Journal of Controlled Release 107 (2005) 276 – 287
`
`www.elsevier.com/locate/jconrel
`
`Cationic lipid saturation influences intracellular delivery of
`encapsulated nucleic acids
`
`James Heyes 1, Lorne Palmer 1, Kaz Bremner, Ian MacLachlan *
`
`Protiva Biotherapeutics Inc., 100-3480 Gilmore Way, Burnaby, B.C., Canada V5G-4Y1
`
`Received 17 January 2005; accepted 15 June 2005
`Available online 28 July 2005
`
`Abstract
`
`An analogous series of cationic lipids (1,2-distearyloxy-N,N-dimethyl-3-aminopropane (DSDMA), 1,2-dioleyloxy-N,N-
`dimethyl-3-aminopropane (DODMA), 1,2-dilinoleyloxy-N,N-dimethyl-3-aminopropane (DLinDMA) and 1,2-dilinolenyloxy-
`N,N-dimethyl-3-aminopropane (DLenDMA)) possessing 0, 1, 2 or 3 double bonds per alkyl chain respectively, was synthesized
`to determine the correlation between lipid saturation, fusogenicity and efficiency of intracellular nucleic acid delivery. 31P-NMR
`analysis suggests that as saturation increases, from 2 to 0 double bonds, lamellar (La) to reversed hexagonal (HII) phase
`transition temperature increases, indicating decreasing fusogenicity. This trend is largely reflected by the efficiency of gene
`silencing observed in vitro when the lipids are formulated as Stable Nucleic Acid Lipid Particles (SNALPs) encapsulating small
`inhibitory RNA (siRNA). Uptake experiments suggest that despite their lower gene silencing efficiency, the less fusogenic
`particles are more readily internalized by cells. Microscopic visualization of fluorescently labelled siRNA uptake was supported
`by quantitative data acquired using radiolabelled preparations. Since electrostatic binding is a precursor to uptake, the pKa of
`each cationic lipid was determined. The results support a transfection model in which endosomal release, mediated by fusion
`with the endosomal membrane, results in cytoplasmic translocation of the nucleic acid payload.
`D 2005 Published by Elsevier B.V.
`
`Keywords: siRNA delivery; RNA interference; Cationic lipid; Phase transition temperature and SNALP
`
`1. Introduction
`
`RNA interference (RNAi) is a recently discovered
`gene-silencing tool. Small interfering RNA (siRNA)
`are short, double stranded RNA molecules that, in the
`
`* Corresponding author. Tel.: +1 604 630 5064.
`E-mail address: ian@protivabio.com (I. MacLachlan).
`1 The first two authors contributed equally to this work.
`
`0168-3659/$ - see front matter D 2005 Published by Elsevier B.V.
`doi:10.1016/j.jconrel.2005.06.014
`
`presence of endogenous RNA-Induced Silencing
`Complex (RISC) unwind and bind to specific
`sequences of messenger RNA (mRNA) subsequently
`mediating the destruction of the target mRNA by
`endogenous cellular machinery. In this way, RNAi
`has the potential to selectively inhibit the expression
`of disease-associated genes in humans.
`Stabilized Plasmid Lipid Particles (SPLP), con-
`sisting of a unilamellar lipid bi-layer encapsulating
`a single copy of plasmid DNA, have been reported
`
`Moderna Ex 1010-p. 1
`Moderna v Protiva
`
`

`

`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`277
`
`as a novel method of systemic nucleic acid delivery
`[1–5]. Originally developed using a detergent dial-
`ysis technique, they are now prepared more quickly
`and uniformly using the spontaneous vesicle forma-
`tion method of Jeffs et al [6]. This method has been
`adapted to encapsulate other nucleic acids such as
`duplex siRNA molecules. The resulting particles,
`referred to as Stabilized Nucleic Acid Lipid Parti-
`cles
`(SNALP), exhibit
`the stability,
`small
`size
`(b 200 nm) and low surface charge required for
`systemic delivery [7].
`The SNALP bi-layer contains a mixture of cat-
`ionic and fusogenic lipids that enable the cellular
`uptake and endosomal release of the particle’s con-
`tents. SNALP also contain a diffusible poly (ethylene
`glycol)-lipid conjugate (PEG-lipid) that provides a
`neutral, hydrophilic coating to the particle’s exterior.
`PEG-lipids both stabilize the particle during the for-
`mulation process and shield the cationic bi-layer,
`preventing rapid systemic clearance. Upon adminis-
`tration, the PEG-lipid conjugate dissociates from the
`SNALP at a rate determined by the chemistry of the
`PEG-lipid anchor,
`transforming the particle into a
`transfection-competent entity [1]. Cellular uptake of
`non-viral transfection reagents occurs primarily via
`endocytosis. Escape from the endosome is known to
`be a limiting step when using lipidic systems to
`deliver nucleic acids. Fusogenic systems readily
`overcome this barrier since they promote the break-
`down of the endosomal membrane leading to cyto-
`plasmic translocation of their nucleic acid payloads.
`Lipidic systems are most fusogenic when arranged in
`the reversed hexagonal phase (HII), as opposed to the
`more stable bi-layer forming lamellar phase (La). A
`low phase transition temperature between the two
`states indicates a lower activation energy for the
`formation of the fusogenic HII phase. Because of
`the highly defined nature of the SNALP particle,
`with its fully encapsulated nucleic acid payload and
`well-characterized mechanism of intracellular deliv-
`ery, SNALP make an ideal system for examining the
`role of individual lipid components in the pharma-
`cology of non-viral vector systems. In this work we
`explore the role of bi-layer fusogenicity in uptake,
`endosomal escape, and gene-silencing efficiency of
`SNALP.
`Fusogenicity is considered to contribute to cyto-
`plasmic delivery of nucleic acids [8–11]. It has been
`
`shown, using gold-labeled particles, that upon uptake
`via endocytosis the majority of lipoplex remain lo-
`calized in the endosome, failing to escape to the
`cytoplasm [12]. The incorporation of fusogenic lipids
`such as dioleoylphosphatidylethanolamine (DOPE)
`improves the efficiency of endosomal
`release by
`encouraging fusion events between the liposomal
`and endosomal bi-layers [12–14]. The resulting dis-
`ruption to the endosomal bi-layer aids in the escape
`of the therapeutic nucleic acid to the cytoplasm. In
`the case of a fully encapsulated system, fusion is
`expected to result in cytoplasmic translocation of the
`nucleic acid payload.
`The fusogenic nature of DOPE-containing bi-
`layers is thought
`to be due to their polymorphic
`nature. Upon formulation, most lipids adopt the bi-
`layer-forming lamellar phase (La). DOPE however,
`has a tendency to form the inverse hexagonal phase
`(HII) [14]. Using video microscopy, Koltover et al.
`studied the interaction of lipoplexes with giant anionic
`vesicles (G-vesicles), a model for the endosomal
`membrane [15]. They demonstrated that while lipo-
`plex comprised of lipids in the La phase would simply
`attach themselves stably to the surface of the G-vesi-
`cles, HII phase forming lipoplex rapidly fused with
`model endosomes. Lipids that adopt the HII phase are
`therefore regarded as dfusogenicT.
`Other researchers have noted that the degree of
`saturation of a lipid hydrophobic domain affects its
`the HII phase [16–19]. All have
`ability to adopt
`reported a trend whereby an increasing number of
`double bonds corresponds with an increasing propen-
`sity to form the non-bi-layer phase. It might therefore
`be possible to increase the tendency of the SNALP bi-
`layer to form the fusogenic HII phase by decreasing
`the degree of saturation in the lipid hydrophobic
`domain of the cationic lipid component. Using the
`SNALP lipid DODMA (containing a single double
`bond per lipid chain) as a starting point, we synthe-
`sized a homologous series of lipids with 0, 1, 2 or 3
`double bonds and studied the effect of these changes
`on physicochemical properties of bi-layers into which
`they were incorporated.
`The lipids were incorporated in SNALP containing
`anti-luciferase siRNA and assessed for both uptake
`efficiency and their ability to inhibit luciferase expres-
`sion in stably transfected Neuro 2A cells. The surface
`pKa was investigated using a toluene nitrosulphonic
`
`Moderna Ex 1010-p. 2
`Moderna v Protiva
`
`

`

`278
`
`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`acid (TNS) assay. The relative influence of each lipid
`on phase transition temperature (Tc) was also studied,
`using 31P-NMR analysis.
`
`2. Materials and methods
`
`2.1. Materials
`
`1,2-Distearoyl-sn-glycero-3-phosphocho-
`DPPS,
`line (DSPC) and cholesterol were purchased from
`Avanti Polar Lipids (Alabaster, AL). TNS was obtained
`from Sigma-Aldrich Canada (Oakville, ON). Ribo-
`Green was obtained from Molecular Probes (Eugene,
`OR). The alkyl mesylates were purchased from Nu-
`Chek Prep, Inc. (Elysian, MN, USA). siRNA (anti-
`luciferase and mismatch control) was purchased from
`Dharmacon (Lafayette, CO, USA). The anti-luciferase
`sense sequence was 5V-G.A.U.U.A.U.G.U.C.C.G.-
`G.U.U.A.U.G.U.A.U.U-3V. The anti-luciferase anti-
`sense sequence was 5V-U.A.C.A.U.A.A.C.C.G.G.A.
`C.A.U.A.A.U.C.U.U-3V. All other chemicals were pur-
`chased from Sigma-Aldrich (Oakville, ON, Canada).
`
`2.2. Synthesis of DSDMA and DODMA
`
`DSDMA and DODMA were synthesized using
`the respective alkyl bromides with methodology
`derived from that of a DOTMA precursor [20]. 3-
`(Dimethylamino)-1,2-propanediol (714 mg, 6 mmol)
`and 95% sodium hydride (NaH, 1.26 g, 50 mmol)
`were stirred in benzene (30 mL) under argon for 30
`min. The correct (either oleyl or stearyl) alkyl bro-
`mide (5.0 g, 15 mmol) was added and the reaction
`refluxed under argon for 18 h. The reaction mixture
`was then cooled in an ice bath while quenching via
`the slow addition of ethanol. Following dilution
`with a further 150 mL of benzene,
`the mixture
`was washed with distilled water (2 150 mL) and
`brine (150 mL), using ethanol
`(~20 mL)
`to aid
`phase separation if necessary. The organic phase
`was dried over magnesium sulphate and evaporated.
`The crude product was purified on a silica gel
`(Kiesel Gel 60) column eluted with chloroform
`containing 0–5% methanol. Column fractions were
`analyzed by thin layer chromatography (TLC) (silica
`gel, chloroform/methanol 9:1 v/v, visualized with
`molybdate) and fractions containing pure product
`
`(Rf = 0.5) were pooled and concentrated. The prod-
`uct was decolorized by stirring for 30 min in a
`suspension of activated charcoal (1 g) in ethanol
`(75 mL) at 60 8C. The charcoal was removed by
`filtration through Celite, and the ethanol solution
`concentrated to typically yield 2.4 g (65%) of
`H 3.65 3.32
`pure product. 1H-NMR (DSDMA): y
`(m, 7H, OCH, 3 OCH2), 2.45 2.31 (m, 2H,
`NCH2), 2.27 (s, 6H, 2 NCH3), 1.611.45 (m, 4H,
`OCH2CH2), 1.40 1.17 (m, 60H, Hstearyl), 0.86 (t,
`6H, CH2CH3). 1H-NMR (DODMA): yH 5.4 5.27
`
`(m, 4H, 2 CH = CH), 3.65 3.35 (m, 7H, OCH,
`3 OCH2), 2.47 2.33 (m, 2H, NCH2), 2.28 (s, 6H,
`2 NCH3), 2.06 1.94 (m, 8H, 4 CH2CH = CH),
`1.38 1.20
`1.61 1.50
`4H, OCH2CH2),
`(m,
`(m, 48H, Holeyl), 0.88 (t, 6H, CH2CH3).
`
`2.3. Synthesis of DLinDMA and DLenDMA
`
`The DLinDMA and DLenDMA were synthesized
`similarly to the DSDMA and DODMA, but used the
`alkyl mesylates instead of alky bromides. The general
`synthetic protocol was identical for those of DSDMA
`and DODMA, substituting the alkyl mesylates for the
`bromides in the same molar ratios. The activated
`charcoal decolorization step was omitted, since the
`products here contain conjugated double bonds and
`activated charcoal is expected to adsorb compounds
`containing such features. Yields were typically 2.0 g
`H 5.43 5.27 (m, 8H,
`(55%). 1H-NMR (DLinDMA): y
`4 CH = CH), 3.65 3.35 (m, 7H, OCH, 3 OCH2),
`2.77 (t, 4H, = CHCH2CH=), 2.47 2.33 (m, 2H,
`NCH2), 2.28 (s, 6H, 2 NCH3), 2.05 (q, 8H,
`4 CH2CH2CH=), 1.62 1.50(m, 4H, OCH2CH2),
`1.40 1.22
`0.89
`32H, Hlinoleyl),
`(t,
`6H,
`(m,
`H 5.44 5.27
`CH2CH3). 1H-NMR (DLenDMA): y
`(m, 8H, 4 CH=CH), 3.62 3.48 (m, 7H, OCH,
`3 OCH2), 2.80 (t, 4H,= CHCH2CH=), 2.43 2.32
`(m, 2H, NCH2), 2.26 (s, 6H, 2 NCH3), 2.12 1.99
`(m, 8H, 4 CH2/3CH2CH=), 1.611.51 (m, 4H,
`OCH2CH2), 1.40 1.22 (m, 20H, Hlinolenyl), 0.98 (t,
`6H, CH2CH3).
`
`2.4. Synthesis of PEG2000-C-DMA
`
`In
`follows.
`synthesized as
`PEG-C-DMA was
`brief, a C14 lipid anchor was prepared by first
`alkylating the hydroxyl groups of 3-allyloxypro-
`
`Moderna Ex 1010-p. 3
`Moderna v Protiva
`
`

`

`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`279
`
`pane-1,2-diol with myristyl bromide. The allyl group
`was subsequently removed via palladium catalysis,
`lipid. The hydroxyl
`resulting in the C14 hydroxyl
`group was converted to the primary amine by mesy-
`lation and amination to yield 1,2-dimyristyloxypro-
`pyl-3-amine, the lipid anchor. Conjugation with PEG
`was effected by treating monomethoxy poly(ethylene
`glycol) (average molecular weight 2000) with an
`excess of diphosgene to form the chloroformate.
`Addition of the C14 amine lipid anchor and stirring
`overnight yielded PEG2000-C-DMA, referred to here
`as PEG-C-DMA.
`
`2.5. SNALP Preparation
`
`SNALP with a lipid composition of DSPC:Chol:-
`PEG-C-DMA:Cationic Lipid (20 : 48 : 2 : 30 molar per-
`cent) were prepared using the spontaneous vesicle
`formation by ethanol dilution method [6]. The sam-
`ples were diafiltered against 100 mL of PBS (20 wash
`volumes) using a cross flow ultrafiltration cartridge
`(Amersham Biosciences, Piscataway, NJ) and sterile
`filtered through Acrodisc 0.8/0.2 Am syringe filters
`(Pall Corp., Ann Arbor, MI). The siRNA concentra-
`tion of final samples was determined using the Ribo-
`Green assay and a siRNA standard curve. Particle size
`and polydispersity was determined using a Malvern
`Instruments Zetasizer 3000HSA (Malvern, UK).
`Nucleic acid encapsulation was determined using a
`RiboGreen assay, comparing fluorescence in the pres-
`ence and absence of Triton X-100. RiboGreen fluo-
`rescence was measured using a Varian Eclipse
`Spectrofluorometer (Varian Inc) with kex = 500 nm,
`kem = 525 nm.
`
`2.6. TNS Assay
`
`20 AM of SNALP lipid and 6 AM of TNS were
`mixed in a fluorescence cuvette in 2mL of 20 mM
`sodium phosphate, 25 mM citrate, 20 mM ammonium
`acetate and 150 mM NaCl, at a pH that was varied
`from 4.5 to 9.5. Fluorescence was determined at each
`pH using a Varian Eclipse Spectrofluorometer (Varian
`Inc) with settings of kex = 322 nm, kem = 431 nm.
`Fluorescence for each system at the various pH values
`was then normalized to the value at pH 4.5. The pKa
`values are the point at which 50% of the molecules
`present are charged. By assuming that minimum fluo-
`
`rescence represents zero charge, and maximum fluo-
`rescence represents 100% charge, pKa can be
`estimated by measuring the pH at the point exactly
`half way between the values of minimum and maxi-
`mum charge.
`
`2.7. 31P Nuclear magnetic resonance spectroscopy
`
`Multilamellar vesicles (MLV) were prepared com-
`prising DPPS and cationic lipid at a molar ratio of
`1 : 1. This was accomplished by drying the lipids from
`chloroform solution,
`transferring to 10 mm NMR
`tubes, and hydrating in 1.5 mL of 10 mM sodium
`citrate,
`pH 4. Free
`induction
`decays
`(FIDs)
`corresponding to 1000 scans were obtained with a
`3.0 As, 608 pulse with a 1 s interpulse delay and a
`spectral width of 25 000 Hz. A gated two-level proton
`decoupling was used to ensure sufficient decoupling
`with minimum sample heating. An exponential mul-
`tiplication corresponding to 50 Hz of line broadening
`was applied to the FIDs prior to Fourier transforma-
`tion. The sample temperature (+/ 1 8C) was regula-
`ted using a Bruker B-VT1000 variable temperature
`unit. Chemical shifts were referenced to 85% phos-
`phoric acid as an external standard.
`
`2.8. In Vitro Transfection
`
`Cells were cultured in MEM (Invitrogen) contain-
`ing 10% fetal bovine serum (FBS) (CanSera) and 0.25
`mg/mL G418 (Invitrogen). Neuro2A-G cells (Neu-
`ro2A cells stably transfected to express luciferase
`[21]) were plated at a concentration of 4 104 cells
`per well in 24-well plates and grown overnight. Cells
`were treated with SNALP at doses of 0.0625 – 1.0 Ag/
`mL nucleic acid (AntiLuc Active or Mismatch Con-
`trol) and incubated for 48 h at 37 8C and 5% CO2.
`Cells were then washed with PBS and lysed with 200
`AL 250mM sodium phosphate containing 0.1% Triton
`X-100. The luciferase activity for each well was de-
`termined using Luciferase Reagent (Promega) and a
`standard luciferase protein (Roche). The luminescence
`for each was measured using a Berthold MicroLumat-
`Plus LB96V plate luminometer. The resulting lucifer-
`ase activity was then normalized for the amount of
`protein using the Micro BCA assay kit (Pierce). Lu-
`ciferase knockdown relative to a control was then
`determined for each system.
`
`Moderna Ex 1010-p. 4
`Moderna v Protiva
`
`

`

`280
`
`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`2.9. Cellular uptake
`
`SNALP were prepared incorporating the non-ex-
`changeable tritium-labeled lipid cholesteryl hexadecyl
`ether (3H-CHE) (11.1 ACi/ Amol
`total
`lipid) [22].
`Neuro2A cells (ATCC, VA, USA) were plated in 12
`well plates at 1.6 105 cells per well
`in minimal
`essential media. The following day, media was re-
`moved and replaced with media containing radiola-
`belled SNALP at 0.5 Ag/mL nucleic acid. After 24 h,
`the media and unincorporated SNALP were removed,
`adherent cells gently washed 4 times with PBS, and
`then lysed with 600 AL Lysis Buffer (250 mM phos-
`phate with 0.1% Triton X-100). The resulting cell
`lysate (500 AL) was added to glass scintillation vials
`containing 5 mL Picofluor 40 (Perkin Elmer) and 3H-
`CHE was determined using a Beckman LS6500 scin-
`tillation counter (Beckman Instruments). The protein
`content of cell lysates was determined using the Micro
`BCA assay (Pierce). Uptake was expressed as a per-
`centage of the total amount of activity applied to the
`cells per mg of cellular protein.
`
`2.10. Uptake of SNALP containing Cy3-labeled
`siRNA
`
`SNALP were formulated as previously described,
`but using siRNA labelled with the fluorophore Cy3
`(Cy3-siRNA was a gift of Sirna Therapeutics Inc,
`Boulder, CO). The encapsulation, siRNA concentra-
`tion, and particle size were determined as described.
`For the uptake study, 8 104 Neuro2A-G cells
`were grown overnight on 4-well chamber slides (BD
`Falcon, Mississauga, ON) in MEM containing 0.25
`mg/mL G418. DSDMA, DODMA, DLinDMA, and
`DLenDMA SNALP containing Cy3-siRNA, as well
`as naked Cy3-siRNA and unlabeled DSDMA SNALP
`were placed on the cells at 0.5 Ag/mL siRNA. After a
`4 h incubation with the transfection media, the cells
`were washed with PBS, then with MEM containing
`G418 and finally with PBS once more. The cells were
`then fixed in a 4% paraformaldehyde solution in PBS
`for 10 min at room temperature. The cells were
`washed with PBS and stained with 300 nM DAPI
`(Molecular Probes, Eugene, OR) in PBS for 5 min.
`The cells were washed with PBS, the mounting media
`ProLong Gold Antifade Reagent (Molecular Probes,
`Eugene, OR) applied and a cover slip added. The cells
`
`were viewed using an Olympus BX60 Microscope
`modified for fluorescence capabilities. Cy3 fluores-
`cence within the cells was visualized using a rhoda-
`mine cube set (Microgen Optics, Redding, CA) and
`the DAPI fluorescence was visualized using a DAPI
`cube set (Carsen Group, Markham, ON). Digital pic-
`tures were captured using an Olympus DP70 camera
`system. Pictures of the cells were taken at exposure
`times of 1 / 4 s when examining Cy3 fluorescence and
`1/80 s when examining DAPI fluorescence.
`
`3. Results
`
`3.1. Formulation characteristics of unsaturated lipids
`are uniform and reproducible
`
`SNALP containing the various cationic lipids were
`prepared as described and encapsulated RNA and
`particle size assessed (Table 1). The three unsaturated
`cationic lipids resulted in formulations that were ap-
`proximately the same size (132 – 140 nm). Polydis-
`persity of all formulations was low,
`indicating a
`narrow distribution of particle size. RNA encapsula-
`tion in the final particles was 84– 85% of the total.
`Attempts to encapsulate siRNA in SNALP using the
`saturated lipid DSDMA resulted in the formation of
`slightly larger particles (180 nm) with encapsulation
`of 67%.
`
`3.2. pKa of cationic lipids is influenced by saturation
`
`The apparent pKa of the cationic lipids was deter-
`mined as described in Materials and Methods. Lipid
`saturation with
`pKa correlated with degree of
`
`Table 1
`Physical properties of SNALP formulations
`
`Cationic lipid
`
`Polydispersity
`
`Diameter
`Percentage
`(nm)
`Encapsulation
`0.15 F 0.03
`182 F 11
`67 F 3
`DSDMA
`0.12 F 0.01
`137 F 4
`84 F 1
`DODMA
`0.11 F 0.02
`140 F 6
`84 F 3
`DLinDMA
`0.13 F 0.03
`132 F 7
`85 F 1
`DLenDMA
`dPercentage EncapsulationT is determined using the RiboGreen
`fluorescence assay to measure the amount of encapsulated nucleic
`acid relative to the total nucleic acid present. Particle diameter and
`polydispersity was measured using a Malvern Zetasizer. Values are
`the mean of 3 separate experiments, the error is standard deviation.
`
`Moderna Ex 1010-p. 5
`Moderna v Protiva
`
`

`

`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`281
`
`DSDMA, DODMA, DLinDMA, and DLenDMA
`exhibiting pKas of 7.6, 7.0, 6.7, and 6.7, respectively
`(Fig. 1).
`
`3.3. The bilayer-to-hexagonal phase transition tem-
`perature increases with alkyl chain saturation
`
`31P-NMR studies have previously shown that
`above certain temperatures (the Phase Transition Tem-
`perature, Tc), lipids may adopt the fusogenic HII phase
`[23,24]. A higher temperature required to convert a
`bilayer (La phase) to the HII phase indicates a less
`fusogenic bilayer. MLV were prepared using the an-
`ionic lipid DPPS in a 1 : 1 molar ratio with each
`cationic lipid. The 31P-NMR spectra of the MLV
`were measured at various temperatures. As can be
`seen in Fig. 2A, for the DSDMA/DPPS system, the
`bilayer pattern occurs from temperatures of 30 to 50
`8C (a high-field peak with a low-field shoulder).
`Therefore, DSDMA would appear to have very little
`
`DSDMA
`DODMA
`DLinDMA
`DLenDMA
`
`5
`
`6
`
`7
`pH
`
`8
`
`9
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`Relative FI
`
`0
`
`4
`
`Fig. 1. Assay to determine the apparent pKa of the cationic lipid
`incorporated in SNALP. An increase in TNS fluorescence correlates
`with an increase of positive charge. pKa is defined as the point at
`which 50% of the molecules are charged, halfway between the
`upper (completely charged) and lower (completely uncharged) lim-
`its of each curve. Error bars represent standard deviation, n = 3.
`
`A. DSDMA
`
`B. DODMA
`
`50°C
`
`40°C
`
`30°C
`
`35°C
`
`30°C
`
`25°C
`
`20°C
`
`100
`
`50
`
`0
`ppm
`
`-50
`
`-100
`
`100
`
`50
`
`-50
`
`-100
`
`0
`ppm
`
`C. DLinDMA
`
`D. DLenDMA
`
`30°C
`
`25°C
`
`20°C
`
`15°C
`
`30°C
`
`25°C
`
`20°C
`
`15°C
`
`100
`
`50
`
`0
`ppm
`
`-50
`
`-100
`
`100
`
`50
`
`0
`ppm
`
`-50
`
`-100
`
`Fig. 2. 31P-NMR analysis to determine the effect of unsaturation on phase transition temperature. Lipids were incorporated into MLV with
`the negatively charged lipid DPPS, in a 1 : 1 ratios. At lower temperatures, bi-layer patterns are observed (a high-field peak with a low-field
`shoulder). As temperature increases, a low-field peak with a high-field shoulder is observed, indicative of the inverted hexagonal phase
`transition. DSDMA-containing bilayers have a higher phase transition temperature and are accordingly less fusogenic.
`
`Moderna Ex 1010-p. 6
`Moderna v Protiva
`
`

`

`282
`
`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`DSDMA
`DLinDMA
`
`DODMA
`DLenDMA
`
`0.05
`
`0.10
`
`0.50
`0.25
`siRNA ( g/mL)
`
`1.00
`
`25
`
`20
`
`15
`
`10
`
`05
`
`% Incorporation
`
`Fig. 4. Cellular uptake. Neuro2A cells were treated with SNALP
`containing 3H-labeled CHE for 24 h. The cells were washed to
`to determination of 3H-
`remove unincorporated SNALP prior
`CHE. Uptake is expressed as a percentage of the total activity
`applied to the cells. Cellular uptake is shown to increase with
`increasing cationic lipid saturation. Error bars represent standard
`deviation, n = 3.
`
`corresponded to the ability of lipids to form the fuso-
`genic inverted hexagonal phase. Formulations com-
`prising the saturated lipid DSDMA demonstrated no
`activity. As unsaturation in the lipid’s alkyl chain
`increased, so did the capacity for RNA interference,
`with DLinDMA particles yielding an 80% knock-
`31P-NMR established
`down in gene expression.
`DLinDMA as having the lowest phase transition tem-
`perature in the series and accordingly, being the most
`fusogenic lipid. Particles comprising DLenDMA, the
`most unsaturated lipid, were slightly less efficient than
`those containing DLinDMA. All results were found to
`be significant by t-Test ( P b 0.05 at siRNA concen-
`tration of 0.5 Ag/mL, and P b 0.01 at siRNA concen-
`tration of 1.0 Ag/mL).
`
`3.5. SNALP uptake is not rate limiting for
`gene-silencing efficiency
`
`The extent to which formulations are taken up by
`cells was measured with SNALP incorporating 3H-
`labeled CHE [22]. After exposing cells to SNALP
`formulations for 24 h, cells were rinsed, lysed and
`3H-CHE uptake determined (Fig. 4). Uptake of each
`individual formulation was independent of SNALP
`concentration, with DSDMA particles exhibiting the
`greatest degree of uptake. SNALP uptake was ob-
`
`ability to form HII phases in conjunction with the
`anionic lipid. The cationic lipid with a single double
`bond, DODMA, possesses a transition temperature
`between 30 and 35 8C (Fig. 2B). The DLinDMA (2
`double bonds) and DLenDMA (3 double bonds) sys-
`tems exhibit somewhat similar transition temperatures
`between 20 and 25 8C (Fig. 2C, D). It should be noted
`that the central, isotropic peak seen in traces 3C and
`3D does not represent the phase transition temperature
`but rather results from small phospholipid vesicles
`that are also present in the preparation. The shift in
`lineshape asymmetry from a high-field peak/low-field
`shoulder (bi-layer phase, lower temperatures) to low-
`field peak/high-field shoulder
`(inverted hexagonal
`phase, higher temperatures) is an indication of phase
`transition. This is exhibited most clearly in trace 3B
`(DODMA). Based on these results it may be assumed
`that the fusogenicity of these systems increases in the
`following order: DSDMA bb DODMA b DLinDMA=
`DLenDMA.
`
`3.4. SNALP containing unsaturated cationic lipids
`show increased gene-silencing activity
`
`The ability of SNALP containing each of the four
`cationic lipids to effect gene silencing in stably trans-
`fected Neuro2A cells was evaluated (Fig. 3). It was
`found that, as hypothesized, knockdown efficiency
`
`DSDMA
`DODMA
`DLinDMA
`DLenDMA
`
`0
`
`0.2
`
`0.6
`0.4
`siRNA ( g/mL)
`
`0.8
`
`1
`
`140
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`% of Mismatch Control SNALP
`
`Fig. 3. SNALP mediated gene-silencing in vitro. Neuro2A cells
`stably transfected to express the luciferase were treated with SNALP
`containing anti-luciferase siRNA for 48 h. Gene-silencing efficiency
`was evaluated by comparing the remaining luciferase activity in
`these cells to that remaining in cells treated with control SNALP
`containing mismatch siRNA. Error bars represent standard devia-
`tion, n = 3.
`
`Moderna Ex 1010-p. 7
`Moderna v Protiva
`
`

`

`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`283
`
`served to decrease with decreasing saturation the
`DLenDMA formulation appearing particularly limited
`in this respect. These results are contrary to our
`expectations, based on the gene silencing results,
`where the DSDMA formulation is found to be least
`effective. They suggest that cellular uptake does not
`limit the gene silencing ability of SNALP, but that
`endosomal escape, mediated by a fusion event with
`the endosomal membrane is critical in SNALP medi-
`ated nucleic acid delivery. Analysis by t-test found all
`results to be significant ( P b 0.05), apart from the
`difference between DODMA and DLinDMA at con-
`centrations of 0.10, 0.50 and 1.00 Ag/mL.
`
`3.6. Visualization of the uptake process
`
`The uptake process was examined further with the
`use of fluorescently labelled SNALP. Neuro2A-G
`
`cells were treated with formulations containing Cy3-
`labeled siRNA for 4 h. After washing and fixing, cell
`nuclei were stained (blue) with the fluorescent marker
`DAPI, to more accurately determine the location of
`the fluorescently labelled siRNA (Fig. 5). In keeping
`with the results of the 3H-CHE uptake experiment, it
`can be seen that the DSDMA formulation is clearly
`the most efficient at delivering siRNA to cells. The
`Cy3 fluorescence (red) is most intense in cells treated
`with DSDMA containing SNALP. Again, in agree-
`ment with the radiolabelled uptake study, as the de-
`gree of saturation of the cationic lipid increases,
`cellular uptake of Cy3 labelled siRNA increases.
`Again, Cy3 fluorescence is extremely faint for the
`DLenDMA formulation, indicating poor uptake. Ne-
`gative controls treated with either naked Cy3-labeled
`siRNA or unlabeled SNALP revealing no cell associa-
`ted Cy3 fluorescence.
`
`A
`
`C
`
`E
`
`B
`
`D
`
`F
`
`DSDMA
`
`DLinDMA
`
`DODMA
`
`DLenDMA
`
`Naked siRNA
`
`Unlabelled SNALP
`
`Fig. 5. Fluorescent microscopy of SNALP mediated uptake of Cy3 labeled siRNA. SNALP labeled with the fluorophore Cy3 were applied to
`cells and incubated for 4 h. After washing and fixing, fluorescence microscopy indicates that siRNA uptake, as measured by Cy3 fluorescence,
`correlates with cationic lipid saturation. Cell nuclei were stained with the fluorophore DAPI (blue). Unlabeled SNALP and naked Cy3-siRNA
`were used as negative controls. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of
`this article.)
`
`Moderna Ex 1010-p. 8
`Moderna v Protiva
`
`

`

`284
`
`J. Heyes et al. / Journal of Controlled Release 107 (2005) 276–287
`
`4. Discussion
`
`Several obstacles beset delivery of genetic drugs.
`These include the ability to safely deliver the therapeu-
`tic to the target cell population, internalization by the
`cell, endosomal escape and protection from nuclease
`degradation. Stable Nucleic Acid Lipid Particles
`(SNALP) are a nucleic acid delivery platform described
`previously that has shown potential for overcoming
`these obstacles both in vitro and in vivo [1–5,25]. A
`diffusible outer PEG layer affords the particles an
`increased half-life in the bloodstream. This, combined
`with the particle’s small, uniform sizes (~130 – 140
`nm), allows them to accumulate passively in solid
`tumors via the fenestrated epithelia characteristic of
`such disease sites [1,3].
`The model
`for SNALP mediated nucleic acid
`delivery suggests that following the accumulation of
`SNALP in the tumor’s interstitial volume, the PEG
`layer slowly diffuses from the particles and the par-
`ticle is taken up by endocytosis [12]. Following
`internalization by the cell,
`the particle must now
`either escape the endosome and release its payload
`into the cytoplasm, or face degradation in the lyso-
`some. The physico-chemical properties of the lipid
`bi-layer will largely determine how effectively the
`nucleic acid payload is delivered to the cytoplasm.
`Fusion events between the SNALP and endosomal
`bi-layers will either promote their destabilization or
`facilitate direct translocation of the nucleic acid pay-
`load to the cytoplasm.
`Fusion events are more likely to occur when lipids
`are able to adopt the non-bi-layer, inverted hexagonal
`(HII) phase [14]. Increasing unsaturation in a lipidTs
`alkyl chains has been reported to increase their HII
`phase forming ability [16–19]. In this study we show
`that
`the transfection efficiency of SNALP can be
`improved by increasing the unsaturation of the cat-
`ionic lipid’s hydrophobic domain and thus, increasing
`the fusogenicity of the system.
`Other researchers have examined the impact of
`lipid saturation on transfection efficiency, usually in
`the context of other studies examining the effects of
`lipid chain length on plasmid DNA delivery. Exami-
`nation of C18 : 0 (oleyl) and C18 : 1 (stearyl) analo-
`gues suggests that the oleyl (C18 : 1) lipids are more
`effective when used to deliver plasmid DNA
`[9,26,27]. However, while these works suggest this
`
`result is related to the transition temperature/fusogeni-
`city of the lipids in question, they have not examined
`this relationship directly (e.g. with 31P-NMR).
`Our study utilizes a series of four lipids of the
`same alkyl chain length (C18) modified with a sys-
`tematic addition of double bonds. Further, as succes-
`sive double bonds are added across the series, the
`position of existing double bonds within the alkyl
`chain is retained. The position of double bonds with
`lipid alkyl chains has previously been noted to have
`a dramatic effect on bi-layer transition temperature
`[28,29].
`lipids examined in this work are
`The four
`DSDMA, DODMA, DLinDMA, and DLenDMA
`(Fig. 6). These lipids possess a protonatable tertiary
`amine head group, C18 alkyl chains, ether linkages
`between the head group and alkyl chains, and 0 to 3
`double bonds respectively. DODMA is of interest for
`nucleic acid delivery since it possesses several desir-
`able characteristics [6]. As well as utilizing the more
`stable ether linkages, its head group is pH titratable.
`At physiological pH the lipid is charge neutral, afford-
`ing SNALP longer systemic circulation time. Inside
`the endosome however, the more acidic environment
`results in head group protonation, resulting in a cat-
`ionic charge that facilitates fusion with the anionic
`lipid containing endosomal bi-layer. We therefore
`used DODMA as a