- Email: [email protected]
Synthesis, in vitro transfection activity and physicochemical characterization of novel N,N′-diacyl-1,2-diaminopropyl-3-carbamoyl-(dimethylaminoethane) amphiphilic derivatives
Chemistry and Physics of Lipids 133 (2005) 135–149
Synthesis, in vitro transfection activity and physicochemical characterization of novel N,N-diacyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane) amphiphilic derivatives Ahmad Aljaberi, Pensung Chen, Michalakis Savva∗ Division of pharmaceutical sciences, Arnold and Marie Schwartz College of Pharmacy and Health Sciences, Long Island University, 75 DeKalb Avenue, Brooklyn, NY 11201, USA Received 2 August 2004; received in revised form 21 September 2004; accepted 21 September 2004 Available online 11 November 2004
Abstract A novel series of N,N -diacyl-1,2-diaminopropyl-3-carbamoyl-(dimethylaminoethane) cationic derivatives was synthesized and screened for in vitro transfection activity at different charge ratios in the presence and absence of the helper lipids DOPE and cholesterol. Physicochemical properties of lipid–DNA complexes were studied by gel electrophoresis, fluorescence spectroscopy and dynamic light scattering. The interfacial properties of the lipids in isolation were studied using the Langmuir film balance technique at 23 ◦ C. It was found that only lipoplexes formulated with the dioleoyl derivative, 1,2lmt[5], mediated significant in vitro transfection activity. Optimum activity was obtained with 1,2lmt[5]/DOPE mixture at a ±charge ratio of 2. In agreement with the transfection results, 1,2lmt[5] was the only lipid found to complex and retard DNA migration as verified by gel electrophoresis. Despite the efficient complexation, no significant condensation of plasmid DNA was observed as indicated by fluorescence spectroscopy measurements. Monolayer studies showed that the dioleoyl derivative 1,2lmt[5] was the only lipid ˚ 2 and 38.7 mN/m, respectively. that existed in an all liquid-expanded state with a collapse area and collapse pressure of 59.5 A This lipid was also found to have the highest elasticity with a compressibility modulus at monolayer collapse of 80.4 mN/m. In conclusion, increased acyl chain fluidity and high molecular elasticity of cationic lipids were found to correlate with improved transfection activity. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Cationic lipids; Gene delivery; Lipofection; 1,2-Diaminopropyl; Monolayer studies; Interfacial compressibility
Abbreviations: DOPE, dioleoylphosphatidylethanoleamine; lmt, lipid monovalent tertiary; 1,2lmt[1], dilauroyl derivative; 1,2lmt[2], dimyristoyl derivative; 1,2lmt[3], dipalmitoyl derivative; 1,2lmt[4], distearoyl derivative; 1,2lmt[5], dioleoyl derivative; MmA, mean molecular area ∗ Corresponding author. Tel.: +1 718 488 1471; fax: +1 718 780 4586. E-mail address: [email protected] (M. Savva). 0009-3084/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.chemphyslip.2004.09.017
136
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
1. Introduction Cationic lipid-mediated gene delivery represents a safer route for delivering plasmid DNA into somatic cells compared to the immunogenic viral vectors. In addition to patient safety, advantages include (1) efficient condensation and protection of plasmid DNA from enzymatic degradation, (2) feasibility of repeating administration, (3) feasibility of targeted delivery and tissue specific accumulation as well as (4) a certain level of simplicity in the handling techniques when compared with the biohazards of viral vectors. Unfortunately, their use as the gene delivery vector of choice is limited by their low efficiency in cell transfection. This serious drawback motivated researchers to investigate the factors controlling the outcome of the lipofection process in order to achieve transfection efficiencies that can be clinically useful. In an effort to generate a meaningful and practical structure–activity relationship, various cationic lipids were synthesized (Ilies et al., 2002; Miller, 1998). Their physicochemical properties in isolation and after they were complexed with DNA were investigated and correlated with their transfection activity. Some of the properties identified to date as being important for transfection activity are fluidity (Akao et al., 1994; Singh et al., 2002) and hydration (Bennett et al., 1997) of cationic lipids and particle size (Rejman et al., 2004; Ross and Hui, 1999), charge density (Lin et al., 2003; Zhang and Anchordoquy, 2004) and structure (Safinya, 2001; Congiu et al., 2004) of lipoplexes. On the other hand, not much has been done with respect to the interfacial properties of cationic lipids. Wang et al. were the first to attempt to correlate such properties of long chain alkyl acyl carnitine esters with their transfection activities (Wang et al., 1998). More recently, the interfacial interactions between DOTAP and DPPC were reported (Bordi et al., 2003). Characterization of the interfacial properties, together with other physicochemical properties, of cationic lipids in isolation is believed to provide important insights into the various lipid–lipid interactions that take place whether between cationic lipids and helper lipids in the formulation or those between the aforementioned lipids and the natural lipids in the cell and endosomal membranes. Therefore, understanding the interfacial properties of cationic lipids and correlating them to their transfection activity will inevitably help
in designing new lipids with enhanced transfection activity. In the current report, a novel series of cationic lipids has been synthesized and tested for their ability to mediate transfection in vitro in the presence and absence of helper lipids. The physicochemical properties of the lipids in isolation and their lipoplexes were studied using various techniques. The interfacial properties of these lipids were studied using the Langmuir film balance technique from which a correlation between these properties and the transfection behavior of these new lipids was determined.
2. Experimental 2.1. Materials 2,3-Diaminopropionic acid monohydrochloride (98%), lauroyl chloride (98%), myristoyl chloride (97%), palmitoyl chloride (98%), stearoyl chloride (99%), oleoyl chloride technical grade (85%), hydrogen chloride (>99%), lithium borohydride 2.0 M solution in THF, triethylamine (99.5%), pyridine (>99%), anh. THF (99.9%), 4-nitrophenyl chloroformate (97%), N,N-dimethylethylenediamine (95%), tris (99.8+%), cholesterol (>99%), 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolam (MTT) and o-nitrophenyl -d-galactopyranoside were purchased from Sigma–Aldrich (St. Louis, MO). Anh. methanol and amylene-stabilized chloroform were from VWR (Bridgeport, NJ). DOPE was from AVANTI Polar Lipids Inc. (Alabaster, AL). Agarose, ethidium bromide, RPMI medium, fetal bovine serum, sodium pyruvate and trypsin–EDTA 1× (10 mg/ml) were from Invitrogen Life Technologies (Carlsbad, CA). Column chromatography was performed with sil˚ (Sigma–Aldrich) and ica gel, 70–230 mesh, 60 A KONTES chromatography columns (VWR). Thin layer chromatography was developed on 0.25 mm silica gel plates (Fisher Scientific) using the following solvent systems: (A) CHCl3 /CH3 OH (20:1, v/v) and (B) CHCl3 /CH3 OH/aqueous NH4 OH (66:33:1, v/v). 1 H NMR spectra were recorded on a Varian Inova 400 MHz spectrometer. High-resolution mass spectroscopy was performed by the Department of Chemistry at Ohio State University. Elemental analysis (C, H,
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
N) was performed by Robertson Microlit Laboratories Inc. (Madison, NJ). 40 mM Tris buffer, pH 7.2 was used in all the studies carried out in this paper unless otherwise stated. Water for buffer preparation was obtained from a Barnstead NANOpure ultrapure water system (Barnstead, Dubuque, IA). The resistivity and surface tension of the ultrapure water were 18 × 106 cm and 72.5 mN/m, respectively. 2.2. Synthesis 2.2.1. Methyl 2,3-diaminopropionate dihydrochloride (1) Fig. 1 represents the synthetic route used to synthesize the various analogs of N,N -diacyl-1,2diaminopropyl-3-carbamoyl-(dimethylaminoethane). Intermediates 1, 2 and 3 were synthesized using a modified procedure initially described by Sunamoto et al. (1990). Briefly, 2,3-diaminopropionic acid monohydrochloride (1 g, 7.11 mmol) was dissolved in 50 ml of absolute methanol saturated with gaseous hydrogen chloride and the reaction was refluxed at 95 ◦ C for 6 h (yield = 90%, calculated from NMR). 1 H NMR (400 MHz, D O, 20 ◦ C, TMS), δ (ppm): 2 H 4.32–4.28 [t, 1H, CH CO ], 3.63 [s, 3H, OCH3 ], 3.40–3.25 [m, 2H, CH2 ]. 2.2.2. Methyl (N,N -dimyristoyl) 2,3-diaminopropionate (2) Compound 1 was dissolved in 150 ml dry DMF. TEA (3.6 g, 35.5 mmol) was added followed by dropwise addition of myristoyl chloride (3.5 g, 14.22 mmol) while stirring at 65 ◦ C. After 3 h, additional TEA (3.6 g, 35.5 mmol) and myristoyl chloride (3.5 g, 14.22 mmol) were added and the reaction was stirred for an additional 12 h at the same temperature. Precipitated triethylammonium chloride salt was removed by suction filtration. DMF was evaporated under diminished pressure and the crude product was taken up with 100 ml CHCl3 , transferred to a separation funnel and washed with 100 ml 1N HCl aqueous solution and 100 ml brine. The organic layer was collected, concentrated and subjected to column chromatography. The column was eluted successively with 100 ml CHCl3 , 100 ml of 0.5%, 1%, 1.5%, 2%, 3%, 5%, 7%, 10%, 15%, 20% CH3 OH:CHCl3 . One hundred milliliter fractions were collected and tested for the presence of the product with TLC and NMR. Fractions
137
4 and 5 were pooled and evaporated under diminished pressure. Precipitation from methanol afforded 2.39 g of compound 2 as a white powder (Yield = 60.2%). Rf = 0.85 (A). 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.71–6.69 [d, 1H, CO NH CH ], 6.02–5.95 [t, 1H, CO NH CH2 ], 4.60–4.58 [m, 1H, CH COO ], 3.73 [s, 3H, OCH3 ], 3.62–3.58 [m, 2H, NH CH2 CH ], 2.22–2.10 [m, 4H, CO CH2 ], 1.62–1.50 [m, 4H, CO CH2 CH2 ], 1.30–1.09 [coherent, 40H, (CH2 )10 ], 0.85–0.82 [m, 6H, CH3 ]. 2.2.3. N,N -Dimyristoyl-1,2-diaminopropan-3-ol (3) Compound 2 (2.3 g, 4.28 mmol) was dissolved in 50 ml anhydrous THF. LiBH4 (12.84 mmol) was added and the reaction was stirred overnight at 60 ◦ C. THF was evaporated under diminished pressure and the dry crude was taken up with 100 ml CHCl3 and washed with 100 ml 1N HCl aqueous solution, 100 ml 1N Na2 CO3 aqueous solution and finally with 100 ml brine. The organic layer was collected, concentrated and subjected to column chromatography. The column was eluted and processed as described in the previous step. Fractions 6 and 7 were pooled and evaporated under diminished pressure to give 1.9 g of compound 3 as a white powder (yield 88.3%). Rf = 0.33 (A). 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.41–6.35 [d, 1H, CO NH CH ], 6.28–6.15 [t, 1H, CO NH CH2 ], 3.82–3.6 [m, 2H, NH CH2 CH ], 3.55–3.46 [m, 2H, CH2 OH], 3.24–3.18 [m. 1H, CH CH2 OH] 2.22–2.15 [m, 4H, CO CH2 ], 1.61–1.53 [m, 4H, CO CH2 CH2 ], 1.32–1.18 [coherent, 40H, (CH2 )10 ], 0.86–0.82 [m, 6H, CH3 ]. 2.2.4. N,N -Dimyristoyl-1,2-diaminopropane-3(4-nitrophenyl)carbonate (4) Compound 3 (1.9 g, 3.8 mmol) was dissolved in 50 ml amylene-stabilized CHCl3 . 4-Nitrophenylchloroformate (0.766 g, 3.8 mmol) was dissolved in 5 ml amylene-stabilized CHCl3 and added to the reaction mixture dropwise followed by addition of pyridine (0.31 ml, 3.8 mmol) while stirring at room temperature (Sheikh et al., 2003). After 2 h, additional 4-nitrophenylchloroformate (0.383 g, 1.9 mmol) and pyridine (0.155 ml, 1.9 mmol) were added and the reaction was stirred overnight. The reaction mixture
138
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Fig. 1. General synthetic scheme of N,N -diacyl-1,2-diaminopropyl-3-carbamoyl-(dimethylaminoethane) cationic derivatives. Briefly, the carboxylic group of the synthon 2,3-diaminopropionic acid monohydrochloride was methyl-protected via fisher esterification to yield methyl 2,3diaminopropionate dihydrochloride (1). Acylation of (1) with varied length hydrophobic fatty acid derivatives resulted in methyl (N,N -diacyl) 2,3-diaminopropionate (2). Reduction of (2) to the corresponding alcohol, N,N -diacyl-1,2-diaminopropan-3-ol (3) was effected with LiBH4 under anhydrous conditions in organic solvent. The exposed hydroxyl group of (3) was further activated with 4-nitrophenyl chloroformate to the corresponding N,N -diacyl-1,2-diaminopropane-3-(4-nitrophenyl)carbonate (4). Finally, nucleophilic attack of (4) by N,N-dimethylethylenediamine resulted into the final tertiary cationic surfactant molecules (5–9).
was diluted with additional CHCl3 to100 ml total and washed once with 100 ml 1N HCl aqueous solution, five times with 100 ml 1N Na2 CO3 aqueous solution and finally with 100 ml alkaline brine. The organic
layer was collected, concentrated and subjected to column chromatography. The column was eluted and processed as described for product 3. Fractions 4 and 5 were pooled and evaporated under diminished
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
pressure to give 2.2 g of compound 4 as a white powder (yield = 75%). Rf = 0.56 (A). 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 8.27–8.24, 7.35–7.28 [2d, 4H, C6 H4 ], 6.63–6.61 [d, 1H, CO NH CH ], CO NH CH2 ], 4.36–4.20 6.07–6.04 [t, 1H, [m, 3H, NH CH CH2 OCONH ], 3.60–3.37 [m, 2H, NH CH2 CH ], 2.20–2.13 [m, 4H, CO CH2 ], 1.64–1.57 [m, 4H, CO CH2 CH2 ], 1.30–1.09 [coherent, 40H, (CH2 )10 ], 0.88–0.82 [m, 6H, CH3 ]. 2.2.5. N,N -Dimyristoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[2] (5) Compound 4 (1 g, 1.15 mmol) was dissolved in 50 ml CHCl3 . N,N-Dimethylaminoethane (0.4 g, 4.54 mmol) was added to the reaction mixture and the solution was stirred at room temperature. After 3 h, the reaction mixture was diluted with 50 ml CHCl3 and washed three times with 100 ml alkaline brine. The organic layer was collected, concentrated and subjected to column chromatography. The column was eluted with 100 ml CHCl3 , 100 ml of 1, 2, 3, 5, 7, 10, 15, 20 and 30% CH3 OH:CHCl3 . Fractions 7 and 8 were pooled and evaporated under diminished pressure to give 0.55 g of 1,2lmt[2] as a white powder (yield = 58.3%). Rf = 0.67 (B). Anal. calcd for C36 H72 N4 O4 (MW 624): C, 69.23; H, 11.54; N, 8.97. Found: C, 69.38; H, 11.72; N, 8.94. MS (ESI) m/z 647.6 [M + Na]+ ; 1 H NMR (400 MHz, CDCl , 20 ◦ C, TMS), δ (ppm): 3 H 6.82–6.80 [d, 1H, CO NH CH ], 6.62–6.59 [t, 1H, CO NH CH2 ], 5.55–5.52 [t, 1H, OCONH ], 4.10–4.08 [m, 3H, CH CH2 OCON ], 3.43–3.32 [m, 2H, NH CH2 CH ], 3.30–3.20 [m, 2H, OCONH CH2 ], 2.42–2.38 [m, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, 2H, CO CH2 and N(CH3 )2 ], 1.60–1.52 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 40H, (CH2 )10 ], 0.85–0.82 [m, 6H, CH3 ]. All the subsequent compounds were synthesized using analogous procedures as described above for the dimyristoyl derivatives. 2.2.6. N,N -Dilauroyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[1] (6) Yield = 72.9%. Rf = 0.69 (B). Anal. calcd for C32 H64 N4 O4 (MW 568): C, 67.60; H, 11.26; N, 9.86. Found: C, 67.89; H, 11.40; N, 9.65. MS
139
(ESI) m/z 591.5 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.81–6.79 CO NH CH ], 6.62–6.59 [t, 1H, [d, 1H, CO NH CH2 ], 5.58–5.56 [t, 1H, OCONH ], CH CH2 OCON ], 4.12–4.09 [m, 3H, NH CH2 CH ], 3.24–3.19 3.4–3.30 [m, 2H, [m, 2H, OCONH CH2 ], 2.41–2.37 [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, CO CH2 and N(CH3 )2 ], 1.60–1.50 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 32H, (CH2 )8 ], 0.85–0.82 [m, 6H, CH3 ]. 2.2.7. N,N -Dipalmitoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[3] (7) Yield = 62%. Rf = 0.66 (B). Anal. calcd for C40 H80 N4 O4 (MW 680): C, 70.59; H, 11.75; N, 8.24. Found: C, 69.99; H, 11.91; N, 8.10. MS (ESI) m/z 703.6 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.70–6.68 [d, 1H, CO NH CH ], 6.48–6.45 [t, 1H, CO NH CH2 ], 5.37–5.35 [t, 1H, OCONH ], CH CH2 OCON ], 4.10–4.08 [m, 3H, 3.43–3.32 [m, 2H, NH CH2 CH ], 3.22–3.19 OCONH CH2 ], 2.41–2.38 [m, [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, 2H, CO CH2 and N(CH3 )2 ], 1.60–1.52 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 48H, (CH2 )12 ], 0.85–0.82 [m, 6H, CH3 ]. 2.2.8. N,N -Distearoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[4] (8) Yield = 32.9%. Rf = 0.64 (B). Anal. calcd for C44 H88 N4 O4 (MW 736): C, 71.74; H, 11.96; N, 7.61. Found: C, 71.6; H, 1214; N, 7.27. MS (ESI) m/z 759.7 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.81–6.79 [d, 1H, CO NH CH ], 6.60–6.58 [t, 1H, CO NH CH2 ], 5.58–5.55 [t, 1H, OCONH ], CH CH2 OCON ], 4.10–4.08 [m, 3H, 3.43–3.32 [m, 2H, NH CH2 CH ], 3.22–3.20 [m, 2H, OCONH CH2 ], 2.42–2.38 [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, CO CH2 and N(CH3 )2 ], 1.60–1.52 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 56H, (CH2 )14 ], 0.85–0.82 [m, 6H, CH3 ].
140
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
2.2.9. N,N -Dioleoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[5] (9) Yield = 49.4%. Rf = 0.65 (B). Anal. calcd for C44 H84 N4 O4 (MW 732): C, 72.13; H, 11.48; N, 7.65. Found: C, 71.84; H, 11.86; N, 7.62. MS (ESI) m/z 755.6 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.78–6.70 [d, 1H, CO NH CH ], 6.55–6.50 [t, 1H, CO NH CH2 ], 5.48–5.42 [t, 1H, OCONH ], 5.33–5.25 [m, 4H, CH CH ], CH CH2 OCON ], 4.10–4.07 [m, 3H, 3.43–3.32 [m, 2H, NH CH2 CH ], 3.25–3.18 [m, 2H, OCONH CH2 ], 2.40–2.35 [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, CO CH2 – and N(CH3 )2 ], 2.00–1.90 [m, 8H, CH2 CH CH CH2 ], 1.62–1.50 [m, 4H, CO CH2 CH2 ], 1.35–1.10 [coherent, 40H, (CH2 )10 ], 0.85–0.70 [m, 6H, CH3 ]. 2.3. Preparation of cationic aqueous dispersions Particle size measurements, gel electrophoresis, ethidium bromide displacement assay and cell transfection studies were conducted with aqueous dispersions prepared at 0.6 mM cationic lipid concentration in Tris buffer. Cationic lipid/DOPE and cationic lipid/DOPE/cholesterol molar ratios were fixed at 6:4 and 6:4:2, respectively. Briefly, aliquots of lipid solutions in chloroform were transferred to borosilicate glass test tubes. Chloroform was first evaporated under a stream of nitrogen and completely removed by high vacuum desiccation for 4 h. Dry lipid films were then hydrated with 1 ml of Tris buffer at 55 ◦ C for 30 min to 1 h with occasional vortexing, followed by sonication for 5 min at room temperature. 2.4. Plasmid preparation A pUC19 plasmid DNA vector containing the betagalactosidase (-Gal) reporter gene driven by the human cytomegalovirus (CMV) immediate-early promoter was isolated and purified as described elsewhere (Sheikh et al., 2003). The concentration of plasmid DNA was determined spectrophotometrically at 260 nm using 1 OD = 50 g/ml of plasmid DNA. The purity and quality of the plasmid DNA was verified by gel electrophoresis and the A260 /A280 ratio was found
to be 1.88 suggesting a high quality plasmid free from protein and salt contamination. 2.5. In vitro transfection assay Cationic lipids formulated in the presence and absence of the helper lipids were screened for activity against B16-F0 mouse melanoma cells (ATCC, CRL6322) at the following cationic lipid/ DNA (±) charge ratios: 1/1, 2/1 and 4/1. All calculations were based on the nucleotide average molecular weight taken to be 330, and assuming complete protonation of the tertiary amine group. Transfection experiments were carried out as described previously. Briefly, approximately 5 × 104 cells were seeded in 1 mm wells (48-well plate) 12 h before transfection. Lipoplexes were prepared 30 min prior to transfection by mixing appropriate amounts of the cationic lipid aqueous dispersions with DNA in serum free media. The cells were transfected with 250 l of lipoplex dispersions and incubated at 37 ◦ C for 4 h. The lipoplexes were then aspirated and fresh RPMI medium containing 10% FBS was added to each well. Cells were tested for -galactosidase activity 48 h after transfection using a 96-well microplate colorimetric assay that employed the substrate o-nitrophenyl -d-galactopyranoside (Sheikh et al., 2003). 2.6. MTT toxicity assay Lipoplex cytotoxicity was assessed in B16-F0 cells using a modified colorimetric MTT toxicity assay described elsewhere (Savva et al., 1999). Briefly, cells were cultured as described before for the in vitro transfection assay. After 44 h of incubation, 50 l of MTT solution in PBS (5 mg/ml) was added to each well and cells were incubated for an additional 4 h at 37 ◦ C. The supernatant was discarded and 250 l of DMSO was added to each well with gentle agitation in order to solubilize the formed purple formazan crystals. Absorption was measured at 630 nm using a multiscan plate reader. The absorption obtained for cells treated with DNA alone was taken to be 100%. 2.7. Agarose gel electrophoresis Lipoplexes at ±charge ratios of 0.5:1, 1:1, 2:1, 4:1, 6:1 and 8:1 were prepared in microfuge tubes by
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
mixing 0.2 g of DNA with the appropriate amount of the cationic lipid aqueous dispersions. The total volume of each microfuge tube was brought to 11 l with 40 mM TAE buffer, pH 7.4. The lipoplexes were briefly centrifuged at 14,000 rpm and incubated at room temperature for approximately 30 min before they were loaded into a 0.8% agarose gel containing 0.5 g/ml of ethidium bromide. Naked DNA was loaded on the outermost lanes as a negative control. The samples were subjected to an electrophoretic field of 5 V/cm for 35 min in 16 mM TAE buffer, pH 7.4. The migration of lipoplexes was visualized using a Biorad mini-transilluminator and pictures were taken with a Biorad photodocumentation system (VWR). 2.8. Dynamic light scattering Particle size distribution of lipid aqueous dispersions and lipoplexes was determined at room temperature using a Malvern Zetasizer Nano ZS particle sizer (Malvern Instruments Inc., MA) operating at 532 nm, at a 90◦ fixed angle. The instrument was validated with polystyrene microspheres (Bangs Laboratories Inc., Fishers, IN). Care was taken during preparation and dilution of the samples to prevent dust contamination. Lipoplexes were freshly prepared by diluting 200 l of 0.6 mM lipid dispersion to a final volume of 1.5 ml with Tris buffer followed by addition of the appropriate amount of DNA to the desired charge ratio. Data presented are the mean raw intensity-weighed particle size distributions obtained from the Gaussian analysis. 2.9. Ethidium bromide displacement assay Fluorescence measurements were collected with a Cary Eclipse fluorescence spectrophotometer (Varian Inc., CA.) at an excitation wavelength of 515 nm (slit width 5 nm) and an emission wavelength of 595–605 nm (slit width 2.5 nm). Briefly, 22.5 g of plasmid DNA and 0.8 g of ethidium bromide (DNA:ethidium bromide complex ratio 34:1) were added to a quartz cuvette and diluted to 3 ml with Tris buffer. Cationic lipids were added in aliquots under continuous stirring at room temperature and fluorescence was monitored continuously until the reading became stable. A sample without ethidium bromide
141
that was run side by side to the original experiment indicated that light scattering effects caused less than 2% change in the fluorescence signal (Borenstain and Barenholz, 1993). Thus, the data presented are raw data that has not been corrected for any light scattering effects. To provide the relative ethidium bromide displacement from plasmid DNA, the fluorescence signal of the ethidium bromide blank solution was subtracted from all measurements and the fluorescence intensity obtained from the plasmid DNA solution containing ethidium bromide in the absence of lipid (F0 ) was assigned a value of 100. The resulting fluorescence intensity after addition of the cationic lipids (F complex) was normalized as follows: Fcomplex % displacement = × 100. F0 2.10. Monolayer studies A computer controlled KSV Minitrough film balance (KSV instruments Ltd., Finland) equipped with two polyacetal made barriers and a platinum Wilhelmy plate was used to construct the pressure–area isotherms of monolayers of the different cationic lipids at the air/water interface. The subphase buffer was a 40 mM Tris buffer, pH 7.2 and was prepared as previously described in the materials section. After each experiment, the Teflon made trough (364 mm × 75 mm, total surface area 24,275 mm2 ) and the barriers were cleaned by rinsing them thoroughly with purified water. Lipid solutions (0.4–0.7 mg/ml) were prepared in chloroform, stored at −20 ◦ C and used within 3 days of preparation. All experiments were carried out with 140 ml of the subphase buffer with the subphase temperature maintained at 23 ◦ C by an external water bath circulator. Each experiment was repeated three to six times in an open-air vibration-free environment with occasional variation in lipid amount spread to ensure isotherms reproducibility (Welzel et al., 1998). Briefly, 20–50 l aliquots of cationic lipidspreading solution were applied carefully drop by drop to the surface of the aqueous subphase with the aid of a Hamilton glass micro syringe. After an initial delay period of 20 min, to ensure chloroform evaporation, the barriers were closed at a constant speed of 9.99 mm/min. Analysis of the constructed π/A
142
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
isotherms was carried out with WinLB 1.61 software provided by the manufacturer and surface pressure (π) was plotted as a function of the mean molecular area ˚ 2 ). Monolayer phase transitions and collapse were (A identified using the first derivative of surface pressure with respect to mean molecular area. All transitions and collapse areas reported here are onset parameters. Monolayer compressibility modulus at a given surface pressure was calculated from the π/A isotherms using a two dimensional form of the bulk compressibility modulus equation (Ali et al., 1998): dπ Kπ = (−Aπ ) dAπ where K is the compressibility modulus in mN/m and ˚ 2 at a certain surface Aπ the mean molecular area in A pressure π.
3. Results
Fig. 2. In vitro transfection activity of 1,2lmt[5] formulations at 1:1, 2:1 and 4:1 ±charge ratios in B16-F0 mouse melanoma cells as compared to DC-Chol/DOPE (6/4) at 2:1 ±charge ratio. All formulations were tested on the same day and the data presented are the average of three wells. -galactosidase levels are expressed as mU/well. Each well is equivalent to 150 l of cell lysate.
3.1. In vitro transfection activity and toxicity The aqueous dispersions of the N,N -diacyl-1,2diaminopropyl-3-carbamoyl-(dimethylaminoethane) analogs were tested for their ability to deliver the pUC19 plasmid DNA containing the -galactosidase reporter gene in B16-F0 cells. All analogs but the dioleoyl derivative 1,2lmt[5] elicited no activity at any of the three charge ratios studied in the presence or absence of the helper lipids (Fig. 2). Lipoplexes made with 1,2lmt[5] in the absence of helper lipids resulted in a moderate increase of -galactosidase levels compared with the DNA control. Inclusion of the helper lipid DOPE resulted in enhanced expression levels of the reporter enzyme with the highest level occurring around a ±charge ratio of 2. Similarly, maximum activity in the presence of cholesterol was observed at a ±charge ratio of 2. However, the activity was much lower compared to the cationic lipid/DOPE formulation. The cytotoxicity of lipoplexes prepared from the different formulations of the five analogs was evaluated in B16-F0 cells using the MTT cytotoxicity assay. No significant cytotoxicity was detected for lipoplexes prepared with the saturated fatty acyl chain analogs (data not shown). As indicated in Fig. 3, lipoplexes prepared with 1,2lmt[5] alone were well tolerated by the cells
with a cell survival percentage exceeding 87%. Formulations containing DOPE resulted in a minor decrease in the cell survival percentage. It is worth noting that all formulations showed a charge and dose dependent cytotoxicity. In other words, increasing the cationic lipid content in the lipoplex resulted in an increased cytotoxicity.
Fig. 3. Cytotoxicity of 1,2lmt[5] formulations at 1:1, 2:1 and 4:1 ±charge ratios in B16-F0 mouse melanoma cells as compared to DC-Chol/DOPE (6/4) at 2:1 ±charge ratio. Data presented are the average of three wells. Results are expressed as percentage survival of treated cell taking cells treated with DNA alone to be 100%.
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
143
Fig. 4. Gel electrophoresis of lipoplexes prepared with different lipid formulations at the following ±charge ratios: 0.5, 1, 2, 4, 6 and 8, respectively. The outermost wells were loaded with plasmid DNA as a negative control. (A) Lipoplexes prepared with cationic lipids alone, (B) cationic lipid/DOPE (6:4 molar ratios) and (C) cationic lipid/DOPE/cholesterol (6:4:2 molar ratios).
3.2. Agarose gel electrophoresis The ability of the N,N -diacyl-1,2-diaminopropyl3-carbamoyl-(dimethylaminoethane) cationic derivatives to complex with plasmid DNA was investigated by a gel electrophoresis retardation assay. As seen in Fig. 4, only 1,2lmt[5] was able to inter-
act with and completely retard plasmid DNA at a ±charge ratio of 4. Interestingly, in the presence of DOPE or DOPE and cholesterol, DNA migration was completely inhibited at a ±charge ratio of 2. Cationic lipids that failed to mediate transfection activity also failed to complex with DNA at all charge ratios.
Fig. 5. Cationic lipid induced DNA condensation monitored by an ethidium bromide displacement assay.
144
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Table 1 Particle size distribution of 1,2lmt[5] liposomal formulations and their lipoplexes 1,2lmt[5]
Liposomes Lipoplexes (1:1) Lipoplexes (2:1) Lipoplexes (4:1)
1,2lmt[5]/DOPE
1,2lmt[5]/d/chol
Mean diameter (nm)
P.I.a
Mean diameter (nm)
P.I.a
Mean diameter (nm)
P.I.a
156.7 194.8 2941 Ab
0.315 0.252 0.557 –
125.4 268.1 1216 275
0.224 0.33 1.0 0.382
136.9 248.2 2360 215
0.250 0.314 0.671 0.241
Particle size measurements were carried out in 40 mM Tris buffer pH 7.2 as described in the experimental section. Lipid dispersions were prepared at 0.6 mM, which is well above their CMC. a Values of polydispersity index (P.I.) above 0.18 denote polydisperse samples. Ab denotes aggregates of size above 6 m. All other lipid formulations showed aggregates >6 m with polydispersity indices exceeding 1.
3.3. Ethidium bromide displacement assay Agarose gel electrophoresis assay shows whether or not the cationic lipid is capable of complexing plasmid DNA, but it does not give any information about the efficiency with which the lipid condenses DNA. Further characterization of the size of the complex and the extent of condensation of DNA requires the use of other techniques. In order to investigate the extent of cationic lipid induced condensation of plasmid DNA, reduction of ethidium bromide fluorescence signal upon titrating the plasmid DNA with aliquots of cationic lipid aqueous dispersions was monitored. All cationic derivatives, including the active 1,2lmt[5] analog, with or without the helper lipid DOPE displaced less than 10% of ethidium bromide at a ±charge ratio of 2 (Fig. 5). Similar results were obtained when an aliquot of either 1,2lmt[5] or 1,2lmt[5]/DOPE equivalent to a ±charge ratio of 2 was added all at once to the plasmid DNA solution suggesting that DNA condensation by 1,2lmt[5] is not efficient. 3.4. Dynamic light scattering The particle size distribution of the dioleoyl derivative dispersions and lipoplexes is summarized in Table 1. Liposomes prepared with the dioleoyl derivative exhibited a particle size distribution below 0.2 m, while the other transfection-inefficient lipids resulted in aggregates of particle sizes well above the measurable range of the instrument. Neither prolonged incubation nor incubation at higher temperatures (∼80 ◦ C) affected the particle size of the aggregates. Upon complexing 1,2lmt[5] formulations with plasmid DNA, no significant increase in particle size was observed with
lipoplexes prepared at a 1:1 charge ratio. Increasing the charge ratio resulted in a tremendous increase in particle size for 1,2lmt[5] lipoplexes in the absence of the helper lipids. Contrary to this, 1,2lmt[5]/DOPE and 1,2lmt[5]/DOPE/chol lipoplexes aggregated at 2:1 charge ratio but exhibited a small particle size distribution when the charge ratio was increased to 4:1. 3.5. Monolayer studies The interfacial properties of the 1,2lmt cationic lipids in isolation were investigated using the Langmuir film balance technique. Aliquots of cationic lipids in CHCl3 were spread on the surface of the same buffer system used to prepare the aqueous dispersions for transfection and for the other physicochemical characterization studies, i.e. 40 mM Tris buffer, pH 7.2. Representative compression isotherms of the five analogues are shown in Fig. 6. At 23 ◦ C, only the unsaturated derivative 1,2lmt[5] formed monolayers that existed in an all liquid-expanded state with ˚ 2 and molecular collapse area and pressure of 59.47 A 38.67 mN/m, respectively (Fig. 7E and Table 2). On the other hand, monolayers of the saturated derivatives either exhibited an all liquid-condensed state or a two-dimensional phase transition toward a liquidcondensed state in their π/A isotherms. More specifically, the compression isotherm of the shortest derivative, 1,2lmt[1], exhibited a typical liquid-expanded to liquid-condensed transition observed at a molecular ˚ 2 and 21.63 mN/m, respecarea and pressure of 65.74 A tively. Further reduction of the surface area resulted in two overlapping maximums in the 1st derivative graph (Fig. 7A). The onset of the first maximum is at ˚ 2 and 42.13 mN/m, and the second at 32.9 A ˚2 39.27 A
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
145
tions (Fig. 7B). Upon compressing the film beyond that transition, surface pressure rose only marginally ˚ 2 , after within an approximate area range 105−75 A which, the surface pressure sharply increased indicating the presence of a chain ordered state. Monolayer collapse was reached at a molecular area and pres˚ 2 and 40 mN/m, respectively (Table 2). sure of 40.74 A The 1,2lmt[3] and 1,2lmt[4] analogs formed all liquidcondensed state monolayers at 23 ◦ C with a collapse ˚ 2 and pressure of molecular area of 39.78 and 42.13 A 39.46 and 39.79 mN/m, respectively (Fig. 7C and D, Table 2). Interfacial compressibility moduli (K) values of monolayers prepared from the different 1,2lmt cationic lipids, calculated at molecular collapse area are presented in Table 2. Comparison of the K values of the different analogues revealed that 1,2lmt[5] monolayers exhibited the greatest compressibility with a K value of 80.36 ± 10.4. For all other saturated derivatives, K was found to increase with increasing chain length, the maximal value being that of the distearoyl derivative 1,2lmt[4]. However, K values cannot be used to judge compressibilities of lipids that have very different molecular dimensions. The final K values could sometimes be misleading in the sense that these values are the slope dΠ/dA multiplied by the area. In Table 2 as well as in Fig. 7, it can be seen that ˚2 1,2lmt[5] monolayer collapsed at an area that is 20 A larger than that for 1,2lmt[1]. In spite of the smaller slope of 1,2lmt[5], its larger collapse area resulted in a K value close to that of 1,2lmt[1]. For comparison purposes between monolayers with varying collapse areas, one should also pay attention to the rate of change of surface pressure with respect to the change of
Fig. 6. Representative π/A isotherms of 1,2lmt cationic derivatives at the air/water interface at 23 ◦ C. Tris buffer (40 mM, pH 7.2) was used as the subphase. Isotherms were constructed using KSV computer-controlled Minitrough film balance with a compression rate of 9.99 mm/minute. (1) 1,2lmt[1]; (2) 1,2lmt[2]; (3) 1,2lmt[3]; (4) 1,2lmt[4]; and (5) 1,2lmt[5].
and 54 mN/m molecular area and surface pressure, respectively. Monolayer collapse was most probably initiated at the first maximum and further increase in surface pressure beyond that area is due to a multilayer, the formation of which was completed at a molecu˚ 2 . The presence of two maximums lar area of 32.9 A in the dΠ/dA versus mean molecular area (MmA) plot was very clear for 1,2lmt[1] and less evident for the other long chain derivatives at the temperature studied (Fig. 7). The compression isotherm of 1,2lmt[2] exhibited a similar two-dimensional phase transition but at a much lower surface pressure (Fig. 6, isotherm 2). In this case, the transition was broad and its onset was difficult to determine by the first derivative calcula-
Table 2 Monolayer parameters of 1,2lmt amphiphilic seriesa ˚ 2) MmA (A 1,2lmt[1] 1,2lmt[2] 1,2lmt[3] 1,2lmt[4] 1,2lmt[5]
39.27 65.47c 40.74 39.78 42.13 59.47
± ± ± ± ± ±
1.6 1.03 0.42 0.78 0.62 2.17
Π c (mN/m) 42.13 21.63c 40.00 39.46 39.79 38.67
± ± ± ± ± ±
1.47 1.24 1.28 1.15 2.06 0.74
dΠ/dAb
K (mN/m) 97.83 60.26c 134.67 157.75 160.2 80.36
± ± ± ± ± ±
15.54 3.88 18.15 8.62 20.8 10.4
2.48 0.91c 3.2 3.85 3.75 1.33
± ± ± ± ± ±
0.39 0.09 0.35 0.27 0.56 0.18
Phase state
˚ 2) ‘Lift off’ area (A
L2
127.99 ± 7.19
L2 L2 L2 L1
113 68.71 78.04 151.89
± 3.76 ± 4.45 ± 6.22 ± 10.08
a Measured in 40 mM Tris buffer (pH 7.2) at 23 ◦ C. All parameters are ‘onset’ transition values. MmA denotes Mean molecular Area expressed ˚ 2 per molecule. L1 denotes liquid-expanded state (Adamson, 1982). L2 denotes liquid-condensed state. in A b Values calculated at monolayer collapse and at other two-dimensional phase transitions. c Phase transition denoted by a maximum in the dΠ/dA plot as a function of mean molecular area.
146
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Fig. 7. Surface pressure (Π) and first derivative of surface pressure (dΠ/dA) as a function of mean molecular area of different 1,2lmt cationic derivatives. (A) 1,2lmt[1]; (B) 1,2lmt[2]; (C) 1,2lmt[3]; (D) 1,2lmt[4] and (E) 1,2lmt[5].
MmA (shown in Table 2 as dΠ/dA) of the monolayer as an indicator of monolayer state and compressibility. Thus, considering both the compressibility modulus and the corresponding dΠ/dA values, the in-plane
compressibility of the five analogues is in the order of 1,2lmt[5] > 1,2lmt[1] > 1,2lmt[2] > 1,2lmt[3] ≥ 1,2lmt[4], with 1,2lmt[3] and 1,2lmt[4] films being the least compressible.
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
4. Discussion The aim of this study was to delineate the physicochemical properties of double-chained cationic surfactants required for high plasmid DNA transfection efficiencies. To this end, a new series of cationic lipids, N,N -diacyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane) derivatives were synthesized by systemically varying the length and degree of unsaturation of the hydrophobic anchor. The fatty acyl chains were linked to the ionizable cationic head group via amide bonds that are biodegradable but significantly more stable than the easily hydrolyzed ester bond found in natural phospholipids. In vitro transfection activity studies conducted showed that only lipoplexes formulated with the dioleoyl derivative 1,2lmt[5] mediated significant in vitro transfection activity. All the saturated derivatives were transfection inefficient at all charge ratios in the presence and absence of the helper lipids. Optimum activity of 1,2lmt[5] was obtained in the presence of DOPE at a ±charge ratio of 2 (Fig. 2). In agreement with the transfection results, 1,2lmt[5] was the only lipid found to complex and retard DNA migration upon gel electrophoresis (Fig. 4). Despite the efficient complexation, no significant condensation of plasmid DNA took place as indicated by fluorescence spectroscopy measurements (Fig. 5). In vitro transfection, gel electrophoresis and fluorescence spectroscopy suggest that inefficient condensation of plasmid DNA by cationic lipids does not abolish in vitro transfection activity, a phenomenon that was previously observed by Banerjee et al. (2001). Loose association of plasmid DNA with cationic lipids is enough to promote significant in vitro transfection activity. However, whether this is also valid for in vivo transfection is doubtful since more stable lipoplexes are required to overcome the inhibitory effect of serum. To gain further insight on the transfection behavior of the different analogues and their formulations, particle size distribution measurements were conducted. Aqueous dispersions of the transfection inefficient lipids exhibited particle size distributions > 6 m. Such large particle size distribution indicates that these lipids failed to hydrate and form liposomes. On the other hand, liposomes prepared from 1,2lmt[5] in the presence or absence of helper lipids exhibited particle size distribution smaller than 200 nm. When these liposomes were complexed with plasmid DNA, aggrega-
147
tion was evident only at a ±charge ratio of 2. Since this charge ratio was the optimum ratio for transfection activity, it appears that a particle size distribution between 1 and 2 m correlates with enhanced in vitro activity. Besides physicochemical characterization of the lipoplexes formed by these novel cationic lipids, the Langmuir film balance technique was used in an attempt to reveal the interfacial properties of cationic lipids pertinent to efficient in vitro transfection activity. First derivative analysis of the π/A isotherms of the saturated derivatives revealed that perturbation of monomolecular films prepared from these lipids started ˚ 2 and 40 mN/m, at an area and pressure around 40 A respectively (Table 2). Interestingly, 1,2lmt[5] monolayer collapsed at a surface pressure similar to that observed for the saturated analogues but its collapse area ˚ 2 larger at 23 ◦ C. Since all was approximately 19–20 A five derivatives share the same head group composition, the origin of this difference in molecular collapse areas between the dioleoyl derivative and the saturated derivatives is the increased fluidity and hydration of 1,2lmt[5] due to the double bond it possesses in its acyl chains. With regard to the saturated derivatives, strong Van der Waals attractions between the acyl chains and intermolecular hydrogen bonding between the polar head group of adjacent lipid molecules in the highly compressed film are expected to lead to removal of the tightly bound interfacial water and tight packing of the dehydrated molecules in the film. On the contrary, the expanded unsaturated acyl chains of the dioleoyl derivative will prevent proper ordering of film components necessary for the exclusion of interfacial water and the subsequent intermolecular hydrogen bonding to take place (Ali et al., 1991). If a water molecule at the interface occupies a surface area Aswater = 9.65 (Feng et al., 1994), then there is at least 2 extra water molecules associated with each 1,2lmt[5] molecule at the collapse relative to the other saturated derivatives. The transfection potent lipid 1,2lmt[5] also formed the most elastic monolayer at 23 ◦ C, as indicated by its low compressibility modulus. All other transfection inefficient saturated derivatives were characterized by large compressibility moduli, which were found to be directly proportional to the chain length, being maximal for 1,2lmt[4] (Table 2). The correlation between molecular fluidity and improved transfection activity was first introduced by
148
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Akao et al. (1994). Their findings suggested that a liquid-crystalline to gel phase transition temperature below 37 ◦ C is required for cationic lipid-mediated delivery of plasmid DNA in vitro. Besides molecular fluidity, Bennett et al. investigated the effect of molecular hydration on in vivo transfection activity with a series of dimyristoyl and dioleoyl derivatives possessing several head group combinations (Bennett et al., 1997). In the presence of DOPE, they found the highly hydrated dioleoyl derivatives to be superior for in vivo transfection activity. They, along with others, have concluded that 14-carbon or oleoyl chains satisfy the hydration, fluidity and bilayer stability requirements for efficient transfection efficiency (Felgner et al., 1994; Ilies et al., 2004; Laxmi et al., 2001; Wang et al., 1998). The results of the current study agree with the findings stated above. The only lipid that satisfied the conditions required for efficient transfection activity was the dioleoyl derivative. On the other hand, the myristoyl derivative 1,2lmt[2] was transfection inefficient even in the presence of DOPE. This can be easily explained by the rigidity and lack of minimum hydration of this lipid aqueous dispersion. As mentioned earlier, particle size distribution measurements showed that this lipid failed to form liposomes upon hydration. Moreover, Langmuir film balance studies performed with this lipid showed that it formed films that exhibited a liquid-condensed phase at a surface pressure value comparable to that usually found in phospholipid bilayers, i.e. 30–35 mN/m (Ali et al., 1998 and references therein). In part, this might be due to the introduction of hydrogen bond functionalities in the polar part of the 1,2lmt[2] that significantly altered its phase behavior making it a more rigid molecule. In summary, five different N,N -diacyl-1,2diaminopropyl-cationic derivatives were synthesized and tested for efficient in vitro transfection activity. By means of the Langmuir film balance technique, an extensive investigation of their interfacial properties was launched in an attempt to correlate these properties with the transfection activity. The current findings suggest that cationic lipids that can interact with plasmid DNA and promote transfection activity are characterized by the following: First, acyl chain fluidity indicated by a liquid-expanded monolayer state and second, high interfacial elasticity indicated by a low compressibility modulus expressed in mN/m.
Acknowledgment The authors are grateful to Dr. Christopher M. Hadad in the Department of Chemistry, Ohio State University, for performing the mass spectroscopy analysis.
References Adamson, R.A., 1982. Physical Chemistry of Surfaces. Wiley, New York. Akao, T., Nakayama, T., Takeshia, K., Ito, A., 1994. Design of a new cationic amphiphiles with efficient transfection ability. Biochem. Mol. Biol. Int. 34, 915–920. Ali, S., Brockman, H.L., Brown, R.E., 1991. Structural determinants of miscibility in surface films of galactosylceramide and phosphatidylcholine: effect of unsaturation in the galactosylceramide acyl chain. Biochemistry 30, 11198–11205. Ali, S., Samby, J.M., Momsen, M.M., Brockman, H.L., Brown, R.E., 1998. Acyl chain-length asymmetry alters the interfacial elastic interactions of phosphatidylcholines. Biophys. J. 74, 338–348. Banerjee, R., Mahidhar, Y.V., Chaudhuri, A., Gopal, V., Rao, N.M., 2001. Design, synthesis, and transfection biology of novel cationic glycolipids for use in liposomal gene delivery. J. Med. Chem. 44, 4176–4185. Bennett, M.J., Aberle, A.M., Balasubramaniam, R.P., Malone, J.G., Malone, R.W., Nantz, M.H., 1997. Cationic lipid-mediated gene delivery to murine lung: correlation of lipid hydration with in vivo transfection activity. J. Med. Chem. 40, 4069– 4078. Bordi, F., Cametti, C., De Luca, F., Gili, T., Gaudine, D., Sennato, S., 2003. Charged lipid monolayers at the air–solution interface: coupling to polyelectrolytes. Colloid Surf. B 29, 149–157. Borenstain, V., Barenholz, Y., 1993. Characterization of liposomes and other lipid assemblies by multiprobe fluorescence polarization. Chem. Phys. lipids 64, 117–127. Congiu, A., Pozzi, D., Esposito, C., Castellano, C., Mossa, G., 2004. Correlation between structure and transfection efficiency: a study of DC-Chol DOPE/DNA complexes. Colloid Surf. B 36, 43–48. Feng, S., Brockman, H.L., MacDonald, R.C., 1994. On osmotic-type equations of state for liquid-expanded monolayers of lipids at the air–water interface. Langmuir 10, 3188–3194. Felgner, J.H., Kumar, R., Sridhar, C.N., Wheeler, C.J., Taai, Y.J., Border, R., Ramsey, P., Martin, M., Felgner, P.L., 1994. Enhanced gene delivery and mechanism studies with a novel series of cationic lipid formulations. J. Biol. Chem. 269, 2550–2561. Ilies, M.A., Seitz, W.A., Balban, A.T., 2002. Cationic lipids in gene delivery: principles, vector design and therapeutical applications. Curr. Pharm. Design 8, 2441–2473. Ilies, M.A., Seitz, W.A., Ghiviriga, I., Johnson, B.H., Miller, A., Thompson, E.B., Balaban, A.T., 2004. Pyridinium cationic lipids in gene delivery: a structure-activity correlation study. J. Med. Chem. 47, 3744–3754.
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149 Laxmi, A.A., Vijayalakshmi, P., Kaimal, T.N.B., chaudhuri, A., Ramadas, Y., Rao, N.M., 2001. Novel non-glycerol-bases cytofectins with lactic acid derived head groups. Biochem. Biophs. Res. Commun. 289, 1057–1062. Lin, A.J., Slack, N.L., Ahmad, A., George, C.X., Samuel, C.E., Safinya, C.R., 2003. Three-dimensional imaging of lipid genecarriers: membrane charge density controls universal transfection behavior in lamellar cationic liposome-DNA complexes. Biophy. J. 84, 3307–3316. Miller, A.D., 1998. Cationic liposomes for gene therapy. Angew. Chem. Int. Ed. 37, 1768–1785. Rejman, J., Oberle, V., Zuhorn, I.S., Hoekstra, D., 2004. Sizedependent internalization of particles via the pathways of clathrin- and caveolae-mediated endocytosis. Biochem. J. 377, 159–169. Ross, P.C., Hui, S.W., 1999. Lipoplex size is a major determinant of in vitro lipofection efficiency. Gene Ther. 6, 651–659. Safinya, C.R., 2001. Structures of lipid–DNA complexes: supramolecular assembly and gene delivery. Curr. Opin. Struc. Bio. 11, 440–448. Savva, M., Duda, E., Huang, L., 1999. A genetically modified recombinant tumor necrosis factor-␣ conjugated to the distal terminal of liposomal surface grafted polyethyleneglycol chains. Int. J. Pharm. 184, 45–51.
149
Sheikh, M., Feig, J., Gee, B., Li, S., Savva, M., 2003. In vitro lipofection with novel series of symmetric 1,3-dialkoylamidopropanebased cationic surfactants containing single primary and tertiary amine polar head groups. Chem. Phys. Lipids 124, 49–61. Singh, R.S., Mukherjee, K., Banerjee, R., Chaudhuri, A., Hait, S.K., Moulik, S.P., Ramadas, Y., Vijayalakshmi, A., Rao, N.M., 2002. Anchor dependency for non-glycerol based cationic lipofectins: mixed bag of regular and anomalous transfection profiles. Chem. Eur. J. 8, 900–909. Sunamoto, J., Goto, M., Iwamoto, K., Kondo, H., Sato, T., 1990. Synthesis and characterization of 1,2-dimyristoylamido1, 2-deoxyphosphatidylcholine as an artificial boundary lipid. Biochem. Biophys. Acta 1024, 209–219. Wang, J., Guo, X., Xu, Y., Barron, L., Szoka Jr., F.C., 1998. Synthesis and characterization of long chain alkyl acyl carnitine esters. Potentially biodegradable cationic lipids for the use in gene delivery. J. Med. Chem. 41, 2207–2215. Welzel, P.B., Weis, I., Schwarz, G., 1998. Sources of error in Langmuir trough measurements: Wilhelmy plate effects and surface curvature. Colloid Surf. A 144, 229–234. Zhang, Y., Anchordoquy, T.J., 2004. The role of lipid charge density in the serum stability of cationic lipid/DNA complexes. Biochem. Biophys. Acta 1663, 143–157.
Synthesis, in vitro transfection activity and physicochemical characterization of novel N,N-diacyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane) amphiphilic derivatives Ahmad Aljaberi, Pensung Chen, Michalakis Savva∗ Division of pharmaceutical sciences, Arnold and Marie Schwartz College of Pharmacy and Health Sciences, Long Island University, 75 DeKalb Avenue, Brooklyn, NY 11201, USA Received 2 August 2004; received in revised form 21 September 2004; accepted 21 September 2004 Available online 11 November 2004
Abstract A novel series of N,N -diacyl-1,2-diaminopropyl-3-carbamoyl-(dimethylaminoethane) cationic derivatives was synthesized and screened for in vitro transfection activity at different charge ratios in the presence and absence of the helper lipids DOPE and cholesterol. Physicochemical properties of lipid–DNA complexes were studied by gel electrophoresis, fluorescence spectroscopy and dynamic light scattering. The interfacial properties of the lipids in isolation were studied using the Langmuir film balance technique at 23 ◦ C. It was found that only lipoplexes formulated with the dioleoyl derivative, 1,2lmt[5], mediated significant in vitro transfection activity. Optimum activity was obtained with 1,2lmt[5]/DOPE mixture at a ±charge ratio of 2. In agreement with the transfection results, 1,2lmt[5] was the only lipid found to complex and retard DNA migration as verified by gel electrophoresis. Despite the efficient complexation, no significant condensation of plasmid DNA was observed as indicated by fluorescence spectroscopy measurements. Monolayer studies showed that the dioleoyl derivative 1,2lmt[5] was the only lipid ˚ 2 and 38.7 mN/m, respectively. that existed in an all liquid-expanded state with a collapse area and collapse pressure of 59.5 A This lipid was also found to have the highest elasticity with a compressibility modulus at monolayer collapse of 80.4 mN/m. In conclusion, increased acyl chain fluidity and high molecular elasticity of cationic lipids were found to correlate with improved transfection activity. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Cationic lipids; Gene delivery; Lipofection; 1,2-Diaminopropyl; Monolayer studies; Interfacial compressibility
Abbreviations: DOPE, dioleoylphosphatidylethanoleamine; lmt, lipid monovalent tertiary; 1,2lmt[1], dilauroyl derivative; 1,2lmt[2], dimyristoyl derivative; 1,2lmt[3], dipalmitoyl derivative; 1,2lmt[4], distearoyl derivative; 1,2lmt[5], dioleoyl derivative; MmA, mean molecular area ∗ Corresponding author. Tel.: +1 718 488 1471; fax: +1 718 780 4586. E-mail address: [email protected] (M. Savva). 0009-3084/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.chemphyslip.2004.09.017
136
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
1. Introduction Cationic lipid-mediated gene delivery represents a safer route for delivering plasmid DNA into somatic cells compared to the immunogenic viral vectors. In addition to patient safety, advantages include (1) efficient condensation and protection of plasmid DNA from enzymatic degradation, (2) feasibility of repeating administration, (3) feasibility of targeted delivery and tissue specific accumulation as well as (4) a certain level of simplicity in the handling techniques when compared with the biohazards of viral vectors. Unfortunately, their use as the gene delivery vector of choice is limited by their low efficiency in cell transfection. This serious drawback motivated researchers to investigate the factors controlling the outcome of the lipofection process in order to achieve transfection efficiencies that can be clinically useful. In an effort to generate a meaningful and practical structure–activity relationship, various cationic lipids were synthesized (Ilies et al., 2002; Miller, 1998). Their physicochemical properties in isolation and after they were complexed with DNA were investigated and correlated with their transfection activity. Some of the properties identified to date as being important for transfection activity are fluidity (Akao et al., 1994; Singh et al., 2002) and hydration (Bennett et al., 1997) of cationic lipids and particle size (Rejman et al., 2004; Ross and Hui, 1999), charge density (Lin et al., 2003; Zhang and Anchordoquy, 2004) and structure (Safinya, 2001; Congiu et al., 2004) of lipoplexes. On the other hand, not much has been done with respect to the interfacial properties of cationic lipids. Wang et al. were the first to attempt to correlate such properties of long chain alkyl acyl carnitine esters with their transfection activities (Wang et al., 1998). More recently, the interfacial interactions between DOTAP and DPPC were reported (Bordi et al., 2003). Characterization of the interfacial properties, together with other physicochemical properties, of cationic lipids in isolation is believed to provide important insights into the various lipid–lipid interactions that take place whether between cationic lipids and helper lipids in the formulation or those between the aforementioned lipids and the natural lipids in the cell and endosomal membranes. Therefore, understanding the interfacial properties of cationic lipids and correlating them to their transfection activity will inevitably help
in designing new lipids with enhanced transfection activity. In the current report, a novel series of cationic lipids has been synthesized and tested for their ability to mediate transfection in vitro in the presence and absence of helper lipids. The physicochemical properties of the lipids in isolation and their lipoplexes were studied using various techniques. The interfacial properties of these lipids were studied using the Langmuir film balance technique from which a correlation between these properties and the transfection behavior of these new lipids was determined.
2. Experimental 2.1. Materials 2,3-Diaminopropionic acid monohydrochloride (98%), lauroyl chloride (98%), myristoyl chloride (97%), palmitoyl chloride (98%), stearoyl chloride (99%), oleoyl chloride technical grade (85%), hydrogen chloride (>99%), lithium borohydride 2.0 M solution in THF, triethylamine (99.5%), pyridine (>99%), anh. THF (99.9%), 4-nitrophenyl chloroformate (97%), N,N-dimethylethylenediamine (95%), tris (99.8+%), cholesterol (>99%), 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolam (MTT) and o-nitrophenyl -d-galactopyranoside were purchased from Sigma–Aldrich (St. Louis, MO). Anh. methanol and amylene-stabilized chloroform were from VWR (Bridgeport, NJ). DOPE was from AVANTI Polar Lipids Inc. (Alabaster, AL). Agarose, ethidium bromide, RPMI medium, fetal bovine serum, sodium pyruvate and trypsin–EDTA 1× (10 mg/ml) were from Invitrogen Life Technologies (Carlsbad, CA). Column chromatography was performed with sil˚ (Sigma–Aldrich) and ica gel, 70–230 mesh, 60 A KONTES chromatography columns (VWR). Thin layer chromatography was developed on 0.25 mm silica gel plates (Fisher Scientific) using the following solvent systems: (A) CHCl3 /CH3 OH (20:1, v/v) and (B) CHCl3 /CH3 OH/aqueous NH4 OH (66:33:1, v/v). 1 H NMR spectra were recorded on a Varian Inova 400 MHz spectrometer. High-resolution mass spectroscopy was performed by the Department of Chemistry at Ohio State University. Elemental analysis (C, H,
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
N) was performed by Robertson Microlit Laboratories Inc. (Madison, NJ). 40 mM Tris buffer, pH 7.2 was used in all the studies carried out in this paper unless otherwise stated. Water for buffer preparation was obtained from a Barnstead NANOpure ultrapure water system (Barnstead, Dubuque, IA). The resistivity and surface tension of the ultrapure water were 18 × 106 cm and 72.5 mN/m, respectively. 2.2. Synthesis 2.2.1. Methyl 2,3-diaminopropionate dihydrochloride (1) Fig. 1 represents the synthetic route used to synthesize the various analogs of N,N -diacyl-1,2diaminopropyl-3-carbamoyl-(dimethylaminoethane). Intermediates 1, 2 and 3 were synthesized using a modified procedure initially described by Sunamoto et al. (1990). Briefly, 2,3-diaminopropionic acid monohydrochloride (1 g, 7.11 mmol) was dissolved in 50 ml of absolute methanol saturated with gaseous hydrogen chloride and the reaction was refluxed at 95 ◦ C for 6 h (yield = 90%, calculated from NMR). 1 H NMR (400 MHz, D O, 20 ◦ C, TMS), δ (ppm): 2 H 4.32–4.28 [t, 1H, CH CO ], 3.63 [s, 3H, OCH3 ], 3.40–3.25 [m, 2H, CH2 ]. 2.2.2. Methyl (N,N -dimyristoyl) 2,3-diaminopropionate (2) Compound 1 was dissolved in 150 ml dry DMF. TEA (3.6 g, 35.5 mmol) was added followed by dropwise addition of myristoyl chloride (3.5 g, 14.22 mmol) while stirring at 65 ◦ C. After 3 h, additional TEA (3.6 g, 35.5 mmol) and myristoyl chloride (3.5 g, 14.22 mmol) were added and the reaction was stirred for an additional 12 h at the same temperature. Precipitated triethylammonium chloride salt was removed by suction filtration. DMF was evaporated under diminished pressure and the crude product was taken up with 100 ml CHCl3 , transferred to a separation funnel and washed with 100 ml 1N HCl aqueous solution and 100 ml brine. The organic layer was collected, concentrated and subjected to column chromatography. The column was eluted successively with 100 ml CHCl3 , 100 ml of 0.5%, 1%, 1.5%, 2%, 3%, 5%, 7%, 10%, 15%, 20% CH3 OH:CHCl3 . One hundred milliliter fractions were collected and tested for the presence of the product with TLC and NMR. Fractions
137
4 and 5 were pooled and evaporated under diminished pressure. Precipitation from methanol afforded 2.39 g of compound 2 as a white powder (Yield = 60.2%). Rf = 0.85 (A). 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.71–6.69 [d, 1H, CO NH CH ], 6.02–5.95 [t, 1H, CO NH CH2 ], 4.60–4.58 [m, 1H, CH COO ], 3.73 [s, 3H, OCH3 ], 3.62–3.58 [m, 2H, NH CH2 CH ], 2.22–2.10 [m, 4H, CO CH2 ], 1.62–1.50 [m, 4H, CO CH2 CH2 ], 1.30–1.09 [coherent, 40H, (CH2 )10 ], 0.85–0.82 [m, 6H, CH3 ]. 2.2.3. N,N -Dimyristoyl-1,2-diaminopropan-3-ol (3) Compound 2 (2.3 g, 4.28 mmol) was dissolved in 50 ml anhydrous THF. LiBH4 (12.84 mmol) was added and the reaction was stirred overnight at 60 ◦ C. THF was evaporated under diminished pressure and the dry crude was taken up with 100 ml CHCl3 and washed with 100 ml 1N HCl aqueous solution, 100 ml 1N Na2 CO3 aqueous solution and finally with 100 ml brine. The organic layer was collected, concentrated and subjected to column chromatography. The column was eluted and processed as described in the previous step. Fractions 6 and 7 were pooled and evaporated under diminished pressure to give 1.9 g of compound 3 as a white powder (yield 88.3%). Rf = 0.33 (A). 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.41–6.35 [d, 1H, CO NH CH ], 6.28–6.15 [t, 1H, CO NH CH2 ], 3.82–3.6 [m, 2H, NH CH2 CH ], 3.55–3.46 [m, 2H, CH2 OH], 3.24–3.18 [m. 1H, CH CH2 OH] 2.22–2.15 [m, 4H, CO CH2 ], 1.61–1.53 [m, 4H, CO CH2 CH2 ], 1.32–1.18 [coherent, 40H, (CH2 )10 ], 0.86–0.82 [m, 6H, CH3 ]. 2.2.4. N,N -Dimyristoyl-1,2-diaminopropane-3(4-nitrophenyl)carbonate (4) Compound 3 (1.9 g, 3.8 mmol) was dissolved in 50 ml amylene-stabilized CHCl3 . 4-Nitrophenylchloroformate (0.766 g, 3.8 mmol) was dissolved in 5 ml amylene-stabilized CHCl3 and added to the reaction mixture dropwise followed by addition of pyridine (0.31 ml, 3.8 mmol) while stirring at room temperature (Sheikh et al., 2003). After 2 h, additional 4-nitrophenylchloroformate (0.383 g, 1.9 mmol) and pyridine (0.155 ml, 1.9 mmol) were added and the reaction was stirred overnight. The reaction mixture
138
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Fig. 1. General synthetic scheme of N,N -diacyl-1,2-diaminopropyl-3-carbamoyl-(dimethylaminoethane) cationic derivatives. Briefly, the carboxylic group of the synthon 2,3-diaminopropionic acid monohydrochloride was methyl-protected via fisher esterification to yield methyl 2,3diaminopropionate dihydrochloride (1). Acylation of (1) with varied length hydrophobic fatty acid derivatives resulted in methyl (N,N -diacyl) 2,3-diaminopropionate (2). Reduction of (2) to the corresponding alcohol, N,N -diacyl-1,2-diaminopropan-3-ol (3) was effected with LiBH4 under anhydrous conditions in organic solvent. The exposed hydroxyl group of (3) was further activated with 4-nitrophenyl chloroformate to the corresponding N,N -diacyl-1,2-diaminopropane-3-(4-nitrophenyl)carbonate (4). Finally, nucleophilic attack of (4) by N,N-dimethylethylenediamine resulted into the final tertiary cationic surfactant molecules (5–9).
was diluted with additional CHCl3 to100 ml total and washed once with 100 ml 1N HCl aqueous solution, five times with 100 ml 1N Na2 CO3 aqueous solution and finally with 100 ml alkaline brine. The organic
layer was collected, concentrated and subjected to column chromatography. The column was eluted and processed as described for product 3. Fractions 4 and 5 were pooled and evaporated under diminished
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
pressure to give 2.2 g of compound 4 as a white powder (yield = 75%). Rf = 0.56 (A). 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 8.27–8.24, 7.35–7.28 [2d, 4H, C6 H4 ], 6.63–6.61 [d, 1H, CO NH CH ], CO NH CH2 ], 4.36–4.20 6.07–6.04 [t, 1H, [m, 3H, NH CH CH2 OCONH ], 3.60–3.37 [m, 2H, NH CH2 CH ], 2.20–2.13 [m, 4H, CO CH2 ], 1.64–1.57 [m, 4H, CO CH2 CH2 ], 1.30–1.09 [coherent, 40H, (CH2 )10 ], 0.88–0.82 [m, 6H, CH3 ]. 2.2.5. N,N -Dimyristoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[2] (5) Compound 4 (1 g, 1.15 mmol) was dissolved in 50 ml CHCl3 . N,N-Dimethylaminoethane (0.4 g, 4.54 mmol) was added to the reaction mixture and the solution was stirred at room temperature. After 3 h, the reaction mixture was diluted with 50 ml CHCl3 and washed three times with 100 ml alkaline brine. The organic layer was collected, concentrated and subjected to column chromatography. The column was eluted with 100 ml CHCl3 , 100 ml of 1, 2, 3, 5, 7, 10, 15, 20 and 30% CH3 OH:CHCl3 . Fractions 7 and 8 were pooled and evaporated under diminished pressure to give 0.55 g of 1,2lmt[2] as a white powder (yield = 58.3%). Rf = 0.67 (B). Anal. calcd for C36 H72 N4 O4 (MW 624): C, 69.23; H, 11.54; N, 8.97. Found: C, 69.38; H, 11.72; N, 8.94. MS (ESI) m/z 647.6 [M + Na]+ ; 1 H NMR (400 MHz, CDCl , 20 ◦ C, TMS), δ (ppm): 3 H 6.82–6.80 [d, 1H, CO NH CH ], 6.62–6.59 [t, 1H, CO NH CH2 ], 5.55–5.52 [t, 1H, OCONH ], 4.10–4.08 [m, 3H, CH CH2 OCON ], 3.43–3.32 [m, 2H, NH CH2 CH ], 3.30–3.20 [m, 2H, OCONH CH2 ], 2.42–2.38 [m, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, 2H, CO CH2 and N(CH3 )2 ], 1.60–1.52 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 40H, (CH2 )10 ], 0.85–0.82 [m, 6H, CH3 ]. All the subsequent compounds were synthesized using analogous procedures as described above for the dimyristoyl derivatives. 2.2.6. N,N -Dilauroyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[1] (6) Yield = 72.9%. Rf = 0.69 (B). Anal. calcd for C32 H64 N4 O4 (MW 568): C, 67.60; H, 11.26; N, 9.86. Found: C, 67.89; H, 11.40; N, 9.65. MS
139
(ESI) m/z 591.5 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.81–6.79 CO NH CH ], 6.62–6.59 [t, 1H, [d, 1H, CO NH CH2 ], 5.58–5.56 [t, 1H, OCONH ], CH CH2 OCON ], 4.12–4.09 [m, 3H, NH CH2 CH ], 3.24–3.19 3.4–3.30 [m, 2H, [m, 2H, OCONH CH2 ], 2.41–2.37 [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, CO CH2 and N(CH3 )2 ], 1.60–1.50 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 32H, (CH2 )8 ], 0.85–0.82 [m, 6H, CH3 ]. 2.2.7. N,N -Dipalmitoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[3] (7) Yield = 62%. Rf = 0.66 (B). Anal. calcd for C40 H80 N4 O4 (MW 680): C, 70.59; H, 11.75; N, 8.24. Found: C, 69.99; H, 11.91; N, 8.10. MS (ESI) m/z 703.6 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.70–6.68 [d, 1H, CO NH CH ], 6.48–6.45 [t, 1H, CO NH CH2 ], 5.37–5.35 [t, 1H, OCONH ], CH CH2 OCON ], 4.10–4.08 [m, 3H, 3.43–3.32 [m, 2H, NH CH2 CH ], 3.22–3.19 OCONH CH2 ], 2.41–2.38 [m, [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, 2H, CO CH2 and N(CH3 )2 ], 1.60–1.52 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 48H, (CH2 )12 ], 0.85–0.82 [m, 6H, CH3 ]. 2.2.8. N,N -Distearoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[4] (8) Yield = 32.9%. Rf = 0.64 (B). Anal. calcd for C44 H88 N4 O4 (MW 736): C, 71.74; H, 11.96; N, 7.61. Found: C, 71.6; H, 1214; N, 7.27. MS (ESI) m/z 759.7 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.81–6.79 [d, 1H, CO NH CH ], 6.60–6.58 [t, 1H, CO NH CH2 ], 5.58–5.55 [t, 1H, OCONH ], CH CH2 OCON ], 4.10–4.08 [m, 3H, 3.43–3.32 [m, 2H, NH CH2 CH ], 3.22–3.20 [m, 2H, OCONH CH2 ], 2.42–2.38 [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, CO CH2 and N(CH3 )2 ], 1.60–1.52 [m, 4H, CO CH2 CH2 ], 1.30–1.18 [coherent, 56H, (CH2 )14 ], 0.85–0.82 [m, 6H, CH3 ].
140
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
2.2.9. N,N -Dioleoyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane), 1,2lmt[5] (9) Yield = 49.4%. Rf = 0.65 (B). Anal. calcd for C44 H84 N4 O4 (MW 732): C, 72.13; H, 11.48; N, 7.65. Found: C, 71.84; H, 11.86; N, 7.62. MS (ESI) m/z 755.6 [M + Na]+ ; 1 H NMR (400 MHz, CDCl3 , 20 ◦ C, TMS), δH (ppm): 6.78–6.70 [d, 1H, CO NH CH ], 6.55–6.50 [t, 1H, CO NH CH2 ], 5.48–5.42 [t, 1H, OCONH ], 5.33–5.25 [m, 4H, CH CH ], CH CH2 OCON ], 4.10–4.07 [m, 3H, 3.43–3.32 [m, 2H, NH CH2 CH ], 3.25–3.18 [m, 2H, OCONH CH2 ], 2.40–2.35 [m, 2H, CH2 N(CH3 )2 ], 2.22–2.10 [m, 10H, CO CH2 – and N(CH3 )2 ], 2.00–1.90 [m, 8H, CH2 CH CH CH2 ], 1.62–1.50 [m, 4H, CO CH2 CH2 ], 1.35–1.10 [coherent, 40H, (CH2 )10 ], 0.85–0.70 [m, 6H, CH3 ]. 2.3. Preparation of cationic aqueous dispersions Particle size measurements, gel electrophoresis, ethidium bromide displacement assay and cell transfection studies were conducted with aqueous dispersions prepared at 0.6 mM cationic lipid concentration in Tris buffer. Cationic lipid/DOPE and cationic lipid/DOPE/cholesterol molar ratios were fixed at 6:4 and 6:4:2, respectively. Briefly, aliquots of lipid solutions in chloroform were transferred to borosilicate glass test tubes. Chloroform was first evaporated under a stream of nitrogen and completely removed by high vacuum desiccation for 4 h. Dry lipid films were then hydrated with 1 ml of Tris buffer at 55 ◦ C for 30 min to 1 h with occasional vortexing, followed by sonication for 5 min at room temperature. 2.4. Plasmid preparation A pUC19 plasmid DNA vector containing the betagalactosidase (-Gal) reporter gene driven by the human cytomegalovirus (CMV) immediate-early promoter was isolated and purified as described elsewhere (Sheikh et al., 2003). The concentration of plasmid DNA was determined spectrophotometrically at 260 nm using 1 OD = 50 g/ml of plasmid DNA. The purity and quality of the plasmid DNA was verified by gel electrophoresis and the A260 /A280 ratio was found
to be 1.88 suggesting a high quality plasmid free from protein and salt contamination. 2.5. In vitro transfection assay Cationic lipids formulated in the presence and absence of the helper lipids were screened for activity against B16-F0 mouse melanoma cells (ATCC, CRL6322) at the following cationic lipid/ DNA (±) charge ratios: 1/1, 2/1 and 4/1. All calculations were based on the nucleotide average molecular weight taken to be 330, and assuming complete protonation of the tertiary amine group. Transfection experiments were carried out as described previously. Briefly, approximately 5 × 104 cells were seeded in 1 mm wells (48-well plate) 12 h before transfection. Lipoplexes were prepared 30 min prior to transfection by mixing appropriate amounts of the cationic lipid aqueous dispersions with DNA in serum free media. The cells were transfected with 250 l of lipoplex dispersions and incubated at 37 ◦ C for 4 h. The lipoplexes were then aspirated and fresh RPMI medium containing 10% FBS was added to each well. Cells were tested for -galactosidase activity 48 h after transfection using a 96-well microplate colorimetric assay that employed the substrate o-nitrophenyl -d-galactopyranoside (Sheikh et al., 2003). 2.6. MTT toxicity assay Lipoplex cytotoxicity was assessed in B16-F0 cells using a modified colorimetric MTT toxicity assay described elsewhere (Savva et al., 1999). Briefly, cells were cultured as described before for the in vitro transfection assay. After 44 h of incubation, 50 l of MTT solution in PBS (5 mg/ml) was added to each well and cells were incubated for an additional 4 h at 37 ◦ C. The supernatant was discarded and 250 l of DMSO was added to each well with gentle agitation in order to solubilize the formed purple formazan crystals. Absorption was measured at 630 nm using a multiscan plate reader. The absorption obtained for cells treated with DNA alone was taken to be 100%. 2.7. Agarose gel electrophoresis Lipoplexes at ±charge ratios of 0.5:1, 1:1, 2:1, 4:1, 6:1 and 8:1 were prepared in microfuge tubes by
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
mixing 0.2 g of DNA with the appropriate amount of the cationic lipid aqueous dispersions. The total volume of each microfuge tube was brought to 11 l with 40 mM TAE buffer, pH 7.4. The lipoplexes were briefly centrifuged at 14,000 rpm and incubated at room temperature for approximately 30 min before they were loaded into a 0.8% agarose gel containing 0.5 g/ml of ethidium bromide. Naked DNA was loaded on the outermost lanes as a negative control. The samples were subjected to an electrophoretic field of 5 V/cm for 35 min in 16 mM TAE buffer, pH 7.4. The migration of lipoplexes was visualized using a Biorad mini-transilluminator and pictures were taken with a Biorad photodocumentation system (VWR). 2.8. Dynamic light scattering Particle size distribution of lipid aqueous dispersions and lipoplexes was determined at room temperature using a Malvern Zetasizer Nano ZS particle sizer (Malvern Instruments Inc., MA) operating at 532 nm, at a 90◦ fixed angle. The instrument was validated with polystyrene microspheres (Bangs Laboratories Inc., Fishers, IN). Care was taken during preparation and dilution of the samples to prevent dust contamination. Lipoplexes were freshly prepared by diluting 200 l of 0.6 mM lipid dispersion to a final volume of 1.5 ml with Tris buffer followed by addition of the appropriate amount of DNA to the desired charge ratio. Data presented are the mean raw intensity-weighed particle size distributions obtained from the Gaussian analysis. 2.9. Ethidium bromide displacement assay Fluorescence measurements were collected with a Cary Eclipse fluorescence spectrophotometer (Varian Inc., CA.) at an excitation wavelength of 515 nm (slit width 5 nm) and an emission wavelength of 595–605 nm (slit width 2.5 nm). Briefly, 22.5 g of plasmid DNA and 0.8 g of ethidium bromide (DNA:ethidium bromide complex ratio 34:1) were added to a quartz cuvette and diluted to 3 ml with Tris buffer. Cationic lipids were added in aliquots under continuous stirring at room temperature and fluorescence was monitored continuously until the reading became stable. A sample without ethidium bromide
141
that was run side by side to the original experiment indicated that light scattering effects caused less than 2% change in the fluorescence signal (Borenstain and Barenholz, 1993). Thus, the data presented are raw data that has not been corrected for any light scattering effects. To provide the relative ethidium bromide displacement from plasmid DNA, the fluorescence signal of the ethidium bromide blank solution was subtracted from all measurements and the fluorescence intensity obtained from the plasmid DNA solution containing ethidium bromide in the absence of lipid (F0 ) was assigned a value of 100. The resulting fluorescence intensity after addition of the cationic lipids (F complex) was normalized as follows: Fcomplex % displacement = × 100. F0 2.10. Monolayer studies A computer controlled KSV Minitrough film balance (KSV instruments Ltd., Finland) equipped with two polyacetal made barriers and a platinum Wilhelmy plate was used to construct the pressure–area isotherms of monolayers of the different cationic lipids at the air/water interface. The subphase buffer was a 40 mM Tris buffer, pH 7.2 and was prepared as previously described in the materials section. After each experiment, the Teflon made trough (364 mm × 75 mm, total surface area 24,275 mm2 ) and the barriers were cleaned by rinsing them thoroughly with purified water. Lipid solutions (0.4–0.7 mg/ml) were prepared in chloroform, stored at −20 ◦ C and used within 3 days of preparation. All experiments were carried out with 140 ml of the subphase buffer with the subphase temperature maintained at 23 ◦ C by an external water bath circulator. Each experiment was repeated three to six times in an open-air vibration-free environment with occasional variation in lipid amount spread to ensure isotherms reproducibility (Welzel et al., 1998). Briefly, 20–50 l aliquots of cationic lipidspreading solution were applied carefully drop by drop to the surface of the aqueous subphase with the aid of a Hamilton glass micro syringe. After an initial delay period of 20 min, to ensure chloroform evaporation, the barriers were closed at a constant speed of 9.99 mm/min. Analysis of the constructed π/A
142
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
isotherms was carried out with WinLB 1.61 software provided by the manufacturer and surface pressure (π) was plotted as a function of the mean molecular area ˚ 2 ). Monolayer phase transitions and collapse were (A identified using the first derivative of surface pressure with respect to mean molecular area. All transitions and collapse areas reported here are onset parameters. Monolayer compressibility modulus at a given surface pressure was calculated from the π/A isotherms using a two dimensional form of the bulk compressibility modulus equation (Ali et al., 1998): dπ Kπ = (−Aπ ) dAπ where K is the compressibility modulus in mN/m and ˚ 2 at a certain surface Aπ the mean molecular area in A pressure π.
3. Results
Fig. 2. In vitro transfection activity of 1,2lmt[5] formulations at 1:1, 2:1 and 4:1 ±charge ratios in B16-F0 mouse melanoma cells as compared to DC-Chol/DOPE (6/4) at 2:1 ±charge ratio. All formulations were tested on the same day and the data presented are the average of three wells. -galactosidase levels are expressed as mU/well. Each well is equivalent to 150 l of cell lysate.
3.1. In vitro transfection activity and toxicity The aqueous dispersions of the N,N -diacyl-1,2diaminopropyl-3-carbamoyl-(dimethylaminoethane) analogs were tested for their ability to deliver the pUC19 plasmid DNA containing the -galactosidase reporter gene in B16-F0 cells. All analogs but the dioleoyl derivative 1,2lmt[5] elicited no activity at any of the three charge ratios studied in the presence or absence of the helper lipids (Fig. 2). Lipoplexes made with 1,2lmt[5] in the absence of helper lipids resulted in a moderate increase of -galactosidase levels compared with the DNA control. Inclusion of the helper lipid DOPE resulted in enhanced expression levels of the reporter enzyme with the highest level occurring around a ±charge ratio of 2. Similarly, maximum activity in the presence of cholesterol was observed at a ±charge ratio of 2. However, the activity was much lower compared to the cationic lipid/DOPE formulation. The cytotoxicity of lipoplexes prepared from the different formulations of the five analogs was evaluated in B16-F0 cells using the MTT cytotoxicity assay. No significant cytotoxicity was detected for lipoplexes prepared with the saturated fatty acyl chain analogs (data not shown). As indicated in Fig. 3, lipoplexes prepared with 1,2lmt[5] alone were well tolerated by the cells
with a cell survival percentage exceeding 87%. Formulations containing DOPE resulted in a minor decrease in the cell survival percentage. It is worth noting that all formulations showed a charge and dose dependent cytotoxicity. In other words, increasing the cationic lipid content in the lipoplex resulted in an increased cytotoxicity.
Fig. 3. Cytotoxicity of 1,2lmt[5] formulations at 1:1, 2:1 and 4:1 ±charge ratios in B16-F0 mouse melanoma cells as compared to DC-Chol/DOPE (6/4) at 2:1 ±charge ratio. Data presented are the average of three wells. Results are expressed as percentage survival of treated cell taking cells treated with DNA alone to be 100%.
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
143
Fig. 4. Gel electrophoresis of lipoplexes prepared with different lipid formulations at the following ±charge ratios: 0.5, 1, 2, 4, 6 and 8, respectively. The outermost wells were loaded with plasmid DNA as a negative control. (A) Lipoplexes prepared with cationic lipids alone, (B) cationic lipid/DOPE (6:4 molar ratios) and (C) cationic lipid/DOPE/cholesterol (6:4:2 molar ratios).
3.2. Agarose gel electrophoresis The ability of the N,N -diacyl-1,2-diaminopropyl3-carbamoyl-(dimethylaminoethane) cationic derivatives to complex with plasmid DNA was investigated by a gel electrophoresis retardation assay. As seen in Fig. 4, only 1,2lmt[5] was able to inter-
act with and completely retard plasmid DNA at a ±charge ratio of 4. Interestingly, in the presence of DOPE or DOPE and cholesterol, DNA migration was completely inhibited at a ±charge ratio of 2. Cationic lipids that failed to mediate transfection activity also failed to complex with DNA at all charge ratios.
Fig. 5. Cationic lipid induced DNA condensation monitored by an ethidium bromide displacement assay.
144
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Table 1 Particle size distribution of 1,2lmt[5] liposomal formulations and their lipoplexes 1,2lmt[5]
Liposomes Lipoplexes (1:1) Lipoplexes (2:1) Lipoplexes (4:1)
1,2lmt[5]/DOPE
1,2lmt[5]/d/chol
Mean diameter (nm)
P.I.a
Mean diameter (nm)
P.I.a
Mean diameter (nm)
P.I.a
156.7 194.8 2941 Ab
0.315 0.252 0.557 –
125.4 268.1 1216 275
0.224 0.33 1.0 0.382
136.9 248.2 2360 215
0.250 0.314 0.671 0.241
Particle size measurements were carried out in 40 mM Tris buffer pH 7.2 as described in the experimental section. Lipid dispersions were prepared at 0.6 mM, which is well above their CMC. a Values of polydispersity index (P.I.) above 0.18 denote polydisperse samples. Ab denotes aggregates of size above 6 m. All other lipid formulations showed aggregates >6 m with polydispersity indices exceeding 1.
3.3. Ethidium bromide displacement assay Agarose gel electrophoresis assay shows whether or not the cationic lipid is capable of complexing plasmid DNA, but it does not give any information about the efficiency with which the lipid condenses DNA. Further characterization of the size of the complex and the extent of condensation of DNA requires the use of other techniques. In order to investigate the extent of cationic lipid induced condensation of plasmid DNA, reduction of ethidium bromide fluorescence signal upon titrating the plasmid DNA with aliquots of cationic lipid aqueous dispersions was monitored. All cationic derivatives, including the active 1,2lmt[5] analog, with or without the helper lipid DOPE displaced less than 10% of ethidium bromide at a ±charge ratio of 2 (Fig. 5). Similar results were obtained when an aliquot of either 1,2lmt[5] or 1,2lmt[5]/DOPE equivalent to a ±charge ratio of 2 was added all at once to the plasmid DNA solution suggesting that DNA condensation by 1,2lmt[5] is not efficient. 3.4. Dynamic light scattering The particle size distribution of the dioleoyl derivative dispersions and lipoplexes is summarized in Table 1. Liposomes prepared with the dioleoyl derivative exhibited a particle size distribution below 0.2 m, while the other transfection-inefficient lipids resulted in aggregates of particle sizes well above the measurable range of the instrument. Neither prolonged incubation nor incubation at higher temperatures (∼80 ◦ C) affected the particle size of the aggregates. Upon complexing 1,2lmt[5] formulations with plasmid DNA, no significant increase in particle size was observed with
lipoplexes prepared at a 1:1 charge ratio. Increasing the charge ratio resulted in a tremendous increase in particle size for 1,2lmt[5] lipoplexes in the absence of the helper lipids. Contrary to this, 1,2lmt[5]/DOPE and 1,2lmt[5]/DOPE/chol lipoplexes aggregated at 2:1 charge ratio but exhibited a small particle size distribution when the charge ratio was increased to 4:1. 3.5. Monolayer studies The interfacial properties of the 1,2lmt cationic lipids in isolation were investigated using the Langmuir film balance technique. Aliquots of cationic lipids in CHCl3 were spread on the surface of the same buffer system used to prepare the aqueous dispersions for transfection and for the other physicochemical characterization studies, i.e. 40 mM Tris buffer, pH 7.2. Representative compression isotherms of the five analogues are shown in Fig. 6. At 23 ◦ C, only the unsaturated derivative 1,2lmt[5] formed monolayers that existed in an all liquid-expanded state with ˚ 2 and molecular collapse area and pressure of 59.47 A 38.67 mN/m, respectively (Fig. 7E and Table 2). On the other hand, monolayers of the saturated derivatives either exhibited an all liquid-condensed state or a two-dimensional phase transition toward a liquidcondensed state in their π/A isotherms. More specifically, the compression isotherm of the shortest derivative, 1,2lmt[1], exhibited a typical liquid-expanded to liquid-condensed transition observed at a molecular ˚ 2 and 21.63 mN/m, respecarea and pressure of 65.74 A tively. Further reduction of the surface area resulted in two overlapping maximums in the 1st derivative graph (Fig. 7A). The onset of the first maximum is at ˚ 2 and 42.13 mN/m, and the second at 32.9 A ˚2 39.27 A
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
145
tions (Fig. 7B). Upon compressing the film beyond that transition, surface pressure rose only marginally ˚ 2 , after within an approximate area range 105−75 A which, the surface pressure sharply increased indicating the presence of a chain ordered state. Monolayer collapse was reached at a molecular area and pres˚ 2 and 40 mN/m, respectively (Table 2). sure of 40.74 A The 1,2lmt[3] and 1,2lmt[4] analogs formed all liquidcondensed state monolayers at 23 ◦ C with a collapse ˚ 2 and pressure of molecular area of 39.78 and 42.13 A 39.46 and 39.79 mN/m, respectively (Fig. 7C and D, Table 2). Interfacial compressibility moduli (K) values of monolayers prepared from the different 1,2lmt cationic lipids, calculated at molecular collapse area are presented in Table 2. Comparison of the K values of the different analogues revealed that 1,2lmt[5] monolayers exhibited the greatest compressibility with a K value of 80.36 ± 10.4. For all other saturated derivatives, K was found to increase with increasing chain length, the maximal value being that of the distearoyl derivative 1,2lmt[4]. However, K values cannot be used to judge compressibilities of lipids that have very different molecular dimensions. The final K values could sometimes be misleading in the sense that these values are the slope dΠ/dA multiplied by the area. In Table 2 as well as in Fig. 7, it can be seen that ˚2 1,2lmt[5] monolayer collapsed at an area that is 20 A larger than that for 1,2lmt[1]. In spite of the smaller slope of 1,2lmt[5], its larger collapse area resulted in a K value close to that of 1,2lmt[1]. For comparison purposes between monolayers with varying collapse areas, one should also pay attention to the rate of change of surface pressure with respect to the change of
Fig. 6. Representative π/A isotherms of 1,2lmt cationic derivatives at the air/water interface at 23 ◦ C. Tris buffer (40 mM, pH 7.2) was used as the subphase. Isotherms were constructed using KSV computer-controlled Minitrough film balance with a compression rate of 9.99 mm/minute. (1) 1,2lmt[1]; (2) 1,2lmt[2]; (3) 1,2lmt[3]; (4) 1,2lmt[4]; and (5) 1,2lmt[5].
and 54 mN/m molecular area and surface pressure, respectively. Monolayer collapse was most probably initiated at the first maximum and further increase in surface pressure beyond that area is due to a multilayer, the formation of which was completed at a molecu˚ 2 . The presence of two maximums lar area of 32.9 A in the dΠ/dA versus mean molecular area (MmA) plot was very clear for 1,2lmt[1] and less evident for the other long chain derivatives at the temperature studied (Fig. 7). The compression isotherm of 1,2lmt[2] exhibited a similar two-dimensional phase transition but at a much lower surface pressure (Fig. 6, isotherm 2). In this case, the transition was broad and its onset was difficult to determine by the first derivative calcula-
Table 2 Monolayer parameters of 1,2lmt amphiphilic seriesa ˚ 2) MmA (A 1,2lmt[1] 1,2lmt[2] 1,2lmt[3] 1,2lmt[4] 1,2lmt[5]
39.27 65.47c 40.74 39.78 42.13 59.47
± ± ± ± ± ±
1.6 1.03 0.42 0.78 0.62 2.17
Π c (mN/m) 42.13 21.63c 40.00 39.46 39.79 38.67
± ± ± ± ± ±
1.47 1.24 1.28 1.15 2.06 0.74
dΠ/dAb
K (mN/m) 97.83 60.26c 134.67 157.75 160.2 80.36
± ± ± ± ± ±
15.54 3.88 18.15 8.62 20.8 10.4
2.48 0.91c 3.2 3.85 3.75 1.33
± ± ± ± ± ±
0.39 0.09 0.35 0.27 0.56 0.18
Phase state
˚ 2) ‘Lift off’ area (A
L2
127.99 ± 7.19
L2 L2 L2 L1
113 68.71 78.04 151.89
± 3.76 ± 4.45 ± 6.22 ± 10.08
a Measured in 40 mM Tris buffer (pH 7.2) at 23 ◦ C. All parameters are ‘onset’ transition values. MmA denotes Mean molecular Area expressed ˚ 2 per molecule. L1 denotes liquid-expanded state (Adamson, 1982). L2 denotes liquid-condensed state. in A b Values calculated at monolayer collapse and at other two-dimensional phase transitions. c Phase transition denoted by a maximum in the dΠ/dA plot as a function of mean molecular area.
146
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Fig. 7. Surface pressure (Π) and first derivative of surface pressure (dΠ/dA) as a function of mean molecular area of different 1,2lmt cationic derivatives. (A) 1,2lmt[1]; (B) 1,2lmt[2]; (C) 1,2lmt[3]; (D) 1,2lmt[4] and (E) 1,2lmt[5].
MmA (shown in Table 2 as dΠ/dA) of the monolayer as an indicator of monolayer state and compressibility. Thus, considering both the compressibility modulus and the corresponding dΠ/dA values, the in-plane
compressibility of the five analogues is in the order of 1,2lmt[5] > 1,2lmt[1] > 1,2lmt[2] > 1,2lmt[3] ≥ 1,2lmt[4], with 1,2lmt[3] and 1,2lmt[4] films being the least compressible.
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
4. Discussion The aim of this study was to delineate the physicochemical properties of double-chained cationic surfactants required for high plasmid DNA transfection efficiencies. To this end, a new series of cationic lipids, N,N -diacyl-1,2-diaminopropyl-3carbamoyl-(dimethylaminoethane) derivatives were synthesized by systemically varying the length and degree of unsaturation of the hydrophobic anchor. The fatty acyl chains were linked to the ionizable cationic head group via amide bonds that are biodegradable but significantly more stable than the easily hydrolyzed ester bond found in natural phospholipids. In vitro transfection activity studies conducted showed that only lipoplexes formulated with the dioleoyl derivative 1,2lmt[5] mediated significant in vitro transfection activity. All the saturated derivatives were transfection inefficient at all charge ratios in the presence and absence of the helper lipids. Optimum activity of 1,2lmt[5] was obtained in the presence of DOPE at a ±charge ratio of 2 (Fig. 2). In agreement with the transfection results, 1,2lmt[5] was the only lipid found to complex and retard DNA migration upon gel electrophoresis (Fig. 4). Despite the efficient complexation, no significant condensation of plasmid DNA took place as indicated by fluorescence spectroscopy measurements (Fig. 5). In vitro transfection, gel electrophoresis and fluorescence spectroscopy suggest that inefficient condensation of plasmid DNA by cationic lipids does not abolish in vitro transfection activity, a phenomenon that was previously observed by Banerjee et al. (2001). Loose association of plasmid DNA with cationic lipids is enough to promote significant in vitro transfection activity. However, whether this is also valid for in vivo transfection is doubtful since more stable lipoplexes are required to overcome the inhibitory effect of serum. To gain further insight on the transfection behavior of the different analogues and their formulations, particle size distribution measurements were conducted. Aqueous dispersions of the transfection inefficient lipids exhibited particle size distributions > 6 m. Such large particle size distribution indicates that these lipids failed to hydrate and form liposomes. On the other hand, liposomes prepared from 1,2lmt[5] in the presence or absence of helper lipids exhibited particle size distribution smaller than 200 nm. When these liposomes were complexed with plasmid DNA, aggrega-
147
tion was evident only at a ±charge ratio of 2. Since this charge ratio was the optimum ratio for transfection activity, it appears that a particle size distribution between 1 and 2 m correlates with enhanced in vitro activity. Besides physicochemical characterization of the lipoplexes formed by these novel cationic lipids, the Langmuir film balance technique was used in an attempt to reveal the interfacial properties of cationic lipids pertinent to efficient in vitro transfection activity. First derivative analysis of the π/A isotherms of the saturated derivatives revealed that perturbation of monomolecular films prepared from these lipids started ˚ 2 and 40 mN/m, at an area and pressure around 40 A respectively (Table 2). Interestingly, 1,2lmt[5] monolayer collapsed at a surface pressure similar to that observed for the saturated analogues but its collapse area ˚ 2 larger at 23 ◦ C. Since all was approximately 19–20 A five derivatives share the same head group composition, the origin of this difference in molecular collapse areas between the dioleoyl derivative and the saturated derivatives is the increased fluidity and hydration of 1,2lmt[5] due to the double bond it possesses in its acyl chains. With regard to the saturated derivatives, strong Van der Waals attractions between the acyl chains and intermolecular hydrogen bonding between the polar head group of adjacent lipid molecules in the highly compressed film are expected to lead to removal of the tightly bound interfacial water and tight packing of the dehydrated molecules in the film. On the contrary, the expanded unsaturated acyl chains of the dioleoyl derivative will prevent proper ordering of film components necessary for the exclusion of interfacial water and the subsequent intermolecular hydrogen bonding to take place (Ali et al., 1991). If a water molecule at the interface occupies a surface area Aswater = 9.65 (Feng et al., 1994), then there is at least 2 extra water molecules associated with each 1,2lmt[5] molecule at the collapse relative to the other saturated derivatives. The transfection potent lipid 1,2lmt[5] also formed the most elastic monolayer at 23 ◦ C, as indicated by its low compressibility modulus. All other transfection inefficient saturated derivatives were characterized by large compressibility moduli, which were found to be directly proportional to the chain length, being maximal for 1,2lmt[4] (Table 2). The correlation between molecular fluidity and improved transfection activity was first introduced by
148
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149
Akao et al. (1994). Their findings suggested that a liquid-crystalline to gel phase transition temperature below 37 ◦ C is required for cationic lipid-mediated delivery of plasmid DNA in vitro. Besides molecular fluidity, Bennett et al. investigated the effect of molecular hydration on in vivo transfection activity with a series of dimyristoyl and dioleoyl derivatives possessing several head group combinations (Bennett et al., 1997). In the presence of DOPE, they found the highly hydrated dioleoyl derivatives to be superior for in vivo transfection activity. They, along with others, have concluded that 14-carbon or oleoyl chains satisfy the hydration, fluidity and bilayer stability requirements for efficient transfection efficiency (Felgner et al., 1994; Ilies et al., 2004; Laxmi et al., 2001; Wang et al., 1998). The results of the current study agree with the findings stated above. The only lipid that satisfied the conditions required for efficient transfection activity was the dioleoyl derivative. On the other hand, the myristoyl derivative 1,2lmt[2] was transfection inefficient even in the presence of DOPE. This can be easily explained by the rigidity and lack of minimum hydration of this lipid aqueous dispersion. As mentioned earlier, particle size distribution measurements showed that this lipid failed to form liposomes upon hydration. Moreover, Langmuir film balance studies performed with this lipid showed that it formed films that exhibited a liquid-condensed phase at a surface pressure value comparable to that usually found in phospholipid bilayers, i.e. 30–35 mN/m (Ali et al., 1998 and references therein). In part, this might be due to the introduction of hydrogen bond functionalities in the polar part of the 1,2lmt[2] that significantly altered its phase behavior making it a more rigid molecule. In summary, five different N,N -diacyl-1,2diaminopropyl-cationic derivatives were synthesized and tested for efficient in vitro transfection activity. By means of the Langmuir film balance technique, an extensive investigation of their interfacial properties was launched in an attempt to correlate these properties with the transfection activity. The current findings suggest that cationic lipids that can interact with plasmid DNA and promote transfection activity are characterized by the following: First, acyl chain fluidity indicated by a liquid-expanded monolayer state and second, high interfacial elasticity indicated by a low compressibility modulus expressed in mN/m.
Acknowledgment The authors are grateful to Dr. Christopher M. Hadad in the Department of Chemistry, Ohio State University, for performing the mass spectroscopy analysis.
References Adamson, R.A., 1982. Physical Chemistry of Surfaces. Wiley, New York. Akao, T., Nakayama, T., Takeshia, K., Ito, A., 1994. Design of a new cationic amphiphiles with efficient transfection ability. Biochem. Mol. Biol. Int. 34, 915–920. Ali, S., Brockman, H.L., Brown, R.E., 1991. Structural determinants of miscibility in surface films of galactosylceramide and phosphatidylcholine: effect of unsaturation in the galactosylceramide acyl chain. Biochemistry 30, 11198–11205. Ali, S., Samby, J.M., Momsen, M.M., Brockman, H.L., Brown, R.E., 1998. Acyl chain-length asymmetry alters the interfacial elastic interactions of phosphatidylcholines. Biophys. J. 74, 338–348. Banerjee, R., Mahidhar, Y.V., Chaudhuri, A., Gopal, V., Rao, N.M., 2001. Design, synthesis, and transfection biology of novel cationic glycolipids for use in liposomal gene delivery. J. Med. Chem. 44, 4176–4185. Bennett, M.J., Aberle, A.M., Balasubramaniam, R.P., Malone, J.G., Malone, R.W., Nantz, M.H., 1997. Cationic lipid-mediated gene delivery to murine lung: correlation of lipid hydration with in vivo transfection activity. J. Med. Chem. 40, 4069– 4078. Bordi, F., Cametti, C., De Luca, F., Gili, T., Gaudine, D., Sennato, S., 2003. Charged lipid monolayers at the air–solution interface: coupling to polyelectrolytes. Colloid Surf. B 29, 149–157. Borenstain, V., Barenholz, Y., 1993. Characterization of liposomes and other lipid assemblies by multiprobe fluorescence polarization. Chem. Phys. lipids 64, 117–127. Congiu, A., Pozzi, D., Esposito, C., Castellano, C., Mossa, G., 2004. Correlation between structure and transfection efficiency: a study of DC-Chol DOPE/DNA complexes. Colloid Surf. B 36, 43–48. Feng, S., Brockman, H.L., MacDonald, R.C., 1994. On osmotic-type equations of state for liquid-expanded monolayers of lipids at the air–water interface. Langmuir 10, 3188–3194. Felgner, J.H., Kumar, R., Sridhar, C.N., Wheeler, C.J., Taai, Y.J., Border, R., Ramsey, P., Martin, M., Felgner, P.L., 1994. Enhanced gene delivery and mechanism studies with a novel series of cationic lipid formulations. J. Biol. Chem. 269, 2550–2561. Ilies, M.A., Seitz, W.A., Balban, A.T., 2002. Cationic lipids in gene delivery: principles, vector design and therapeutical applications. Curr. Pharm. Design 8, 2441–2473. Ilies, M.A., Seitz, W.A., Ghiviriga, I., Johnson, B.H., Miller, A., Thompson, E.B., Balaban, A.T., 2004. Pyridinium cationic lipids in gene delivery: a structure-activity correlation study. J. Med. Chem. 47, 3744–3754.
A. Aljaberi et al. / Chemistry and Physics of Lipids 133 (2005) 135–149 Laxmi, A.A., Vijayalakshmi, P., Kaimal, T.N.B., chaudhuri, A., Ramadas, Y., Rao, N.M., 2001. Novel non-glycerol-bases cytofectins with lactic acid derived head groups. Biochem. Biophs. Res. Commun. 289, 1057–1062. Lin, A.J., Slack, N.L., Ahmad, A., George, C.X., Samuel, C.E., Safinya, C.R., 2003. Three-dimensional imaging of lipid genecarriers: membrane charge density controls universal transfection behavior in lamellar cationic liposome-DNA complexes. Biophy. J. 84, 3307–3316. Miller, A.D., 1998. Cationic liposomes for gene therapy. Angew. Chem. Int. Ed. 37, 1768–1785. Rejman, J., Oberle, V., Zuhorn, I.S., Hoekstra, D., 2004. Sizedependent internalization of particles via the pathways of clathrin- and caveolae-mediated endocytosis. Biochem. J. 377, 159–169. Ross, P.C., Hui, S.W., 1999. Lipoplex size is a major determinant of in vitro lipofection efficiency. Gene Ther. 6, 651–659. Safinya, C.R., 2001. Structures of lipid–DNA complexes: supramolecular assembly and gene delivery. Curr. Opin. Struc. Bio. 11, 440–448. Savva, M., Duda, E., Huang, L., 1999. A genetically modified recombinant tumor necrosis factor-␣ conjugated to the distal terminal of liposomal surface grafted polyethyleneglycol chains. Int. J. Pharm. 184, 45–51.
149
Sheikh, M., Feig, J., Gee, B., Li, S., Savva, M., 2003. In vitro lipofection with novel series of symmetric 1,3-dialkoylamidopropanebased cationic surfactants containing single primary and tertiary amine polar head groups. Chem. Phys. Lipids 124, 49–61. Singh, R.S., Mukherjee, K., Banerjee, R., Chaudhuri, A., Hait, S.K., Moulik, S.P., Ramadas, Y., Vijayalakshmi, A., Rao, N.M., 2002. Anchor dependency for non-glycerol based cationic lipofectins: mixed bag of regular and anomalous transfection profiles. Chem. Eur. J. 8, 900–909. Sunamoto, J., Goto, M., Iwamoto, K., Kondo, H., Sato, T., 1990. Synthesis and characterization of 1,2-dimyristoylamido1, 2-deoxyphosphatidylcholine as an artificial boundary lipid. Biochem. Biophys. Acta 1024, 209–219. Wang, J., Guo, X., Xu, Y., Barron, L., Szoka Jr., F.C., 1998. Synthesis and characterization of long chain alkyl acyl carnitine esters. Potentially biodegradable cationic lipids for the use in gene delivery. J. Med. Chem. 41, 2207–2215. Welzel, P.B., Weis, I., Schwarz, G., 1998. Sources of error in Langmuir trough measurements: Wilhelmy plate effects and surface curvature. Colloid Surf. A 144, 229–234. Zhang, Y., Anchordoquy, T.J., 2004. The role of lipid charge density in the serum stability of cationic lipid/DNA complexes. Biochem. Biophys. Acta 1663, 143–157.