Characterization of liposomes
Our previous study showed that peptide liposomes with tri-ornithine as the headgroup (LOrn3) could transfer DNA and DNA very efficiently into tumor cells, such as NCI-H460 and Hep-2 cells. Also, it was able to deliver combined DNAs against c-Myc and VEGF for silencing distinct oncogenic pathways in lung tumors of mice, with little in vitro and in vivo toxicity [19]. To develop further applications of peptide lipids for gene delivery, a comparison with the lipid LOrn1, with one ornithine as the headgroup, was used to obtain detailed information about the interaction between peptide lipids and genes and the intracellular transport process. Our previous study showed that liposomes had the highest transfection efficiency with equivalent molar ratio of DOPE, accordingly the liposomes LOrn1 and LOrn3 were prepared with DOPE as a co-lipid at a molar ratio of 1:1. Figure 2a shows that the sizes were around 80 nm for both liposomes LOrn1 and LOrn3 (Fig. 3a). And the Zeta potentials of LOrn1 and LOrn3 were 30 mV and 45 mV respectively (Fig. 3b), which ensured the stability of the liposomes. The morphological characteristics were visualized by transmission electron microscopy (TEM) (Fig. 3c). Images obtained from LOrn1 and LOrn3 liposomes revealed well-defined spheres structures, regularly shaped and uniformly distributed.
DNA transfection in vitro
In our previous work, we reported that peptide liposomes delivered DNA and siRNA into various tumor cells with Lipofectamine 2000 as control. We discovered that the transfection efficiency of liposomes had a similar performance in HeLa, Hep-2, A549, MCF-7, H460 cells, though HeLa and Hep-2 cells showed higher transfection efficiency. Therefore, to evaluate further transfection efficiency of the two liposomes, we used them to deliver the GFP into HeLa and Hep-2 cells. The expression levels of GFP are shown in Figure S1, Figure 4a and 4b after the qualitative measurement of fluorescence and the quantitative determination of the fluorescent intensity. The charge ratios had a great impact on DNA transfection, LOrn1 and LOrn3 possessed the highest transfection efficiency at charge ratios of 4:1 and 3:1, respectively. And the expression of GFP of the LOrn3-treated cells was significantly stronger than that of the LOrn1-treated cells at the charge ratios from 1:1 to 8:1 in two cells. And the transfection efficiency against HeLa cell was higher than that against Hep-2 cells. Therefore, we chose HeLa cells to do the subsequent experiments. And it's noted that LOrn1/DNA and LOrn3/DNA lipoplexes had much lower toxicity to HeLa cells at charge ratios of 4:1 and 3:1, respectively. The cell viability was also over 80% after treatment with lipoplexes for 72 h (Additional file 1: Fig. S2).
Then, we investigated the ability of liposomes to complex DNA at charge ratios from 0.5:1 to 8:1, findings in the gel retardation assay revealed that DNA binding amount increased with the charge ratios increase. LOrn1 and LOrn3 completely bound DNA at charge ratios more than 4:1 and 3:1, respectively (Fig. 4c and d). Meanwhile, the Zeta potentials of lipoplexes were detected, the results show that their Zeta potentials increased from negative to positive with the increase of charge ratios (Fig. 4e). For LOrn1 and LOrn3, the values became positive just at charge ratios of 4:1 and 3:1, respectively, corresponding with the charge ratios of complete DNA complex, these results accorded well with the above gel retardation assay results. This study indicates that LOrn3 was more effective than LOrn1 in terms of the interaction with DNA. In addition, synchrotron small-angle x-ray scattering (SAXS) was used to analyze the structures of the liposome/DNA lipoplexes, and the liposome was composed of peptide lipid (LOrn1 or LOrn3) and DOPE at ratio of 1:1. For LOrn1/DNA and LOrn3/DNA lipoplexes, though the LCα and HC‖ structures coexisted (Additional file 1: Fig. S3), LOrn3/DNA lipoplex formed more HC‖ structures than LOrn1/DNA, which was considered as one reason for higher transfection efficiency of LOrn3. SAXS scans showed sharp peaks at q1 = 0.106 Å-1 and q2 =0.239 Å-1 from the lamellar periodic structure (d = 2π/q1 = 59.27 Å) and at q3= 0.126 Å-1 from hexagonal structure. The peptide lipid-DOPE bilayer thickness was dm = 40 Å [17], the water gap between bilayer was 19.27 Å (dw = d – dm). While the hexagonal lattice constant a is 57.58 Å (a = 4π/(3)0.5q3). During this process, DOPE is a fusogenic phospholipid that presents a super synergistic effect when used with cationic liposomes, decreasing the cytotoxicity imposed by the cationic lipids, destabilizing lipid bilayers by increasing the fusion with the cellular and endosomal membranes and allowing the release of DNA into the cytosol. Lipoplexes containing DOPE may facilitate the transition from lamellar to hexagonal structures promoting high gene transfection efficiency [17].
Binding kinetics of liposome with DNA
To obtain detailed information about the interaction between liposomes and DNA, we utilized agarose gel electrophoresis and fluorescent resonance spectroscopy to examine the time-dependent interaction of cationic peptide liposomes with DNA. As shown in Figure 5a and b, the binding amount of DNA significantly increased with time, and nearly 100% DNA binding was reached after 30 min and 24 min for LOrn1 or LOrn3 liposomes, respectively. The data also demonstrated that the binding speed between liposome LOrn3 and DNA was faster than that between LOrn1 and DNA. The interaction between liposomes and DNA labeled with a nucleic dye (GelRed) was also monitored by fluorescence spectroscopy [26, 27]. GelRed fluoresces strongly when bound to DNA, and the displacement of GelRed from DNA by liposomes results in a decrease in fluorescence intensity that correlates with the amount of liposome binding DNA. The results show that the displacement of GelRed from DNA by liposome LOrn3 was much faster than that by LOrn1 (Fig. 5c). Moreover, the fluorescence intensity of GelRed/DNA was lower for LOrn3 when the displacement reached the equilibrium. The results substantiate the fact that the DNA binding of liposomes is faster at the initial stage, and LOrn3 has a stronger DNA binding ability than LOrn1. To further investigate the binding process of the two liposomes and DNA, a stopped-flow fluorimeter (SX-20MV, UK) setup was employed to monitor the formation of liposome/DNA lipoplexes using the fluorescent probe GelRed. The data indicate that GelRed fluorescence decreased much more rapidly for LOrn3 than LOrn1 (Fig. 5d), and the K values of LOrn1 and LOrn3 were 0.102795 and 0.003472, respectively. By fitting the binding data, it was found that the binding process between liposomes and DNA conformed to the kinetics equations: y = 0.235141 × exp (-0.102795x) + 3.461248 and y = 1.663631 × exp (-0.003427x) + 6.278163 for LOrn1 and LOrn3, respectively. It means that the binding rate of LOrn3 and DNA was faster than that of LOrn1 and DNA. Then the binding affinity of liposomes to DNA was identified by using MicroScale Thermophoresis (MST) performed on the Monolith NT.115Pico system. In this method, the affinity of liposome to DNA is represented by Kd value, which is the liposome concentration required to bind 50% DNA [28, 29]. The dose response curves generated by titrating LOrn1 or LOrn3 liposomes to DNA solutions (Fig. 5e) show that the Kd values of LOrn1 or LOrn3 were 0.296 and 0.136 µM, respectively. Therefore, we speculate that high affinity of LOrn3 to DNA led to fast binding rate between them, and then resulting in more efficient transfection. Therefore, we further evaluated the effect of interaction kinetics between the liposomes and DNA on the transfection efficiency. The expression of GFP gradually increased with binding time, and then reached to the maximum values at about 24 min and 32 min for LOrn3 and LOrn1 liposomes, respectively (Fig. 5f and g). At this time the interaction between liposomes and DNA also reached equilibrium (Fig. 5b and c). We confirmed that the interaction kinetics is a key factor for DNA transfection efficiency.
Stability of lipoplexes in blood mimicked environment
Blood is rich in negatively charged macromolecules, such as serum protein and heparin, and these materials are known to dissociate DNA from lipoplexes by the interaction between liposomes and serum proteins [30–33]. Therefore, we first examined the effect of serum on the dissociation of lipoplexes, as shown in Figure. 6a and b, which depicts DNA release of lipoplexes in a medium with varying serum concentrations. A serum concentration of 10% had nearly no obvious effect on the stability of both lipoplexes, with a DNA release of about 2%. With increasing serum concentrations, the stability of both lipoplexes decreased, and the LOrn1/DNA lipoplex showed much more DNA dissociation than the LOrn3/DNA lipoplex. When the serum concentrations were greater than 25%, the release of DNA from the LOrn1/DNA lipoplexes increased to 20–35%; in contrast, the percentage of DNA released from the LOrn3/DNA lipoplexes under the same conditions was much lower. Our previous study had already revealed that lower concentrations of serum (5% or 10%) resulted in only a small reduction in transfection efficacy; with 20% serum, the transfection efficacy was slightly lowered [34]. Therefore, we hypothesize LOrn3 could well protect DNA in the lipoplexes from the interaction with serum when they are in the bloodstream.
Heparin sodium can competitively bind to liposomes from lipoplexes, which may lead to the dissociation of lipoplexes in blood. Therefore, the stability of lipoplexes was further examined by adding heparin sodium to simulate the blood environment. Figure 6c and d showed that the brightness and the area of the bands corresponding to free plasmid DNA increased with the increase in heparin concentration. Heparin sodium affected the stability of the LOrn1/DNA lipoplex to a much greater extent than the LOrn3/DNA lipoplex in the concentration range of 0.1–2.0 µg/µL. At a heparin concentration that is common in the blood (0.1 µg/µL), the release of DNA from the LOrn1/DNA lipoplex was 40%, and it was 20% for the LOrn3/DNA lipoplex. Therefore, the LOrn3/DNA lipoplex could show great advantages over LOrn1/DNA when used for gene delivery.
Cellular uptake of lipoplexes
Cellular uptake is required for successful transfection of genes. The cellular uptake of liposomes/FAM-DNA lipoplexes was first evaluated by FACS within 6 h. As Trypan Blue could quench FAM-DNA fluorescence outside cells, it was used to differentiate internalized vs extracellular fluorescence markers. As displayed in Figure 7a, the cellular uptake increased gradually with the transfection time; cellular uptake rates of LOrn1 and LOrn3 were about 73% and 94% after 4 h, respectively. LOrn3 showed a stronger interaction with the negatively charged cell membrane than LOrn1, so the LOrn3/DNA lipoplex could more easily cross the cell membrane, resulting in a higher cellular uptake rate compared with LOrn1/DNA. And we examined its morphological features at the ultra-structural level in HeLa cells incubated with LOrn3/DNA lipoplexes for different periods of time (Fig. 7b). The majority of LOrn3/DNA lipoplexes were found within early endocytic structures (endosomes) and multi-vesicular bodies in the early phase of uptake (2 h). They were detected within lysosomes 4 h after exposure of cells to LOrn3/DNA lipoplexes.
To further explore the kinetics of LOrn3/DNA lipoplex uptake, cellular uptake of the liposome/DNA lipoplexes was also investigated by laser scanning confocal microscope (LSCM) (Fig. 7c). As illustrated in the merged images, LOrn3/DNA lipoplexes displayed visible signals in the cytoplasm at 30 min post-transfection, but only a small amount of LOrn1/DNA lipoplexes were internalized into the cells. Though more LOrn1/DNA lipoplexes were localized in the cytoplasm at 2 and 4 h post-transfection, it was found that the fluorescent signals of LOrn3/DNA lipoplexes were distributed throughout the entire cytoplasm at 2 h post-transfection and mainly gathered around the nuclei at 4 h post-transfection. Compared to liposome LOrn1, liposome LOrn3/DNA lipoplexes exhibited a faster and more uniform uptake in HeLa cells. And combined with the interaction data of liposome and DNA (Fig. 5), it's found that the more the liposome-DNA interaction was, the higher the cellular uptake of liposome/DNA lipoplex was.
Dissociation kinetics of liposome and DNA
After the assembly with DNA, the stimuli-responsive features and DNA dissociation behavior of the lipoplexes were tested in vitro in a simulated intracellular microenvironment (pH 5–5.5) using a gel retardation assay [35, 36]. As can be seen in Figure 8a and b, the DNA dissociated from lipoplexes was time-dependent at pH 5.5, and the DNA dissociation rate of the LOrn3/DNA lipoplex was faster than that of the LOrn1/DNA lipoplex. Seventy percent of DNA could dissociate from LOrn3/DNA lipoplex and only 51% from LOrn1/DNA lipoplex at 72 h. In contrast, little DNA dissociated from the two lipoplexes at the similar time point (8% at 72 h) at pH 7.0. The pH-induced structural destruction of lipoplexes contributed to the DNA release by exposing carbamate bonds in the lipids to the reductive environment for cleavage (Additional file 1: Fig. S4).
A major intracellular barrier for DNA delivery is endosomal entrapment followed by trafficking and lysosomal degradation or exocytosis [37, 38]. The escape of DNA from late endosomes/lysosomes into the cytosol is thought to be a rate-limiting step for many delivery approaches [39]. The lipoplexes are designed to facilitate late endosomal/lysosomal escape of DNA by synergistic effects of the proton sponge effect [40, 41] and the degradation of the peptide lipid LOrn1 or LOrn3 by carbamate bond breaks in late endosomes/lysosomes. To assess intracellular DNA release, confocal microscopy was used to investigate the distribution of FAM-labeled DNA (green) in lipoplexes and LysoTracker Red labeled late endosomes and lysosomes against HeLa cells. As shown in Figure 8c, we observed that DNA was gradually released from the both of lipoplexes over time. For LOrn3/DNA lipoplex, most FAM signals of the LOrn3/DNA lipoplex were merged in late endosomes/lysosomes at 12 h, as shown by the yellow color. Much DNA could be released from late endosomes/lysosomes at 24 h after the transfection, and at 48 h, strong green fluorescence in the images for LOrn3/DNA lipoplex showed an obvious release of DNA from late endosomes/lysosomes (indicated with white arrows).
The co-localization ratio (%) is a commonly used index to measure DNA release [42, 43], which was carried out by the following formula:
Signalyellow means the fluorescence signal of FAM-DNA (green) co-localized with lysosomes (red), and Signalgreen means the fluorescence signal of FAM-DNA in cytoplasm. A low co-localization ratio means quick DNA release from the late endosomes/lysosomes and a better distribution of the DNA in cytoplasm. The results show that the co-localization ratio gradually decreased with time; they were 49% and 58% at 48 h for LOrn3 and LOrn1, respectively (Fig. 8d). According to the results of the lipoplexes dissociation study in a simulated tumor microenvironment (Fig. 8b), the response to the acidic environment was probably caused by the break of carbamate bonds in the lipids. In addtition, DNA release of LOrn3/DNA lipoplex was more than that of LOrn1/DNA lipoplex at 48 h, this was consistent with the result in Figure 8b. Although LOrn1 and LOrn3 have a similar chemical constitution, LOrn3, containing more amino groups, promoted not only cellular internalization [44] but also displayed a higher proton buffering capacity. Therefore, LOrn3/DNA lipoplex showed an enhanced capability in late endosomes/lysosomes escape and DNA release compared to LOrn1/DNA lipoplex. That also explains the reason why LOrn3 had more efficient transfection than LOrn1. Accordingly, FRET was utilized to further confirm the intracellular disassembly process of DNA from the LOrn3/DNA lipoplex in HeLa cells, by using Rhodamine-DNA and NBD-PE-LOrn3 liposome to constitute a Rhodamine/NBD FRET system (acceptor/donor) [45]. After the HeLa cells were transfected with lipoplexes for 4 h, they were incubated for 2, 4, 6, 8, 12, 24, 36, and 48 h. Figure 8e shows the fluorescence emission intensity of Rhodamine-labeled DNA decreased with time. The data indicate that less resonance energy was transferred from donor to acceptor due to lipoplexes dissociation in HeLa cells, and much of the DNA was released from lipoplexes at 48 h after transfection.
Together, these results strongly demonstrated that both of the lipoplexes could exist in a comparatively stable structure under normal physiological conditions (pH 7.0), while they rapidly release a significant amount of DNA due to the destabilization resulting from the acidic microenvironment in endosomes or lysosomes of tumor cells (pH 5.5) [46, 47].