watson C (1982). The rat brain in stereotaxic coordinates. Academic Press,
New York, 83pp.
8.
Chen MY, Hoffer A, Morrison PF, Hamilton JF, Hughes J, Schlageter KS, et al. (2005).
Surface properties, more than size, limiting convective distribution of virus-sized
particles and viruses in the central nervous system. J Neurosurg 103: 311-319.
48
MacKay et al. 2008 supplemental
9.
Krauze MT, Saito R, Noble C, Tamas M, Bringas J, Park JW, et al. (2005). Reflux-free
cannula for convection-enhanced high-speed delivery of therapeutic agents. J Neurosurg
103: 923-929.
10.
Mamot C, Nguyen JB, Pourdehnad M, Hadaczek P, Saito R, Bringas JR, et al. (2004).
Extensive distribution of liposomes in rodent brains and brain tumors following
convection-enhanced delivery. J Neurooncol 68: 1-9.
11.
Saito R, Bringas JR, McKnight TR, Wendland MF, Mamot C, Drummond DC, et al.
(2004). Distribution of liposomes into brain and rat brain tumor models by convectionenhanced delivery monitored with magnetic resonance imaging. Cancer Res 64: 25722579.
12.
MacKay JA, Deen DF, Szoka FC, Jr. (2005). Distribution in brain of liposomes after
convection enhanced delivery; modulation by particle charge, particle diameter, and
presence of steric coating. Brain Res 1035: 139-153.
13.
Tyagi M, Rusnati M, Presta M, Giacca M (2001). Internalization of HIV-1 tat requires
cell surface heparan sulfate proteoglycans. J Biol Chem 276: 3254-3261.
14.
MacKay JA, Szoka FC (2003). HIV TAT protein transduction domain mediated cell
binding and intracellular delivery of nanoparticles. Journal of Dispersion Science and
Technology 24: 465-473.
49