Stearylamine, sorbitan monooleate (Span^® 80), and poly(oxyethylene sorbitan monooleate) (Tween^® 80) were obtained from Sigma Chemical (St. Louis, MO, USA). Medium-chain triglycerides (Captex^® 355) were donated by Abitech (Columbus, OH, USA). The water used in this study was freshly purified by a Milli-Q Gradient A10 system (Millipore, Molsheim, France). All reagents were used without further purification.

The plasmid pIRES2-EGFP was amplified in Escherichia coli (strain DH10B) and extracted by MidiPrep [26]. The purity of the plasmidial DNA (pDNA) was measured by OD260/OD280 (1.85∼1.90; OD, optical density) and electrophoresis in a 0.7% agarose gel. pDNA was resuspended in pure water (Milli-Q) and frozen in aliquots.

The nanoemulsions were produced by the sonication method. Briefly, the aqueous phase components were weighed and mixed with pure water (Milli-Q). The mixture was sonicated for 3 minutes using a probe sonicator at 30 W (VCX600, Sonics and Materials, Danbury, CT, USA). This mixture was added into the oily phase, which was 50 ng of Captex^® 355 per mg of nanoemulsion, and sonicated for 4 min in an ice-water bath and, then, stored at 4 ºC. In order to evaluate possible changes in the physicochemical properties of the nanoemulsions containing stearylamine (CLN) and consequently variations in the complexation efficiency, two different CLNs were formulated, in which the cationic lipid stearylamine was dispersed into the water phase (CLN₁) or into the oily phase (CLN₂), as described in Table 1.

The mean particle size and zeta potential of the CLN and lipoplexes (at the best pDNA/CLN ratio, which was 2.3 (2.3 ng of pDNA per 1 ng of CLN₁) and 0.92 (0.92 ng of pDNA per 1 ng of CLN₂, found by the compaction process assay)) were measured by Dynamic Light Scattering (DLS) and phase analysis light scattering, respectively (Brookhaven ZetaPALS, Holtsville, NY,USA). The formulations were diluted with distilled water by 300-fold before the measurement.

The complexes were loaded onto 0.7% agarose gel containing ethidium bromide (0.5 mg/mL) using 40 mM Tris acetate and 1mM EDTA as buffer. Gel was run at 80 mV for 60 min and the trapping efficiency of pDNA by CLNs was determined as the DNA was no longer accessible to ethidium bromide intercalation by the assay of gel electrophoresis [20].

DLS analysis of CLNs and lipoplexes ensured that the produced systems were at nanoscale dimensions (Table 3). It is important to note that the pDNA in stock solution presented a mean particle size of 52.2 ± 0.6nm. It should be emphasized that this size is given as a purpose of comparison with the CLN size. In fact, hydrodynamic volume (which is measured by the DLS) changes according the kind of coiled in which the plasmid has been presented.

CLNs presented particle sizes smaller than 200 nm, for both CLN₁ and CLN₂. After lipoplex formation, no significant increase in droplet size was found for either CLN₁ or CLN₂, probably because the pDNA was not only adsorbed, but also condensed on the droplet surface, and its negligible concentration (460 ng) does not induce enough surface volume charge to improve the nanoemulsion droplet size.

The Polydispersity Index (PI) was around 0.2 for all formulations (Table 3). Moreover, the droplet size distribution of the systems was quite homogeneous and not influenced by the lipoplex formation generated between pDNA and CLN.

As can be observed by the positive values of zeta potential, stearylamine gives a positive charge to the droplet surface. Zeta potential results were around +40 mV for both CLN₁ and CLN₂, which not only provides enough positive charges for lipoplex formation, but also provides CLN stability by electrostatic repulsion. When lipoplexes were formed, zeta potential decreased (Table 3). This could be explained by the negative charge from the pDNA being added to the system, thereby turning the zeta potential less positive.

This result is quite interesting because although the complexation study revealed that CLN₁, which has stearylamine at the outermost oil/water interface, possesses different behavior concerning lipoplex formation compared to CLN₂, which has stearylamine at the innermost oil/water interface, both products presented the same zeta potential. This is probably due to the migration of the stearylamine from the innermost to the outermost interface induced by the dilution process of 300 fold, mandatory for this assay.

Although at different pDNA/CLN ratios, 100% of DNA compaction was achieved by diluting the pDNA solution with both CLN₁ and CLN₂. The presence of the CLN droplets at pDNA/CLN ratio above 2.30 for CLN₁ and above 0.92 for CLN₂ (Table 2), in which a high amount of droplets exists, induces totally complex formation and no free pDNA could be seen in the gel bend. Complexation is possible due to electrostatic interactions between the positively charged stearylamine head group (which is located on the oil/water interface) and the negatively charged phosphate anchor from the pDNA [5,6,8,16].

Matulis et al. have demonstrated by thermodynamic assays that condensation is dependent on DNA concentration [25]. In this way, from the data set revealed by CLN₁, 4 μL of CLN₁ was enough to completely condense the genetic material. Figure 1 shows that by increasing the DNA/CLN ratio higher than 2.30 pDNA/CLN₁ ratio, no free pDNA can be seen, which means that the 460 ng of pDNA was completely condensed with 4 µL of CLN₁.

It is known that polyethylene glycol (PEG) domains present in Tween^℘ 80 avoid oil droplet aggregation in emulsions [10]. This is also related to a steric effect (because of those mentioned PEG domains) for Tween^℘ 80, which could make the entrapment of the pDNA molecule [27] difficult. When stearylamine is placed into the aqueous phase, Tween 80^℘ might be placed on the oil/water surface and the stearylamine could be located more externally at the outermost surface of the drop, minimizing the furtive effect from PEG. On the other hand, when stearylamine is loaded into the oil phase, the charge is there, but located in the innermost regions of the drop, which allows the steric effect from PEG to participate, and makes the complete complexation of pDNA difficult. It is important to note that, in this case, a probable partition of stearylamine into the two phases was negligible because the steric effect, which is dependent on the phase where stearylamine was placed, was demonstrated. This is in agreement with the works of Matulis et al., who described a binding model where cationic head groups are near DNA phosphates while the aliphatic tails stand perpendicular to the DNA surface [25].

Concerning CLN₂, it was found that a larger amount of CLN was necessary to compact the same quantity of pDNA compared to CLN₁ (Figure 2). In this case, around 10 μL of CLN₂ (0.92 pDNA/CLN₂ ratio) was necessary to completely compact the 460 ng of the genetic material, which represents more than twice the volume of CLN₁.

Most of the time, complexation is done by adsorption of the genetic material to the cationic system by electrostatic interaction [16,17,19,20,21]. However, other methods could be used to ensure that the genetic material is loaded on the carrier. Nanoencapsulation of genetic material is another related approach [28]. This was demonstrated by Badea et al. [29], who produced nanoparticles where the genetic material was put into the formulation during the manufacturing process. Also, Kwon et al. [23] previously prepared nanoemulsions, added pDNA, and sonicated the mixtures in order to form lipoplexes. All these approaches are in agreement with our results. Moreover, a combination of two mechanisms, named electrostatic interaction and nanoencapsulation assembly, can play an important role in the pDNA complexation in CLN systems.

Concerning non-viral gene delivery systems, the literature reports different times of complexation. It seems that there are no rules governing this variation and many different times are related to the complete complexation of genetic material. Time of complexation changes not only concerning the kind of genetic material, but also the different pharmaceutical systems. Tang et al. [20] worked with two complexation times. They started complexing for 30 min and then reduced the time for 10 min of incubation at RT. Kim et al. [14] did nanoparticle complexing of pDNA for 1 hour at RT but Munier et al. [19] used 10 min as the time of complexation for their nanoparticles. Other authors working with nanoparticles have described achieving complete complexation in just 5 min [2,3,21]. Teixeira et al. [16] associated oligonucleotides with cationic nanoemulsions for 12 h. Martini et al. [17] used a similar system and associated the genetic material for 30 min and Bruxel et al. [11] for just 15 min. Concerning liposomes, there is also no pattern of complexation time [5,30,31]. Zhang et al. [15] evaluated complexation of both pDNA and siRNA for 20 min using a unique system and found different ratios of complexation according to the kind of genetic material. Therefore, it could be seen that each researcher chooses the best time of association experimentally, according to the kind of both formulation and genetic material [11,15,16,17,20].

This study reveals that for CLN 120 min is the best time of complexation (Figure 3). It was observed that after just adding the pDNA in contact with the CLN, lipoplexes were immediately generated in spite of the important amount of DNA that remained uncomplexed. With the time increasing, the intensity of the bend diminished and, after 120 min, free pDNA was no longer observed, meaning that 100% of the pDNA was complexed and condensed.

Usually, complexation is done at RT [14,16,17,19,20]. In this study a comparison of three different temperatures [RT, “cold” temperature (4 °C), and “hot” temperature (37 °C)] was performed to observe whether there is any influence of temperature on the complexation efficiency (Figure 4). It was observed that the same amount of pDNA in the same time of complexation had differences in the complexation range. Naked pDNA was also analyzed at the three different temperatures, as a control, to ensure that there would be no difference in the gel visualization when the pDNA was maintained at different temperatures.

As shown in Figure 4, at RT some compaction is achieved. However, it seems that by increasing the temperature to 37 °C less pDNA is packed. On the other hand, at 4 °C all the genetic material was packed by the CLN₁. A crucial point is that at this temperature, naked DNA, although at lower mass concentration (1,000 ng), presented a bend on the running gel (Figure 4a). The reason is that, according to the temperature, plasmids should assume, more or less, the supercoiled form [32,33]. In fact, at high temperature, the pDNA assumes more relaxed forms and gains more volume. Inversely, at cold temperature it assumes less volume and becomes supercoiled [34] and more pDNA can be packed by the CLN. Furthermore, Matulis et al. demonstrated that at a higher temperature, the interaction between the cationic lipids with DNA is slightly weaker, due to the less favorable electrostatic and hydrophobic entropies and electrostatic enthalpy [25]. For a clinical approach, it is important to note that the transfection efficiency is highly dependent on the quality and conformation of the used DNA [35].