The lipids DOTAP and DOTAP-cholesterol (DC) were chosen for the preparation of cationic lipid DNA complexes since they gave good transfection efficiency in cell culture and in animal models [12,30,34-36]. To determine the optimal ratio of lipid to DNA, one μg of a plasmid encoding the luciferase reporter gene was complexed with increasing amounts of DOTAP (Fig. 1A) or DC (Fig. 1B). The functionality of the complexes were analyzed by transfection of 293A target cells and subsequent measurement of the luciferase activity in the cell extracts of the transfected cells. For both lipids a dose dependent increase in transfection efficiency was observed. Since protamine sulphate was reported to enhance transfection efficiency of cationic lipid DNA complexes [33], it was added in increasing concentrations to 1 μg of plasmid DNA prior to the addition of the lipids at all the previously analyzed lipid to DNA ratios. Protamine sulphate enhanced transfection efficiency by a factor of 10 to 100 at all lipid to DNA ratios tested. The strongest increase was observed at low lipid to DNA ratios. For DOTAP, the highest luciferase activities were obtained with 4 to 8 μg protamine sulphate at high DOTAP to DNA ratios (Fig. 1A). For high DC to DNA ratios, the transfection efficiency did not vary considerably in the presence of 0.5 to 16 μg protamine sulphate indicating that a wide range of protamine sulphate concentration could be used for complex formation. Similar experiments with a plasmid encoding the GFP reporter gene revealed that at the optimal ratios of lipid to DNA to protamine sulfate about 30–40% of the cells were transfected in the presence of 10% fetal bovine serum and up to 70% of the cells were transfected when transfection was done in the absence of serum (data not shown). Similar high transfection efficiencies were obtained with selected complexes in other immortalized cell lines (data not shown).

Since toxicity of the lipid DNA complexes could limit their use, cell viability was analyzed after transfection with various ratios of lipid to DNA to protamine sulphate. For ratios of up to 64 nmol DOTAP/μg DNA cell viability was higher than 80% of the viability of untransfected cells independent of the amount of protamine sulfate added (Fig. 2A). At higher DOTAP to DNA ratios (> 64 nmol DOTAP/μg DNA), toxicity peaked at 8 μg protamine sulfate/μg DNA. High cell viabilities were observed for complexes containing DC at ratios ≤ 64 nmol lipid/μg DNA with limited correlation of the amount of protamine sulfate added to the complexes.

Serum has been previously reported to interfere with transfection efficiency of lipid DNA complexes [12,28,36]. In these studies, lipid DNA complexes were usually optimized for transfection efficiency in the absence of serum, and small number of complexes were subsequently analyzed for transfection efficiency in the presence of high serum concentrations. To determine, whether DOTAP and DC lipid DNA complexes could be obtained with increased transfection efficiency in the presence of high serum concentrations, the 162 different complexes analyzed above were also studied for transfection efficiency after incubation with 66% serum. Lipid DNA complexes composed of ≤ 16 nmol lipid/μg DNA and ≤ 2 μg protamine sulfate/μg DNA were strongly inhibited by high serum concentrations (compare Fig. 1A and 3A), while complexes with higher amounts of DOTAP were inhibited to lesser extent or not at all. Peak transfection efficiencies of DOTAP complexes with less 16 nmol lipid/μg DNA in the presence of high concentrations of serum were obtained with 4 μg protamine sulfate/ μg DNA (Fig. 3A). At higher DOTAP concentrations protamine sulfate affected transfection efficiency in the presence of high serum concentrations to a lesser extent. Transfection efficiencies of DC DNA complexes were only inhibited by high serum concentrations at high ratios of protamine sulfate to DNA (Fig. 1B versus 3B). Surprisingly, high serum concentrations increased transfection efficiency of DC complexes in the absence of protamine sulfate. Transfection efficiencies of complexes containing ≤ 4 nmol DC/μg DNA were also increased by high serum concentrations at low amounts of protamine sulfate (0.5 to 4 μg protamine sulfate/μg DNA). However, the highest transfection levels in the presence of high serum concentrations were obtained with DC complexes composed of 64 nmol lipid/μg DNA in the presence of a wide range of protamine sulfate to DNA ratios (0.5 to 64μg protamine sulfate/μg DNA).

On the basis of these findings, two ratios of lipid to DNA to protamine sulphate, which seem to cover the range of efficient lipid DNA complexes, were chosen for each lipid from the 162 different complexes analyzed for further characterization of the effect of serum on transfection efficiency. Complexes with low lipid to DNA ratios consisted of 16 nmol DOTAP (DOTAP-low) or DC (DC-low)/1 μg DNA with 1 μg protamine sulphate, while complexes with a high lipid to DNA ratio were formed with 64 nmol DOTAP (DOTAP-high) or DC (DC-high)/1 μg DNA with 16 μg of protamine sulphate. The in vitro transfection efficiencies at low serum concentrations were very similar for DOTAP-low, DOTAP-high, DC-low, and DC-high, ranging from approximately 1 to 2 × 10⁷ RLU/μg protein extract (Fig. 4) and did not change notably during preincubation of the complexes for up to four hours prior to their addition to the cells (data not shown). However, mixing of the four selected complexes with 2 volumes of serum for less than 10 minutes, for 1 hour, or for 2 hours prior to transfection and subsequent determination of luciferase activities revealed striking differences. The DC-high complex was only marginally affected by 2 volumes of serum (Fig. 4). Preincubation with serum for two hours inhibited transfection with DC-low and DOTAP-high complexes by approximately 10- and 20-fold, respectively. Transfection with DOTAP-low complexes was almost completely blocked by preincubation with serum (Fig. 4).

To determine, the transfection efficiency of the four selected complexes in vivo, 25 μg luciferase expression plasmid were mixed with the lipids and protamine sulfate at the indicated ratios and injected into the tail vein of mice. Twenty four hours later the luciferase activity was determined in the lung extracts, the organ with the highest reporter gene activity following intravenous injection of cationic lipid DNA complexes ([34,37]; data not shown). The highest luciferase activities were obtained with the DC-high complex (Fig. 5), which was also least inhibited by high serum concentrations (Fig. 4). Intermediate levels of transfection efficiencies were obtained with DC-low and DOTAP-high, both of which also showed intermediate levels of inhibition by high serum concentrations. DOTAP-low transfection complexes, which were strongly inhibited at high serum concentrations, only gave low luciferase activities in vivo. Thus, the in vivo transfection efficiency seems to correlate well with the transfection efficiency in cell culture in the presence of high serum concentrations.

The size and the charge of the four selected lipid DNA complexes was examined by light scattering and laser electrophoresis to determine the importance of these parameters for in vivo transfection efficiency and serum inhibition. The four selected complexes were positively charged with charges ranging from 13.1 mV for DOTAP-low to 19.6 mV for DC-high (Table 1). With the exception of DC-low complexes, lower zeta potentials were observed when the complexes were formed in the absence of protamine sulfate (data no shown). The largest size with 0.44 μm was predicted for DC-high complexes (Table 1), while the other three complexes had a predicted size of 0.12 to 0.19 μm. For these three complexes, protamine sulfate reduced the size of complexes since control complexes formed at the same lipid to DNA ratios in the absence of protamine sulfate had sizes ranging from 0.42 μm to 0.45 μm (data not shown), confirming the previously reported condensing effect of protamine sulfate [31,33]. However, DC-high complexes were not condensed by addition of protamine sulfate, suggesting that a small size of the complexes is not required for good transfection efficiency in vivo and in the presence of high serum concentrations.

The sensitivity of the complexes to serum seems to be an important parameter for in vivo transfection efficiency. To directly assess the effect of serum on the structure of the lipid-DNA complexes, the turbidity of the complexes was determined before and after addition of two volumes of serum. After addition of serum the turbidity increased within 10 minutes but remained rather constant for the rest of the observation period in case of DOTAP-low, DC-low and DC-high complexes (Figure. 6). This might indicate formation of rather stable complexes with little tendancy to additional aggregation. For DOTAP-high complexes we observed a continous increase in turbidity after addition of serum for at least 150 minutes, indicating aggregation of the complexes. Since the transfection efficiency of DOTAP-low complexes was inhibited to a larger extent by high serum concentrations than the transfection efficiency of DOTAP-high complexes (Fig. 4), stable turbidity values after addition of serum were not predictive for resistance to inhibition of transfection efficiency by high serum concentrations.

A second approach, based on fluorescence energy transfer (FRET) was used to study lipid interactions during complex formation and to assess the direct effect of serum on the interaction between the lipids. During preparation of the liposomes, the fluorescent molecules N-NBD-PE (donor) and RH-PE (acceptor) are incorporated at a low molar ratio (1% mol/mol of both NBD-PE and RH-PE) in the vesicle bilayer. Under steady state conditions with appropriate ratios of NBD-PE to RH-PE, the fluorescence emitted by NBD-PE is reduced by RH-PE because the emission wavelength of NBD-PE overlaps with the excitation wavelength of RH-PE. When a perturbation of the lipid bilayer occurs as a result of interaction with either DNA or serum components a decrease in the surface densities of RH-PE and NBD-PE occurs resulting in an increase in NBD-PE fluorescence which was measured in this experiment. An increase in NBD-PE emission intensity therefore indicates a reduction in FRET probably due to changes in the conformation of the liposomes or complexes.

For DC-low, addition of DNA and protamine sulphate to the labelled liposomes led to a >10% increase in the emission intensity of NBD-PE, while the increase was less striking during formation of DC-high, DOTAP-low, and DOTAP-high complexes (Fig. 7). Fluorescence intensity increased after addition of serum for all complexes (> 40%) with minor differences between different protocols (Fig. 7). To disrupt the complexes completely, Triton X-100 was added, which led to a further increase in fluorescence intensities. The highest levels of fluorescence intensity obtained after addition of Triton X-100 was very similar for all the complexes (data not shown).

1,2-Dioleoyl-3-trimethylammonium-propane (DOTAP), cholesterol, NBD-PE (N-[7-nitrobenz-oxa-1,3-diazoyl-4-yl]-phosphatidylethanolamine) and Rh-PE (N-[lissamine-rhodamine-B-sulfonyl]-phosphatidylethanolamine) were purchased from Avanti Polar Lipid (Alabaster, AL, USA). Protamine sulphate and MTT (Dimethyl-thiazol-diphenyl-tetrazolium-bromide) were from Sigma (Darmstadt, Germany). The Luciferase assay kit was purchased from Promega (Manheim, Germany). All other chemicals were of reagent grade.

The reporter plasmid containing the firefly luciferase cDNA protein was cloned into the backbone of pcDNA3.1 plasmid (Invitrogen, Groningen, Netherlands). The expression of the reporter gene is under the control of Cytomegalovirus (CMV) immediate early promoter. The plasmid pEGFP-C1 expressing the enhanced green fluorescent protein cDNA driven by the CMV promoter was obtained from Clonetech (Heidelberg, Germany). Plasmid DNA was prepared with the Qiagen plasmid Giga kit (Valentia, CA, USA) according to the manufacturer's instruction and dissolved in phospate buffered saline (PBS, pH 7.4).

Cationic liposomes containing DOTAP or DOTAP and cholesterol in a 1:1 molar ratio were prepared using a previously described method [39]. The lipid was initially dried from chloroform, subsequently dispersed in PBS (pH 7,4) and shaken at a temperature above the gel-to-liquid-crystalline transition temperature of the lipid (50°C) for 10 min. The milky solution was sonicated for 10 min under temperature control using a Branson sonifier 250 (Branson Sci., USA) at minimum ultrasound power to avoid extensive bubble formation.

The plasmid DNA was diluted to a concentration of 1 μg/μl in PBS (pH 7.4) prior to complex formulation. All reagents were sterile and were adjusted to room temperature before complex preparation. Protamine sulphate was added to DNA diluted in PBS at a weight ratio of 1 μg protamine sulphate per μg DNA (DC- and DOTAP-low) or 16 μg protamine sulphate per 1 μg DNA (DC- and DOTAP-high), the mixture was vortexed briefly and incubated at room temperature for 5 minutes. The respective amounts of DOTAP (DOTAP-high: 64 nmol/μg DNA; DOTAP-low: 16 nmol/μg DNA) or DC (DC-high: 64 nmol/μg DNA; DC-low: 16 nmol/μg DNA) were added to the solution at a final volume of 50 μl, the mixture was vortexed briefly and further incubated for ten minutes at room temperature. Stable fluorescence intensity values were obtained within minutes after addition of compounds (Fig. 7) indicating rapid completion of mixing.

The electrophoretic mobilities of the complexes were measured using a Zetasizer 4 device (Malvern, U.K.). Lipid DNA complexes were formed with 4 μg DNA and the respective amounts of Lipid and protamine sulfate in a volume of 200 μl. The volume was adjusted to 4 ml by addition of PBS. The mobilities of the particles were measured using a modulation frequency of 1000 Hz, the electrode current was 5 mA. Light scattering at 90° was measured using the same sample. Multimodal analysis was used for the deconvolution of the data. The number distribution mode were used for the data analysis.

About 40000 293A cells (Quantum Biotechnology Inc., Montreal, Canada) were seeded into a 24 well plate and grown in a 37°C incubator in a 5% CO₂ in a final volume of 1 ml DMEM from GIBCO (Karlsruhe, Germany). The next day, the cells (60–70% confluent) were then transfected by adding 50 μl complex solution containing 1 μg of DNA to each well of the plate without removing the 10% fetal bovine serum containing medium from the cells. The cells were then incubated for 48 hours without removing the medium. When transfections were set up in the absence of serum, cells were incubated with the complexes for 4 hours in serum free medium before the medium was replaced with fresh medium containing 10% fetal bovine serum. To study serum inhibition, complexes formed as described above were incubated with two volumes of heat inactivated fetal calf serum (GIBCO; Karsruhe, Germany) prior to addition to 293A cells. Two days after transfection cells were washed with PBS and lysed with 200 μl lysis buffer (0.05% Triton X-100, 2 mM EDTA, 0.1 M Tris, pH 7.8) through one freeze-thaw cycle. The cell lysate was then transferred to microcentrifuge tubes and centrifuged for 10 minutes at 14000 g at 4°C. The supernatant was diluted one to twenty and 20 μl of the diluted sample was used for the luciferase assay according to the manufacturer's instructions. The luminescence was determined in a luminometer from Murex (Germany) The protein concentration was determined using a multiplate assay from Biorad (Hilden, Germany) according to the manufacturer's instructions. For quantification, a standard containing known amounts of BSA in luciferase lysis buffer was included. Reporter gene expression was expressed as relative light units (RLU) per μg protein extract. Each data point represents the mean ± the standard deviation of triplicate experiments.

293A cells transfected as described above were assayed for cell survival using the MTT reagent. The MTT reagent was diluted in PBS to a final concentration of 2.5 mg/ml then filtered using a 0.22 μm filter. 50 μl of the stock solution was then added to each well of the tranfected plates and incubated for 1 hour at 37°C under 5% CO₂. After an hour the medium was removed and 200 μl of 0.04 N HCl was used to dissolve the resulting blue formazan crystals in living cells. The extract was then diluted 1:4 in 0.04 N HCl and the optical density was determined at 540 nm. Untransfected 293A cells were used as control.

The animal experiments were approved by the local authorities. Female Balb-C mice 6–8 weeks old were obtained from the experimental animal services of the University of Leipzig and housed in accordance with institutional guidelines. Individual mice in groups of five were injected via the lateral tail vein with 25 μg of DNA formulated in the complexes in a total volume of 300 μl. Twenty-four hours following i.v. injection, the mice were killed by cervical dislocation and the lungs were collected. The organ was then washed with PBS and homogenized in 500 μl of lysis buffer (Promega, Mannheim, Germany). After two cycles of freeze-thaw, the homogenates were centrifuged at 14000 g at 4°C for 10 minutes. The supernatant was analyzed for luciferase activity as described above.

The relative absorbance of the complexes was determined at 600 nm according to a method described [38]. Complexes were formed with 4 μg DNA and the respective amounts of lipids and protamine sulfate in a volume of 200 μl. After addition of 400 μl PBS the absorbance of the complexes was determined for 10 minutes in the absence of serum with PBS as the blank control. A volume of 400 μl serum was added and the kinetics of the turbidity was measured with stirring at 37°C for 5 hours. Serum in PBS was used as the blank control.

The multi-step lipid mixing assay was adapted from the methods described by Wasan [21]. Structural reorganization of the complexes after binding of DNA and serum was followed by the fluorescence resonance energy transfer method, using NBD-PE and Rh-PE. The liposomes were composed of DOTAP or DC and 1% (mol/mol) of both NBD-PE and Rh-PE. The liposomes with the incorporated fluorophores were suspended in 2 ml of PBS (pH 7.4). The fluorescence measurements were carried out at 25°C with an excitation wave length of 465 nm. The emission was recored at a wavelength of 520 nm using a Perkin-Elmer LS50B spectrofluorometer. The recording started after the system had almost reached equilibrium. Initially, fluorescence readings of the respective liposomes (DOTAP-low and DC-low: 64 nmol; DOTAP-high and DC-high: 256 nmol) in 2 ml PBS were taken continuously at 25°C for 5 minutes, 4 μg DNA mixed with protamine sulfate (4 μg for DOTAP-low and DC-low; 64 μg for DOTAP-high and DC-high) diluted in 500 μl of PBS (pH 7.4) were added and changes in the fluorescence emmission of the complexes at 520 nm wave length of were monitored while stirring for another 5 minutes. To correct for the decrease in fluorescence by increasing the volume, the fluorescence intensities obtained after addition of PBS were subtracted from the values obtained after addition of PBS with DNA. The resulting change in fluorescence intensities was added to the fluorescence value obtained with liposomes alone. To determine the effect of serum, 800 μl of FBS was added and mixed briefly. Fluorescence values were monitored and corrected for dilution as before. Finally 66 μl of Triton X-100 (0.2% v/v) was mixed with the complexes to completely solubilize the vesicles. The mean of the fluorescence intensity values obtained after addition of Triton X-100 was set as 100% and all values were expressed as percentage of this mean.