Polyethylenimine of various molecular weights as adjuvant for transfection mediated by cationic liposomes

Abstract

Over the last years significant progress has been made in non-viral gene delivery mediated by cationic liposomes. However, the results obtained are still far from being satisfactory regarding transfection efficiency, particularly when compared to that achieved using viral vectors. We have previously demonstrated that association of transferrin with cationic liposomes significantly improves transfection in a large variety of cells, both in vitro and in vivo. In this work, several strategies have been explored in order to further improve transfection mediated by transferrin-associated lipoplexes. To this regard, the effect on transfection of pre-condensation of DNA with polyethylenimine of low MWs (2.7, 2.0 and 0.8 KDa) at various N/P ratios, lipid composition, cationic lipid/DNA (+/-) charge ratio and the presence of a surfactant in the lipoplexes was investigated. Two different modes for preparing the liposomes were tested and the extent of cell association of their complexes with DNA as well as their capacity to protect the carried DNA were evaluated. Our results show that complexes generated from cationic liposomes prepared by the ethanol injection method in which the carried DNA was pre-condensed with low MW polyethylenimine are highly efficient in mediating transfection. The differential modulating effect observed upon association of transferrin to various liposome formulations on transfection mediated by the polyethylenimine-complexes suggests that these complexes enter into the cells through different pathways (involving clathrin versus caveolin), most likely by taking advantage of their intrinsic biophysical properties to escape from the endosome to the cytosol.

• Non-viral vectors
• cationic liposomes
• transfection
• low molecular weight PEI

Introduction

The success of gene therapy depends mainly on the development of vectors capable of delivering genetic material into target cells, both efficiently and in a safe manner. There are two major groups of vectors that have been studied and developed to mediate gene delivery: Viral and non-viral vectors. Despite exhibiting high gene transfer efficiencies, viral vectors have their use limited due to serious side effects, like immunogenicity and oncogene activation [1]. The disadvantages presented by viral vectors have prompted investigators to develop alternatives to surpass these limitations. One of the most viable alternatives involves the use of non-viral vectors, namely those based on cationic liposomes.

Over the last years, our laboratory has shown unequivocally that ternary complexes prepared by associating the ligand transferrin (Tf) with cationic liposomes and DNA [transferrin-associated lipoplexes (Tf-complexes)] present an increased capacity to mediate intracellular gene transfer compared to plain lipoplexes (lacking Tf), both in vitro and in vivo[2–8]. However, gene transfer efficiency mediated by Tf-associated lipoplexes is still far from that achieved with viral vectors. Therefore, efforts to further develop Tf-complexes towards the production of more efficient particles in terms of transfection should be made in order to move on to gene therapy protocols which can compete with those involving viral vectors, while avoiding the safety concerns inherent to these systems. Several recent studies have reported a strong synergistic effect of cationic liposomes and polyethylenimine (PEI) on the process of intracellular gene transfer [9–12]. PEIs are synthetic polymers with a high cationic charge density and a protonable amino group in every third position [13]. The observed enhancement of transfection has been related with the ability of PEI to condense and compact the carried DNA [14]. In addition, PEI has been described to have the capacity to buffer cellular compartments like endosomes and lysosomes (‘proton sponge hypothesis’), which can protect DNA from degradation and help in its escape from the endosome to the cytoplasm [13], [15]. In the present work, PEI was chosen as a DNA pre-condensing agent, since among other cationic polymers, such as poly-L-lysine or protamine, PEI has been described as being the most efficient in terms of gene delivery [12]. Besides, PEI has the ability to enter the nucleus [16] and accelerate nuclear translocation of DNA from the cytoplasm [17], which may constitute a significant advantage over the other DNA pre-condensing agents.

On the other hand, the successful use of surfactants as non-viral vectors is related to their capacity to induce destabilization of the endosomal membrane, which could facilitate DNA release from the complexes into the cytoplasm and lead to enhanced transfection efficiencies [18]. The efficiency of gene transfer mediated by cationic liposomes can also be strongly influenced by their mode of preparation [19]. Recently, it was reported that when cationic liposomes are prepared through the ethanol injection method an enhanced gene expression mediated by the resulting lipoplexes is observed [20]. The preparation of cationic liposomes by the ethanol injection method has several advantages, like the generation of very narrow and small size distribution in a very reproducible manner, avoiding an extrusion step and thus preventing lipid loss [21]. In addition, this method allows the preparation of cationic liposomes in saline solution (PBS) at neutral pH without aggregation, as opposed to the hydration-extrusion method which may lead to extensive aggregation. Aiming at identifying properties desirable to further improve gene delivery mediated by the extensively reported ternary complexes composed of cationic lipid, DNA and transferrin, we took advantage of the above-mentioned findings and screened a panel of PEI molecules with low molecular weight (MW) as DNA pre-condensing agents. Low MW PEIs were chosen because of their low toxicity compared to that of high MW PEIs, such as the very commonly used 25 KDa PEI [22]. We have generated quaternary complexes composed of DNA, cationic lipid, transferrin and PEI (PEI-Tf-complexes) and evaluated several biological parameters (transfection, DNA protection, cell association and cytotoxicity) using different modes of liposome preparation (hydration-extrusion versus ethanol injection). We have tested the effect on these parameters of different low MW PEIs (non-commercial 2.7 KDa and commercial 2.0 and 0.8 KDa from Sigma), lipid composition (DOTAP:Chol and DOTAP:CHEMS:DOPE:Chol) and the presence of the surfactant octyl β-D-glucopyranoside (OGP) in the composition of the liposome formulations. The effect of cationic lipid (CL)/DNA charge ratio as well as that of PEI/DNA (N/P) ratio on the biological activity of the resulting complexes was also investigated. Finally, results from the modulating effect of Tf on transfection in combination with cell association studies raise some questions on how PEI-Tf-complexes may interact with and enter into the cells, which will also be addressed here.

Materials and methods

PEI 2.7 KDa was a kind gift from Dr T. Merdan (Department of Pharmaceutical Technology and Biopharmacy, Philipps University of Marburg, Germany). Low-molecular weight PEIs (2000 and 800 Da) were purchased from Sigma (St Louis, MO, USA). 1,2-dioleoyl-3-(trimethylammonium) propane (DOTAP), dioleoyl phosphatidylethanolamine, (DOPE), cholesterol (Chol) were purchased from Avanti Polar Lipids (Avanti Polar Lipids, Alabaster, AL, USA). Colesteryl hemisuccinate (CHEMS) was purchased from Sigma (St Louis, MO, USA). Stock solutions of all lipids were prepared in chloroform and kept at -20°C until needed.

Plasmid DNA

The plasmid expression vector pCMVSPORT- LacZ (Gibco BRL, Gaithersburg, MD, USA) (kindly provided by Professor Patrick Aebischer, École Polytechnique de Lausanne, Switzerland) contains the gene coding for Escherichia coli β-gal, under the control of the CMV promoter. pDNA was dissolved in a saline solution (20 mM Hepes buffer, pH 7.4 and 100 mM NaCl) and its concentration was determined spectrophotometrically considering that for an absorbance of 1, at 260 nm, a solution of dsDNA has a concentration of 50 µg/ml [23]. The ratios in absorbance at 260 and 280 nm of the stock solutions were found to be between 1.7 and 1.9, confirming the DNA purity [24].

Liposome preparation

1. Hydration-extrusion. Small unilamellar cationic liposomes (SUVs) were prepared from a 1:1 (mol ratio) mixture of DOTAP and Chol by extrusion of multilamellar liposomes (MLVs). Briefly, DOTAP and Chol dissolved in CHCl₃ were mixed at a 1:1 molar ratio and dried under vacuum in a rotatory evaporator. The dried lipid films were hydrated with deionised water to a final lipid concentration of 6 mM and the resulting MLVs were then sonicated (in an ultrasonic bath) for 5 min and extruded 21 times, through two stacked polycarbonate filters of 50 nm pore diameter using a Lipofast device (Avestin, Toronto, Canada). The resulting liposomes were then diluted five times with deionised water and filter-sterilized utilizing 0.22 µm pore-diameter filters (Schleicher & Schuell, BioScience, Germany).

2. Ethanol injection. Alternatively, liposomes were prepared by the ethanol injection method [25–27]. Briefly, a lipid film containing DOTAP:Chol at 1:1 mol ratio or DOTAP:CHEM:DOPE:Chol at 50:14.8:22.2:13 mol ratio was obtained by mixing the lipids and drying them from chloroform solution under nitrogen flow in a rotatory evaporator. The dried lipid film was dissolved in ethanol and injected into a saline solution (HBS; 100 mM NaCl, 20 mM HEPES, pH 7.4) with or without 20 mM octyl β-D-glucopyranoside (OGP) (Sigma, St Louis, MO, USA) under vortex for 5 min in order to obtain a final concentration of ethanol of 7% (v/v). The liposomes were then sonicated for 10 min and filter-sterilized utilizing 0.22 µm pore-diameter filters (Schleicher & Schuell, BioScience, Germany). The suspension was stored at 4°C until use.

PEI-Tf-complex preparation

The PEI-Tf-complexes (quaternary complexes) were prepared by sequentially mixing the plasmid DNA, PEI (at the desired N/P ratio which is expressed as the PEI amine nitrogen/DNA phosphate ratio [13]) and holo-transferrin (Sigma Chemical, St Louis, MO) followed by incubation of 15 min prior to the addition of the liposomes (at the desired lipid/DNA ratio (+/-) charge ratio). The resulting mixture was further incubated for 15 min. The complexes were used immediately after being prepared.

Cells

HeLa cells, an epithelial-like adherent cell line derived from human tissue, were obtained from American Type Culture Collection, MD. Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂ and maintained in Dulbecco's Modified Eagle's Medium with High Glucose (DMEM-HG) (Sigma) supplemented with 10% heat inactivated fetal bovine serum (FBS) (Biochrom, Berlin, Germany), 0.5 g sodium bicarbonate/l, 1.10 g HEPES/l and 100 µg/ml of streptomycin and 100 units/ml of penicillin (Sigma). Cells were propagated by diluting the cell suspension 1/10 every 3–4 days. For transfection studies, HeLa cells were seeded in 48 multiwell culture dishes (Corning Costar Corp., Cambridge, MA, USA) in a final volume of 1 ml, 1 day before transfection at a density of 25×10³ cells/well. Cell viability was determined by Trypan blue exclusion.

Transfection activity

HeLa cells were covered with fresh 0.3 ml of DMEM-HG containing 10% serum before the complexes were added. The complexes were added gently to the cells in a volume of 0.2 ml per well. After 4 h incubation (in 5% CO₂ at 37°C), the medium was replaced with fresh DMEM-HG containing 10% FBS. The β-galactosidase activity was determined 48 h post-transfection by measuring o-nitrophenyl β-D-galactopyranoside (ONPG) cleavage using a photometric assay. After washing with PBS, the cells were lysed by the addition of lysis buffer (0.1% w/v Triton-X, 0.25 M Tris base, pH 8, 70 µl/well) and freeze-thawing once. PBS (50 µl) and a solution of ONPG (150 µl from a stock of 1.5 mg ml⁻¹) in buffer (60 mM sodium dibasic phosphate; 1 mM magnesium chloride; 10 mM potassium chloride; 50 mM β-mercaptoethanol, pH 8) were added to each well. After incubation at 37°C for 45 min, absorbance at 405 nm was determined in a plate reader and compared to the on-plate β-galactosidase standard curve. The expression of β-galactosidase was normalized based on the protein content of the lysates, which was quantified by the Sedmak method [28] using bovine serum albumin as standard.

Ethidium bromide intercalation assay

The accessibility of ethidium bromide to the DNA associated with the complexes was monitored at 37°C, for 2 min, in a SPEX Fluorolog 2 fluorometer (SPEX Industries, Edison, NJ, USA). The fluorescence was read at excitation and emission wavelengths of 518 and 605 nm, respectively, using 1 mm excitation and 2 mm emission slits. The sample chamber was equipped with a magnetic stirring device, and the temperature was controlled with a thermostatted circulating water bath. The fluorescence scale was calibrated such that the initial fluorescence of EtBr (20 µl of a 2.5 mM solution added to a cuvette containing 2 ml HBS solution) was set at residual fluorescence. The value of fluorescence obtained upon addition of 1 µg DNA (control) was set as 100%. Complexes containing 1 µg DNA were added to the cuvette containing 2 ml HBS solution followed by addition of 20 µl of EtBr. The amount of DNA available to interact with the probe was calculated by subtracting the values of residual fluorescence from those obtained for the complexes (1 µg DNA) and the results were expressed as the percentage of the control.

Cell association

Cell association experiments were performed at 37°C under the same conditions described for transfection. HeLa cells were incubated in a final volume of 1 ml with complexes containing cationic lipid labelled with 5 mol% of the fluorescent probe rhodamine-phosphoethanolamine (Rh-PE) (Avanti Polar Lipids, Alabaster, AL, USA). After incubation, the medium containing the non-associated PEI-Tf- complexes was collected and diluted to a final volume of 2 ml DMEM-HG medium. Fluorescence was measured at 37°C following addition of Triton X-100 (Merck) at a final concentration of 0.5% (v/v). To assess the fluorescence associated with the cells, cells were rinsed with DMEM-HG medium and detached from the culture dishes and then suspended in 2 ml of medium. The fluorescence of the cell suspension was measured in the presence of Triton X-100 as described above. The extent of cell association was determined according to the following equation:1

where F_cells is the value of fluorescence associated with the cells and F_{non-associated} is the value of fluorescence of non-associated complexes. For microscopy fluorescence studies the cells were transfected with the complexes prepared as described above. After 4 h of incubation, the cells were fixed in 4% paraformaldehyde and visualized under an epifluorescence microscope (Zeiss Axioscope).

Cell viability assay

Following transfection under the different experimental conditions, cell viability was assessed by a modified Alamar Blue assay [29]. The assay measures the redox capacity of the cells due to the production of metabolites as a result of cell growth and allows determination of viability over the culture period without the detachment of adherent cells. Briefly, 1 ml of 10% (v/v) Alamar Blue dye in complete DMEM-HG medium was added to each well 45 h following the initial transfection period (4 h). After 3 h of incubation at 37°C, 0.2 ml of the supernatant were collected from each well and transferred to 96-well plates. The absorbance at 570 and 600 nm was measured in a Mediators PhL luminometer (Mediators Diagnostika). Cell viability (as a percentage of control cells) was calculated according to the formula (A₅₇₀ - A₆₀₀) of treated cells ×100/(A₅₇₀ - A₆₀₀) of control cells.

Results

Effect of mode of liposome preparation and its lipid composition on transfection activity of quaternary complexes (PEI-Tf-complexes)

Low MW PEI (2.7 KDa) was initially selected to pre-condense plasmid DNA (pDNA), which was followed by addition of Tf and cationic liposomes to generate quaternary PEI-Tf-complexes. Figure 1 shows the effect of the presence of PEI on transfection mediated by Tf-associated cationic liposomes containing DOTAP:Chol (1:1), prepared at different cationic lipid/DNA (CL/DNA) charge ratios by the hydration-extrusion method. The presence of low MW PEI increased the transfection capacity several orders of magnitude for all CL/DNA charges ratios tested, the nitrogen/phosphate (N/P) ratios of 1/1, 4/1 and 8/1 being the most efficient.

Figure 1.  Effect of cationic lipid/DNA (+/-) charge ratio, PEI/DNA ratio [nitrogen/phosphate (N/P)] (2.7 KDa PEI synthesized in the laboratory) of PEI-Tf-complexes on β-galactosidase (β-gal) gene expression in HeLa cells. PEI-Tf-complexes were obtained from DOTAP:Chol liposomes prepared by the hydration-extrusion method, as described in Materials and methods. The liposomes were complexed, at the indicated lipid/DNA charge ratios, with 1 µg of plasmid DNA, pre-condensed with 2.7 KDa PEI, at the indicated PEI/DNA in the presence of transferrin (32 µg/µg DNA). After 4 h incubation of PEI-Tf-complexes with HeLa cells in DMEM-HG containing 10% FBS, the medium was replaced with fresh medium also in the presence of 10% FBS. The cells were further incubated for 48 h and the levels of gene expression were evaluated as described in Materials and methods. The data are expressed as µg of β-gal per mg of total cell protein (mean±SEM obtained from triplicates) and are representative of, at least, three independent experiments.

The mode of liposome preparation (hydration-extrusion versus ethanol injection) had a significant effect on transfection mediated by the PEI-Tf-complexes as can be concluded by comparing the results shown in Figures 1 and 2A. The presence of PEI (2.7 KDa) in PEI-Tf-complexes resulted in levels of transfection that are several orders of magnitude higher when the ethanol injection rather than the hydration-extrusion method was used in the liposome preparation, the most efficient formulations being those prepared at high N/P ratios (10/1 and 12/1). The influence of the lipid composition of PEI-Tf-complexes on their transfection activity was also investigated using two different liposome formulations, DOTAP:Chol (1:1 mol ratio) (DC liposomes) and a pH-sensitive liposome formulation, DOTAP:CHEM:DOPE:Chol (50:14.8:22.2:13 mol ratio (CatpH-liposomes), prepared by the ethanol injection method. Furthermore, we examined the effect of the incorporation of the surfactant octyl β-D-glucopyranoside (det) in both DC liposomes (DC-det) and CatpH-liposomes (CatpH-det), as well as the influence of the CL/DNA and N/P ratios on transfection activity mediated by the complexes prepared from both lipid formulations.

Figure 2.  Effect of the lipid composition, cationic lipid/DNA (+/-) charge ratio, PEI/DNA ratio [nitrogen/phosphate (N/P)] (2.7 KDa PEI synthesized in the laboratory) of PEI-Tf-complexes on β-galactosidase (β-gal) gene expression in HeLa cells. PEI-Tf-complexes were obtained from liposomes (DOTAP:Chol or DOTAP:CHEMS:DOPE:Chol) prepared by the ethanol injection method, in the presence or absence of the detergent octyl β-D-glucopyranoside (OGP), as described in Materials and methods. Experiments were performed as described in the legend to Figure 1. The PEI-Tf-complexes were produced from (A) DOTAP:Chol liposomes prepared in the absence of OGP (DC); (B) DOTAP:Chol liposomes prepared in the presence of 20 mM OGP (DC-det); (C) DOTAP:CHEMS:DOPE:Chol liposomes (CatpH) and (D) DOTAP:CHEMS:DOPE:Chol liposomes prepared in the presence of 20 mM OGP (CatpH-det). The data are expressed as µg of β-gal per mg of total cell protein (mean±SEM obtained from triplicates) and are representative of, at least, three independent experiments.

For all tested conditions, pre-condensation of pDNA with low MW PEI resulted in a dramatic increase of transfection activity mediated by the PEI-Tf-complexes. As observed, the N/P ratio had a more pronounced effect on transfection than the cationic lipid/DNA charge ratio. The highest levels of transfection were achieved for PEI-Tf-complexes prepared from DC liposomes, the most efficient being those prepared at N/P ratios between 8/1 and 12/1 (Figure 2A). Although the PEI-Tf-complexes prepared from CatpH- liposomes presented a decreased transfection activity as compared with those prepared from DC liposomes, it is interesting to note that lower N/P ratios (1/2 and 1/1) resulted in higher levels of transfection (Figure 2C). The inclusion of surfactants in liposome composition has been described to enhance their capacity to mediate transfection [30], [31]. Therefore, the effect of incorporating the surfactant octyl β-D-glucopyranoside (det) in the formulations DOTAP:Chol (DC-det) and CatpH (CatpH-det) was tested aiming at achieving higher transfection activities. Surprisingly, the transfection activity mediated by PEI-Tf-complexes prepared from DC-det liposomes was significantly decreased for all the CL/DNA charge ratios and N/P ratios as compared to that observed for complexes prepared from plain DC liposomes (Figure 2A, 2B). On the other hand, the use of CatpH-det liposomes in the preparation of PEI-Tf-complexes showed some advantage over that of CatpH liposomes in transfection, since complexes prepared with the latter formulation were less active for all tested conditions (Figure 2C, 2D).

pDNA condensation using commercial PEIs with low molecular weight (0.8 and 2.0 KDa)

Based on the significant enhancement of transfection mediated by Tf-associated complexes achieved upon pre-condensation of plasmid DNA with 2.7 KDa PEI, synthesized in the laboratory, we selected the most efficient CL/DNA charge ratios and respective N/P ratios from the tested formulations and evaluated their biological activity using commercial low molecular weight PEIs (0.8 and 2.0 KDa). Our goal was to evaluate whether the resulting formulations would exhibit a similar capacity to improve transfection mediated by Tf-associated complexes. For this purpose, we started by examining the ability of the two low molecular weight PEIs to condense and complex DNA using the ethidium bromide exclusion assay. Figure 3 shows the results for DNA complexation by both 2.0 and 0.8 KDa PEIs at different N/P ratios, ranging from 1/2 to 80/1. As expected, for both PEIs tested, increasing the amount of PEI, which corresponds to higher ratios N/P ratios, led to a higher degree of DNA protection from ethidium bromide access. However, it is clear that PEI with a MW of 2.0 KDa exhibited a higher capacity to condense and protect DNA than 0.8 KDa PEI. As shown, the maximal protection capacity for the former was observed at the 2.5 N/P ratio, while for the latter this was achieved only at the 40 N/P ratio.

Figure 3.  Effect of the N/P ratio and molecular weight of PEI on its capacity to complex DNA. 1 µg of DNA was complexed with PEI (0.8 or 2 KDa) at the indicated N/P ratios in the presence of ethidium bromide (EtBr), as described in Materials and methods. The amount of DNA available to interact with EtBr was calculated by subtracting the values of residual fluorescence from those obtained for the samples and expressed as the percentage of the control. Control corresponds to the fluorescence of the probe in the presence of 1 µg DNA (100% of EtBr accessibility). The results correspond to the mean±SEM obtained from two independent experiments.

Biological activity of PEI-Tf-complexes prepared with low MW PEIs

Figure 4 shows the results obtained for the transfection activity mediated by PEI-Tf-complexes prepared using the commercial 2.0 KDa PEI. Similarly to that observed for PEI-Tf-complexes prepared with the non-commercial 2.7 KDa, the biological activity of PEI-Tf-complexes prepared with the commercial 2.0 KDa PEI was strongly dependent on the lipid composition of the liposome formulation, the highest transfection levels being achieved for complexes prepared from DC liposomes. However, the biological activity of PEI (2.0 KDa)-Tf-complexes prepared from CatpH-det liposomes was considerable high. Regarding the effect of the presence of octyl β-D-glucopyranoside on transfection mediated by PEI-Tf-complexes, it is interesting to note that when cationic liposome formulations contain 2.0 KDa PEI no significant change in the transfection capacity was observed upon incorporation of the surfactant in the liposome formulation, as opposed to what was obtained for transfection mediated by Tf-complexes when non-commercial 2.7 KDa PEI was used to pre-condense DNA. However, when DC-det liposomes containing the surfactant were used to prepare PEI-Tf-complexes, a differential modulating effect of the surfactant on transfection was observed depending on the molecular weight (2.0 KDa vs. 2.7 KDa) of PEI incorporated in the liposome composition. Figure 5 shows the results obtained for the transfection activity mediated by PEI-Tf-complexes prepared from DC, DC-det, CatpH and CatpH-det liposomes using the commercial 0.8 KDa PEI to pre-condense DNA. As observed, all tested PEI-Tf-complexes presented a considerably lower transfection activity compared to those prepared with 2.0 KDa (commercial) or 2.7 KDa (non-commercial) PEIs, which shows that small differences in the molecular weight of PEI strongly influence transfection mediated by PEI-Tf-complexes. In order to test the influence of transferrin on transfection activity, the most efficient complexes (prepared from DC or CatpH-det liposomes incorporating 2.0 KDa PEI) were selected and their transfection activity was evaluated in the absence of transferrin at the various CL/DNA charge ratios and N/P ratios. Curiously, complexes lacking transferrin (PEI-complexes) revealed a differential behaviour in terms of transfection activity when compared to that observed in the presence of transferrin, depending on the liposome formulation used in their preparation (compare Figures 4 and 6). In general, PEI-complexes prepared from DC liposomes exhibited a higher transfection activity (Figure 6A) as compared to that for PEI-Tf-complexes (Figure 4A), while the opposite was observed when the complexes were prepared from CatpH-det liposomes (Figure 4D and 6B), thus indicating that the role of Tf in modulating transfection is dependent on the liposome composition.

Figure 4.  Effect of the lipid composition, cationic lipid/DNA (+/-) charge ratio, PEI/DNA ratio [nitrogen/phosphate (N/P)] (2.0 KDa PEI) of PEI-Tf-complexes on β-galactosidase (β-gal) gene expression in HeLa cells. PEI-Tf-complexes exhibiting the highest levels of transfection when prepared with 2.7 KDa PEI (Figure 2) were selected to further test their transfection activity upon pre-condensation of the plasmid DNA with commercialized 2.0 KDa PEI. ]Preparation of the PEI-Tf-complexes, at the indicated lipid/DNA charge ratios and PEI/DNA ratios, and transfection experiments were carried out as described in the legend to Figure 1. (A) DOTAP:Chol liposomes prepared in the absence of OGP (DC); (B) DOTAP:Chol liposomes prepared in the presence of 20 mM OGP (DC-det); (C) DOTAP:CHEMS:DOPE:Chol liposomes prepared in the absence of OGP (CatpH) and (D) DOTAP:CHEMS:DOPE:Chol liposomes prepared in the presence of 20 mM OGP (CatpH-det). The data are expressed as µg of β-gal per mg of total cell protein (mean±SEM obtained from triplicates) and are representative of, at least, three independent experiments.

Figure 5.  Effect of the lipid composition, cationic lipid/DNA (+/-) charge ratio, PEI/DNA ratio [nitrogen/phosphate (N/P)] (0.8 KDa PEI) of PEI-Tf-complexes on β-galactosidase (β-gal) gene expression in HeLa cells. PEI-Tf-complexes exhibiting the highest levels of transfection when prepared with 2.7 KDa PEI (Figure 2) were selected to further test their transfection activity upon pre-condensation of the plasmid DNA with the commercialized 0.8 KDa PEI. Preparation of the PEI-Tf-complexes, at the indicated lipid/DNA charge ratios and PEI/DNA ratios, and transfection were carried out as described in the legend to Figure 1. (A) DOTAP:Chol liposomes prepared in the absence of OGP (DC); (B) DOTAP:Chol liposomes prepared in the presence of 20 mM OGP (DC-det), (C) DOTAP:CHEMS:DOPE:Chol liposomes prepared in the absence of OGP (CatpH) and (D) DOTAP:CHEMS:DOPE:Chol liposomes prepared in the presence of 20 mM OGP (CatpH-det). The data are expressed as µg of β-gal per mg of total cell protein (mean±SEM obtained from triplicates) and are representative of, at least, three independent experiments.

Figure 6.  Effect of the lipid composition, cationic lipid/DNA (+/-) charge ratio, PEI/DNA ratio [nitrogen/phosphate (N/P)] (2.0 KDa PEI) of PEI-complexes on β-galactosidase (β-gal) gene expression in HeLa cells. The most efficient liposome formulations using 2.0 KDa PEI (Figure 4) were selected to examine the effect of transferrin on transfection. PEI-complexes lacking transferrin were obtained from liposomes containing DOTAP:Chol (DC) or DOTAP:CHEMS:DOPE:Chol with OGP (CatpH-det) prepared by the ethanol injection method, as described in Materials and methods. Transfection was carried out as described in the legend to Figure 1. (A) DOTAP:Chol liposomes (DC), (B) DOTAP:CHEMS:DOPE:Chol liposomes prepared in the presence of 20 mM OGP (CatpH-det). The data are expressed as µg of β-gal per mg of total cell protein (mean±SEM obtained from triplicates) and are representative of, at least, three independent experiments.

Protection of DNA by PEI-Tf-complexes

The capacity of non-viral vectors to protect the associated genetic material from degradation by DNases present in blood circulation and/or in the cell cytoplasm constitutes a very important property for their in vivo use, since DNA should be intact when reaches the target cells in order to be transcribed [32], [33]. Among the complexes tested, those prepared from DC liposomes in the presence or absence of transferrin and the ones prepared from CatpH-det liposomes with Tf were the most efficient in mediating transfection and, therefore were chosen to be further evaluated regarding their capacity to protect the carried DNA. The ability of these complexes, to protect DNA was tested using the fluorescent probe ethidium bromide. This probe binds specifically to double strand DNA, under which conditions its fluorescence is drastically enhanced, allowing the determination of the degree of DNA protection conferred by the complexes. The results are presented as a percentage of protected DNA after normalization to the fluorescence of the probe obtained in the presence of free DNA in the same amount as that associated with the complexes. Figure 7 shows that, in general, all the tested complexes were able to protect efficiently the associated DNA, although a little tendency for a more efficient DNA protection was observed for complexes prepared at both higher cationic lipid/DNA charges ratios and N/P ratios. In addition, the degree of DNA protection conferred by the complexes depended slightly on the liposome formulation and the presence of Tf, the complexes prepared from CatpH-det liposomes being slightly less efficient to protect the DNA compared to those prepared from DC liposomes.

Figure 7.  Effect of cationic lipid/DNA charge ratio (+/-), N/P ratio and liposome formulation on the access of ethidium bromide to DNA associated with the PEI-complexes (2.0 KDa PEI) in the presence or absence of transferrin. PEI-complexes, containing 1 µg of DNA, were obtained from DOTAP:Chol liposomes (DC) associated with (A) or without (B) transferrin, or from DOTAP:CHEMS:DOPE:Chol liposomes containing OGP (CatpH-det) associated with transferrin (C), prepared by the ethanol injection method. Complexes were incubated with EtBr as described in Materials and methods. The amount of DNA available to interact with the probe was calculated by subtracting the values of residual fluorescence from those obtained for the samples and expressed as the percentage of the control. Control corresponds to the fluorescence of the probe in the presence of 1 µg DNA (100% of EtBr accessibility). The results correspond to the mean±SEM obtained from triplicates, and are representative of two independent experiments.

Evaluation of N/P ratio and cationic liposome/DNA charge ratio of PEI-Tf-complexes on cell association

The influence of liposome composition, cationic lipid/DNA charge ratio and N/P ratio of the PEI-Tf-complexes, which had the optimum transfection profiles (Figures 4 and 6), were studied in terms of the extent of their association with HeLa cells, as assessed by both fluorimetry and fluorescence microscopy. Cell association encompasses cell binding, fusion with the plasma membrane, endocytosis and fusion with the endosomal membrane [32].

Figure 8 shows the extent cell association for PEI-Tf-complexes prepared from DC or CatpH-det liposomes and for PEI-complexes prepared from DC liposomes. As observed, for all tested formulations the extent of cell association depended on the N/P ratio and cationic lipid/DNA charge ratio. This effect was more pronounced for PEI-Tf-complexes prepared from DC liposomes. It is interesting to note the differential effect of the N/P ratio on the cell association profile for complexes prepared from DC liposomes with and without Tf (Figure 8A, 8B). Whereas for PEI-Tf-complexes prepared from DC liposomes the extent of cell association increased for higher N/P ratios, complexes lacking Tf associated less extensively with increasing the N/P ratio. Most interestingly, complexes exhibiting the highest transfection capacity (PEI-complexes, lacking Tf, prepared from DC liposomes, Figure 6A) were those presenting the lowest extents of cell association. On the other hand, PEI-Tf-complexes prepared from DC or CatpH-det liposomes showed a positive correlation between the extent of cell association and their transfection activitiy. Thus, although not exhibiting the best transfection profiles, complexes containing 2.0 KDa PEI and prepared from DC liposomes in the presence of Tf were those showing the highest capacity to associate to HeLa cells (Figure 8A). Figure 9 illustrates the results from fluorescence microscopy studies following transfection, using a cationic lipid/DNA charge ratio of 3/2 and N/P ratios of 8/1, 10/1 and 12/1 (the most effective in mediating transfection as well as in terms of cell association) for PEI (2.0 KDa)-complexes prepared from DC liposomes in the presence of Tf (A), DC liposomes in the absence of Tf (B) and CatpH-det liposomes (C). As observed, transfected cells presented fluorescent particles which should correspond to the complexes that gained access to the cell interior. Intriguingly, although there was not a clear difference between the number of fluorescent particles associated to the cells for the different complexes (particularly for PEI-Tf-complexes prepared from CatpH-det liposomes and PEI-Tf-complexes prepared from DC liposomes), a significant difference was observed for the transfection activity mediated by the complexes. Moreover, HeLa cells transfected with PEI-complexes (absence of Tf) prepared from DC liposomes were those presenting fewer fluorescent particles although this formulation was the best in terms of mediating the highest values of transfection.

Figure 8.  Effect of lipid/DNA (+/-) charge ratio, N/P ratio and lipid composition on the extent of association of PEI-complexes (2.0 KDa PEI) with HeLa cells. PEI-complexes were obtained from DOTAP:Chol liposomes (DC), associated with (A) or without (B) transferrin, or DOTAP:CHEMS:DOPE:Chol liposomes containing OGP (CatpH-det) and associated with transferrin (C), prepared by the ethanol injection method. Liposomes were labeled by incorporating Rh-PE into the lipid bilayer, at a concentration of 5 mol% of total lipid, and the extent of cell association was measured after 4 h incubation at 37°C of the complexes with the cells, as described in Materials and methods. The data are expressed as a percentage of the total fluorescence and represent the mean±SEM from two independent experiments carried out in triplicate.

Figure 9.  Representative images obtained from fluorescence microscopy for cell association of PEI-complexes (2.0 KDa PEI) prepared by the ethanol injection method from DOTAP:Chol (DC) liposomes associated or not with transferrin, or DOTAP:CHEMS:DOPE:Chol (CatpH-det) liposomes associated with transferrin, at the 3/2 lipid/DNA (+/-) charge ratio and at the indicated N/P ratios (8/1, 10/1 and 12/1. Liposomes were labeled by incorporating Rh-PE into the lipid bilayer, at a concentration of 5 mol% of total lipid and PEI-complexes were visualized after 4 h incubation with HeLa cells under an epifluorescence microscope (Zeiss Axioscope). This figure is reproduced in colour in Molecular Membrane Biology online.

Evaluation of the cytotoxicity induced by PEI-Tf-complexes

Since the application of PEI or cationic liposomes for gene delivery is usually associated with some cytotoxicity, cell viability studies were performed after transfection for the most efficient Tf-PEI-complexes. The results presented in Figure 10 show that some toxicity was induced at 24 h or 48 h after the initial 4 h incubation of the cells with PEI-Tf-complexes prepared with 2.0 KDa PEI and DOTAP:Chol cationic liposomes, while PEI or cationic liposomes alone did not cause significant cytotoxicity (approximately 95% of viability, data not shown).

Figure 10.  Effect of the lipid/DNA (+/-) charge ratio and N/P ratio PEI-Tf-complexes (2.0 KDa PEI) on the viability of HeLa cells. PEI-Tf-complexes were obtained from DOTAP:Chol (DC) liposomes prepared by the ethanol injection method at lipid/DNA (+/-) charge ratios and N/P ratios that exhibited the highest levels of transfection. Cell viability was measured by the Alamar blue assay at (A) 24 h and (B) 48 h after transfection, as described in Materials and methods and was expressed as the percentage of the untreated cells. The results correspond to the mean±SEM obtained from triplicates, and are representative of, at least, two independent experiments.

Discussion

To compete with viral vectors for use in therapeutic applications, gene delivery systems based on non-viral technology should be improved to achieve more efficient transfection rates. Among non-viral vectors, those based on cationic liposomes have been the most extensively studied. Here, our aim was to optimize ternary complexes of DNA, cationic liposomes and transferrin, previously developed in our laboratory, by exploring several strategies described in the literature, in an attempt to generate a vector with high transfection efficiency, while lacking cytotoxicity. For this purpose, the effect of several parameters, including the mode of liposome preparation, lipid composition, the presence of a surfactant, the DNA pre-condensation using PEIs with different low molecular weights, as well as, cationic lipid/DNA charge ratio and N/P ratio, were analysed in terms of their effect on the biological activity of the complexes, extent of their association to cells, cytotoxicity and capacity to protect the associated DNA.

The mode of preparation of cationic liposomes used to generate the PEI-Tf-complexes strongly affected their efficiency of transfection. From the results illustrated in Figures 1 and 2A, it is clear that when the ethanol injection method was used to prepare the DC liposomes to generate the PEI-Tf-complexes, transfection was significantly enhanced as compared to that observed when the DC liposomes were prepared by the hydration-extrusion method, which is in agreement with previous results [20]. This enhancement could be explained by the presence of ethanol in the transfection medium, which is known to exhibit the capacity of fluidizing the biomembranes [34], [35] when present at low concentrations, thus facilitating the entry of the complexes into cells [36]. Moreover, the presence of ethanol in the liposome membranes increases their propensity to bend, allowing the formation of non-bilayer intermediates which can trigger vesicle fusion [37]. The formation of non-bilayer structures can result in more efficient transfection [38], [39], which further explains the supremacy of PEI-Tf-complexes to mediate gene delivery when the liposomes were prepared by ethanol injection. In this context, it is important to mention that ethanol has been reported to confer increased capacity of the liposomes to condense DNA [40]. Lipoplexes generated from liposomes prepared by the ethanol injection method have been described to exhibit longer circulation in the blood than the ones prepared by the hydration-extrusion method, presenting even a tendency to accumulate in tumours [19]. Although this could constitute an advantage for in vivo use, the presence of ethanol in the liposome composition cannot be the sole component responsible for the observed enhanced transfection. In fact, complexes produced from different cationic liposome formulations prepared by ethanol injection exhibit very different transfection efficiencies (Figure 2).

The results presented in Figures 4 and 5 show that high transfection levels can be achieved using commercial low molecular weight PEIs as DNA pre-condensing agents in combination with transferrin-associated complexes produced from DC liposomes prepared by ethanol injection. As observed, a small difference in the molecular weight of the PEI has a dramatic effect on the capacity of PEI-Tf-complexes to mediate gene delivery. The poor toxicity of low molecular weight PEIs compared to high molecular weight PEIs (such as the commonly used 25 KDa PEI) allows the use of higher N/P ratios [22]. Moreover, complexes prepared with low molecular weight PEIs seem to be less sensitive to the inhibitory effect of serum on transfection when compared to those prepared with high molecular weight PEIs [41]. In this context, it should be emphasized that transfection mediated by all tested PEI-Tf-complexes has been evaluated in the presence of 10% of serum. However, it seems that PEIs with molecular weights above a certain threshold should be used in order to achieve efficient gene transfer. In fact, the results presented in Figures 4 and 5 clearly show the drastic reduction in the transfection efficiency mediated by PEI-Tf-complexes when prepared with 0.8 KDa PEI as compared to those prepared with the 2.0 KDa PEI. Therefore, it appears that the advantage of using PEIs in combination with cationic liposomes to promote transfection is lost when such low MW PEIs, as 0.8 KDa, are used. A possible explanation for this could be the lower capacity of the 0.8 KDa PEI to condense DNA (Figure 3), in agreement with previous observations [42], [43]. The effect of very low MW PEIs on the size of the complexes [43], [44], may partially explain the observed decrease in the biological activity of PEI-Tf-complexes when prepared with 0.8 KDa PEI.

Another parameter that strongly affects transfection activity mediated by PEI-Tf-complexes is the lipid composition, as concluded from our results showing that, among the different liposome formulations tested, the highest transfection levels were achieved when DOTAP:Chol liposomes were used to prepare the complexes (Figure 4). An interesting conclusion is the differential effect of the presence of a surfactant (octyl β-D-glucopyranoside-OGP) on transfection depending on the liposomal formulation. In fact, whereas the presence of OGP reduces transfection capacity of PEI-Tf-complexes prepared from DOTAP:Chol liposomes, an opposite effect is observed when CatpH liposomes were used to prepare the complexes, particularly at the highest cationic lipid /DNA charges ratios (Figure 4), which may suggest a synergistic effect of the CatpH lipids and the surfactant. The pH-sensitive liposome formulation (CatpH) contains DOPE which is known to undergo a conformational change from a lamellar to an hexagonal phase at acidic pH [45]. It is possible that the presence of a surfactant, which has the natural tendency to promote the formation of non-lamellar structures, in combination with that of pH-sensitive lipids will constitute a promising strategy to enhance transfection. Therefore, in the case of the complexes prepared from the DC formulation, which lacks pH-sensitivity, the presence of OGP may not be sufficient to allow their escape from the endosome. Intriguingly, PEI-Tf-complexes prepared from CatpH or CatpH-det liposomes, designed to be more efficient to promote DNA escape from the endosomes, presented a decreased biological activity compared to those prepared from DOTAP:Chol liposomes. A possible explanation for this could be that PEI-Tf-complexes utilize different intracellular trafficking pathways depending on the presence of cationic or pH-sensitive lipids in their composition.

Another interesting finding from the present work is the effect of the presence of transferrin on transfection, which is illustrated in Figure 6 for the most efficient complexes (prepared from 2.0 KDa PEI and DC or CatpH-det liposomes). Previous results from our laboratory have demonstrated that association of Tf with lipoplexes increases their capacity to interact with cells and to mediate transfection [5–7]. It is interesting to note that when Tf was taken out of the composition of PEI-containing complexes transfection activity was enhanced when DC liposomes were used in their preparation, whereas in the case of complexes prepared from CatpH-det their transfection was drastically reduced. This observation reinforces the idea that in the case of PEI-Tf-complexes prepared from CatpH-det liposomes, the presence of a component with fusogenic activity triggered by low pH, like Tf, is crucial to achieve high transfection levels, whereas transfection mediated by PEI-complexes prepared from DC liposomes is relatively insensitive to its presence, or even more efficient in the absence of such a ‘fusogenic’ contribution. It is possible that these complexes have different properties allowing them to use different internalization pathways other than that mediated by clathrin-coated vesicles, which is used by Tf. Although PEI-complexes lacking Tf, prepared from DC liposomes presented very similar capacities to those prepared from CatpH-det liposomes to protect the associated DNA (Figure 7), the extent of their association to cells was significantly lower (similar to that for complexes prepared from CatpH-det liposomes) than that observed in the presence of transferrin (Figure 8A, 8B and 8C). This indicates that other factors, besides the capacity of the complexes to interact with the cells or to protect DNA, should contribute to such different observed transfection levels. Recently, van der Aa and colleagues reported that the uptake of PEI/DNA polyplexes is mediated via different uptake routes, including both the clathrin- and caveolae-mediated pathways [46], which is in agreement with previous reports by Rejman et al. [47]. However, it was shown that only the DNA delivered into caveosomes was expressed. Although caveosomes exhibit a neutral pH, PEI/DNA polyplexes should face a low pH such as that in the endosomes, for successful transfection [48]. In addition, some studies provided evidence that when cells were incubated with PEI/DNA polyplexes, these accumulated mainly in the lysosomes [49]. A clear mechanism for the cellular internalization of the polyplexes allowing a successful transfection was not established yet. However, it seems that polyplexes can use the different available endocytic pathways to enter into the cell and depending on their ability to escape from the endosome/caveosome, they will have more or less success in mediating gene transfer. The differential effect of the presence of Tf on transfection mediated by PEI-complexes prepared from DOTAP:Chol or the pH-sensitive CatpH-det liposomes can be understood, taking into account the fact that this protein is internalized into the cells by clathrin-mediated endocytosis and therefore, when associated with the PEI-complexes may drive them to the same cell entry pathway. Since clathrin-mediated endocytosis is a process followed by a decrease of pH, any property of the complexes that can take advantage of this phenomenon should increase transfection. Based on our previous findings that under acidic conditions Tf may assume fusogenic properties [6], CatpH-det-PEI-Tf-complexes would lose the capacity to escape from the endosomes when this protein is taken from their composition, thus inhibiting their transfection activity (Figures 4 and 6). In addition, PEI itself should play a role in transfection, by buffering the endosome pH and thus promoting the escape of the complexes or DNA from the endosomes [13], [14], [50]. Alternatively, it should be reasonable to assume that the presence of cholesterol in the composition of PEI-complexes lacking Tf would allow their preferential interaction with the lipid rafts, which are rich in cholesterol, and trafficking via this cell entry pathway. In fact, recent studies have shown that liposomes containing cholesterol are internalized by a caveolae-mediated pathway [51]. If these PEI-complexes follow the caveolin-mediated pathway, they will not have to face the problem of, at least partially, ending up in the lysosomes, where they would be destroyed. Therefore, the complexes could remain in the caveosome for more prolonged times before gaining access to the cell cytoplasm. Besides, DOTAP:Chol liposomes when concentrated, as they would be in the caveosomes, have the capacity to exhibit hexagonal structures [52], which may explain how their complexes with DNA can escape from the caveosome. In addition, the main role for endocytosis has been assumed to be the transport of DNA from near the plasma membrane to the perinuclear zone [53]. In this case, the caveolin-mediated pathway would be ideal and escape from the caveosome would be a question of time. Once in the cytoplasm, the PEI-complexes could take advantage of the presence of an NLS sequence in PEI which could facilitate DNA translocation into the nucleus [17]. Preliminary studies carried out in our laboratory demonstrated that PEI-complexes prepared from DOTAP:Chol liposomes, with or without Tf, are able to mediate efficient transfection of melanomas induced in Balb/c mice when injected directly into the tumor. Although both types of complexes (with or without Tf) presented capacity to mediate transfection in this animal model, it was clear that the PEI-complexes lacking Tf exhibit an increased capacity to transfect the tumor cells. The different role played by transferrin in transfection of cultured cells mediated by the PEI-complexes prepared from DC or CatpH-det liposomes may indicate that these complexes enter into the cell through different pathways (involving clathrin vs. caveolin), taking advantage of their intrinsic biophysical properties to escape to the cytosol and thus gain access to the nucleus where DNA can be transcribed. Studies are currently in progress to address the physical properties of these systems that will be responsible for increasing their transfection activity.

Acknowledgements

Nuno Penacho is the recipient of a fellowship from the Portuguese Foundation for Science and Technology (SFRH/BD/6135/2001). This work was partially supported by a grant from the Portuguese Foundation for Science and Technology (PTDC/BIO/65627/2006). Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.