Development of ROS-Triggered Masked Liposomes for Activated Cellular Delivery using a Charge Balance Strategy

Abstract

We report a charge balance strategy for reactive oxygen species (ROS)-triggered activation of liposome cell delivery by unmasking cationic membranes. Zeta potential experiments, microplate-based vesicle interaction assays, and cellular delivery studies confirmed that modification of anionic lipid 1 by ROS led to the uncaging of cationic liposomes, thereby driving cellular association.

Table of content figure:

While nanocarriers have emerged as powerful platforms for the delivery of therapeutic agents, further work is needed to maximize their potential. In particular, despite recent advancements, further improvements are needed in the ability to target diseased cells while avoiding healthy cells. Among the different classes of nanocarriers, liposomes, composed of spherical lipid bilayer membranes, exhibit promising drug delivery characteristics.¹ Indeed, at least fifteen FDA-approved liposomal formulations are currently in clinical use.² A promising approach for achieving targeted cell delivery is to develop responsive liposomes that are only activated in the presence of target stimuli that are amplified in diseased cells, such as acidity, redox conditions, and enzymes.³ These strategies typically employ lipid switches that are activated in the presence of the overexpressed stimulus in order to control activity.

A promising family of stimuli for targeting diseased cells that has been underexplored consists of reactive oxygen species (ROS), including superoxide, hydrogen peroxide, singlet oxygen and hydroxyl radicals. ROS are normally produced in low amounts within cells and play significant roles in metabolism and regulation of physiological functions. However, an overabundance of ROS can lead to oxidative stress and cellular degeneration that are associated with diseases such as cancer, hypertension, asthma, Alzheimer’s Disease, and Parkinson’s Disease.⁴ Furthermore, the concentrations of ROS within diseased cells (~100 μM) have been determined to be significantly higher than those of healthy cells (~20 nM),⁵ resulting in their emergence as strategic triggers for differentiating cells.

Our group previously reported ROS-triggered activation of liposome cellular entry employing a neutral lipid conjugate containing a caged guanidine.⁶ In the presence of hydrogen peroxide, boronate caging groups were removed, producing cationic guanidinium-presenting liposomes, thereby activating cell entry. While this initial approach was effective, there are aspects of the design that could be improved. In particular, generation of our caged guanidine lipid required a lengthy synthesis, making this compound difficult to access. To develop a streamlined strategy for ROS-triggered activation of cell delivery, herein we report an alternative approach based on charge balance for masking liposomes (Scheme 1A). Rather than direct caging of a cationic lipid, this strategy instead employs an unmodified cationic lipid within liposomes. Liposome charge is then balanced using an ROS-responsive anionic lipid to initially render approximately neutral liposomes. Encountering ROS leads to conversion of the anionic lipid to a neutral product, and therefore the overall liposome charge becomes positive. We present this approach as an efficient and accessible avenue for achieving ROS-activated programming of cell entry.

Scheme 1.

ROS-triggered uncaging of cationic liposomes using a charge balance strategy. A. Cartoon depicting liposome uncaging. Initially, cationic DOTAP is balanced through the inclusion of anionic lipid 1 to produce approximately neutral liposomes. The uncaging of 1 by ROS (H₂O₂) produces the neutral lipid DOPE, thereby revealing the native positive charge of DOTAP on the liposome surface, which can be exploited to enhance cellular delivery. B. Structure of boronate-caged lipid 1 and the mechanism of uncaging in the presence of ROS, which includes boronate oxidation and linker self-immolation to produce DOPE.

For the cationic lipid in this formulation, we selected 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), which has been extensively used for gene delivery and transfection, RNA interference studies, and drug delivery.⁷ However, DOTAP-containing liposomes rarely exhibit cell-type selectivity⁸ since their positive charge induces general infiltration of cell membranes,⁹ a problem that we seek to remedy by rendering liposomes ROS-responsive. The ROS-responsive lipid in this platform is compound 1 (Scheme 1B). While we previously reported this boronate-caged lipid for ROS-triggered release of encapsulated cargo from liposomes,¹⁰ in the current work we repurpose this compound to instead control membrane charge. Lipid 1 is designed to undergo ROS-driven boronate oxidation followed by linker self-immolation to generate the neutral (zwitterionic) lipid 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE, Scheme 1B). Conveniently, the synthesis of 1 requires only two synthetic steps starting from DOPE itself.

To investigate the unmasking of liposomes using 1/DOTAP formulations, we first conducted Zeta Potential (ZP) experiments to test surface charge before and after H₂O₂ treatment. Liposomes containing variable percentage of 1 (0–30%), DOTAP (0–25%), and l-α-phosphatidylcholine (PC, 50–100%) as a bulk lipid, were prepared using thin-film hydration, including extrusion through a 200 nm membrane to control liposome size.¹¹ Dynamic light scattering (DLS) studies were conducted before and after H₂O₂ treatment to analyze liposome average vesicle size (Z-Ave) and polydispersity index (PDI)to assess stability. Formulations tested included equal percentages of both 1 (0–25%) and DOTAP (0–25%) or a slightly higher percentage of 1 (25–30%) than DOTAP (20%), each doped into PC (50–100%). ZP changes following 40 μM H₂O₂ treatment are shown in Figure 1A, for which the most significant increases in positive charge, while maintaining vesicle uniformity, were exhibited by liposomes containing 20/20/60 1/DOTAP/PC (~0 to ~28 mV) and 25/20/55 1/DOTAP/PC (~−13 to ~33 mV). Both formulations maintained Z-Ave values less than 200 nm and PDI values less than ~0.2 before and after H₂O₂ (Figures S1 and S2C-D). While 25/25/50 and 30/20/50 1/DOTAP/PC liposomes developed positive charge upon H₂O₂ incubation, these formulations were less effective for yielding uniform vesicle sizes. (Figures S1 and S2E-F). This indicates that higher percentages of 1/DOTAP lead to the formation of larger and non-uniform vesicles that may result from fusion.¹² To further probe stability, we encapsulated the hydrophobic dye Nile red within successful liposome formulations and found that cargo release upon H₂O₂ treatment was less than PC control liposomes (Figure S3). These experiments identified two promising formulations that are effective for ROS-driven uncaging of lipid switch 1 to form cationic liposomes.

Figure 1.

ZP results for unmasking of 1/DOTAP/PC liposomes A. End-point results for liposome containing 0–30% 1, 0–20% DOTAP and 50–100% PC. Liposomes consisting of 1 and DOTAP generally showed a significant increase in positive charge after H₂O₂ treatment. B. End-point results for 20/20/60 or 25/20/55 1/DOTAP/PC liposomes treated with different concentrations of H₂O₂ (0–100 μM). Statistical analysis (t test) was performed, where n.s. (p > 0.05), *(p < 0.05), **(p < 0.01), ***(p < 0.001). C. Time-dependent studies show a significant increase in ZP values for 20/20/60 or 25/20/55 1/DOTAP/PC liposomes in comparison with 100% PC (negative control) and 20% DOTAP/PC (positive control) liposomes. Error bars indicate standard error from at least three independent trials.

We next tested the effects of increasing H₂O₂ on these effective formulations. Here, we observed a dose-dependent increase in positive charge following treatment with 0–40 μM H₂O₂ for both liposome formulations (Figure 1B). When the H₂O₂ concentration was further increased to 100 μM, both samples showed slightly diminished positive charge compared to 40 μM treatment. PDI values and changes in DLS raw data upon increasing [H₂O₂] are shown in Figures S4 and S5, respectively. Next, time-dependent studies of ZP changes H₂O₂ addition were performed (Figure 1C). The previous liposomes were tested alongside two controls, including 100% PC (negative control lacking 1 and DOTAP) and 20/80 DOTAP/PC (positive control lacking 1). For both 20/20/60 and 25/20/55 1/DOTAP/PC liposomes, ZP values increased over time after 40 μM H₂O₂. Both control formulations showed minimal change in charge over time, as expected. These data are also replotted to show ΔZP values before and after H₂O₂ incubation (Figure S6). Additionally, liposome formulations containing cholesterol and PEG-PE in line with clinical drug formulations¹³ were tested and showed stability over several days (Figure S7 and S8). These results confirm the uncaging of 1 by H₂O₂ and the formation of cationic liposomes.

Promising formulations were next tested using a fluorescence-based vesicle binding/fusion assay involving the immobilization of intact liposomes onto streptavidin-coated microplates (Figure 2A).¹⁴ Each experiment entailed treatment of microplates with two samples of liposomes. The first sample contained Biotin−PE (1%) for streptavidin immobilization as well as l-α-phosphatidylserine (Brain PS, 5%), an anionic lipid that is overexpressed in the outer membrane of diseased cells¹⁵ and is used here as a target for cationic liposomes. These initial liposomes were incubated with microplates for surface coating through biotin-streptavidin binding. Next, the second liposome sample was added, containing different percentages of 1 and DOTAP in PC, also doped with 18:1 Lissamine Rhodamine PE (Rd-PE) as a fluorescent lipid marker for detection (0/0/99/1, 20/20/59/1, or 25/20/54/1 1/DOTAP/PC/Rd-PE). These liposomes were either pre-treated with 10 mM H₂O₂ or instead treated with buffer as a negative control. Following thorough washes to remove unbound components, detection of Rd-PE fluorescence provides a readout of interactions of the two liposomes on the plate surface through fusion and/or association. Fluorescence analysis showed that only 1/DOTAP/PC liposomes treated with H₂O₂ showed a significant increase in Rd-PE fluorescence in comparison with negative control liposomes that were not treated with H₂O₂ or PC liposomes lacking 1 and DOTAP (Figure 2B). These data support that H₂O₂ caused uncaging of 1, resulting in cationic liposomes that associated with and/or fused to immobilized PS-liposomes.

Figure 2.

Schematic to show the microplate assay to evaluate ROS-driven liposome binding/fusion through in a microplate assay. A. Cartoon illustrating the fluorescence-based streptavidin-coated microplate assay showing vesicle interaction. Anionic PS and biotin-PE doped liposomes were first immobilized onto the microplate, followed by the incubation with liposomes containing 0/0/99/1, 20/20/59/1, or 25/20/54/1 1/DOTAP/PC/Rd-PE with or without H₂O₂ treatment. B. Relative fluorescence intensity (F-F₀) of Rd-PE showed a significant increase only when liposomes containing 20/20/59/1 or 25/20/54/1 1/DOTAP/PC/Rd-PE were incubated with PS liposomes with H₂O₂ treatment. Error bars indicate standard error from at least three independent trials.

We next tested this platform for ROS-activated cellular delivery using fluorescence microscopy.^{6, 16} Human malignant melanoma (A375) cells were treated with different liposomes labelled with 0.08% Rd-PE as a fluorescent marker to detect their association with cells. Study liposomes (20/20/59.92/0.08 1/DOTAP/PC/Rd-PE, 2 mM) were pre-treated with H₂O₂ (40 μM) and were studied alongside negative controls that were not pre-treated with H₂O₂ (replaced with buffer) or liposomes that did not contain 1 or DOTAP (0/0/99.92/0.08 1/DOTAP/PC/Rd-PE). The formulation with a lower percentage of 1 (20%) was selected since our prior results showed that higher percentages led to instability. All liposomes were diluted to 1 mM in DPBS⁺⁺ or media and incubated with cells for 1 hour (37 °C and 5% CO₂), after which the cells were washed, fixed, stained with DAPI, and then imaged using fluorescence microscopy (Figure 3).

Figure 3.

Fluorescence images showing enhanced liposomal association with A375 cells only when 1/DOTAP/PC/Rd-PE liposomes were treated with H₂O₂. 20X fluorescence images of cells treated with DPBS⁺⁺ buffer only (A), 40 μM H₂O₂ (B), 0/0/99.92/0.08 1/DOTAP/PC/Rd-PE liposomes without (C) and with (D) 40 μM H₂O₂ pre-treatment, and 20/20/59.92/0.08 1/DOTAP/PC/Rd-PE liposomes without (E) and with (F) 40 μM H₂O₂ pre-treatment. Blue: DAPI; Red: Rd-PE; Scale bar: 100 μm. (G) Quantified relative Rd-PE fluorescence of cells treated with no liposomes, PC/Rd-PE liposomes or 20/20/59.92/0.08 1/DOTAP/PC/Rd-PE liposomes with and without H₂O₂ treatment. Statistical analysis (t test) was performed, where n.s. (p > 0.05), *(p < 0.05), **(p < 0.01), ***(p < 0.001). Error bars indicate standard error from at least three independent trials.

The resulting images show that the negative control samples did not result in appreciable fluorescence signal, including cells not incubated with liposomes without or with H₂O₂ (Figure 3A and 3B, respectively), and 0/0/99.92/0.08 1/DOTAP/PC/Rd-PE liposomes without or with H₂O₂ treatment (Figure 3C and 3D, respectively). While 20/20/59.92/0.08 1/DOTAP/PC/Rd-PE liposomes without H₂O₂ treatment showed no fluorescence (Figure 3E), upon pre-treatment with H₂O₂ fluorescence labelling of cells was observed (Figure 3F), supporting cell association for this sample. Quantified fluorescence intensities from images using Gen5 are shown in Figure 3G, which indicate that H₂O₂ pre-treatment of 20/20/59.92/0.08 1/DOTAP/PC/Rd-PE liposomes led to an approximately ten-fold increase in fluorescence (Figures S9 and S10). Similar experiments with slightly modified protocol using media containing catalase from fetal bovine serum (10% of cell culture medium) resulted in a similar fluorescence labelling of cells only when 1/DOTAP/PC/Rd-PE liposomes were pre-treated with H₂O₂ (Figure S11). Fluorescence microscopy results for 20/20/59.92/0.08 1/DOTAP/PC/Rd-PE liposomes treated with H₂O₂ for uncaging were comparable to experiments that were performed on positive control samples lacking 1 (20/79.92/0.08 DOTAP/PC/Rd-PE liposomes) with or without H₂O₂ treatment, which were therefore not, which were therefore not ROS-responsive, as expected (Figures S12 and S13). Studies were also performed at 4^oC, during which all active liposomes retained activity, supporting that liposomes do not enter cells via an energy-dependent endocytosis pathway (Figures S14 and S15). Cell viability experiments were also conducted to probe for cytotoxicity of ROS-responsive liposomes, in which they were comparable to controls containing naturally existing lipids (Figure S16). Collectively, these experiments demonstrate that 1/DOTAP/PC/Rd-PE liposomes are activated for cell association only after encountering H₂O₂, thereby enabling programmed delivery based on ROS abundance.

In conclusion, we have demonstrated an efficient and accessible platform for ROS-activated liposomal delivery to cells using charge balance to mask DOTAP-containing liposomes. We found that ROS-driven cleavage of anionic lipid switch 1 was effective at unveiling positively charged liposomes that are activated for cell delivery. Our work shows that smart liposomal delivery systems can be developed through careful control of charge states within liposomes, in which the membrane surface charge is altered through chemical reaction with a target stimulus. This approach shows strong prospects for enabling selective targeting of diseased cells by harnessing chemical triggers that are overexpressed during disease onset. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under award number NIH R15GM146193. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Cartoons were created in BioRender. We also thank Dr. Mahshid Mokhtarnejad and Dr. Bamin Khomami for assistance with DLS experiments.

Data availability

All data from this work is available upon request.