Cationic Lipid Polymerization as a Novel Approach for Constructing

Bioconjugate Chem. 2001, 12, 251−257

251

Cationic Lipid Polymerization as a Novel Approach for Constructing New DNA Delivery Agents Jian Wu,† Mike E. Lizarzaburu,‡ Mark J. Kurth,‡ Li Liu,† Henning Wege,† Mark A. Zern,† and Michael H. Nantz*,‡ Department of Internal Medicine, Transplant Research Institute, University of CaliforniasDavis Medical Center, Sacramento, California 95817, and Department of Chemistry, University of CaliforniasDavis, One Shields Avenue, Davis, California 95616. Received August 10, 2000; Revised Manuscript Received January 22, 2001

In vivo gene delivery mediated by cationic lipids is often compromised by aggregation due to complexation with proteins in the blood. To improve the stability of cationic lipid-DNA complexes, the present study aimed to develop a novel approach in which a poly(cationic lipid) (PCL) is utilized to form stable cationic polyplexes for gene transfection. Hydrogenation of the acrylamide analogue of βAE-DMRI, the polymerizable precursor of PCL, provided a monomeric lipid derivative (MHL) which was used for direct comparison of corresponding lipoplex stability, toxicity, and transfection activity. Various formulations of cationic liposomes, such as MHL, MHL-cholesterol (Chol), PCL, PCL-Chol, DOTAP-Chol, and commercially available lipofectamine were generated and examined in this study. The new poly(cationic lipid) did not display any significant toxicity to rat hepatocytes or Hep G2 cells as indicated by an LDH leakage assay. Furthermore, PCL was significantly less toxic than MHL, DOTAP-Chol or lipofectamine. Suspensions of PCL were resistant to aggregation even after 24 h of exposure to solutions containing 50 and 100% fetal bovine serum (FBS). In contrast, suspensions of lipofectamine extensively aggregated after 24 h of exposure to 50% FBS. To examine the influence of lipid polymerization on gene transfer activity, liposome-mediated transfections of a luciferase vector (pGL3) were performed in Hep G2 and Alexander cell lines. The luciferase activity of the PCL formulations in Hep G2 cells were similar to those of the MHL, DOTAP-Chol and lipofectamine formulations, demonstrating that lipid polymerization does not compromise transfection activity. In comparison to the monomeric precursor MHL and to the industry transfection standards DOTAP and lipofectamine, the novel poly(cationic lipid) exhibited the lowest cytotoxicity, was the most resistant to serum-induced aggregation and had comparable transfection activity when coformulated with cholesterol. This novel polymerization approach for the development of stable and active polyplexes may prove a valuable alternative for in vivo gene delivery.

INTRODUCTION

The prospect for improving nonviral methods of DNA complexation and transfer to mammalian cells continues to motivate researchers in a variety of disciplines. Among the many constructs that have been used to facilitate the intracellular delivery of DNA, cationic lipids and cationic polymers have received the greatest attention (1, 2), and consequently, the technology has since advanced to clinical trials (3, 4). Over the course of the past several years it has become clear that the use of cationic lipid systems for in vivo DNA transfection is compromised by substantial problems that are uniquely dependent on the route of lipid-DNA complex (lipoplex) administration. For example, the systemic administration of lipoplex preparations results in their rapid clearance due to destabilization arising from interactions with serum proteins (5). This obstacle prompted the development of stabilized plasmid lipid particles wherein poly(ethylene glycol) (PEG) chains are conjugated to the surface of lipoplex preparations to provide a measure of steric protection (6). Issues of practicality also point toward the development of more stable lipoplex preparations. For * To whom correspondence should be addressed. Phone: (530) 752-6357. Fax: (530) 752-8995. E-mail: [email protected]. † University of CaliforniasDavis Medical Center. ‡ University of CaliforniasDavis.

instance, the majority of transfection protocols require that lipoplexes are freshly prepared prior to use to avoid problems associated with lipoplex aggregation (7). Given these and other challenges, we have focused much of our attention on developing new lipid-based constructs to improve lipoplex transfection activity (8, 9). One approach that has been used to enhance stability and prevent aggregation of lipid-based assemblies is liposome polymerization (10, 11). We have recently developed methods for the polymerization of a novel cationic acrylamide lipid and reconstitution of the resultant poly(lipid) to yield stable cationic vesicles (12). We now report our observations on the ability of this cationic polymer to mediate gene delivery. Although cationic polymers such as poly(L-lysine) (13, 14), poly(ethyleneimine) (15, 16), poly(methacrylate) (17), and polyamidoamine dendrimers (18, 19) have been well studied, to our knowledge the use of a poly(cationic lipid) as a transfection agent remains unexplored. We report herein the first examples of poly(lipid)-mediated DNA transfection and the results of stability and toxicity studies using the new cationic polymer. The attributes of lipid polymerization are defined by comparison of the poly(cationic lipid) (PCL)1 with lipoplex preparations derived from structurally analogous monomeric cationic lipids.

10.1021/bc000097e CCC: $20.00 © 2001 American Chemical Society Published on Web 03/06/2001

252 Bioconjugate Chem., Vol. 12, No. 2, 2001 Scheme 1a

a H , 10% Pd‚C, EtOAc, room temperature, 6 h. b Polymer2 ization: (1) H2O, dodecylmercaptan (10% w/w), sonication, 50 °C, 20 min; (2) AAPH (0.1 eq), 80 °C, 15 h.

MATERIALS AND METHODS

General Experimental Procedures. CH2Cl2 was distilled from CaH2 prior to use. After reaction workup, solutions were dried using Na2SO4 and solvents were subsequently removed by rotary evaporation. NMR spectra were recorded with a General Electric QE-300 spectrometer (1H at 300 MHz, 13C at 75 MHz). Infrared spectra were recorded with a Mattson Genesis II FTIR spectrometer. Melting points are uncorrected. Mass analysis was performed by UCR Mass Spectrometry (Riverside, CA). All sonications were performed in a bath sonicator (Laboratory Supplies Inc., Hicksville, NY). Lipids. The cationic liposomal formulation lipofectamine (DOSPA-DOPE, 3:1 w/w) was obtained from Gibco Life Technologies (Gaithersburg, MD). N-(2,3(Dioleoyloxy)propyl)-N,N,N-trimethylammonium chloride (DOTAP) was purchased as a chloroform solution from Avanti Polar Lipids (Alabaster, AL). Cholesterol (Chol) was purchased from Sigma Chemical Company (St. Louis, MO). Synthesis of the polymerizable acrylamide lipid (AL) (Scheme 1) proceeded from commercially available 3bromo-1,2-diol by bis(esterification) with myristoyl chloride (12). Bromide displacement using N,N′-dimethylethylenediamine (15 equiv, DMF, 70 °C, 45 min) installed the amine-linker required for subsequent attachment of the polymerizable acrylamide group. Direct treatment of the crude displacement product with acryloyl chloride (1.5 equiv, CH2Cl2, room temperature, 2 h) followed by chromatographic purification afforded the corresponding acrylamide lipid. Methyl iodide quaternization and subsequent chloride counterion exchange using an established protocol (20) gave cationic acrylamide lipid AL. Synthesis of PCL (Scheme 1) was accomplished by first preparing a 5 mL aqueous liposome suspension containing AL (50 mg, 73.0 µmol) and n-dodecylmercaptan (5 mg, 25.0 µmol). The lipid mixture was vigorously mixed for 30 s and then sonicated for 20 min at 50 °C to afford 1 Abbreviations: AL, acrylamide lipid; Chol, cholesterol; DOPE, L-R dioleoyl phosphatidylethanolamine; DOTAP, N-(2,3(dioleoyloxy)propyl)-N,N,N-trimethylammonium chloride; FBS, fetal bovine serum; LDH, lactate dehydrogenase; MHL, monomeric hydrogenated lipid; PCL, poly(cationic lipid).

Wu et al.

a clear suspension. To the liposome suspension was added 2,2′-azobis(2-methylpropionamidine) dihydrochloride (AAPH) (2.0 mg, 7.3 µmol) as a solution in 0.5 mL of H2O. The resultant mixture was degassed using nitrogen and then placed in a constant-temperature bath heated to 80 °C. After 15 h, the solvent was removed and the polymerized product was analyzed by IR, 1H, and 13C NMR spectroscopy; IR (neat) 2917, 2856, 1747, 1638, 1169 cm-1; 1H NMR (CDCl3) δ 0.83 (t, J ) 6.5 Hz, 6H), 1.25 (m, 40H), 1.57 (m, 4H), 2.32 (m, 4H), 3.39 (m, 3H), 3.53 (m, 6H), 3.87 (m, 4H), 4.10 (m, 2H), 4.31 (m, 1H), 4.53 (m, 1H), 5.66 (m, 1H); 13C NMR (CDCl3) δ 14.0, 22.6, 24.7, 29.0-29.6 (four signals) 31.9, 33.8, 34.1, 36.6, 42.3, 51.7, 62.1, 63.1, 64.0, 65.5, 172.7, 173.1. In all cases examined, the vinyl proton signals of monomer AL (δ 6.57, 6.29, and 5.73 ppm) were absent in the product spectra, suggesting at least a 95% conversion of monomer to polymer. The poly(cationic lipid) was stored at 4 °C without detectable decomposition for periods up to six months. AL was hydrogenated to obtain a more representative monomer for direct comparison with PCL. A solution of AL (50 mg, 79 µmol) in ethyl acetate (2 mL) at room temperature was treated with 10% palladium on carbon (20 mg). The suspension was placed under an atmosphere of hydrogen and stirred 6 h at which time the suspension was diluted with ethyl acetate and filtered through a short pad of Celite. Removal of the solvent afforded the monomeric hydrogenated lipid (MHL) (46 mg, 92%) as an amorphous white solid, mp ) 109-110 °C; TLC, Rf ) 0.32 (9:1, CH2Cl2:MeOH); IR (neat) 2921, 2853, 1747, 1635, 1172 cm-1; 1H NMR (CDCl3) δ 0.83 (t, J ) 6.2 Hz, 6H), 1.06 (t, J ) 6.7 Hz, 3H), 1.22 (m, 40H), 1.54 (m, 4H), 2.29 (m, 6H), 3.16 (s, 3H), 3.38 (s, 3H), 3.41 (s, 3H), 3.86 (m, 2H), 4.02 (m, 4H), 4.34 (d, J ) 13.5 Hz, 1H), 4.46 (dd, J ) 1.5, 13.2 Hz, 1H), 5.57 (m, 1H); 13C NMR (CDCl3) δ 5.0, 13.9, 22.5, 24.7, 25.7, 29.5-29.0 (four signals), 31.8, 33.8, 34.1, 35.5, 42.3, 51.0, 61.5, 63.3, 63.9, 65.8, 171.5, 172.7, 175.0; HRMS (C39H77O5N2, M+) calcd, 653.5832; found, 653.5865. Liposome Formulations. Monomer MHL and polymer PCL were hydrated separately using 10 mM TrisHCl buffer containing NaCl (150 mM) at a concentration of 1 mg/mL (1.5 mM cationic construct). The hydrated lipid suspensions were vigorously mixed and then sonicated for 5 min at 50 °C to afford turbid liposome suspensions. To generate the cholesterol-containing liposome formulations of MHL, PCL, or DOTAP, the lipids were dissolved separately in chloroform and mixed with cholesterol at a 1:1 molar ratio, except for DOTAP, which was mixed with cholesterol in a 3:1 molar ratio. The chloroform was evaporated and the thin lipid films were dried using a stream of nitrogen followed by drying under vacuum for 30 min. The MHL-Chol and DOTAP-Chol mixtures were subsequently hydrated using 10 mM TrisHCl buffer containing NaCl (150 mM) at a concentration of 1 mg/mL (1.5 mM cationic lipid) and then briefly sonicated at 50 °C. The PCL-Chol mixture was hydrated in identical fashion and at the same concentration, but was sonicated 20 min at 50 °C. Lipofectamine was purchased from Gibco Life Technologies and used as received. Cytotoxicity Study. Spraque-Dawley rats were purchased from Harlan Laboratory, Inc. (Indianapolis, IN), kept at 12-h light/dark cycles and fed a commercial chow diet. The protocol for rat hepatocyte isolation was approved by the UCD Institutional Animal Use and Care Administrative Advisory Committee. Rat hepatocytes were isolated by two-step collagenase digestion according

Polymerized Liposomes for Gene Transfection

Bioconjugate Chem., Vol. 12, No. 2, 2001 253

Figure 1. Cytotoxicity of liposomal formulations to isolated rat hepatocytes. Rat hepatocytes were exposed to cationic liposomes (final concentration of lipofectamine ) 14 µg/mL, final concentration of all other lipid agents ) 20 µg/mL) for 24 h. LDH release from the cells, an indication of cytotoxicity, expressed as percentage of cell death based on the control levels. Each data point represents the mean value of n ) 4 and the standard deviation from the mean. (**) p < 0.01 compared to MHL; (#) p < 0.05 compared to PCL. Chol ) cholesterol, DOTAP ) N-(2,3-(dioleoyloxy)propyl)-N,N,N-trimethylammonium chloride, LDH ) lactate dehydrogenase, Lipofect. ) lipofectamine, PCL ) poly(cationic-lipid), MHL ) monomeric hydrogenated lipid.

to a method reported previously (21) and cultured at 37 °C in 5% CO2 and air with Williams’ medium E with 10% fetal bovine serum (FBS). The cells were cultured overnight and washed with 10 mM PBS, pH 7.4, before being subjected to the cytotoxicity experiments. Hep G2 cells were cultured in MEM medium with 10% FBS, and the cell passage was performed one or 2 days before analysis in Costar 12-well plates (250 000 cells/well) until the cell culture was 80-90% confluent. All liposome formulations except lipofectamine were added directly to the cultured rat hepatocytes to reach a final cationic lipid concentration of 20 µg/mL. In the case of lipofectamine, the liposome formulation was added directly to the cultured rat hepatocytes to achieve a final concentration of 14 µg/mL. Experiments in Hep G2 cells were performed using several concentrations of cationic lipids (10, 20, and 40 µg/mL). The culture medium was sampled 24 h after the cells were exposed to the lipid agents. Lactate dehydrogenase (LDH) leakage from the cells was determined by a commercially available kit from Roche Molecular Biochemicals (Indianapolis, IN), and the cytotoxicity was calculated based on a formula provided by the manufacturer and expressed as percentage of cell death based on the control levels. Particle Size Determination: Stability Study. The particle sizes of liposome formulations were measured using a laser-based, submicron particle size analyzer from Beckman Coulter, Inc as previously described (22). To test the stability of these liposome suspensions against culture medium containing serum, 200 µL of each liposome suspension was added to 0.5 mL of Willians' medium E containing 10% FBS and then diluted with Tris-HCl buffer to a total volume of 3.5 mL. After equilibration for 2 min, the particle size was measured. To test the stability of MHL, PCL and lipofectamine at higher serum concentrations and longer time intervals, a series of three FBS solutions were prepared containing 10, 50, and 100% FBS by diluting FBS with MEM medium to the appropriate concentration. To these solutions was added a 50-100 µL aliquot of the cationic lipid suspensions. Particle size measurements and size distributions were recorded at 5, 30, 50, 75, and 120 min. A measurement was also taken at 24 h to examine the long-term effect of serum exposure. Transfection Protocols. Plasmid vector pGL3-Control which encodes firefly luciferase was purchased from

Promega, Inc. (Madison, WI) and transformed (amplified) into competent Escherichia coli cells from Gibco Life Science Technologies (Grand Island, NY). The plasmid DNA was extracted from overnight cultures of the competent cells and purified by affinity chromatography with an Endofree plasmid extraction kit from Qiagen, Inc. (Valencia, CA). The quality of the DNA was determined by UV spectroscopy and agarose gel electrophoresis (1.0% agarose gel) with 0.5 mg/mL ethidium bromide after cleavage by specific restriction endonucleases (23). The DNA concentration was quantitated spectrophotometrically. The plasmid DNA was frozen at -20 °C until immediately prior to use. The protocol for tansfection followed that recommended by the manufacturer for the use of lipofectamine. Hep G2 cells and Alexander cells were plated at 250 000 cells/ well on a standard Costar 12-well plate 48 h prior to transfection. Cells were approximately 60-70% confluent at the time of transfection. Solutions of plasmid DNA in 100 µL of Opti-MEM were mixed gently with 100 µL OptiMEM solutions containing cationic agents, and then the lipid-DNA suspensions were diluted to a final volume of 0.5 mL. Typically (e.g., using PCL and MHL), 20 µg of the cationic lipid was used to complex 2.0 µg of the plasmid DNA, representing a cationic:DNA phosphate charge ratio of 5:1. Lipofectamine was used at the lipid: DNA weight ratio of 14 µg of lipofectamine to 2 µg of DNA, the ratio determined for maximum expression in Hep G2 cells. Cationic lipid-DNA complexes were incubated for 15-25 min. The cells were washed twice with 2 mL of 10 mM PBS, pH 7.4, and then the 0.5 mL lipidDNA solutions were added to the cells. Six hours after adding the lipid-DNA complexes to the cell culture, the Opti-MEM medium was replaced with MEM containing 10% FBS. The transfected cells were lysed following two washes with PBS 48 h after transfection. The luciferase activity in the lysates was measured with a luciferase kit from Promega Inc. and a luminometer from EG&G Wallac (Gaithersburg, MD). Statistical Analysis. Particle size data were taken directly from the laser light scattering instrument. LDH leakage data and luciferase activity were evaluated by means of the one-way variance test and Newman-Keuls test for multiple comparisons between groups. A p-value of less than 0.05 was considered as statistically significant.

254 Bioconjugate Chem., Vol. 12, No. 2, 2001

Wu et al. Table 1. Particle Size Distribution of PCL Liposomes time (min)

no serum

10% serum

50% serum

100% serum

5 30 50 75 120 1440

333 ((76) 406 ((160) 349 ((130) 420 ((170) 301 ((37) 341 ((61)

347 ((60) 300 ((25) 291 ((36) 413 ((48) 359 ((49) 363 ((58)

309 ((45) 331 ((52) 348 ((40) 328 ((53) 312 ((54) 277 ((25)

372 ((66) 375 ((66) 458 ((72) 376 ((66) 390 ((72) 314 ((26)

Table 2. Particle Size Distribution of MHL Liposomes time (min)

Figure 2. Cytotoxicity of liposomal formulations to Hep G2 cells. Hep G2 cells were exposed to cationic liposomes of final cationic lipid/polymer concentrations ranging from ) 0-40 µg/ mL. LDH release from the cells, an indication of cytotoxicity, was measured after 24 h and is expressed as a percentage of cell death based on the control levels. (O) PCL; (b) PCL-Chol; (0) MHL; (9) MHL-Chol. Each data point represents the mean value of n ) 4, and the standard deviation from the mean. RESULTS AND DISCUSSION

To examine the influence of cationic lipid polymerization on cytotoxicity, we compared the liposome formulations of monomer MHL, MHL-cholesterol, polymer PCL, PCL-cholesterol, DOTAP-cholesterol, and lipofectamine in a cytotoxicity assay using rat hepatocytes (Figure 1). We examined the influence of cholesterol in these formulations since the addition of cholesterol to cationic liposome preparations such as DOTAP has been reported to improve both transfection activity (24) and resistance to serum (25). Isolated rat hepatocytes in primary cultures were treated with the cationic agents at the concentrations indicated in Figure 1, which are the concentrations that were found to be optimal for DNA transfection (vide infra). The culture medium was sampled after 24 h and LDH activity in the medium was assayed using a commercially available kit. Relative cytotoxicity was calculated based on the control experiment and expressed as the percentage of cell death. It is clear from Figure 1 that MHL alone caused significantly greater cell death compared to the control, and that the polymer PCL did not display obvious cytotoxicity to isolated hepatocytes at 20 µg/mL. Comparison of PCL with its monomeric analogue MHL reveals that lipid polymerization significantly reduced cytotoxicity to rat hepatocytes. This result is notable since cationic polymers are generally believed to elicit greater cytotoxic responses (26, 27). The cytotoxicity of the popular transfection formulations DOTAP-Chol (20 µg/ mL) and lipofectamine (14 µg/mL) also led to significant LDH leakage from the isolated hepatocytes, and their toxicity was more profound than PCL or PCL-Chol. To further examine the influence of cationic lipid polymerization on cytotoxicity, we exposed Hep G2 cells to liposome formulations of MHL, MHL-Chol, PCL, and PCL-Chol at varying concentrations (Figure 2). In each case, the culture medium was sampled at 24 h and LDH activity in the medium was assayed as previously described. All formulations were relatively nontoxic (