Cationic Liposomes as Vectors for Nucleic Acid and Hydrophobic Drug Therapeutics

Abstract

Cationic liposomes (CLs) are effective carriers of a variety of therapeutics. Their applications as vectors of nucleic acids (NAs), from long DNA and mRNA to short interfering RNA (siRNA), have been pursued for decades to realize the promise of gene therapy, with approvals of the siRNA therapeutic patisiran and two mRNA vaccines against COVID-19 as recent milestones. The long-term goal of developing optimized CL-based NA carriers for a broad range of medical applications requires a comprehensive understanding of the structure of these vectors and their interactions with cell membranes and components that lead to the release and activity of the NAs within the cell. Structure-activity relationships of lipids for CL-based NA and drug delivery must take into account that these lipids act not individually but as components of an assembly of many molecules. This review summarizes our current understanding of how the choice of the constituting lipids governs the structure of their CL-NA self-assemblies, which constitute distinct liquid crystalline phases, and the relation of these structures to their efficacy for delivery. In addition, we review progress toward CL-NA nanoparticles for targeted NA delivery in vivo and close with an outlook on CL-based carriers of hydrophobic drugs, which may eventually lead to combination therapies with NAs and drugs for cancer and other diseases.

Keywords: affinity targeting; cationic liposomes; gene therapy; homing peptide; hydrophobic drug delivery; nanoparticles; nucleic acids; small-angle X-ray scattering.

Figure 1

Schematic depiction of a unilamellar liposome consisting of a self-assembly of amphiphilic lipid molecules. The liposome can transport cargo at three distinct sites: within its hydrophobic bilayer (red spheres), within its hydrophilic aqueous interior (green ellipsoid), and at the amphipathic interface (yellow/blue rods).

Figure 2

Structures of the cationic lipids DOTAP (1+; 1,2-dioleoyl-trimethylammonium propane chloride) [83], MVL5 (5+) [84], DLin-DMA [85], and the cationic lipids used in FDA-approved CL–NA nanoparticle formulations. Tozinameran and mRNA-1273 are names for the COVID-19 vaccines developed by Pfizer/BioNTech and Moderna, respectively. The positive charge on DOTAP is present independent of pH, while the amine groups of the other lipids acquire their charge by protonation below a certain pH. To highlight this fact, those lipids are sometimes called “ionizable” lipids. MVL5 is a commercially available multivalent ionizable lipid.

Figure 3

Structures of selected neutral (“helper”) lipids and PEG-lipids used in CL-based NA vectors. The molecular weight of the PEG chain of all depicted lipids is 2000 g/mol (n = 45). The phospholipids DOPC, DOPE, and DLinPC are zwitterionic, while GMO and cholesterol are uncharged. The blue highlighting represents the large difference in hydration for DOPE and DOPC. DLinPC: 1,2-dilinoleoyl-sn-glycero-phosphocholine; DO: dioleoyl; DOPC: 1,2-dioleoyl-sn-glycero-phosphocholine; DOPE: 1,2-dioleoyl-sn-glycero-phosphoethanolamine; GMO: glycerol monooleate.

Figure 4

Relation of lipid shape and the resulting self-assemblies. Inverse-cone-shaped lipids give rise to inverse micellar structures (left); cylindrical lipids (center), where the headgroup area approximately matches the projected area of the tails, prefer (nearly) flat membranes (lamellar phases and large vesicles, depending on water excess); cone-shaped lipid (right) prefer micellar structures such as cylinders of varying length and spheres. Blue and red lines indicate the membrane curvatures C₁ and C₂, respectively, of the assemblies. Purple lines behind the schematic depictions of lipids with different shapes indicate the spontaneous curvature C₀ of their assemblies. See Figure 2, Figure 3 and Figure 4 for the chemical structures of the displayed lipids.

Figure 5

Chemical structure and maximum charge of custom synthesized multivalent lipids ((T)MVLs) mentioned in the text [14,17,30,84].

Figure 6

(a) Structures of a cubic phase with Im3m symmetry and two inverse bicontinuous cubic phases, the gyroid (Q_II^G, space group Ia3d) and the diamond (Q_II^D, space group Pn3m) cubic phases. These structures consist of a single continuous lipid bilayer interface (with saddle-shaped negative Gaussian curvature C₁C₂ < 0) dividing space into two disconnected water channels. The drawn surfaces, with one side colored gray and the other colored green, represent the shape of the water–bilayer interface [119,120]. (b) Schematic depiction of a membrane pore, illustrating its saddle-shaped negative Gaussian curvature. Adapted with permission from [121]. Copyright 2011 American Chemical Society.

Figure 7

(a) Schematic of the L_α^C phase of CL–DNA complexes, which consists of lipid bilayers of thickness δ_m alternating with DNA monolayers of thickness δ_w. The interlayer spacing, which gives rise to the series of peaks labeled q_00i in SAXS, is d = δ_w + δ_m = 2π/q₀₀₁. From [13]. Reprinted with permission from AAAS. (b) Example of the typical SAXS pattern resulting from CL–DNA complexes in the lamellar (L_α^C) phase. The Bragg reflections at q₀₀₁ and q₀₀₂ result from the multilamellar structure (see part b). The broad DNA–DNA correlation peak at q_DNA reflects the ordered arrangement of the DNA rods (see part b) with an average interaxial spacing d_DNA = 2π/q_DNA. Complexes were formed from a DOTAP/DOPC (53:47, mol:mol) lipid mixture and λ-phage DNA. Reprinted with permission from [12]. (c) Schematic of the lamellar phase of CL–siRNA complexes. Note the lack of orientation order for the short siRNA rods. Reprinted with permission from [18]. Copyright 2007 American Chemical Society.

Figure 8

(a) Schematic of the inverted hexagonal (H_II^C) phase of CL–DNA complexes. In this phase, inverse cylindrical micelles containing DNA (i.e., DNA coated with a lipid monolayer) are arranged on a hexagonal lattice. The average spacing between the inverse micelles, a, can be obtained from the SAXS profile as a = 4π/√3q₁₀. From [13]. Reprinted with permission from AAAS. (b) Example of the characteristic SAXS pattern of CL–DNA complexes in the H_II^C phase (top profile). Also shown are characteristic SAXS patterns of CL–DNA complexes transitioning from the L_α^C phase (bottom profile) to coexisting L_α^C and H_II^C phases (middle profile) and eventually the H_II^C phase (top profile) as the content of DOPE in the membranes of the DOTAP/DOPE–DNA complexes increases. From [13]. Reprinted with permission from AAAS. (c) SAXS profile of complexes of DOTAP/DOPE (69 mol% DOPE) with plasmid DNA, revealing that the complexes are in the H_II^C phase. Reprinted from [16], Copyright 2003, with permission from the Biophysical Society. (d) SAXS of DOTAP/DOPE–siRNA complexes reveals the formation of the lamellar phase at low content of DOPE (top), the inverse hexagonal phase at high content of DOPE (bottom), and coexistence of the two phases in a narrow regime of intermediate DOPE content. Reprinted with permission from [18]. Copyright 2007 American Chemical Society.

Figure 9

(a) Schematic of the H_I^C phase of CL-DNA complexes. In this phase, cylindrical micelles formed by membranes containing strongly cone-shaped lipids such as MVLBG2 are arranged on a hexagonal lattice and surrounded by the oppositely charged DNA chains which form a honeycomb structure. (b) Synchrotron SAXS pattern of MVLBG2/DOPC–DNA complexes at 25 mol% of the highly charged lipid MVLBG2 (Figure 5). As described in the text, the SAXS peaks index to a 2D hexagonal lattice. Reprinted with permission from [14]. Copyright 2006 American Chemical Society.

Figure 10

(a) Schematic depiction of the double-gyroid cubic phase of CL–siRNA complexes labeled (Q_II^{G, siRNA}). The two intertwined but independent water channels are shown in green and orange. For clarity, the lipid membrane separating the two water channels is represented by a gray surface corresponding to its center (see inset). Note the negative Gaussian curvature of the bilayer, C₁C₂ < 0. (b) Synchrotron SAXS data obtained for DOTAP/GMO–siRNA at a DOTAP/GMO molar ratio of 15/85 (top) and 25/75 (bottom). The large number of peaks reveals the body-centered gyroid cubic structure (space group Ia3d). Reprinted with permission from [15]. Copyright 2010 American Chemical Society.

Figure 11

The membrane charge density (σ_M, the average charge per unit area of the membrane) is a universal parameter for the transfection efficiency (TE) of lamellar CL–DNA complexes, but not for nonlamellar (H_II^C or H_I^C) complexes. (a) TE plotted as a function of molar fraction cationic lipid for DNA complexes of MVL2, MVL3, MVL5, TMVL5, and DOTAP mixed with DOPC (see Figure 2, Figure 3 and Figure 5). (b) The same data as in (a), but plotted as a function of σ_M collapses onto a universal, bell-shaped curve as a function of σ_M (the solid line is a Gaussian fit to the data). TE data for DOTAP/DOPE complexes (open circles, H_II^C phase) deviates from the universal curve, indicative of a distinctly different transfection mechanism for the inverted hexagonal phase. Three regimes of transfection efficiency are highlighted as described in the text. The membrane charge density can be written as σ_M = [1 − Φ_nl/(Φ_nl + rΦ_cl)]σ_cl. Here, r = A_cl/A_nl is the ratio of the headgroup areas of the cationic and the neutral lipid; σ_cl = eZ/A_cl is the charge density of the cationic lipid with valence Z; Φ_nl and Φ_cl are the mole fractions of the neutral and cationic lipids, respectively. The membrane charge density was calculated using A_nl = 72 Ų, r_DOTAP = 1, r_MVL2 = 1.05 ± 0.05, r_MVL3 = 1.30 ± 0.05, r_MVL5 = 2.3 ± 0.1, r_TMVL5 = 2.5 ± 0.1, Z_DOTAP = 1, Z_MVL2 = 2.0 ± 0.1, Z_MVL3 = 2.5 ± 0.1, Z_MVL5 = Z_TMVL5 = 4.5 ± 0.1 [17]. (c) TE for DNA complexes of MVLG2 (4+), MVLBisG1 (8+), MVLBisG2 (16+), and DOTAP mixed with DOPC (see Figure 2, Figure 3 and Figure 5) plotted as a function of membrane charge density. Filled symbols are for lamellar complexes, while empty symbols are for complexes in the H_I^C or distorted H_I^C phases. Again, the data for nonlamellar complexes deviates from the universal curve for lamellar complexes, indicating different transfection mechanisms. Parts (a,b) adapted from [17] with permission from John Wiley & Sons, Ltd. Part (c) adapted with permission from [115]. Copyright 2009 American Chemical Society.

Figure 12

Sequence-specific gene silencing of CL–siRNA complexes incorporating the cubic-phase forming lipid GMO (Figure 2) is strongly improved compared to DOTAP/DOPC–siRNA complexes. Optimal silencing corresponds to K_T (total (specific and nonspecific) gene knockdown; black lines and symbols) approaching 1 while K_NS (nonspecific gene knockdown, red lines, and symbols) is minimal. DOTAP/GMO–siRNA complexes (squares) are in the gyroid cubic phase (Q_II^G,siRNA) at a high mole fraction of neutral lipid (Φ_NL) where K_T is high and K_NS is low. In contrast, lamellar (L_α^siRNA) DOTAP/DOPC–siRNA complexes (circles) show low K_T at high Φ_NL. The increased K_NS at low Φ_NL, when both formulations form lamellar complexes because of their high content of DOTAP (C₀ ≈ 0), indicates an undesirable onset of vector toxicity. Reprinted with permission from [15]. Copyright 2010 American Chemical Society.

Figure 13

Two strategies to enhance transfection efficiency (TE) of PEGylated CL–DNA complexes. TE in murine CCL-1 cells is plotted versus ρ (lipid/DNA charge ratio) for DOTAP/DOPC/PEG-lipid–DNA complexes (80 mol% DOTAP(1+)) and control complexes without PEG-lipid. TE drops strongly upon the inclusion of 10 mol% PEG-lipid. However, complexes containing RGD-PEG2K-lipid or HPEG2K-lipid instead show partial recovery of TE, which is due to distinct mechanisms as discussed in the text. (a) Comparison of the TE of complexes without PEG2K-lipid (black), with 10% PEG2K-lipid (red), and with 10 mol% acid-labile HPEG2K-lipid (blue). Adapted from [179], Copyright 2012, with permission from Elsevier. (b) Comparison of the TE of complexes without PEG2K-lipid (black), with 10% PEG2K-lipid (blue), and with 10 mol% RGD-PEG2K-lipid (green). Adapted from [180], Copyright 2014, with permission from Elsevier.

Figure 14

(a) Structure of the HPEG2K-lipid. The acid-labile acylhydrazone moiety is underlain in red, the lipophilic tails in tan, and PEG in blue. (b) Schematic depiction of the proposed mechanism of TE recovery by the low-pH-sensitive HPEG2K-lipid. During the maturation of endosomes, acidification cleaves the PEG chains from the lipid tails. This unmasks the positive charge of the CL–DNA NP, allowing electrostatically mediated recruitment to, and fusion with, the negatively charged endosomal membrane, facilitating endosomal escape [179].

Figure 15

(a) Structure of the RGD-PEG2K-lipid as an example of a ligand-PEG-lipid. The peptide ligand is highlighted with a red triangle, the lipophilic tails in beige, and PEG in blue. (b) Schematic depiction of ligand-tagging of CL–NA NPs. (c) PEGylation reduces cellular uptake of NPs, reducing efficacy. Functionalization of the distal end of a PEG-lipid with an appropriate ligand induces receptor-mediated binding and increases cellular uptake (and thus efficacy) in cells expressing the peptide’s receptor [180].

Figure 16

Peptide-tagging for specific targeting of CL–DNA NPs in vitro and in vivo. NPs were formulated at a lipid/DNA charge ratio of 1.5 and a molar ratio of 10/70/10/5/5 of MVL5/DOPC/cholesterol/PEG-lipid/x, where x = PEG-lipid (control) or peptide-PEG-lipid. (a,b) Fluorescence from bound and internalized NPs containing labeled DNA in two cell lines (M21 and PC-3) measured by flow cytometry. The graphs compare several tagged (peptide-PEG-lipid) and untagged (PEG-lipid only) NPs with free DNA (no lipid) as a control. Binding and uptake were differentiated by the addition of Trypan Blue, a membrane-impermeable dye that quenches the fluorescence of NPs outside the cells. NPs tagged with cRGD detached a large number of cells from the substrate. These cells were measured separately. *: too few cells remained attached to allow measurement. (c) In vivo biodistribution of intraperitoneally (i.p.) administered CL–DNA NPs. Mice bearing intraperitoneal MKN-45P tumors were i.p. injected with either PBS (control) or ~0.5 mg of CL–DNA NPs. After 24 h the tumors and organs of interest were excised and the fluorescent signal from the Cy5-labeled DNA was imaged (inset) and quantified (bars; normalized to control; n = 3). The vast majority of the fluorescent DNA is found in the tumor, and peptide-tagged NPs show higher selectivity for the tumor than untagged NPs. (d–i) Confocal microscopy images showing CL–DNA NPs in sections of the tumor nodules. Parts (g–i) are enlarged views of the marked areas in parts (d–f), respectively. Cy5 (DNA-label, i.e., NPs): red, DAPI (cell nuclei): blue. Tumor nodules from mice treated with untagged (control) PEG2000-lipid NPs (d,g) show NPs on the nodule surface, while iRGD- (e,h) and cRGD-tagged NPs (f,i) penetrated into the tissue of smaller tumor nodules (diameter ~300 μm). Scale bars: 500 μm (d–f) and 200 μm (g–i). Adapted from [242], Copyright 2018, with permission from Elsevier.

Figure 17

Top: Structure of DOTAP, with oleoyl (C₁₈ with a single cis double bond) tails, and the corresponding lipid DLinTAPwith linoleoyl tails (with an additional cis double bond). Bottom: Space-filling molecular models of the ground-state structure of the lipid tails and PTX solubility kinetic phase diagrams for the corresponding DOTAP/DOPC and DLinTAP/DLinPC formulations. Formulations of increasing PTX content (x-axis) were monitored over time (y-axis) for PTX crystallization (red color). Blue color indicates the absence of PTX crystals. See Figure 3 for the structures of DOPC and DLinPC. Solubility phase diagram data reprinted with permission from [109].

Figure 18

(a) Schematic illustration of the structural transitions observed upon PEGylation of PTX-carrying CLs. Unilamellar vesicles of varying sizes are replaced by small vesicles and discoidal micelles (bicelles). (b) Cytotoxicity of PTX-carrying CLs as a function of increasing PEGylation (at a constant amount of PTX). The efficacy of the CLs against cancer cells increased (cell viability decreased) with the extent of PEGylation. (c) Cryogenic electron microscopy image of a formulation of DOTAP/DOPC/PEG2000-lipid/PTX at a molar ratio of 50/37/10/3. Small vesicles of varying size (white arrows) and edge-on (green arrowhead), tilted (green arrow), and top-down (green dashed arrow) views of discoidal micelles are discernible. Adapted with permission from [112]. Copyright 2020 American Chemical Society.